MENS CEST hate

ed vy

Lis

hit Bl Lh Wh Tc hl |

Dil ste areas

IM EN ONTO EA ION BAD LOS Th eV PA

belie a a

PCa La

ANON he tay

Che penny. TOR

ry 1. my WA ne hee

i ¥ we POOR ENT VDA DS Pa Meta VE DKA R SSDS SIE Bait Mit KW! Hb TEYASN VED oad yh Waly MY TVW NA dM tals iL tial

ew BA ‘ S . ! > R f APs WS Bho lg

fa Korve alg ‘ it SAE NOAVE YOR Oe oh ota ret MAMA ph Meet phy PAS An Pri rat on rh Me ty Dear erty wiitin Ty te VEK ls VEN. type aa ey atta ttl wh ras ren "

La atthe tind POM NNO Wey Rm anO rin en ore 5 ; LLP ML 2A YES SEN AVR WRN Cb y An by ered reer g PV NEE Cuabiustoneenves Ida ead Shops eh SMA ew TL ATR M Da CATE SEAN AN MELT, aS RIN MONT PAE y WAY Te oom enn

Vern HOP He MEM 2iEWhe be pt TN eh a ce a at Poot SOP oer vey

NVI ea ys

AV Ca DY Ma Cees se ale ili, Gat marae tall ews rad UL) INAS e

vite nedy yy AADAC ADS SAWN ¥IL OEM UA LURE AMG

eral DyT Qe B

AUN NSA Chy Thar Ye ge! Ay

F y De SDN yea Ve Bia Poe I BEAM ify Mea aN g Mi fa

rs eee eC) aay ey Lah od thay, py La ie T rity

er POLE BLUE Ey riewets (ues Malbah ys VME Vaan vve% PREnwEeny AYMLAW Merny nce nne aan y AeaTelsin hy vtLanstaa atalitaty arin yh

a be Nese gral eee mpi nt adie cine ahs Sie aa LN SAME A HAM A AAD AS VINE IGN Pd Waar Beene L Ag ot Be My Phe Ya Ue SNe HiEMP Ah de yl sMytey THE rae ha yA ¥iMM yb aM

Sara IAM yd anna A A esi viseis ath retry pi He a terre nee Ate be hast by ‘ ie Ya VAL GUA NA PIM BOI HEIWAD #5) NES ANNE ANS Ay MSDs Ve WH IME MBs ath aa sf

feds nite antan Nala ta 2 etatorlarhe AY Ne Wey OR Une OTT Ete eat ms se ay “y eye oby oes eh

UE MUTA AY Oa Gehiowens Met en eR ee MCE MELEE SOLE Ly MMOD Get # Heth; tad Se Wefe ren F i Site ARES SLES RO ane, SYA Gy EAB MUS Ah ate ©

A vdietic ta hte Net yo Tatlall ewe A Ud ree wile Mw Nth hi Ayatiatetid MRE DN ITT MEO OTS, ny POUR UR a) tywordid Hoe ie irene baa

Men EWEN. MRO IT TAMER Ally OMNI tOL OU enentie ao mat TTA SAN Fe yee ay tet, Oe

ioe fo Mates he eiaerry ed etieti tad YT it wing 7

ELI PoP ye ENTS ed

Lye eee

po Crttee QRS Aa of Faure fou tt

Rear Ty RW ee

YOU AD ph ene

Naya aioe, : SH A Sed ed 5 ie SDT eer her re eas

Katie net cet eT MNGE DLL Cate Gen OD alate Mag ladles eS WK Ly seONneretsy eri ‘et ; . DEVAL EE EY ERISA LY ONE INR ALOU ALY Mt MC or ERNE 5 ISSO NT gyi (WEN Ae ST

Se ne Te Shy trate te bete tia Gamat: SIAL rmsSl TH Teast Hive dae Hee

PE eee LAME IN Goplow ne Cat ad el ctielesioha hey tied ky obit Went LO eOt pe

SEY OS ote

EY Li aM titey oe fatlad emely PRN Hey AYO M iy Nad Aner aie gt fice

Ayan [lm ete Ph Datel Hat eA bahd atid a Na a SRT BE

es PLA DEIR Mee he UAE aU Bef Vatte ay in

wthals Cig ted ee Pen Wen Deana

UN OMAR Spey figs t. Ved VIP MY, ‘ Ly

Leis able t H I . NEVA Wadley 2%

SO STII a Ea ah ot i Zahn ml ilba Sims toe ae

sd vo eEN est ‘

De Trae ay none rates

Sere erento en Whine

ets Sor ouieR Whe beh al

‘ €

AVP 0 LOIN ADA ED I Moe

ES EMISES COE INS BOR Md DS ek a ee ee ed

Fave si

ab PDE eM Ne

Bye eh TL es a ESL SRI TEEPE ART MC oh MW NIEE G Sori t eg \letyy

ey ar eityd Vit Ware did Soot ot Ep ene c Sanus j ON NE ale eve oe CE RY Pel a het WEY Went dite @® ome Wat Mab UM FF Nw Ve te we

ra Pe Re atin eae MP et cash ds afte at th SPAN ee eR ee INCA ANGE CAEN EYS RIDER LAY Coty ONY PEE MW Ber TLE VVC TM LU Ry Sree

FO SBA Lyte Mona ate eae y ce NMS OW Sey OP Mee : ‘

Nerd Charen Tae)

te Wa Dagan

PEMNAN EG NN AY ag

AONE by OC nde By ond el tite

DL ed VI IN eet ew

VENETO OH tN ey cyt

fe tiny

freee

Z VE SINE Enh eres #

eo Ohay UN COTY

OES ENOL Sone ts

BAH rary Al mivigih i WWI

eS EE VIE IN SANE Tp stor Med ©

De Nge Sten taths

“elvan

are ons

NOt A III A a See pap et

vei

wwe

Sarre

AY INOS HPs eet ye

MANNS oye

WAS

AO Wee

sheet

; i rina ns

Meath VAN CoA heues ab Aw bt a Tt MVS, Sa arhy Ot a years SVN Ny Danby at 5 ir “

: : P # Ne Day fip AE abana vb) ye sey et

ay adil goo apaty Pee Re th ns Witte Le henr ra Te ee We Ae wren eaety a iy atti “ i val! F peta

Crean era) A0ey ru Sante ane? ae : Cy INSa Ke he mye H v7 id ee! ' ies ee ea Re Dan 0 ea LY AUD KE DADS wPom gga Warr A ASE MAE NP wees Wika

Sere ‘Bein garagyteenges Sawega omer ytgey funds devant daewas TT Sa fate as rane Merah eh Ce ieanscdagdane ben unytan x on-enaunadarmsiotonrniNoegem vamos RiRass ek Lo RTE

ee te dg tee Saw tye Mi) oy ww wea UNA n ' . vv yet eae PLB bia RN ee Arn EARN OWI Darniey AbD toe oaths wing vey Malte why a bgp,

Bact Borat al el Tote HN D Sob MEW a wid Cun liWe BU gia otET YEMEN BM OEE EAD AES LDN INT IO NV WV a Peers Hote MAY od lik LST HNDM UDB GMANG, Meaty VEL ot Lumen te Fe 0 toe thule Mey emi nh shai,

oe eg Cat ONY bn ee Ce ee LOR eet ee ed TT OT me Ride Dd SCC Oe) De Te ek i

LECH LNCS CP eat OS dN AP UPR ae SMR MERE ET IN OYE SIR OI Me Lab Y® Vet he ate

OPE NON eh Lae gh tt TO Tae

: PEN ad eS BEES EN IE Ay te UE eh ND ONE TY oe DT SANUS #

Wet ee TINA Cutt yt Ot UI EN ORY Md et eth ete g tp eee

SEER gm re ey i

atthe (mids AM aan

CY ONENESS

ws nahi

NYatiate wiht w” alsbag nyt n

HATA NI BUS W WLW IG ftw, fo Nh Ha

: Q DE DNC eae net ome

ee et ivy Ae ANY Ct Coe eum ye ne”

SLE APRN AV EES, Sea V ANNES LW Lee Ty MANE NENT TCT IN Gated Sa ELON,

ry .

Nee)

testy

Wad at rode bey

AWN EE pep hy EN Re

Np be Fina he

‘ Voorn sh gf pV at

DN RORY YN A

VOR Sua Re ae

NOLES OD oe by A Oh et

‘ vot CSAWHIEL OF Mey renee WS se le 7 eet ra Father

WOU PE Fes v ort NP ANP IN AE eA WINN DDI IMA yd dae CFR AMD Sy OAS Sd OL ee OO ee Pfs ey va fa EMI traiaAbig halve hayeg i Vai NBs altro dap Bete h wtp

ia Gunien tens Pe WEEN NEMA Style AF ICOM EM WH AA OR SIPLE THY vg eed VL TU VE MALONE E ENE NEMS WIM PET aval eVatorrbe ed coreg oes RHE QUAD TAS VRERYS Y errant SN verninnl owner leit red™ see

hen CMe CANE Mee tig tee eS Woon Teen ha eee ee eT ty AON eee 2 RPC LF NE dE OA SANCTUS Vd De Aes EE NCHS Bee iROe Bye Huey rrteye AVDA OTA Medica gg tite Le ade

PON ER EAE " SAMO Ag 14 P89 EPIL TD Pomp hare Abe DVINOP HEN ARs g Wy a Rap ee Arr eres Lael eS y rary Ye NPY SU aS 2 oy my

Veer eouy

ky te vey ie eny

Aaa ee eee ne he re a

ren ces eRe

rrebey

wha

; Awha ths

COMM TENA TSAO NS diy tia ee BAYA 7

we VEINS

Ine

POLL ALVES INCOR E RES Mags ey

Hoyer ieeaye

ANP NWT Ld

’ \t > N , : oy ce FONSI Wet Ye SSN So arene ry

PA ated Cera ak a see SUP ieee a et eee ey Oro Uy 1H HN dog Me SAMAR Woe fg ie Low eave y

C eee Tristat gt er, a

a ey Coe ea ern avai otiord sty SOLU esate NE ENE Lo ANS A MEN AV ERIN EN HANAN Mow Vey INS FYE as ach hi Si er ee

FEU Ye SY wl oy CRT Coen e a AW TIN NEVE LIM DN Sy ay ha AON Uae ee aes RUSE EAT en] Oe ee hee ee Or APAS ST en it

ANS Bete ee, Srenrait ie t Tee mev IN Oper ed Ve pee ates ee elena PET Vy ety eee Mhee NUNC OB Voie

ometig het yy Mav tee ae SPARE Wat pet Cee thel Oe Ft nen

one

CONT he we

i WW Mp a9

Seouhers

betta dtl

warn ree ee ee OT een 1 ONES HP ON t, y

aces Mesut wikigg C25 meee

We eR hat Dwth er aes

rate »

‘

SET Vb Te Na ely ee

WOM WAT Ny bal bia SY wi MPa! eaten bys aw She oer beat y

: “re ERC feel pa bp oll ner AAS ergs 2 oS b yan ANY SPEYER ET PON be SORE ms Paper tencpenbis PEAS Mey PORTIA ve nba ESI br manatee bat Ll iwihat RmAa ete

Panett anne wie tn dra Htafin Wiss Veute¥ ny ove2 Padi Am eV MSU NET Ob ASOSMIVN LIP Ales Pend f WES VERDE Nay he we A ered bitin! WLR EY Ng ns Mie Seated va EN TAY LDA Ud Mill ag Ynvtses 1a Yang Te Dg MaPS gala Ny

ty" WENO tee ve arerny Fete at Wt Ye ee WA fy eee eed ede Huten in “yu i, * h; ‘ of AVIS Mt LVDS ET ot HOPING og pe ehe bs pang MEME ATCT FN SC MALY Ah thd eh al, \ %

1a Meese he! erat rn ORO LMR acon LADEN mehr MPC MENE LIED TS ed AEM EERE AN SINAN hy en ye ke FREAD OMAN

le ket

Me EN te nial nbn atte thy og tot

Vy Tenge Ne wtiigita ley

CV LAE OPH TV ND AD ate SEPLEN Gao WEN OY? Uwe

ware wey

rvae eee

DIAC YN ANY ete

Park Sen Ae RPT (es

INYS OSG edo edies

Se eer

ee at We ee edie dy

rer at

owt

Pena Pratik rene

ene

fatty

ayn,

WS ed LA OF ee heal 5a VAD IY ANB VB hy

CPI ls hae PINON DV de ow Oe ne

belay

. b Ay PD vow Very WN em a Bo FM Le ti

sl her CNP ONERENS SIOPY Ree ATEN AICI PMN rey ANTE BE ba ATE Meat e¥DOLIODM GINS Wa tincar

em tienes NSF nay Nba Sone ee Ha tiety PRM TMA NE Oot ety GENE eed D PN VOY MUSE Oe Ny Dew es " “ . 3

‘ eu Feb FANN 1 1t sa ’ . rat Pon

WSa eyalcruer Chigse Ms ORME Dt Le ttent ye Pat ed ded DN . ' fp IN eigen e lh wh si Seeenraalvat aging

WL Me Mette MN Pe Tete EUS ay ASN ONDA Yells pues De ¢ a ee 2 PON hy vad ep U WM vee

is a al err Te it Ase ty aw ae ar SORT See he rie NOt ea bed Ry ais Ni) vie) “7.4ih

eet. vi a f my NDS va NY ak

eI an yer, vente Th oe eet Vit bs SEER Rear y i beable ei se Ty

an fe Pow “

WAN rime it

NW TRAMs bd very

Pra Lee Ek ee

sytem hittin NNN ee

UV ASING sy Me bye

PEN bay Se Sete

PhO

Peony

ero

Yom eRe HE’

(My Mas ptt Pov Dag 9%

pom entnn

ADM ako be Tree

94 antares ey hid

MTA et re)

PENT Wee

Ae ANF aA add ew vogt

yore

IETS 6 LOTRE LOU ge

ieee ae

ee venents

Semen

Feit wens

STIL tty

eed el DMG yes

wie

tafe

COeeny

HE HEM NO aS,

ea Ve bes

ey sinave

> ving!

Ne at aL

ta AALS de eee

a mad BEE

ets open

oy caret

nalts

ee Ree Hh ee

weedy

Ontos

Sr ond

my.

a ys

Wa tp Nath

Se bttats

Woe wow tye

r ; . ; ‘ Sa ee VB beg

yeururn onew “ othe he SIR ot aiuerage

a ed

Wn Peete DS ee

Meaty wre ie"

Ce ee

Fp g Nt

AE (SM arto abipatte vals

ENS why owy

TIEN INOS

vee eee eh ae

Seg ate ar

7 tye ded : ;

Celtet pe a ee oan

ny Nested heehee . ys etetyy

aie Ra iog’ he

Te thant as

Feeney ovy oy

beg ts ndied el Neebbaly whe nth N vlgy!

vy HUN mia be

8 eM elige ioe al

any amet

te owas

WSS Mgr d ut

Pe ihe Sane we SEN

hy tyeany

mete

aN ee erin ne

Ve INLAY

Son atiee!

a Neate

Veit

Pe id

Were diy

eS et

dae

MMeace Minette were

us oy waren a ae

Fae rea vaca ab Abamun tae

ei ette

Ce...

as) ft 7

WN + ; VOLUME 81

Number 1

Jour nal of the aes , March, 1991

WASHINGTON

ACADEMY... SCIENCES

ISSN 0043-0439

Issued Quarterly

at Washington, D.C.

Nov 021998

LIBRARIES

CONTENTS

Articles:

MAXWELL H. MILLER, JEFFREY L. HARPSTER, & JAMES H.

HOWARD, JR., “An Artificial Neural-Network Simulation of Auditory

Intensity Perception and Profile Analysis”

i ey

WILLIAM B. TAYLOR & FREDERICK J. EDESKUTY, “Evaluation of St.

Lucia’s Geothermal Resource”

“CORRIGENDUM”

JOHN J. O'HARE, “Perceptual Integration”

ee ie s .ehe! 070 lene 6) 0s 00] eee © ).0)\0 ©. 0,0 06 «ase. e|

Washington Academy of Sciences

Founded in 1898

EXECUTIVE COMMITTEE

President

Armand B. Weiss

President-Elect

Walter E. Boek

Secretary

F. K. Mostofi

Treasurer

Norman Doctor

Past President

Robert H. McCracken

Vice President, Membership Affairs

Marie Bourgeois

Vice President, Administrative Affairs

Grover C. Sherlin

Vice President, Junior Academy Affairs

Marylin F. Krupsaw

Vice President, Affiliate Affairs

Edith L. R. Corliss

Board of Managers

~R. Clifton Bailey

Jean K. Boek ©

James W. Harr

Betty Jane Long

Thomas N. Pyke

T. Dale Stewart

REPRESENTATIVES FROM

AFFILIATED SOCIETIES

Delegates are listed on inside rear cover

of each Journal.

ACADEMY OFFICE

1101 N. Highland Street

Arlington, VA 22201

Phone: (703) 527-4800

EDITORIAL BOARD

Editor:

John J. O'Hare, CAE-Link Corpora-

tion

Associate Editors:

Albert G. Gluckman, University of

Maryland

Marc Rothenberg, Smithsonian Insti-

tution

Marc M. Sebrechts, Catholic Univer-

sity of America

Edward J. Wegman, George Mason

University

The Journal

This journal, the official organ of the Washing-

ton Academy of Sciences, publishes original

scientific research, critical reviews, historical

articles, proceedings of scholarly meetings of

its affiliated societies, reports of the Academy,

and other items of interest to Academy

members. The Journal appears four times a

year (March, June, September, and De- |

cember). The December issue contains a di-

rectory of the current membership of the

Academy.

Subscription Rates

Members, fellows, and life members in good

standing receive the Journal without charge.

Subscriptions are available on a calendar year

basis, payable in advance. Payment must be

made in U.S. currency at the following rates:

Claims for Missing Issues

Claims will not be allowed if received more ~

than 60 days after the day of mailing plus time

normally required for postal delivery and

claim. No claims will be allowed because of

failure to notify the Academy of a change of

address.

Notification of Change of Address

Address changes should be sent promptly to

the Academy office. Such notifications should

show both old and new addresses and zip-code

numbers, where applicable.

Published quarterly in March, June, September, and December of each year by the

Washington Academy of Sciences, 1101 N. Highland Street, Arlington, VA 22201.

Second-class postage paid at Arlington, VA, and additional mailing offices.

Journal of the Washington Academy of Sciences,

Volume 81, Number 1, Pages 1-21, March 1991

An Artificial Neural-Network Simulation

of Auditory Intensity Perception and

Profile Analysis

Maxwell H. Miller, Jeffrey L. Harpster, & James H. Howard, Jr.

Department of Psychology, The Catholic University of America

ABSTRACT

This paper describes a computer simulation of human auditory intensity discrimination.

There are currently two different views of how intensity discrimination is carried out by

human listeners. The traditional view holds that a listener successively compares the acoustic

energy between two sounds and selects the louder of the two. A more recent view, called

profile analysis, suggests that a listener simultaneously compares the spectral profile within

each sound individually. In other words, the timbre or the perceived spectral shape of the

sound is also considered. The computer simulations replicated a study by Green, Kidd, &

Picardi (1983) in which the interval between the sounds was varied. Results from the simula-

tions are consistent with results obtained from human data.

Introduction

In the past five years there has been a growing interest in a class of adaptive

computer models known as artificial neural networks. The discipline dedicated

to the study of neural networks is a multidisciplinary field which involves re-

searchers from diverse backgrounds including computer science, neuroscience,

cognitive science and psychology. The application of neural networks is

- uniquely defined within each discipline. For example, computer scientists are

interested in applying neural networks as a technology to solve difficult engineer-

ing problems encountered in robotics, computer-assisted pattern recognition

and artificial intelligence. Neuroscientists are principally interested in building

computational models of neurophysiological systems, and artificial neural net-

works provide them with a useful tool for modeling.

Cognitive science and psychology have embraced this technology as a theoreti-

cal framework to help explain, understand, and predict human performance.

Symbolic models such as expert systems have enjoyed much success when ap-

2 MILLER, HARPSTER, AND HOWARD

plied to rule-based decision making problems. However, these same symbolic

models have failed abysmally when confronted with perceptual pattern recogni-

tion and classification problems—the very problems at which artificial neural

networks excel. The neural network paradigm offers a powerful alternative to

the traditional symbolic models so widely applied in the artificial intelligence

and cognitive psychology disciplines.

A defining characteristic of artificial neural networks is their parallel architec-

ture. Unlike conventional digital computers, which have a single central process-

ing unit executing instructions sequentially, artificial neural network models

have many simple processors acting together concurrently. Each processing

element is capable of computing only a few very simple logical or algebraic

operations. The real strength of neural network models comes from the interac-

tion of large assemblies of processing elements acting together in parallel.

Processing elements are arranged in layers. The connections between process-

ing elements are assigned numerical values which are referred to as weights. The

architecture of a neural network is tailored to each individual application or

problem. The simplest neural network architecture consists of two layers of

processing units. More complex architectures support three or more layers. Due

to the fact that computers with true parallel architectures—sometimes referred

to as neurocomputers—are not yet widely available, neural network simulations

of parallel architectures are carried out on digital computers with conventional

von Neumann architectures.

Another important distinction between the processing in neural networks as

compared with conventional computing is the way that informatiom is repre-

sented. In neural networks, knowledge is not stored locally or associated with an

address in memory. Rather, concepts are represented implicitly by a pattern of

activation over a large number of processing units. Information, or knowledge is

encoded in the connections themselves, and not in the processing elements

which serve only as computing devices. This results in a system which is very

fault tolerant, and degrades gracefully if a subset of processing units fails or if the

input data become corrupted or degraded by noise.

Perhaps the most remarkable aspect of neural networks is their ability to learn

by example, without being programmed by a human. Through the application

of an iterative training algorithm, a neural network can learn to associate one set

of patterns with another set of patterns. In essence, a function is computed by

the network which maps a set of input patterns onto an output pattern. The

connection weights of a neural network implement this mapping. The proper set

of weights to compute this function is usually never known a priori. So various

statistical training algorithms have been developed to adjust the weights.

Perceptual psychologists have become interested in experimenting with

ARTIFICIAL NEURAL NETWORK | 3

neural networks for several reasons. First, due to the fact that the parallel archi-

tectures of neural networks lend themselves to perceptual pattern recognition

problems, they provide psychologists with a superior computational model of

the underlying perceptual processes. Second, parallel models not only provide a

better theory of the processes, their architecture is based on gross biological

principles understood about the brain. And third, neural networks can learn by

being trained, can adapt to changing input parameters, and can generalize to

novel patterns never seen before. This is one of the cornerstones of human

intelligence and performance, and has been neglected in the past with the statis-

tical models so pervasive in the psychophysical literature. In sum, neural net-

works are able to demonstrate the type of flexible and adaptive performance

which conventional symbolic models lack, while also accounting for perceptual

learning (Clark, 1989).

In the present research program an artificial neural network was applied to a

classic problem studied by perceptual psychologists for over one hundred years;

How can we explain the ability of a listener to discriminate a difference in

intensity between two simple sounds? In a typical intensity-discrimination ex-

periment, two sounds are presented successively, separated by a brief intersti-

mulus interval. Both sounds contain a pure tone (known as the standard) and

one contains the standard with a small increment (known as the signal). The

observer is forced to decide which sound pattern contains the standard plus

signal. The traditional explanation given to account for this phenomenon states

that a listener performs the task by choosing the tone with the greater acoustic

energy.

In contrast to the traditional view is a more current view of intensity discrimi-

nation and signal detection which David Green and his colleagues have referred

to as auditory profile analysis (Green, 1988). The stimuli used in profile analysis

experiments are complex broadband sounds, in contrast to the pure-tones typ1-

_ cally used in psychoacoustic experiments. In a profile analysis task, the sound

with the signal component may have less overall energy than the standard;

hence, intensity discrimination will not work. The judgment of whether a signal

is present or absent must be made by considering the internal shape of the

spectrum as opposed to comparing the differences in energy between sounds.

Despite the fact that this phenomenon has been studied so extensively, there is

still not a comprehensive theory of intensity perception and spectral shape dis-

crimination which can account for the entire body of empirical data. This pro-

vided the motivation for performing the computer simulations in the present

study.

In this study an artificial neural network was trained in both an intensity

discrimination and profile analysis task. The independent variable of interest

- MILLER, HARPSTER, AND HOWARD

was the duration of the interstimulus interval (ISI) between the presentation of

each sound pattern in a two-alternative forced choice task. The computer simu-

lations carried out here replicated a study originally done by Green, Kidd, &

Picardi (1983). Four separate computer simulations were performed, each of

which corresponded to one of the four experimental conditions in the original

Green et al. study.

Artificial Neural Networks

The early roots of artificial neural network models were as simple pattern

associators. The first neural networks widely discussed in the scientific literature

were known as “‘Perceptrons,” a name coined by Frank Rosenblatt in the late

1950’s. Perceptron models were often applied to pattern-recognition and classifi-

cation problems. Through training with an iterative learning algorithm, percep-

trons could learn to associate a set of input patterns with an output pattern

which coded category membership.

Initially, perceptrons held much promise for modeling perceptual pattern

recognition processes. Rosenblatt proved an important mathematical theorem,

known as the perceptron convergence theorem (Rosenblatt, 1962). This

theorem guaranteed that if a set of input patterns are learnable by a perceptron,

then this learning procedure would converge on a-set of connection weights

which enable a perceptron to adequately represent the problem. This proof was

an important contribution to both the machine learning field in engineering and

learning theory in psychology.

However, as researchers began to experiment with perceptrons it soon be-

came apparent that there was a class of problems that a simple linear perceptron

could not solve. A severe limitation placed on the perceptron by the learning

algorithm enabled it to adjust only one layer of adaptive weights. As a result, the

perceptron could only solve problems which were linearly separable, meaning

that a perceptron could only perform a linear mapping between the set of input

patterns and output or “target” pattern. This limitation was formally stated in a

book called “‘Perceptrons” written by Marvin Minsky and Seymour Papert

(1969). The classic example given in ““Perceptrons”’ was the exclusive-or (XOR)

logic operation, which a perceptron could not compute. As a result of Minsky

and Papert’s critical evaluation of the perceptron model, coupled with its inher-

ent limitations in the types of problems which could be represented, research

with neural networks died out in the late 1960’s.

In the past five years there has been a resurgence of interest in adaptive neural

network models. This is due to the fact that a powerful new learning algorithm

has been discovered which allows neural networks composed of multiple layers

of adaptive weights to be trained (Rumelhart, Hinton, & McClelland, 1986).

ARTIFICIAL NEURAL NETWORK 5

Multi-layer networks with three or more layers incorporate a middle layer of

‘hidden’ units between the input and output layers. By adding an additional

hidden layer, the neural network can recode the input patterns into a higher-

order internal representation. The multi-layer networks have been able to solve

many of the problems that a simple linear perceptron was unable to solve previ-

ously, such as the XOR problem.

Network architecture. The architecture of the neural networks used in the

present simulations is illustrated in Figure |. A two-layer fully-interconnected

feed-forward artificial neural network was used. There are two layers of process-

ing elements, an input and an output. All input units are fully-connected to the

output unit. There is no feedback from the output unit back to the inputs;

activation can only flow forward.

Only two layers of processing elements were used in the simulations discussed

in this study, as a multi-layer network with hidden units was not needed to

perform the task. More complex architectures often support recurrent connec-

tions where activation is passed-backwards from the output layer to the input

layer, or laterally to other units within a layer. Connections between non-adja-

cent layers are also possible. The connection weights themselves are stored in a

weight matrix. There is an additional weight associated with each output unit

called a bias. The bias value may be thought of as threshold term that influences

the amount of input needed to elicit a response by that particular unit. The bias

values are stored in the weight matrix as well.

Network output function. The transfer function for each neuron is known as

an activation function. The activation function defines an input to output rela-

tionship for a processing element by establishing an output value for a given

input value. The output for any unit O; is a non-linear function of the weighted

sum of its inputs plus a threshold or bias value

where W,; is the strength of the connection between units i and j and B,; is the bias

value or threshold for unit j. S(x) is a nonlinear squashing function which re-

maps the sum of inputs into the range 0.0 to 1.0. In the present research a

sigmoidal squashing function was used

S(x) = 1/(1 + exp(—x)).

This function “squashes” positive values into the range 0.5-1.0 and negative

values into the range 0.0-0.5. As a result of the squashing function, the output

elicited from a processing unit will not be at its maximum unless it receives a net

positive input greater than its bias value.

6 MILLER, HARPSTER, AND HOWARD

Discrimination Response (0.01 =First Sound, 0.99 = Second Sound)

1 Output Node

® O 2 Input Nodes

First Sound Second Sound

a. Schematic ofa 2-1 feed-forward architecture used to

detect the presence of a signal increment to a sinusoid.

Discrimination Response (0.01 = First Sound, 0.99 = Second Sound)

1 Output Node

22 Input Nodes

First Sound Second Sound

b. Schematic of a 22-1 feed-forward architecture used to

detect the presence of a signal ina complex sound.

Fig. 1. Schematic of 2-1 & 22-1 neural-network architectures trained with the back-propagation learning

algorithm.

ARTIFICIAL NEURAL NETWORK 7

The learning algorithm. The learning algorithm used to adjust the connec-

tion weights of the network during training is the popular back-propagation

algorithm. Back-propagation is a training technique which allows multi-layer

networks to establish an optimal mapping between input and output units

(Rumelhart, Hinton, and McClelland, 1986).

Training involves two steps. At the outset all of the connection weights are

initialized to small random values. In the first step, input patterns are applied to

the network. The input vector is fed forward through the network and an output

value is computed. This output value (or observed value) is then compared with

a target value, which is the desired output value. If there is a discrepancy be-

tween the observed value and the target value then an error signal is generated.

The second step in back-propagation involves a backward pass through the

network. The error signal is propagated backwards through the network, and

each of the weights between the output unit and the input units are adjusted by

an amount proportional to the error term. A similar adjustment is made to the

bias term.

During the testing phase, the connection weights are fixed and cannot be

modified. Testing patterns are applied as input and the output or response of the

network is measured. In the case of the simulations discussed in this study, the

network was only required to perform two-category classification. The task

required of the network was to select the interval which contained the signal

increment.

Stimulus Conditions

Two methods were used to select the level of the background or masker

components of the stimulus patterns. In each case, the levels of the maskers were

sampled randomly from a uniform distribution. However, the critical distinc-

tion is whether the level of the maskers remains fixed across trials, or is varied

randomly within trials.

The first method used a between-trial variation. In this case the amplitudes of

the masker components in the first and second sound-intervals are equivalent.

The second method used a within-trial variation. The levels of the maskers in a

stimulus pair are chosen independently of one another, and the levels of the

maskers for the first and second patterns usually differ.

Three different sets of stimulus patterns were generated for the between-trial

variation method. They are referred to as the single-sinusoid condition, the

uniform-spectrum condition, and the multi-component condition. In the sin-

gle-sinusoid condition (Figure 2a), there is only one frequency component, the

signal increment was always added to this. In the uniform-spectrum condition,

(Figure 2b) there are multiple frequency components. The signal in this instance

8 MILLER, HARPSTER, AND HOWARD

Frequency

a. Single-Sinusoid Condition

Spectral Magnitude

Frequency

Spectral Magnitude

b. Uniform-Spectrum Condition

?dmM— M+S

<q M

Background Alone (Standard) Background plus Signal

Fig. 2. Schematic line spectra of test sounds for the single-sinusoid and uniform-spectrum condi-

tions. S = Signal and M = Masker. ;

is always added to each of the background components in the tonal complex. In

both of the above conditions the overall shape of the stimulus remains the same

across intervals of a trial.

In the multi-component condition the signal increment 1s always added to the

center component of a multitonal complex (Figure 3a). The amplitudes of the

maskers in both sounds are identical. However, a critical difference between this

condition and the conditions in Figure 2 is that information regarding the spec-

tral shape of the stimulus is also available to the listener. The signal can now be

depicted as a bump occurring in the center of a symmetric tonal complex.

A fourth set of sounds was generated using the within-trial variation proce-

dure. This will be referred to as the profile-analysis condition (Figure 3b). As in

the multi-component condition, the profile stimuli are composed of multiple

frequency components which provide information on the spectral shape of the

stimulus. However, unlike the above conditions, the levels of the maskers are

ARTIFICIAL NEURAL NETWORK 9

a. Multi:Component Condition

» tb)

= =

i= = nearer en ie

= > | € dageeecen Fi cha

£ fe

B 3

a Frequency wn Frequency

b. Profile-Analysis Condition

M+S

Case 1 ¢ M

<q M

M+S

Case 2 ‘4

Background Alone (Standard) Background plus Signal

Fig. 3. Schematic line spectra of test sounds for the multi-component and profile conditions. In the profile

analysis condition the masking level will be of less (Case 1) or greater (Case 2) value than the amplitude of the

background plus signal. S = Signal and M & M’ = Masker Components.

selected randomly within trials (as well as between trials). Case 1 describes the

~ task where the amplitude of the standard is less than the background-plus-signal

whereas in Case 2 the amplitude of the standard is greater than the background-

plus-signal. To select the interval with the signal component, amplitude cues are

no longer sufficient and the shape or profile of the stimulus must be analyzed.

Traditional View

The traditional view of intensity perception has evolved from both empirical

studies as well as classical psychoacoustic theory. This view of auditory process-

ing assumes three serial stages (Green, 1988). In the first stage, auditory signals

are passed through a bank of linear, critical-band filters, each of which is tuned

to pass a limited band of acoustic energy. Each auditory filter is assumed to have

10 MILLER, HARPSTER, AND HOWARD

a critical bandwidth. Energy which falls within the boundaries of a critical band

is passed and energy falling outside of a critical band is rejected.

In the next stage of processing a single channel is selected, and an energy

detector measures the acoustic energy present within the band at the signal

frequency in each sound interval. A sensory representation of each stimulus is

encoded and stored in a rapidly decaying short term memory. Durlach & Braida

(1969) have referred to this form of representation as “‘sensory encoding.” Sen-

sory decay results from the tendency of the memory trace to become corrupted

with internal neural noise and deteriorate with time. In the final stage of process-

ing, a successive comparison is performed between the two sensory traces in

order to report whether they are same or different, or one is louder or softer.

A shortcoming of the traditional explanation of intensity discrimination is

that it may tell us more about the psychoacoustic task itself, rather than about

the underlying psychological processes. In an intensity discrimination task, a

listener need only select the sound with the greater overall energy. This perspec-

tive is consistent with the literature on critical band theory. Only energy which

falls inside of a critical band filter is attended to, and energy falling outside of the

critical band is ignored.

While research into pure-tone intensity discrimination has been useful in

developing an understanding of basic auditory processes, the stimuli used in

those experiments are highly synthetic and arbitrary in their structure. Most

importantly, these types of sounds are rarely encountered in the real world, and

this has placed a severe constraint on the development of psychoacoustic theory.

The work on auditory profile analysis has begun to address this weakness, and

has put in place a foundation for development of auditory models which have

explanatory power that can capture the way that the auditory system deals with

complex signals encountered in the real world.

Profile Analysis View

The profile analysis view assumes the same psychophysical signal processing

stages as the traditional view. However, in a profile analysis task both the listen-

ing strategies and memory processes differ from those used in pure-tone inten-

sity discrimination. Profile analysis emphasizes a simultaneous or broadband

comparison process within a sound across frequency channels rather than a

successive comparison between critical bands. Green has referred to this broad-

band spectral analysis as a global comparison process (Green, 1983).

Some of the most dramatic evidence of broadband listening comes from

experiments which varied the bandwidth of the stimulus out to regions beyond

one critical band. It has been demonstrated that there 1s a consistent improve-

ment in the listener’s performance in signal detection tasks as the bandwidth

ARTIFICIAL NEURAL NETWORK 11

and component density of the tonal complex is increased in remote critical

bands (Bernstein & Green, 1987). These results have also been simulated in

experiments using neural networks carried out in our laboratory (Howard,

Harpster, & Miller, 1989).

Perhaps most exciting is new evidence reported by Berg and Green. Berg has

developed a technique for estimating the subjective weights that a listener as-

signs to each component of a tonal complex (Berg, 1989; Berg & Green, 1990). A

full discussion of this research is beyond the scope of the present report, how-

ever, two relevant issues will be highlighted.

First, Berg’s findings have supported the empirical data gathered in previous

profile analysis experiments, and has quantitatively shown that listeners do

indeed assign relevance or salience weights to remote frequency regions of a

_ broadband sound. This work seems particularly important and builds on the

earlier work done by Gilkey and his associates (Gilkey, 1987). Second, there

seem to be some interesting parallels between the recent work of Berg and his

concept of human weights, and the weights derived from artificial neural net-

work models.

A further distinction between profile analysis and pure-tone intensity discrim-

ination is the memory process involved in the two types of tasks. Unlike the

sensory representation of the stimulus believed to come into play with pure-tone

intensity discrimination, the encoding of the stimulus in profile analysis is hy-

pothesized to be a higher-order symbolic representation. This qualitative encod-

ing of the stimulus has been characterized by Green as “signal like” or “not

signal like” and is less susceptible to sensory decay over time.

Durlach & Braida (1969) have referred to this type of internal representation

as “context coding.’ The sensory percept is transformed into a symbolic or

categorical representation of sorts. The context-coding mode of memory sug-

gests that a higher-level encoding of the stimulus occurs which is more robust,

and is relatively immune to the effects of the interstimulus interval between the

pair of auditory patterns presented to the listener. As Durlach and Braida have

pointed out, ““whereas in the trace mode the effects of the noise are dynamic and

change with time, in the context mode the effects are independent of time”’

(Durlach & Braida, 1969; p. 374). This will be discussed in more detail in the

following section.

To summarize, the optimal listening strategy in a psychoacoustic experiment

will vary as a function of the demands of a given listening task. In profile analysis

experiments the sound with the signal increment may have less overall energy

than the comparison sound; hence a listening strategy which relies on comput-

ing the overall level of energy to detect the signal wiil be ineffective. The listener

in a profile analysis task is forced to listen for a qualitative change in the shape or

12 MILLER, HARPSTER, AND HOWARD

* Single-Sinusiod

® Multi-Component

CJ Profile-Analysis

LEGEND

SIGNAL LEYEL! BACKGROUND LEYEL (DB)

INTERSTIMULUS INTERVAL (SEC)

Fig. 4. Auditory discrimination under four conditions (Green et al., 1983). Value plotted along the ordinate

is the signal level/background level required 1n order to discriminate 70.7% of the patterns correctly.

profile of the stimulus, as opposed to a quantitative change in the acoustic

energy per se. In theory, well-trained listeners in a profile analysis task could

listen to one interval and base their judgments on the presence or absence of a

signal solely on this one sound. Unlike the traditional view of pure-toné inten-

sity discrimination, which strictly emphasizes a narrowband energy detection

mechanism, profile analysis emphasizes a broadband listening strategy which

involves a simultaneous or global comparison of the frequency components

within a single sound.

The Role of The Interstimulus Interval (IST)

Figure 4 illustrates the results from the Green et al. study (1983) on the effect

of the ISI in an auditory discrimination task. ISI is plotted against the signal-to-

masker ratio required for a listener to perform at a 70.7% level of correct detec-

tion performance. The neural network simulations discussed below replicate the

experiments discussed in the above study. Four simulations were conducted;

two where only level cues were available, and two others where spectral shape

information was also available. .

The two upper curves of Figure 4 represent the conditions where no profile

information is available to the listener. These were referred to earlier as the

single-sinusoid and the uniform-spectrum conditions. In these conditions, as

ARTIFICIAL NEURAL NETWORK | 13

the duration of the ISI was increased, listener performance on the task declined

(as revealed by the higher signal levels that were needed to maintain the same

level of correct performance). As the sensory traces of the signals in memory

begin to decay, successive comparison of the traces becomes increasingly diff-

cult.

The two lower curves represent the conditions where information on the

spectral shape of the stimuli was available. These were referred to as the multi-

component and the profile-analysis conditions. If the encoding of the stimuli is

symbolic or qualitative as is proposed in a profile analysis task, then the duration

of the ISI should have relatively little effect on listener performance, and indeed

this is the case.

Of particular interest is the multi-component condition. In this situation the

listener could have used a narrow band listening strategy to attend to the part of

the spectrum where the signal is added and essentially listened for differences in

the overall energy of the sounds. Or the listener could have used a broadband

strategy and attended to other background components in the spectrum to

analyze the shape of the sound. In this situation, it appears as if the latter strategy

was chosen. As Green points out, “it means that for sounds of changing absolute

level the ability to make simultaneous comparisons leads to better intensity

discrimination than does successive comparison” (Green et al., 1983; p. 641).

Method

Four neural network simulations were carried out, each one replicating an

experiment from the original Green et al. study. There were two stages involved

in each simulation. In the first stage, each neural network was trained to perform

a psychoacoustic task to a specified criterion in common use for this type of task.

After each network had learned to perform the task, the connection weights

were fixed so that they could not be modified during the testing phase. In the

second stage, each network was tested with an adaptive psychophysical proce-

dure (Levitt, 1971). The measurement used to judge performance was the ratio

of the signal level relative to the masker level (SMR). Good performance is

indicated by a low SMR (low threshold), where poorer performance is indicated

by a higher SMR (higher threshold).

Stimulus encodings. All stimulus patterns presented to the network for

training and testing were encoded as line spectra as illustrated in Figures | and 2.

The stimuli were generated digitally, and the values encoded in each input

vector corresponded to acoustic pressure in the psychoacoustic experiment we

are modeling. A unique set of sound patterns was created for each of the four

14 MILLER, HARPSTER, AND HOWARD

network simulations, and 100 patterns were used per condition for training.

Different sets of patterns were used for the training and testing phases.

In all of the conditions, except for the single-sinusoid condition, the stimuli

consisted of 11 frequency components equally spaced from one another. The

signal component was always an increment to the center frequency of the com-

plex—except in the uniform-spectrum condition where all components were

incremented equally. In the single-sinusoid condition the signal always con-

sisted of an increment to a single frequency/masker component.

Auditory discrimination networks. The neural network used in each simula-

tion was a two-layer, fully-interconnected, feed-forward artificial neural net-

work trained with a back-propagation algorithm. There were twenty-two input

units’ and one non-linear output unit as shown in Figure 1b. The values of the

stimulus patterns encoded in the input vector are clamped to the nodes in the

input layer. The output unit is non-linear, and its transfer function corresponds

to a sigmoid.

Each sound pattern was encoded as an input vector. Each input vector con-

sisted of 23 elements,* encoding the stimuli from the first and second intervals,

plus a single target value. The first 11 values in each vector encoded the first

sound, and the second 11 values encoded the second sound. The target value

was either 0.01 or 0.99. A value of 0.01 indicated that the signal occurred in the

first interval, and a value of 0.99 indicated that the signal occurred in the second

interval. All input vectors were sampled from the full set of vectors, and ran-

domly presented to the network during training.

Decay function. Since the neural networks used in this study have no ex-

plicit means for representing time, the decay associated with increasing ISI

durations was simulated as follows. Sensory decay was modeled by differentially

incrementing the values of the weights taken from the networks which were

trained and had converged on a solution. Each weight from a particular network

was divided by the same value or constant. The function relating ISI to the value

of the divisor is shown in Figure 5. The smaller the value of the divisor, the

greater the simulated ISI. The rationale for this approach is the belief that sen-

sory decay relates to some weakening or decrement of a sensory response to a

physical event.

During the testing phase, the asymptotic weight matrices were tested at 6

different divisor constant or scaling levels ranging from 0.89 to 0.60. Twenty

' Except for the single-sinusoid network where there were only two input units.

* Again, the single-sinusoid network was an exception. In this case each input vector consisted of three

elements, the value of the two frequency components, plus a target value.

ARTIFICIAL NEURAL NETWORK 15

SCALING YALUE

INTERSTIMULUS INTERVAL (SEC)

Fig. 5. Decay function relating ISI to a scaling constant.

psychophysical adaptation runs were performed at each constant level and an

average of the twenty values was taken.

Results and Discussion

The psychophysical adaptation curves from each of the four computer simula-

tions are plotted in Figure 6. In each of the four computer simulations, the

networks learned to perform the task without difficulty. In general, the simula-

tion data fit the human data (Figure 4). This holds for both the shapes of the

curves as well as their absolute dB levels. The one exception to this is the multi-

component condition.

In the two conditions where a successive energy comparison is the only strat-

egy available to detect the signal (single-sinusoid and uniform-spectrum condi-

tions), discrimination performance declined as ISI increased. In contrast, for the

condition in which a simultaneous, broadband spectral comparison was re-

quired (profile-analysis condition), discrimination performance was relatively

constant, independent of ISI. In all three of the above conditions the network’s

performance mimics the performance of human listeners.

16 MILLER, HARPSTER, AND HOWARD

LEGEND

# Uniform Spectrum

@ Multi:-Component

0 Single Sinusiod

SIGNAL I MASKER LEYELS (dB)

| | 10

INTERSTIMULUS INTERVAL (SEC)

Fig. 6. Discrimination thresholds obtained from neural network simulations that varied ISI (scaling values)

for each of the four experimental conditions.

In the multi-component condition, the network’s performance diverged from

the human data. Unlike the profile-analysis condition, performance declined as

the ISI was increased. In this condition, either of two listening strategies are

effective for performing the task: a successive comparison across intervals, or a

simultaneous comparison within an interval. While it can be inferred from the

psychoacoustic data that the human uses the latter strategy, the neural network

model used the former. What can account for this difference? This will be

discussed 1n the next section.

Analysis of Asymptotic Weights

Next, an analysis of the asymptotic weight matrices was performed. By exam-

ining the pattern of weights from each network after it has been trained, insights

can be drawn into how each network performed the discrimination task. It was

observed that in all cases the strongest connection weights were from the input

units where the signal component was clamped. Conclusions about the atten-

tional aspects of the neural networks and the type of listening strategies which

were used will be discussed.

Single-sinusoid network. Figure 7 illustrates the weights for the network

trained in the single sinusoid condition. There are three values shown, one for

ARTIFICIAL NEURAL NETWORK | 17

<?\——_-_ -0 42 (Bias Term) Output Node

7 <4 166

€ O Input Nodes

Fig. 7. Asymptotic weights from a neural network trained in the single-sinusoid condition.

_ each input node, plus a bias term for the output node. The values of the weights

were: left input node = —15.7, right input node = 16.6, bias term = —0.42. The

opposite but equivalent magnitudes of the two input weights, with a bias value

close to 0.5, suggests two points. First, it’s clear that both units were used by the

network in performing the discrimination task. This is due to the non-zero

weights assigned to each node.’ Second, the bias value indicates that the network

was equally likely to respond to a signal either on the left or the right side of the

input vector.

Uniform-spectrum network. Figure 8 plots the magnitude of the weights for

the 22 input nodes for the uniform-spectrum condition. There are two impor-

tant observations. First, as indicated by the non-zero weights in both intervals,

the network used a successive comparison strategy to compute the difference in

energy between the sounds processed at separate intervals. This successive com-

parison strategy is consistent with human auditory discrimination data. Second,

the entire frequency spectrum was attended to by the network. In other words,

the network used a broad-band strategy to select its response, and attended to all

channels where energy was present.

The discrimination thresholds for human listeners are no better in the uni-

form-spectrum condition than in the sinusoidal condition. It appears that hu-

man listeners were using narrowband listening strategy. This may have occurred

for two reasons.

The first possibility is that by attending to only a narrow bandwidth of energy,

less attentional effort is required of the listener to perform the task. Clearly, any

single channel attended to in each sound is sufficient for making a decision. A

second possibility is due to internal noise which is correlated across channels in

> It is not important if the weights are negative values in order to assess the contribution of an individual unit

to the overall decision arrived at by the network. The absolute value of the weight is what counts.

18 MILLER, HARPSTER, AND HOWARD

2 s

156

ep le

ia bE

Ne gS -

_ =

Oo 0 Set ee ee ee

LU :

es ele

= f

S She

1234 56 7 8 910 11 121314 151617718 192021 22

INPUT NODES

Fig. 8. Asymptotic weights from a neural network trained in the uniform-spectrum condition.

the auditory periphery. The advantage of monitoring more than a single channel

is minimized, so the listener locks into a single band.

The network on the other hand is adjusting its weights in response to anything

which correlates with the response (1.e., the presence of the signal). Since each

masker component was incremented by a constant signal amount, all input

nodes contained information which correlated with the signal and provided

reliable data which are used by the network. This explains the 1.5 dB improve-

ment in performance seen (Figure 6) with the network trained in the uniform-

spectrum condition as contrasted with the networks which had a signal incre-

ment added to only a single component.

Profile-analysis network. Figure 9 plots the weights from the network

trained in the profile-analysis condition. As can be seen, the weight of the input

unit centered at the signal frequency is largest in magnitude, while the weights of

the nodes clamped to the masker components are relatively small non-zero

values. This indicates that the principal comparison which is occurring is within

each interval, where the signal component is being compared with the maskers.

As in the uniform-spectrum condition, the network used a broad band compari-

son strategy to perform the task. Unlike the previous condition, the network is

computing a function which determines the difference between the value of the

signal component against the sum of the masker values. This is entirely consis-

tent with the human data which argue for a simultaneous comparison in a

ARTIFICIAL NEURAL NETWORK 19

MAGNITUDE OF WEIGHTS

io 3 4 6 8 oO OT 12 ts 1415:16 1718 1920 21 22

INPUT NODES

Fig. 9. Asymptotic weights from neural network trained in the profile-analysis condition.

profile analysis task. Further, the network’s performance with larger ISI’s sup-

ports the view of a more robust internal representation of the stimulus which is

less sensitive to sensory decay over time.

Multi-component weight matrix. Figure 10 plots the weights from the

multi-component condition. As above, the magnitudes of the weights from the

input unit centered at the signal frequency is largest in magnitude. However,

unlike the above network trained in the profile condition, the values of the

weights clamped to the nodes for the masker components are all zero. This

explains the performance decrement observed in Figure 6 for this condition. It

appears that the network was performing narrowband listening, and succes-

sively comparing energy across intervals of a trial.

This result is inconsistent with Green’s inference that in the case of profile

analysis, a simultaneous or successive comparison may be possible, but that a

simultaneous comparison within a sound is more likely. In this situation, the

network could have performed a simultaneous comparison of the components

within an interval, but did not.*

“In other simulations carried out in our laboratory using multi-layer feed-forward networks with a single

hidden layer, the performance of the network trained in this condition was very similar to the profile-analysis

network. Another distinction involved the decay function. Time or sensory decay was modeled by adding

Gaussian noise with zero mean and different standard deviations.

20 MILLER, HARPSTER, AND HOWARD

MAGNITUDE OF WEIGHT S

1234 5 6 7 8 910111213 1415 1617 18 139202

INPUT NODES

Fig. 10. Plot of asymptotic weights from neural network trained in the multi-component condition.

Conclusions

The purpose of the computer simulations carried out in this study.was to

understand the psychological phenomena of intensity perception and auditory

profile analysis. These studies have shown that neural networks can be a useful

tool for modeling and evaluating the existing body of psychoacoustic data.

Further insights into the more cognitive aspects of auditory perception were

discovered. These included both the listening strategies as well as attentional

issues involved in auditory processing.

Acknowledgements

The authors wish to acknowledge Dr. J. J. O’Hare for his careful review of this

manuscript. This research was supported by the Office of Naval Research.

References

Berg, B. G., (1989). Analysis of weights in multiple observation tasks. J Acoust Soc Am, 86:1743-1746.

Berg, B. G., & Green, D. M. (1990). Spectral weights in profile listening. J Acoust Soc Am, 88:758-766.

Bernstein, L. R., & Green, D. M. (1987). The profile analysis bandwidth. J Acoust Soc Am, 81:1888-1895.

Clark, A. (1989). MicroCognition: Philosophy, cognitive science, & parallel distributed processing. Cambridge,

MA: MIT Press.

ARTIFICIAL NEURAL NETWORK 21

Durlach, N. I., & Braida, L. D. (1969). Intensity perception I. Preliminary theory of intensity resolution. J

Acoust Soc Am, 46:372-383. '

Gilkey, R. H. (1987). Spectral and temporal comparisons in auditory masking. In W. A. Yost & C. S. Watson

(Eds.), Auditory processing of complex sounds (pp. 26-36). Hillsdale, NJ: Erlbaum.

Green, D. M. (1983). Profile analysis: A different view of auditory intensity discrimination. American Psychol-

ogist: 38, 133-142.

Green, D. M. (1988). Profile analysis: Auditory intensity discrimination. New Y ork: Oxford University Press.

Green, D. M., Kidd, G., Jr., & Picardi, M. C. (1983). Successive versus simultaneous comparison in auditory

intensity discrimination. J Acoust Soc Am, 73:639-643.

Howard, J. H. Jr., Harpster, J. L., & Miller, M. H. (1989). Profile analysis by parallel distributed networks. J

Acoust Soc Am, Supplement 1, 85, N18, Spring.

Levitt, H. (1971). Transformed up-down methods in psychoacoustics. J Acoust Soc Am, 49:467-477.

Minsky, M. L., & Papert, S. A. (1969) [1988]. Perceptrons: Expanded edition. Cambridge, MA: MIT Press.

Rosenblatt, F. (1962). Principles of neurodynamics. New York: Spartan Books.

Rumelhart, D. E., Hinton, G. E., & McClelland, J. L. (1986). A general framework for parallel distributed

processing. In D. E. Rumelhart & J. L. McClelland (Eds.), Parallel distributed processing. Vol. 1: Founda-

tions (pp. 45-76). Cambridge, MA: MIT Press.

Journal of the Washington Academy of Sciences,

Volume 81, Number |, Pages 22-42, March 1991

Evaluation of St. Lucia’s

Geothermal Resource

William B. Taylor, P.E.

Alexandria, Virginia 22309

and

Frederick J. Edeskuty, Ph.D.*

Los Alamos, New Mexico 87545

ABSTRACT

The Caribbean island of St. Lucia has a strong geothermal energy resource. Under USAID

funds, the Los Alamos National Laboratory made measurements of St. Lucia’s geothermal

field and conducted an engineering and economic evaluation of its potential for meeting

electricity and industrial process heat needs of St. Lucia. This paper, presented at a 1984

international energy conference in Caracas, Venezuela, summarizes the Los Alamos study

and its recommendations for development of St. Lucia’s geothermal resource. Subsequent

drillings of geothermal wells at St. Lucia under United Nations funding have confirmed the

Los Alamos geotechnical measurements. St. Lucia’s negotiations with several private firms

for developing the geothermal field and associated electrical and process heat plants have not

yet produced agreement to undertake the work.

Introduction

St. Lucia is a volcanic island on the eastern rim of the Caribbean Sea. The

southwestern portion of the island is dominated by mountainous terrain, with

steam fumaroles and boiling pools attesting to its volcanic origin. The last vol-

canic eruption probably occurred about 20,000 years ago, and geologic and

geophysical studies suggest the presence of a large geothermal resource beneath

the Qualibou Caldera. British geologists explored this resource and drilled seven

shallow wells in 1970. Four of these wells encountered steam in limited quanti-

* This work was performed under the auspices of the°U. S. Department of Energy.

EVALUATION OF ST. LUCIA’S GEOTHERMAL RESOURCE 25

ties. A subsequent Italian assessment concluded that development of St. Lucia’s

geothermal resource is probably economically feasible.:

In 1983, the Government of St. Lucia commissioned the Los Alamos Na-

tional Laboratory to conduct a study of the possibilities for generating electricity

and industrial process heat from the geothermal resource beneath the Qualibou

Caldera. In April 1984, Los Alamos reported to the Prime Minister of St. Lucia

that new measurements and analysis of available data now justify beginning the

drilling phase of a program to develop the full geothermal potential of St. Lucia.

This report summarizes results of the Los Alamos assessment of St. Lucia’s

geothermal resource, and it discusses possibilities for generating electricity and

process heat, both for existing St. Lucian industries as well as for new industries

which cheap energy might attract.

_ The work performed for this study was funded by the Trade and Develop-

' ment Program of the United States Government.

Geology and Geophysics

St. Lucia is a former British colony which is now an independent member of

the British Commonwealth. An island nation in the eastern Caribbean Sea, it is

located at 61° west longitude and 14° north latitude. The island is approxi-

mately 25 miles long and has a maximum width of 15 miles. St. Lucia covers an

area of 238 square miles and had a population in 1981 of 122,000 persons.

Geology

St. Lucia has a volcanic geology, and the Qualibou Caldera near Soufriere in

the southwest exhibits many of the characteristics of an underlying geothermal

resource (Figure 1). This caldera, first identified in 1964 during detailed geologic

mapping, is estimated to be between 40,000 and 300,000 years old. The geother-

mal aspects of the area have long been recognized in the Sulphur Springs and the

nearby hot springs, which have been used for mineral baths at least since the

early European settlement in the 17th century. More recent investigations have

resulted in new geologic and economic studies of the Sulphur Springs area which

describe a steam-rich geothermal system within a relatively shallow reservoir.

The volcanic events which formed St. Lucia are estimated to have occurred

approximately one million years ago. The most spectacular volcanic domes of

that period of activity are the plug-domes known as Petit Piton and Gros Piton,

dated at 250,000 years. The Pitons are located along the western side of the

caldera.

Two major northeast-southwest faults straddle the Qualibou Caldera. There

are also minor southeast-northwest faults. Faulting has controlled, for the most

24 WILLIAM B. TAYLOR AND F. J. EDESKUTY

4 Mount Tabac

sd “at

aCe ae EN N eb , 4 Mount Gimie

: aN ee as WA

iN <a 2

C < % GEE Soytriére

setees

use

Soutridre Bay\

> ne aril a x

A / i Ge

72 \ ee Ji

CARIBBEAN SEA

FAULT, BAR AND BALL ON DOWNTHROWN SIDE

TOPOGRAPHIC Rid, QUALIBOU CALDERA

\y. oe LINES OF CROSS-SECTIONS

e f eo* Cy " PHREATIC OR PHREATOMAGMATIC CRATERS

Nite Sea!

nee L'ivrogne ° GEOTHERMAL WELL

Uz,

Yi, AREAS RECOMMENDED AS DRILLING TARGETS

DACITE OF BELFOND

SCALE 1:25,000

(cao iias wi: baleen a ae in €) OACITE OF TERRE BLANCHE

)

ANDESITES OF MORNE BONIN, BOIS D'INDES FRANCIOU

STRUCTURE MAP, QUALIBOU CALDERA

ST. LUCIA DACITE OF THE PITONS

~~ SPRINGS

Fig. 1. Structural map of the Qualibou caldera showing regional NE-SW and NW-SE linear faults and

curvilinear caldera faults. The last eruptive units are shown by the dacites of Belfond and Terre Blanche.

part, the location of thermal springs. Regional faults and caldera faults are

important in providing pathways to the surface for thermal waters.

The Qualibou Caldera appears to be over a magma body or bodies located at a

depth of three to four kilometers. The temperature gradient of 220° C per km

measured in wells drilled by the British (numbered vertical lines on Figure 2)

suggest temperatures higher than 600° C at a depth of 3 km. Geothermal fluids

and vapors should be found at a depth of 1 to 2 km under the entire caldera area

and in abundance where permeable rocks allow for greater fluid movement.

EVALUATION OF ST. LUCIA’S GEOTHERMAL RESOURCE

“PIOP[BS OY} JO OPIS UJ9}SOM IY} UO M\—J SI -D pur ‘ayouryg oa],

We puod & UM MAS-AN S!_@-€ “ASS-MNN S!_V-V :BJopjeo noqyengd ay1 Jo suonsas-sso19 o180]09H) *z “B14

siwsva (2)

SALISSJGNY NOSIWNO-JAud iy

SNOLId SHI JO ALIOVAG

2 "ios p)

JHONV 18 3dyud1 4O 41IOVG

43NL NOANWNO @)

QNO41]44 4O 311D0VvVa

nA pee

io 208

Aeti

auiv19 INIAWY SONIUdS UNHAINS =

JONVSIV 1d

3HONW 19 | 3uuzL NOLId SOUD ei

aon

21W1D “LW

an3a@ 8

;

iDe gapeet> Se HS\ IG

1 Cy CY: a @o 9 6 *% vee [= ry a 1"

oe SpE eons a SSS ESS 5 ee ee os

=. Ee « wf i> ~ r=

4 ORs —. > SS = Ce]

eae Se SS SSS JHONV 18 3HH3L pe

ano41a9 NOSY ALY

SONIHdS UNHAINS

26 WILLIAM B. TAYLOR AND F. J. EDESKUTY

Geotechnical Measurements

Quantitative assessment of a geothermal resource requires geotechnical mea-

surements and analysis to provide confidence in estimates of fluid temperatures

and of the precise location of promising subterranean regions. Two proven

techniques for acquiring such data are measurement of the electrical resistivity

of the soil and the chemical analysis of samples of steam or fluid from the

suspected geothermal regions. The Los Alamos investigation employed both of

these techniques. |

Electrical Resistivity

Direct current (DC) electrical resistivity methods have been employed for

geothermal exploration in many countries and have proven to be a valuable

adjunct to the drilling of shallow and deep holes. Numerous case studies indicate

that high quality (greater than 200° C), liquid dominated geothermal reservoirs

are characterized by resistivities of less than 10 ohm-meters (o-m) This has been

proven true regardless of the resistivity of the host rock, which may be typically

in the range from 100 to 1000 o-m. Rocks such as granite, basalt, limestone and

sandstone have essentially infinite resistivity at 450 to 500° C.

Shallow resistivity measurements had been made in the Qualibou Caldera

region prior to the Los Alamos effort. The earlier studies developed resistivity

profiles throughout the region to a maximum depth of 700 m. Low resistivity

values were found in the vicinity of Soufriere and nearby points generally to the

north of Sulphur Springs. Additional lows were centered in the south near

Etangs. Resistivity highs were associated with the Belfond area and beneath

Sulphur Springs at depths below 600 m. The low values are associated with the

geothermal system, while the higher values measured beneath Sulphur Springs

may be due to a steam zone.

After evaluating all available data, Los Alamos geophysicists ran a 5.2 km,

north-south trending resistivity profile (Figure 3). A nominal dipole length of

200 m was selected to obtain high resolution data from the survey, which con-

tained a total of 32 electrode stations. In order to obtain resistivity measure-

ments to a depth of 2 km, the Los Alamos team used a 35 kilowatt DC transmit-

ter capable of generating currents up to 80 amperes from up to 438 volts.

The resistivity data from the dipole soundings are plotted as a function of

depth in Figure 4. These data are consistent with the shallow resistivity measure-

ments made during earlier investigations. The regions containing material of

less than 10 o-m most probably contain hot brines. The zone beneath Etangs,

containing material of less than | o-m, may contain hot water rising along a

fault. The deep high resistivity zone beneath Sulphur Springs can be reasonably

EVALUATION OF ST. LUCIA’S GEOTHERMAL RESOURCE

Aho CASTRIES

61°05'00''W )

FOND CANNES

hy i \ | ae e1

a= ROSE =

2 BNE ri

sot ren TAPEARLE~

WY 4

Gd, ; ,

Yi DIAMOND.

8 :

be -

\\ 3

x\ SAPHIRE L,

» CUENNGN

Ss SSS is

——S —

CARIBBEAN SE\ SS = =

S ZF

S Ze

> 7

~ RABOT e130)

‘| = \ ESPERANCE

\a, | aS SULPHUR 14 ESTATE

AN S SPRINGS \ Joys FOND

ES = > T.JA

13°50'00"'N 1 nie a Well \r 28 xX ST. JACQUES

eS ae = (am) o16

S So DASHEENE \ HERMITAGE $17

z= 4 RS HOTEL 019 F

ZZ SENG ON 5 OND

Yi = °20 LLOYD

} = 221

= Well \r |

022 ii, =

= 23 BELFOND ae ke ~

= s BELLEPLAINE —

= 3 024 ay URNA EL Ny

= 025 E <s

= LA DAUPHINE % aS sS

Zz ESTATE #26 oe’ =

= ETANGS 027 \

Sy Well Nr 3@e29

= ST. REMY

\ —

yt z DE VILLE

WIS ae We = ; ESTATE

> SNS Ze —

SR, N <Z= VICTORIA

= WY == — ; ) JUNCTION

Ss =-

LOSS Tans

ee SS

age Viapnaal | AN

OG Votdlin, \\eorens 1 2 KILOMETERS

yh |' XO 1/4 1/2 3/4 1 MILES

SCALE

Fig. 3. Location of the 5.2 km-long resistivity profile line across the Qualibou Caldera containing a total of

32 electrode stations.

interpreted in two ways. The first is that the high resistivity layer is due to a very

| hot, dry steam field. The second interpretation is that the higher resistivity

| indicates the presence there of less permeable rock. This would suggest that the

| region of low resistivity between Belfond and Sulphur Springs represents a fault

along which thermal fluids are rising and moving from south to north, emerging

at Sulphur Springs.

WILLIAM B. TAYLOR AND F. J. EDESKUTY

“UOTINS SSOIO d1BO[OINS oyeLdo1dde ay} YJeIUIq UMOYS oe PUP SI9}9U-WYO

UI aie SaNyeA APANSISAY “Yap Jo UoNouNy & se ponoyd Aaains a[od1p-ajodip oy} Woy eyep ALAN SISAY “p “314

: — Gz

‘S.

— 02

— si

is \

— oO!

ot 3 \

O, vA he Nc

f¢ Ne

000t sa 9 SO)

t OV | ‘

aa ( \ Ot

e= IN

Were 9? 4 92 we C2 ce 2 on bt! ol «wx Ot vb et =) rat th Ob 6 1] 1 t°) 9 v c z t

SONIYdS

YNHd INS

TTI Lodd pT TTT.

yy]

SJHONVIE 3yYuSL

SONIYdS YNHdINS

QNOJ148

C# L C#

T1SM T1SM TVEM

NOSVWW 1VW

EVALUATION OF ST. LUCIA’S GEOTHERMAL RESOURCE 29

Hydrogeochemistry

The first comprehensive geochemical studies of the Sulphur Springs area of

the Qualibou Caldera were done during the British geothermal drilling effort in

1970. Data on outlying springs were reported recently by the Italian firm,

Aquater. Los Alamos geochemists augmented this earlier work by obtaining and

analyzing fluid samples from outlying hot springs and cold waters within the

caldera and from numerous sources in the Sulphur Springs area.

The most common thermal features are broad pools or caldrons filled with

fountaining muddy water. The lavas are altered to clay. Sulphur, gypsum and

pyrite are concentrated near fumaroles and steaming ground, while iron oxides

stain the surrounding rocks to various hues of red and orange.

_ Hot spring waters at Sulphur Springs display chemical characteristics com-

~ mon to acid-sulfate systems. Acid-sulfate water does not necessarily indicate

that a steam-dominated system lies at depth. The available drilling data indicate

variable conditions in the steam zone beneath Sulphur Springs. Temperatures

approach 220° C and pressures are less than 444 pounds per square inch (psi),

which suggest a steam-dominated zone beneath Sulphur Springs. The British

well #4 alternately produced geothermal brine and steam. Chemical analysis of

that brine showed wide variations of concentrations of elements but essentially

constant ratios of certain elements. This indicates steam loss from brine in a

fracture system of limited volume.

Analysis of available data points to extremely high temperatures beneath

Sulphur Springs. The three best indicators are:

—Stable isotopes of water, where the brine shows deuterium and oxygen compositions

that indicate extensive exchange of oxygen between water and rock in the reservoir.

The pronounced “‘shift’’ in the measured isotopic oxygen composition is a character-

istic observed of deep reservoir fluids with temperatures in excess of 250° C.

—Brine chemical analyses (Table 1), which point to Sulphur Springs having a consistent

subsurface reservoir temperature of 280° C.

—Temperature of vapors from superheated fumaroles (171° C), which calculates isen-

tropically to 260° C at the source depth.

Table 1.—Gas Temperature Indicators, Sulphur Springs

Bubbling Pool Superheated Vent

Brine components IDE KTGIS F) 171° € (340° F)

H, (vol-%) 5.04 5.63

O, 0.02 0.54

N, ett Des)

CH, 0.64 0.69

HS 1.67 1.09

CO, 91.93 89.59

Total 100.4 100.1

Gas geothermometer 283eeC, (941 2) 280%E (93694)

30 WILLIAM B. TAYLOR AND F. J. EDESKUTY

Summary

The Los Alamos analysis of the deep resistivity and geochemical measure-

ments supports a recommendation to drill geothermal wells at the following

locations (Figure 4):

—The craters of Belfond, which could determine if the main area of geothermal upflow

is centered within the area of youngest volcanism. This well is expected to encounter

dry rocks to a depth of 600 to 900 m, below which geothermal brine is expected.

-—The valley of Sulphur Springs, near the British wells #4 and #7, but targeted for a

depth of 2 km to pass through the shallow vapor zone and penetrate the postulated

deep brine reservoir. This well is expected to encounter very hot, dry steam between

600 and 1700 m and a geothermal brine at approximately 1800 m. It is also possible

that a less permeable fluid bearing rock will be encountered instead of the steam zone.

—The Etangs area, south of the road near the Nutmeg Bar. This well should pass

through the southern caldera-bounding fault and is expected to encounter geothermal

fluids as shallow as 1000 m.

Applications of St. Lucian Geothermal Energy

The heat energy produced by geothermal wells in St. Lucia could be converted

to electricity or industrial grade process heat by heat exchangers and steam

turbines of proven design similar to those now in use in many countries. In the

following sections the current and projected demand for electricity and indus-

trial process heat on St. Lucia, and concepts and budgetary cost estimates for

meeting those demands with geothermal energy from wells near Sulphur

Springs are described. The costs of generating electricity and process heat from

the St. Lucian geothermal source are compared with costs of energy generated

from other heat sources, such as fuel oil, wood and coal. The possibilities for

using geothermal heat in some existing St. Lucian industries are discussed and

some new industries which might be attracted to St. Lucia by the availability of

cheap energy described. Finally, a discussion is provided on the possible impact

on the economy of St. Lucia which could result from the development of its

geothermal resources.

Electricity

Electricity on St. Lucia is now generated by St. Lucia Electricity Services

Limited (LUCELEC), using large, British, low speed diesel engine generators. A

total of approximately 16 megawatts (MW) is installed in two stations: one in

the north near the capital, Castries, and the second near the southern city of

Vieux Fort. Two separate, 11 kilovolt (KV) transmission grids serve LUCE-

LEC’s 14,814 domestic and industrial customers in the northern and southern

regions of the island. Plans exist to replace these two grids with a single, inte-

CONSUMPTION

peak power consumption (12 MW) and average

AVERAGE POWER

and it plots the LUCELEC projection of annual

EVALUATION OF ST. LUCIA’S GEOTHERMAL RESOURCE

LUCELEC projects a continuation of current demand growth of 5% per year

in the northern region and 4% per year in the south. Figure 5 shows the 1983

installed capacity (16 MW)

capacity is then assumed for the remainder of the period studied. To maintain

grated, 66 KV grid in the near future in order to reduce the excessive transmis-

growth in power demand, which they assume will continue to grow at the same

rate through the year 2016. Figure 5 also reflects the 4 MW reserve capacity

currently in the system and projects increasing the reserve at the assumed de-

mand growth rate until a 5 MW reserve capacity is achieved. A 5 MW reserve

that level of reserve capacity, LUCELEC must add new capacity soon, and the

sion losses being experienced with the current system. |

power consumption (8 MW),

Wi) }/ PA /N,

NG i) i

KKK KKK KOK RIKKI KIKI

K wteaeteattaletettatteetcaltaeteetteletaettaeteteltaertattenitaittnere

Measatlsltaletaltaleteletalteltettatetettaetaeteetaetealeatsttcdtebtcitsitdt)

Da] PRX RX RXR ROA RX RR X ALA RL ARLE RRL RE ERK KL RE KT

wstealelatteateattattaltattataeteatteletettaltaeteeltatetaeteetetattaiteat

f eteedtarebetabtcdtcietattstetaitnitcii retaatetettaeteitat taal: Eat

Wueetallcaltabetaltceteeteetetetes AURRLRLY]

RX RXR XR RX KRY Or ORONO

\// RRA RRA RAR ANN NG RO OO

i Weeatitoaieloniaieleiattataeteiias AXE

dhndendtadtantcrtartcrteitcnb ui tudte RX KXRA RXR

RXR X RR AAA LKR)

A PH CS aeararareronnne Ras

< Uf, atccteteteiies LK KER)

(| Hededetecaceecatacorecers KARR ARK R XRG

| SRA AND easrlctlaitaitaitcite

xq "i

) WY VV VV YY Y Tavatatatata' atatatatata™, @ VV V AIA SAAS, VV

Pens se neeeeens atetererenen 86000004 006060205060 050000000 00000000000 0000000, Raed

Haniel wesetatetabtattahcahee wisietadtadctcehcit miseleuteatettsieteateatigs "=

| AAA TARO RRR ay ARR

‘

Mr i) ne ARXAX ALARA RAY, KERR AA LALA

P Meseceeateteee reatcleatateetatattatet

b¢7// Washtatatelatetet KKK)

y RNS ON 4 OD

\ y G Be XXAK , ‘ SAAR Oo

CMU) vis PS

eltete PRK KRK 1

oe ROXKRAND masetsatelaree

KK LS Wn

> BN) LRT CO)

Parerratare KE

! y Wrote AER PR EARL RL

Ms EARNS varororocarnenenarnrnrirn’

q yy © //////PRern nesitibtateeteletettite Koad

» /] LKR KER CO

eeslelattiitaers

] ninieiieienits

SKK KK]

PEAK POWER

CONSUMPTION

<sS

a FS

—_

Siete

AANA RN O

b SKK RK XEN

‘ aealaltatetaiey

ararscorocococorororocococorer

WYK X RRR LILY

nrararararneniioee les)

LTR RAR RELL]

asacetartartadtaitrtetatettattritries Coz

RRR RXR RRL RY RXR RYH KY

KEN REAR RX RX AL RL ARAL

MPA HOMOORERNONNNNN 99

// Prrrceiierlaccitetctetc’s band

RAUB OAXOR OAK A OD

CAPACITY

DESIRED INSTALLED

CAPACITY

TOTAL INSTALLED

N oO vt oO © N © +t oO © N 0 t

(MW) GNVW3SG YSMOd

CALENDAR YEAR

Fig. 5. Geothermal electric power demand in St. Lucia for 1983 (actual) through 2016 (estimated).

32 WILLIAM B. TAYLOR AND F. J. EDESKUTY

total new capacity required to meet new demand will be approximately 10 MW

by 1996 and approximately 32 MW more by the year 2016. Figure 5 shows 10

MW of geothermal power coming on line during 1986-88 and 20 MW more

during 1992 through 2011.

The geothermal system assumed for generating electricity on St. Lucia con-

sists of a field of production wells, each producing brine at 250° C and with an

assumed flow of 270,000 pounds per hour. The 50,000 pounds of steam per

hour required to drive a 2.5 MW, well-head steam turbine-generator would be

flashed from the brine in a steam separator, leaving residual heat totalling ap-

proximately 44 million BTU’s per hour in the brine. This waste heat can be used

for industrial process heat applications, as will be discussed later. Typical tur-

bine designs have an inlet steam pressure of 75 pounds/in’ in absolute units

(psia) and a back pressure of 3 in. of mercury (1.5 psia). Steam condenser

cooling can be by air flow in a cooling tower or by cooling water, perhaps using

sea water.

This modular approach of adding one 2.5 megawatts of power available to the

electrical grid (MWe) of the geothermal well and one steam turbine-generating

set at a time to the LUCELEC grid permits the geothermal system to operate at

near rated capacity for maximum economy. It also permits using the existing

diesels for standby and peaking operation and thus allows them to remain in

service longer before reaching their design life of 100,000 hr. of operation.

Using the planning schedule reflected in Figure 5, estimates of the capital and

operating expenditures which would be required over that 33-year period were

developed. Costs were itemized by year and a discounted cash-flow computer

code calculated the Levelized Life Cycle Cost (LLCC) for the system. The LLCC

is the unit cost of electricity in dollars per kilowatt-hour which, if charged to

consumers throughout the life of the plant, would result in revenues equal to the

total capital and operating costs of the plant during that same period.

The major assumptions on which our estimate of the capital cost of the

geothermal electric power system are based were:

(a) Values in 1983 $(US)

(b) LUCELEC power demand projection

(d) Integrated grid of geothermal and diesel power

(e) 30 yr. of operation

(f) 30 MW of geothermal power

(g) $(EC) = $(US) x 2.7

The well-head steam-turbine system is of a type currently available from several

U.S. competitors. A set rated at 2.5 MW can be skid-mounted and moved over

rough terrain to the well-head. Such equipment has a design life of 30 yr. and can

EVALUATION OF ST. LUCIA’S GEOTHERMAL RESOURCE 33

be purchased in the U.S., shipped to St. Lucia and installed, all for approxi-

mately $1000 (US) per kilowatt ($2700 (EC)/KW). The production wells are

assumed to require drilling to a depth of 2000 m and to produce brine or steam

with a wellhead temperature of 250° C. Each well is assumed to cost $2 million

US ($5.4 million EC). A 50% success ratio for bringing in producing wells, each

of which will have a 20 yr. life, is also assumed.

In calculating operating costs, all funds were borrowed at 10.5% interest and

inflation was assumed to be at a constant 5% annual rate throughout the period

being considered. Well maintenance would be 10% of the initial well cost and

generator maintenance would be 2% of the initial cost. A three-shift operation

with a ten-person crew plus three supervisors per shift totals to an annual cost of

$540,000 (EC) per yr. Additional wells would have to be drilled and additional

-well-head steam turbine-generators added as needed, at the costs cited above. In

view of the preliminary nature of the system design and the uncertainties asso-

ciated with a new venture of this type in St. Lucia, 25% was added for contingen-

cies to the estimates of both initial capital costs and annual operating costs.

Using these assumptions, a LLCC was estimated at $0.17 (EC)/K Wh ($0.063

US) for the geothermal steam turbine-generator system. The corresponding cost

of electricity produced by the existing diesel system was calculated to be $0.243

(EC)/KWh ($0.09 US). The cost of producing the much smaller amounts of

peaking power by the diesel system is slightly higher, or $0.275 (EC)/KWh

($0.102 US).

Figure 6 compares the LLCC of geothermal and diesel power with the costs

calculated for modern steam plants fueled by imported Bunker C fuel oil and by

Illinois No. 6 coal shipped from the Gulf Coast of the U.S. The conclusion to be

drawn from Figure 6 is that, even with the very conservative capital and operat-

ing costs assumed for the geothermal system, it offers over 30% cost savings

compared to the existing system and is also significantly cheaper than electricity

. from new coal- or oil-fired steam plants. However, electricity costs of 6 US cents

per kilowatt-hour are not low enough to attract such energy-intensive industries

as aluminum smelting, which are able to buy electricity from large power plants

in the U.S. for 3 to 4 cents per kilowatt-hour. Nevertheless, the 30 MW geother-

mal system assumed here for St. Lucia would result in a reduction of oil imports

totalling 231 million gallons over 30 yr. The difference in cost between all-diesel

generation and geothermal base-load plus diesel peaking indicates a saving of

$250 million (EC) over that 30-yr. period.

Industrial Process Heat

The heat available from the St. Lucian geothermal resource can be converted

to a form useful for industrial processes in two ways: (1) direct heat transfer at

34 WILLIAM B. TAYLOR AND F. J. EDESKUTY

LEVELIZED LIFE CYCLE COST

$(EC)/KWh

0.0 0.1 0.2 0.3

1. GEOTHERMAL

NO U.N. FUNDS

2. GEOTHERMAL

WITH U.N. FUNDS

3. DIESEL (PEAKING)

4. DIESEL (TOTAL)

5. OIL-FIRED STEAM

6. COAL-FIRED STEAM

0.00" 9/02" "0.040106 - Cloe 78 OTe soar

$(US)/KWh

Fig. 6. Cost comparison of electricity from geothermal generation and from alternative technologies.

the well-head to a pressurized hot water loop, or (2) capture of heat in the

residual brine from the steam separator in the power generation system.

The direct transfer of heat from the base-case well assumed for a 2.5 MW

turbine-generator could be used instead to deliver over 95 million BTU’s per hr.

in water heated to 222° C (433° F) to industrial plants requiring process heat

and located within two miles of the well. For more distant plants, the delivered

temperature would be slightly lower. Using waste heat from the power genera-

tion process, approximately half that heat flow at a lower temperature of 140° C

(285° F) could be delivered to the same plants.

Figure 7 indicates the comparative costs of process heat from the geothermal

heat source in St. Lucia with costs from Bunker-C fuel oil in St. Lucia and with

oil and natural gas in the U.S. The upper range of costs for the St. Lucian

geothermal case (approximately $3.7 (US) per million BTU) is less than half the

cost of process heat derived from Bunker-C fuel in a modern boiler in St. Lucia.

The lower range of St. Lucian geothermal process heat costs (approximately

$0.8 (US) per million BTU) is less than one tenth of the St. Lucian Bunker-C

case. |

The principal factors causing the wide range of costs of geothermally pro-

duced process heat for industrial applications are the temperature required at

the factory, the distance of the factory from the geothermal plant, the plant’s

EVALUATION OF ST. LUCIA’S GEOTHERMAL RESOURCE 35

Cost of Geothermal Process Heat

(3 Shift, 5 Day/Wk Operation)

$EC/Million Btus

0 3 6 9 12 15 18 21 24

Bunker C in St. Lucia XXXX

Oil in U.S. XXXXXXXXKKK

Natural Gas U.S. XXXXKXKKKKK KK KK

Geothermal in St. Lucia XXXXXXXXXK XK XXX XXX XXX

0 il 2 3 4 5 6 7 8

$US/Million Btus

Fig. 7. Process-heat cost comparisons in St. Lucia.

operational plan (numbers of shifts per day and days per week), and the percent-

age of available heat per well delivered to the factory. Figure 8 reflects these

variables, with the lowest cost being for 140° C (285° F) heat delivered to a plant

. Operating 3 shifts per day, 7 days per week, located 2.4 km from the well-head

and using all the heat available from the supplying well. The highest cost of

geothermal process heat shown (or over four times the lowest cost) is for 222° C

(433° F) heat delivered to a plant operating 5 days per week and using only 10%

of the heat available from the well. Figure 8 also shows, as did Figure 7, that the

cost of Bunker-C heat in St. Lucia is over twice the cost of the costliest geother-

mal process heat.

The conclusion to be drawn from these estimates is that industrial process

heat from the St. Lucian geothermal resource promises to be so cheap that

existing industries would probably switch to geothermal heat and new industries

would be attracted to the island.

WILLIAM B. TAYLOR AND F. J. EDESKUTY

$EC $uS

sh 8.00 cosToFB

UNKER-C HEAT IN ST.LUCIA

20.00

700 (j= 3SHIFTS, 5d/wk

18.00 —

FJ= 3 SHIFTS, 7d/wk

14.00

COST OF OIL HEAT IN THE U.S.

1200 ~The

- 400

—_s oe oe eeeee cee eeoees= ces eee ees aes

HEAT COST ($/million Btu)

2.00

4.00 \b

1.00 Af

200) - | Ni

NY:

0.00 - wt

TEMPERATURE, °C 190 140 222 222

DELIVERED TO (°F) (285) (285) (433) (433) (270)

A FACTORY

DISTANCE km 2.4 2.4 2.4 2.4 21.0

DELIVERED

% OF AVAILABLE fete) 10 100 10 100

HEAT PER WELL

DELIVERED TO A FACTORY

Fig. 8. Comparative costs of process heat (vertical bars represent costs for geothermal heat source).

Potential industries that could use process heat and are candidates for apply-

ing the St. Lucian geothermal energy to profitable operation would include:

copra plants; timber processors; concrete-block factories; breweries; manufac-

turers of alcohol from sugar; manufacturers of dry ice; producers of banana

EVALUATION OF ST. LUCIA’S GEOTHERMAL RESOURCE 37

chips; tourist hotels; producers of potable water from steam; and alumina pro-

ducers. Some of these industries are already represented on the island, while

others might be attracted there by the availability of large supplies of cheap heat

in a readily usable form.

Copra Plant

An existing copra processing plant in Soufriere produces 3300 gallons of

coconut oil per day from 22 tons of coconuts. The plant employs about 100

people and consumes 135,000 Imperial gallons of Bunker-C fuel oil per yr. to

provide approximately 3,000 pounds of steam per hr. at 142° C (288° F) for the

refinery and 104° C (220° F) for the copra presses. The oil-fired boiler could be

replaced by a heat exchanger connected to an insulated, 2-in. diameter, steel

pipeline capable of pumping 30,000 pounds per hr. of hot water through a closed

loop containing another heat exchanger at the well-head. The cost of the hot

loop and heat exchangers could be prorated among other industrial customers

in the Soufriere area for increased economy. This would, of course, require

increasing the diameter of the insulated supply line to accommodate the flow

requirements of the other plants.

Timber Processing

Over three quarters of the St. Lucian land-area is forested and produced the

113,730 board feet of lumber sold in 1982. Modern kilns for drying timber

operate at temperatures between 38-93° C (100—200° F) and can dry approxi-

mately 15,000 board feet of lumber in a drying cycle, which varies from a few

days to several weeks in length. A 2-in. insulated steel pipeline could deliver

process heat from the geothermal plant to a timber drying kiln in the amounts

required for continuous operation. A geothermally powered timber-processing

industry in St. Lucia might also attract the import of raw hardwood from South

and Central America, and the dressed lumber might be highly marketable for

manufacture of furniture in the U.S. and Europe.

Concrete Block

Concrete block construction is used in most of the residential and commercial

buildings built in St. Lucia and neighboring islands. St. Lucian construction

consumes approximately 650,000 blocks per year. A modern, atmospheric pres-

sure steam kiln for curing concrete blocks operates at 88° C (190° F) and

requires an average of 1.2 million BTU per hr. An insulated, 2-in. diameter steel

pipe heat-loop could deliver enough process heat from the geothermal field to a

kiln suitable for meeting the concrete block requirements of St. Lucia and sev-

eral of her neighbors.

38 WILLIAM B. TAYLOR AND F. J. EDESKUTY

Beer Production

A brewery in Vieux Fort, 21 km southeast of Soufrere, employs 109 people,

produces approximately 6,000 bottles of beer per day and is sized for twice that

rate. It currently consumes approximately 80,000 K Wh of electricity and 5,000

Imperial gallons of Bunker-C fuel oil per month to generate steam at 150 psi and

161° C (322° F). An insulated pipeline from the geothermal field would cost an

estimated $1.6 million (US), which might not be justified by savings in oil cost

unless other industries in the Vieux Fort area shared the cost of the pipeline and

the consequent savings from avoidance of conventional energy costs.

Alcohol from Sugar

By partially reviving its dormant sugar industry, St. Lucia could apply its

geothermal energy to the production of enough industrial grade alcohol to re-

place approximately 500,000 gallons of motor fuel per yr. by adding alcohol to

gasoline to form gasohol. Typical sugar cane yields for that region from 1500

acres in St. Lucia would be enough to feed a combined sugar-mill and distillery

requiring about 4 million BTU per hr. at temperatures up to 218° C (425° F) to

meet the gasohol production rate. An insulated, underground, 6 in.-diameter

steel pipeline to deliver that amount of heat would cost approximately $30 (US)

per ft., and an uninsulated return line installed in the same excavation would

cost about $7.5 (US) per ft.

Dry Ice

The production of dry ice in St. Lucia could provide a readily transportable

source of refrigeration for enlarging the fishing industry as well as for exporting

dry ice to neighboring countries. Noncondensable gases contained in geother-

mal steam typically contain about 1% carbon dioxide, which would amount to

approximately 1400 tons per yr. from each producing well. Dry ice production

requires about 20,000 pounds of steam per ton of dry ice for driving compressors

to increase the gas pressure to 1100 psi. The only other resource required is labor

at the rate of 8 person-hr. per ton.

Banana Chips

The production and processing of 4.3 million banana stems per year is a

major export industry in St. Lucia. The manufacture of banana chips from the

estimated 10% waste from the harvesting and boxing operations might be an

attractive new industry to capitalize on the availability of cheap process heat in

St. Lucia. Deep frying the 10 million pounds of waste fruit per yr. to produce

banana chips would require a plant costing about $750,000 (US) and consum-

ing process heat at the rate of approximately 2 million BTU’s per hr. Locating

EVALUATION OF ST. LUCIA’S GEOTHERMAL RESOURCE 39

such a plant at either Soufriere or Vieux Fort would allow it to share the costs

and benefits of heat loops from the geothermal field to either the existing copra

plant or the brewery now operating in those cities.

Tourist Hotel

A typical, 256-room tourist hotel in a tropical climate consumes 3.5 million

KWh of electricity, 18.2 million gallons of water and 4.2 million standard cubic

feet of natural gas for kitchen and laundry services per yr. The availability of

competitively priced electricity and cheap process heat might induce existing

hotels to convert to geothermal energy and attract new tourist hotels to St.

Lucia.

Water from Geothermal Steam

The price of water treated adequately for human consumption and for indus-

trial processes such as brewery operations in St. Lucia is approximately $2.25

(US) per thousand gallons, compared to less than $1 (US) per thousand gallons

in the U.S. Geothermal energy in St. Lucia could be used as process heat in a

multistage flash distillation system or as electricity to power a reverse Osmosis

system for converting seawater to potable water. However, neither of these

approaches nor any other yet developed appears likely to be competitive in St.

Lucia in the foreseeable future.

Another possible approach for producing fresh water could be to use conden-

sate from the turbine exhaust and residual hot brine from the steam separator.

This brine could be flashed again to separate a lower grade steam which could be

condensed and combined with the condensed turbine exhaust to produce ap-

proximately 7,000 gallons of potable water per hr. from each 2.5 MWe unit. A

12-well geothermal complex using this approach could produce over 13 gallons

per capita per day in St. Lucia for both domestic and industrial use.

Aluminum Production

Aluminum metal is produced from bauxite ore in two sequential, energy-in-

tensive processes, with the intermediate production of white, powdered alu-

mina. The conversion of two tons of bauxite to one ton of alumina requires

steam containing 15 million BTU per hr. and 200 KWh of electricity. Prelimi-

nary estimates indicate that alumina might be produced in St. Lucia using

geothermal energy for less than the current US market price. More detailed

analysis after geothermal development has begun might prove the profitability

of processing South American or Jamaican bauxite in St. Lucia to produce

alumina as the feedstock for aluminum smelters. However, the production of

aluminum metal in St. Lucia does not appear economically feasible in the

foreseeable future, since large blocks of electricity (hundreds of megawatts)

40 WILLIAM B. TAYLOR AND F. J. EDESKUTY

would be required at 3 to 4 cents (US) per K Wh in order to be competitive with

large producers in the U.S.

Table 2 summarizes the energy requirements for several process heat applica-

tions described above. These five applications require a total of approximately

33 million BTU’s per hr., or about one third of the assumed capacity of one

producing well in the St. Lucia field. The production of alumina from imported

bauxite ore has the potential for using all the heat delivery capacity from one

such well. An energy intensive industry of this type could result in lower costs for

all users serviced by the common system, since the smaller users would benefit

from the economies of scale in transmission pipe diameters, pumping power,

and delivered heat temperatures.

Economic Impact

Los Alamos developed an econometric model focused on the consumption

and output components of the St. Lucian economy. The output equations dealt

with the important private sectors of the economy, namely: agriculture, tour-

ism, industry and services. The consumption equations considered food and

beverages, fuel and light, and durable goods.

Exercise of this model resulted in forecasts of increased economic activity in

these seven sectors over the next 30 yr. Figure 9 shows these results, measured as

the predicted increase in consumption if St. Lucia converts from diesel to geo-

thermal as the primary energy source. Two cases are shown: oil prices remaining

constant throughout the period, and oil prices increasing at the rate of 2% per yr.

through 2017. The constant oil price case shows an increase in total consump-

tion of more than $20 million (US) in 1983 dollars during the 30-yr. period, and

Table 2.—Energy Requirements of Geothermal Process Heat Applications, St. Lucia

Heat Req’d Temperature Flow

Location/Industry (Million BTU/hr) C@) (Lbs. of Steam/hr)

Soufriere

Sugar-Alcohol Mill & Distillery 11.99 218 (max) 119,900

Copra Mfg. Ltd. SOO, 142 (max) 30,000

71 (min)

Timber Drying Kiln 0.40 (avg) 93 (max) 4,000 (avg)

Choiseul

Concrete Block Manufacturing 13.30 (max) - 88 (max) 133,000 (max)

8.10 (avg) 88 (min) 81,000 (avg)

Vieux Fort .

Windward & Leeward Brewery 4.70* 47,000*

ROS Fin 161 (max) 10,500**

Total 33.39 (max) 218 (max) 333,900 (max)

* Assumes 40-hr. work week.

** Assumes 168-hr. work week.

EVALUATION OF ST. LUCIA’S GEOTHERMAL RESOURCE 41

MILLIONS OF 1983 DOLLARS

87 92 97 02 07 12 Av 87 92 97 02 07 12 17

YEARS YEARS

PRICE OF OIL PRICE OF OIL INCREASES

REMAINS CONSTANT BY 2% PER YEAR

Fig. 9. Increase in annual consumption due to installation of geothermal power.

the 2% annual increase in oil prices more than doubles the projected increase in

consumption.

Conclusions

Geotechnical, engineering, industrial and economic analyses (Altseimer et al.,

1984; Hanold et al., 1984; Taylor, 1983) of all available data indicate that the

geothermal resources underlying the island of St. Lucia can be developed by

drilling production wells in the Qualibou Caldera near Soufriere and installing

modular power generating units and process heat exchangers, insulated pipe-

lines and power transmission grids to energy consumers. Based on very conser-

vative estimates of construction and operating costs, the cost of geothermal

electricity in St. Lucia should be more than 30% below current costs of diesel-

generated electricity. The estimated costs of industrial process heat from the

geothermal source range from 50% to 90% less than heat from Bunker-C fuel oil.

Several existing St. Lucian industries and a number of new ones could benefit

from the abundance of cheap energy which development of the geothermal

42 WILLIAM B. TAYLOR AND F. J. EDESKUTY

potential would provide. The St. Lucian economy is projected to grow at a

significantly faster rate if geothermal energy becomes the prime source of elec-

tricity and industrial process heat on the island, even if the world price of oil

remains constant over the next 30 yrs.

References

1. Altseimer, J. H., Edeskuty, F. J., Taylor, W. B., & Williamson, K. D., Jr. (August, 1984). Evaluation of the

St. Lucia geothermal resource: Engineering investigation and cost estimate (Rep. LA-10209-MS). Los

Alamos, NM: Los Alamos National Laboratory.

2. Hanold, R., et al. (April, 1984). Evaluation of the St. Lucia geothermal resource: Summary report (Unpub-

lished). Los Alamos, NM: Los Alamos National Laboratory.

3. Taylor, W. B. (October, 1983). Cost of energy alternatives for St. Lucia (Unpublished). Los Alamos, NM:

Los Alamos National Laboratory.

Journal of the Washington Academy of Sciences,

Volume 81, Number |, Page 43, March 1991

Corrigendum

Corrigendum should be noted within the article: Gluckman, A. G. (1990). The

discovery of oscillatory current. Journal of the Washington Academy of

Sciences, 80(1), 16-25.

(a) pg. 18, line 9 from bottom, change to. . . is the germinal /dea in Savary’s

Notice

In issue number 4 of volume 80 (December, 1990), pg. 187, the corrigenda that

were listed were not entirely correct:

(a) this item appeared on pg. 20 not on pg. 21.

have included. . . “ADDITION au Mémoire de M[onsieur]. Savary sur l’Ani-

mantation”’.

Journal of the Washington Academy of Sciences,

Volume 81, Number |, Pages 44-59, March 1991

Perceptual Integration

John J. O’Hare

CAE-Link Corporation, Dallas, TX 75261-9490

ABSTRACT

The concept of perceptual integration and evidence for that phenomenon is assessed with

a review of representative experimental studies with complex activities in the areas of haptic,

auditory, and visual performance by human observers. The phenomenon can be defined asa

process of the sensory input systems that facilitates object recognition both within and across

modalities. It is concluded that there are a sufficient number of positive behavioral findings

in support of an integrative process, and experimental techniques are available that show

promise for the direct observation of it mechanisms.

Introduction

The immediacy and ease of perception belies the complex analytic processing

that operates on sensory information to enable human observers to achieve

veridical knowledge about objects and events in the external world around

them. Some theorists (Kubovy & Pomerantz, 1981) have adopted the view that

the process is holistic, i.e., that the entire object is perceived first, and in later

stages it 1s progressively analyzed into its constituent parts. This notion is some-

times called top-down to characterize a hierarchical analysis from the whole

object that branches down to its particular parts. Yet the term top-down is used

by other theorists to affirm that expectations about the object also guide an

individual toward recognition of the object that could begin with either the

whole object or its individual features. A different view, called ecological (Shaw

& Bransford, 1977), states that the perceptual system detects higher-order, ac-

tion-relevant properties of objects, such as their gradients and ratios, rather than

their detailed features. There is also a considerable body of evidence (Treisman,

1986) that leads to the conclusion that perception arises through the decomposi-

tion of an object into its features followed by an information-processing stage

that maps the physical representations into subjective experience. Fodor (1983)

has posited the existence of input modules where complex objects are initially

decomposed into their components by a set of independent, feature-analyzer

systems. Perceptual experience occurs after the features are re-assembled.

PERCEPTUAL INTEGRATION 45

In feature-recognition theory (Treisman, 1988; Treisman & Gelade, 1980) it

is proposed that the partly independent modules extract, in parallel, the elemen-

tary features of an object. To illustrate that process, imagine the letter “T”’

printed on a card in red ink. There would be a “hue map” for color, a “‘shape

map” for the horizontal line, and a “shape map” for the vertical line. When

attention is focussed on the letter’s location on the card, the information from

the three ““maps’’ becomes a composite representation of a red letter ““T”’. Ex-

periments on sorting time with those three features have demonstrated that each

represents an independent dimension of the object.

The integration of features to form a percept have lead to studies that seek to

understand the operative factors in that process within single and multiple sen-

sory-systems and that objective is the primary concern of this review. Perceptual

integration will be defined as a process of physiological input systems that facili-

tates detection, precedes and enhances recognition of objects, and operates both

within and across sensory systems. The process is assumed to be the result of an

amplification of sensory signals, the availability of additional and/or redundant

information, and/or the combination of sensory signals that leads to the develop-

ment of a new representation of the source object. Ryan’s summary (1940) of

the long history of research studies directed toward an understanding of the

process of sensory facilitation, and the opposite effect of inhibition, provided

sufficient evidence to encourage further investigations of those processes. Re-

search on this topic is still at the stage of demonstrating the phenomenon rather

than pursuing systematic studies to define its important parameters, to describe

lawful effects of experimental conditions, and to provide theoretical constructs

for a general understanding of the process.

If these latter stages of research progress are successfully completed, the pro-

cess might be useful for the development of techniques to achieve more effective

and efficient human performance in a variety of settings which require the

detection and recognition of target signals of low amplitude, when there is

‘ minimal attention on the part of the observer. One task domain where the

effective combination of sensory signals became a practical question was sonar

watchstanding. It was asked whether the success of a sonar operator who listens

for faint, acoustic signals of a target, could be improved if, at the same time, an

image of the same target was displayed visually. In one example of studies that

were directed at conditions that might lead to the enhancement of the sonar-

operator task, O’Hare (1956) found significant shifts in acuity for pure tones (up

to 3.9 db) during concurrent exposure to a visual image that was varied in color,

as tested by standard procedures. No simple statement could describe the effects

observed. A yellow color patch (300 cd/m?) always induced an increase in audi-

tory sensitivity if any shift occurred but that effect could have been related to the

46 JOHN J. O7HARE

greater brightness of the yellow patch. However, that possibility did not fully

account for the data because the yellow patch exerted less and less effect from

the low to high tones. It was concluded that the feature of color can be of

importance in intersensory effects on auditory detection thresholds.

Integration of Complex Signals

Though the phenomenon of audio-visual interaction has been established, its

practical importance is diminished when it is realized that an increment in

auditory acuity of 3 db could result from sending the same auditory signal to

both ears instead of one (Pollack, 1948). Nonetheless, despite the prospect of

meager practical gains based on studies of concurrent stimulation, using highly

controlled stimuli, there has been a continued interest in the process as poten-

tially useful with more complex, naturally occurring stimuli. A group of experts

on the Naval Studies Board (1988) has recommended that research resources be

committed to that topic. In addition, a recent research study (Doll & Hanna,

1989) has provided positive evidence of audio-visual interactions, and it 1s indi-

cated that research interest is motivated by a search for operational techniques

to enhance detection in the sonar watchstanding task, as underwater sounds

from machinery noise emitted by other platforms and from other sources are

more effectively attentuated.

Another impetus for looking again at the process of integration is the willing-

ness of investigators to examine complex events rather than to confine their

attention to the perception of highly controlled, unidimensional sensory events.

The term complexity refers to many characteristics of information, such as

numerosity, density, number of control paths and linkages, and levels of ab-

straction. With visual figures, complexity is stated as increasing with the number

of edges, corners, contours, and turns (Chipman, 1977). This review examines

representative behavioral studies that have provided positive evidence of inte-

grative processes in order to assess the promise of further research effort on those

phenomena. Studies that have used complex activities have been reported in the

domains of haptics, audition, and vision, and they will be the focus of this

analysis.

Haptics

Haptic systems incorporate information from sensors in the skin, muscles,

tendons, and joints, to achieve purposive touch. The evidence for haptic inte-

gration is focussed on cross-modality effects between the cutaneous and motor

systems when an individual grasps an object. Studies have examined how the

PERCEPTUAL INTEGRATION 47

Table 1.—Haptic Procedures that are Associated with Acquiring Knowledge of Objects (Lederman &

Klatzky, 1987)

Object Dimensions Explanatory Procedures

1. Substance

texture lateral motion (LM)

hardness pressure (PR)

temperature static control (SC)

weight unsupported holding (UH)

2. Structure

weight unsupported holding (UH)

volume enclosure (EN); contour following (CF)

global shape enclosure (EN)

exact shape contour following (CF)

3. Function

part motion part motion test (PMT)

specific function function test (FT)

dimensions of an object are combined and what happens when one of those

dimensions is withdrawn. Those outcomes have been confirmed by observa-

tions of the exploratory procedures employed by the same participants during

object recognition. The theoretical explanation for the occurrence of integration

rests on the notion of motor and/or regional compatibility of the two sensory

systems.

There are stereotypical hand movements, 1.e., exploratory procedures, which

are associated with the recognition of an object’s dimensions. Those dimensions

have been grouped (Lederman & Klatzky, 1987) into the three classes of infor-

mation that blind-folded observers utilized to recognize an object: substance,

structure, and function (Table 1). Lederman and Klatzky asked their observers

to explore the objects freely and to match them to a standard object; the ob-

served motor patterns were consistent for each dimension. In the process of

making those observations, the investigators noted that some of the exploratory

procedures led to above-chance performance for more than one object dimen-

sion. This finding led to the subjective impression that the object dimensions

formed a coherent whole and suggested to Lederman and Klatzky that an inte-

grative mechanism for haptics was at work.

The integration of three object dimensions is seen in Figure | (Klatzky, Le-

derman, & Reed, 1989). Response time, on the ordinate, is the interval (in ms)

between touching an object and the verbalization of a correct recognition of a

standard object. The observers used the dimensions of shape (S), hardness (H),

and texture (T), to classify the objects. The use of those dimensions is inferred by

the experimenter from the exploratory procedure(s) employed by the observers.

It is evident that a given procedure can be used to perceive more than one

dimension (Table 1) and exercising a given procedure allows the observer to

48 JOHN J. O7HARE

1600 +

= ae

= 1500 \

— Sy

Lo a

wn xs

io) Ase es

Ce a—— a T

~ TEST

S/T

T/H

BLOCK (12 trials/block)

Fig. 1. Speed (msec) in the identification of objects of one, two, and three dimensions, over a series of trials

(Klatzky, Lederman, & Reed, 1989).

sense other dimensions of the object, so that the research studies have been

designed to minimize those confounding factors. Smooth curves can be drawn

through the mean data points (12 blocks of 12 trials each) for the three curves

that represent single dimensions. Such curves would depict the expected effects

of practice and the attainment of asymptotic performance. Where pairs of di-

mensions (S/H, S/T, T/H) were used by the observers, response times were

significantly lower and also reached asymptotic levels. However, there were no

additional gains when three dimensions (represented by T/S/H) were employed

by these observers.

These investigators devised another experiment on the observer’s use of

paired dimensions to recognize an object. After the observers had reached an

asymptotic level of performance, one of the dimensions was made constant, or

effectively, withdrawn (W/D). If performance remained unchanged, it could be

inferred that the observer had not really been using the constant dimension, 1.e.,

no integration had occurred.

The clearest results are shown in Figure 2 where texture and hardness were the

paired dimensions (Klatzky, Lederman, & Reed, 1989). In the upper curve,

removal of the texture dimension led to an immediate increase in response time

which did not return to its former asymptotic level in subsequent trials. The

same results occurred, but with greater effect, in the lower curve, when hardness

was withdrawn as an object dimension. This is clear evidence for the presence of

integrative action.

PERCEPTUAL INTEGRATION 49

1300 TEXTURE/HARDNESS

1200

=E 1100 +

2 |

A= o

r 1000 eNGe 4 a

S x Om a— 23 —o

08 IX “~p aes

i—-a

rl areas

706 —

LEMRN sete Ste WGRE-LRN! OSS

=o aes Eas

zx” xa!

We De

o =

Period

Fig. 2. Speed (msec) in the identification of objects that vary in the dimensions of texture and hardness, and

after one of those dimensions remains constant (Klatzky, Lederman, & Reed, 1989).

Another view of the respective roles of these two object-dimensions can be

seen during an examination of the exploratory movements of the same ob-

servers. Recall that for the recognition of texture, the initial studies (Table 1)

showed that lateral movement (LM) was the exploratory procedure used; and

for the recognition of hardness, pressure (PR) was identified as the procedure

that would be employed. The procedures of enclosure (EN) and contour follow-

ing (CF) were observed when shape recognition was required.

Figure 3 depicts the results of a study in which the texture and hardness

dimensions were paired (Klatzky, Lederman, & Reed, 1989) and the observers

reached asymptotic performance in the recognition of the objects (designated as

pre W/D). It can be seen that lateral movement (LM) and pressure (PR) were

used about equally in those judgments. At a later interval (designated EXPL),

new objects were introduced and freely explored by the blind-folded observers.

Enclosure (EN) and contour following (CF) movements were elicited but LM

and PR movements predominated. In the third interval (labelled TEST), object

hardness was made a constant dimension and PR movements dropped sharply

but LM activity rose. At the last interval, asymptotic performance was reached.

These curves reflect the separation of the integrated movements when one of

them no longer remains as an effective basis for the observer’s judgments.

The same experimental design is shown in Figure 4 where texture became the

constant dimension (at TEST) and the frequency of LM movements plum-

meted (Klatzky, Lederman, & Reed, 1989). The same inference can be made

that a previously integrated movement had become separated.

50 JOHN J. O HARE

TEXTURE /HARDNESS-> TEXTURE

ae |

v 7

OG i

“ a

3s eal

5 0.4

Fost = ER

AR!

eo ae

Sac :

ee Get

if a

aac aes ee —— CF

= <

sf 8 3 3

Interval

Fig. 3. Proportion of four exploratory procedures that are utilized in the identification of objects that vary in

the dimensions of texture and hardness, and after the hardness dimension remains constant (Klatzky, Leder-

man, & Reed, 1989).

These investigators discovered in subsequent studies that the use of the extra

dimension was spontaneous even when the observers were instructed to attend

to only one object dimension.

The strongest experimental evidence was for a linkage of the exploratory

TEXTURE /HARDNESS-> HARDNESS

8) i ——9O PR

eke pee ah

Prop. Exploratory Procedure

ro)

Interval

Fig. 4. Proportion of four exploratory procedures that are utilized in the identification of objects that vary in

the dimensions of texture and hardness, and after the texture dimension remains constant (Klatzky, Leder-

man, & Reed, 1989).

PERCEPTUAL INTEGRATION 51

Table 2.—Mean Pointing Error (Degrees and Percent) During Auditory and Visual Localization as a Func-

tion of Angular Separation of the Target and Competing Signal (Bertelson & Radeau, 1981)

Mean Pointing Error

Visual Bias Auditory Bias

Separation

(Degrees) Degrees Percent Degrees Percent

7 3299 2/0) 34 4.8

15 6.27 41.8 47 3H

25 8.16 32.6 61 2.4

procedures associated with texture and hardness. The explanation for the results

was that there was a motoric compatibility, 1.e., the motor pattern incorporated

both exploratory procedures. Other exploratory procedures which led to signifi-

cant integrative results were explained as regional compatibility, 1.e., the motor

patterns focused on the center, thickness, or edges of the object.

Audition

The auditory system processes speech and non-speech acoustic signals but

there is evidence that motor components are integrated into the perception of

both types of signals. Three areas where evidence of perceptual integration have

been pursued include: (a) auditory and visual interaction; (b) phonetic and

visual location; and (c) compression of acoustic features into a single image.

They will be examined in turn.

Interaction. One approach that has been employed in the study of intersen-

sory interaction has been to ascertain whether visual stimulation biases an ob-

server's response to an auditory signal, and vice versa. In Table 2 are shown

some experimental data on pointing errors as angular separation between a

target signal and a competing signal in a different sensory modality was varied

_ randomly (Bertelson & Radeau, 1981). When the target signal was visual, a

visual-bias condition is created because the observer was instructed to attend to

that target; and when the target was changed to an auditory signal, an auditory

bias was obtained, in the same manner. Results indicated that when the angular

separation between the target and the competing signal was increased, absolute

pointing errors increased significantly for both auditory and visual conditions.

However, pointing errors were much larger (8 degrees) when the target signal

was visual than when the target signal was auditory (less than | degree).

A second study examined whether the bias from a competing signal depended

on the signal’s perceived source. In that experiment, observers not only pointed

at the source of the target signal but also indicated whether the competing signal

came from the same source.

ey JOHN J. OPHARE

Table 3.—Mean Pointing Error (Degrees and Percent) During Auditory and Visual Localization as a Func-

tion of Angular Separation of the Target and Competing Signal, and Accuracy of Source Identification (Percent

Same) for the Target Signal (Bertelson & Radeau, 1981)

Mean Accuracy of Source Identification and Mean Pointing Error

Separation Source Identification

(Degrees) (Percent Same) Degrees Percent

Visual Bias

7 79 3.39 48.4

i) 48 4.36 29.1

25 12 3.88 15.5

Auditory Bias

Tl 74 19 2

15 38 31 ait

25 p| 18 5

In the visual mode (Table 3) the observers showed a significant error for all

three separation levels. Although the absolute-error in degrees was similar across

the three separation-levels, the percent bias decreased with increasing separa-

tion. In the auditory mode, non-significant levels of absolute error were ob-

tained. The percentage of trials on which the observers reported that the target

and competing signals were from the same source declined with increases in

angular separation.

When the data attributed to the same source are calculated separately from

the data where the observers said that the signals were from a different’source,

pointing errors did not increase with separation on “different” trials but in-

creased markedly on “‘same”’ trials. This outcome was interpreted to mean that

in the decision process, there is not just a visual modality bias but that there is a

shift in the location of the criterion for perceptual fusion based on the spatial

data from a competing auditory signal.

Phonetic location. A vivid result has been reported on the integration of

phonetic and visual location by McGurk & MacDonald (1976) when speech and

visual information were placed in conflict. A video recording showed a talker

producing a consonant-vowel syllable while a dubbed sound track produced a

different consonant-vowel syllable. Thus, when the sound of “‘ba-ba” was mixed

with a video recording of “ga-ga”, 98% of the adult listeners reported a fused

sound, “‘da-da’’. If they closed their eyes, the sound reverted to “‘ba-ba”’. These

authors do not provide a strong explanation for these results but Summerfield

(1979) suggested that the dimensions may be amodal, i.e., the neural code for

the visual and auditory properties may have the same representation in both

modalities.

An alternative explanation to that provided by Summerfield can be derived

PERCEPTUAL INTEGRATION 53

from modularity theory (Fodor, 1983) which argues that for perception there are

independent input systems or modules with distinct properties. One of those

properties is that the analyzers within the input systems have shallow outputs,

i.e., they tend to solve the tasks they were designed to solve. The modules of the

auditory system have been divided (Liberman & Mattingly, 1989) into two

classes: open and closed. Open modules are for pitch, loudness, and timbre; they

are adapted for the perception of a large number of acoustic events. Closed

modules provide specialization and require synthesis for the formation of their

more complex outputs, e.g., phonetic perception or echo-ranging. It is likely that

integrative effects would be seen more readily among functions mediated by

closed systems. Summerfield conjectured that simple energy properties were

interacting at a level equivalent to open modules; however, the magnitude of

effect at that level would be small, 1.e., in the case of audition, no greater than a

few decibels. The concept of closed modules is more suitable for an understand-

ing of the occurrence of the qualitative changes reported by McGurk and Mac-

Donald.

Acoustic images. Echo-ranging data have provided another source of evi-

dence on integration. Recent research findings (Simmons, Moss, & Ferragamo,

in press) on echo-ranging in the brown bat have revealed some of the processes

that are employed by that organism in capturing, with the aid of acoustic images,

a food object rather than an inedible sphere of the same size. The echo-locating

bat integrates information on the absolute distance of the target and its shape

into an unified image of the food object.

Brown bats perceive images of targets that explicitly represent the location

and spacing of discrete glints from those targets, using echo delay and echo

spectral representations that taken together resemble a spectrogram of the

echoes (Figure 5). Two glints, A and B, from the moth reflect the acoustic

emissions from the bat and the resultant range distance. They are received by the

_bat as two echoes. Echo A at time “‘t”’, and echo B at time “‘t + delta t’”. The bat’s

auditory system encodes these echoes so that time, frequency, and amplitude

dimensions of the spectrogram are compressed into an image that has only time

and amplitude dimensions (Figure 6).

The spectrogram-delays at each frequency are represented as a target-range

map in the auditory cortex of the bat for glint A and at a slightly later time for

glint B. Through this transformation, the bat synthesizes a perceptual dimen-

sion of “target range”’ that is an integration of absolute distance and the shape of

the target.

Visual Recognition

A few classical experiments will be considered to provide evidence of integra-

tive action in the visual system. The integration of visual dimensions has been

54 JOHN J. O HARE

target . a

ranae ir) : Gs

FM EMISSION

l 1 HSes

Fig. 5. Acoustic events with frequency-modulated (FM) emissions during the detection and recognition of

prey by the brown bat (Simmons, Moss, & Ferragamo, in press).

the object of study by a large number of investigators (Garner, 1970, 1974;

Treisman, 1986, 1988) and has led to the definition of those dimensions as

either integral or separable. A dimension is separable when the level along one

dimension can be expressed without requiring the specification of the level of

the other. Integral dimensions, on the other hand, are dependent or correlated

with each other.

Speed in sorting or classifying objects (Garner & Felfoldy, 1970) has been

employed to demonstrate the existence of integrality. An example is seen in

Table 4 where the results of an experiment in which colored chips, which varied

along the dimensions of brightness and hue, were sorted into two categories.

Mean sorting times (s) are shown in the first column when only one dimen-

sion was used as the basis for placement into one of two classes of brightness or

hue. The second column shows the outcomes when the observers were in-

structed to sort by one dimension but were not informed that the other dimen-

sion was correlated with that dimension, 1.e., it was redundant. The observers

could have sorted correctly by either dimension. There is a significant reduction

in sorting time for both dimensions. This facilitation of sorting time 1s a property

of integral dimensions. Finally, the observers were instructed to sort on one of

the dimensions but were not informed that the other dimension would be chang-

ing, in an uncorrelated or orthogonal fashion, with that dimension. The result

was a Significant increase in mean sorting time. This interference with sorting

PERCEPTUAL INTEGRATION 55

AUDITORY DISPLAY

delay tuning

Bae OG kines amnesia sinreet, hear aberration’

<a) ((o) eum SO b 111 bees Gremeees

Pp eT ey Eee rss Logo agement

i. £50 VaR AC SS Ve egy er eertpe sy, ¢ She dyseate tie) PMN Se sys

2. ate see a corre ce ag

a ame ||| Hie Wee

Gy igor te eats Bria sates fing! hedonic

bls HHH

fi bee Ake Geass Peek

- Fue

eter dt

# TARGET

ss IMAGE

neural

response

SARL inee ese,

se booy =

range scale

Fig. 6. Formation of an image of “target range” by the brown bat from echoes of target distance and shape

elicited from prey (Simmons, Moss, & Ferragamo, in press).

88)

time is a property of integral dimensions under these conditions, and provides

converging evidence for the definition of integrality.

Another example of this same pattern of effects is shown in Table 5. Mean

sorting times (s) have been obtained from observers on the basis of the horizon-

tal or vertical position of a single dot. The next column shows the results when

the observers were instructed to sort by one dimension but were not informed

that the other dimension was correlated with that dimension (Garner & Fel-

foldy, 1970). There was a significant reduction in sorting time for both dimen-

6 JOHN J. O7HARE

Table 4.—Mean Sorting Time (sec) for Objects that Vary in Brightness and Hue, and Type of Relationship

(Garner & Felfoldy, 1970)

Mean Sorting Time (sec)

Character of Dimensions on Visual Image

Dimension to be Sorted One Correlated Orthogonal

1. Brightness 15.09 ise78 18.55

2. Flue 14.22 13.24 17.49

sions which is a sign that the dimensions are integral. When the observers were

instructed to sort on one dimension while the other dimension was changed

randomly, the result was a significant increase in mean sorting time. Both facili-

tation and interference effects are consistent with integral dimensions.

There are other studies on gains in performance when redundant information

is added to a visual image and further evidence for the existence of the integra-

tion of information. Some cautions have been raised on the need to control for

limitations in the process, 1.e., that the dimension being sorted is discriminable

to an ordinary observer. And in addition, controls need to be made in those

experiments for state limitations, e.g., that the conditions when the dimension 1s

being sorted are of sufficient contrast to be seen, or not of such a short duration

that they are missed. The foregoing findings have supported the notion that

integrality is a property of the physical characteristics of an object but there are

other studies (Pachella, Somers, & Hardzinski, 1981) that indicate that integra-

lity is, in addition, a property of the psychological processes of the observer.

Pachella et al. established that additional fact by using the statistical method of

multidimensional scaling to map the physical specifications of a set of signals

into a set of psychological attributes. As judged by human observers, when the

correspondence between the specifications and attributes was high, those di-

mensions had the property of integrality; however, when that match was low,

they did not.

Table 5.—Mean Sorting Time (sec) for Dots That Vary in Horizontal and Vertical Position, and Type of

Relationship (Garner & Felfoldy, 1970)

Mean Sorting Time (sec)

Character of Dimensions on Visual Image

Dimension to be Sorted One Correlated Orthogonal

1. Honzontal 18.30 16.91 19.36

2. Vertical les) 16.24 17.95

PERCEPTUAL INTEGRATION 57

A Direct Methodological Approach

The evidence for perceptual integration has been based on behavioral data

and the concept would be more persuasive if concurrent, direct observation of

brain functioning congruent with those other findings could be obtained. One

approach that shows considerable promise toward that goal is the technique of

positron emission tomography scanning which measures activity-related

changes in regional cerebral blood-flow that identifies brain areas that are more

highly active during a mental task (Marshall, 1988). A radioactive isotope that

can attach itself to red blood cells is introduced into one of the main arteries that

supply the brain. A gamma-camera monitors the radioactivity counts from the

different brain regions and those counts are transformed into measures of re-

gional cerebral blood-flow that reflect the differential activity of brain areas

when a patient or volunteer is engaged in a mental task. The data from the

camera are processed in a computer and images are displayed on a color moni-

tor in which different hues are assigned to different levels of blood-flow. The

radiation dose is low, the half-life of the isotope (e.g., oxygen-15) is brief (123s),

and the data acquisition interval (40s) fast enough so that repeated measure-

ments on diverse tasks can be obtained from the same person. The spatial

resolution of the adjacent point sources is poor (about | cm) but a single source

can be localized fairly accurately (about 5 mm). Image-analytic strategies, that

lead to increased cerebral blood-flow as in a task that is intense and focal, can

generate functional zones separated by less than 3 mm (Fox, Mintun, Raichle,

Miezin, Allman, & Van Essen, 1986).

For those tomographic studies, a “subtraction”’ methodology has been de-

vised to isolate component mental operations (Petersen, Fox, Posner, Mintun,

& Raichle, 1988; Posner, Petersen, Fox, & Raichle, 1988) in a stimulus-target

state that was not present in a control or resting state. Each new behavioral task

adds a further processing requirement to the prior one, and blood-flow measure-

‘ments from the simpler task are subtracted to obtain an image of the successive

stages of the more complex tasks. This experimental technique showed that

there were distinct cortical areas for: (a) passive processing of visual forms; (b)

generation of a use for a word after it is heard; and (c) monitoring the frequency

of occurrence for a target type. Results indicated that the locations of the sepa-

rate brain areas involved in the visual and auditory coding of words, each with

independent access to supramodal articulatory and semantic systems. These

findings fit in well with parallel models of perception. Other studies with this

technique (Corbetta, Miezin, Dobmeyer, Shulman, & Petersen, 1990) have dem-

onstrated sensitivity to the visual attributes of shape, color, and velocity of visual

58 JOHN J. OHARE

stimuli. This technology represents an effective means for a fresh approach to

the investigation of integrative mechanisms.

Summary

The foregoing review supports these conclusions:

(a) Positive evidence is available of integrative processes in haptic recognition, across

sensory and motor processes. The processes led to a significant enhancement of

performance with two dimensions but no additional improvement occurred when

three dimensions were made available.

(b) In the auditory domain, integration has been shown to occur with the visual modal-

ity for the localization of sound.

ity for the recognition of a phonetic sound.

(d) In the echo-locating brown bat, the spectral and temporal features of an echo have

been shown to be compressed into a single integrated quantity for the recognition of

its prey.

(e) In the visual domain, studies of multi-dimensional signals have developed tech-

niques for identifying integral and separable dimensions of a visual image.

(f) Positron emission tomography scanning has emerged as a feasible technique for the

isolation and identification of potential mechanisms that participate in integrative

processes.

These results have demonstrated that there are sufficient bases for systematic

studies of perceptual integration across the modalities of haptics, audition, and

vision, in which the focus would be on exploring processes at several stages of

perception:

(a) interaction across modalities;

(b) interaction of signal parameters within a given modality that leads to the formation

of a new representation; and

information about that signal.

There is sufficient theory to guide research toward a plausible explanation of

how sensory input systems mediate signals and the conditions under which

those systems could interact with each other. Systematic studies of these findings

could have a practical value for an improved understanding of the role of redun-

dant, irrelevant, and compressed dimensions in the formatting of information.

Moreover, general principles of integration could be useful for developing pro-

cedures for data fusion that would reduce complexity in information-processing

tasks.

References

Bertelson, P., & Radeau, M. (1981). Cross-modality bias and perceptual fusion with auditory-visual spatial

discordance. Perception & Psychophysics, 29:578-584.

PERCEPTUAL INTEGRATION 59

Chipman, S. F. (1977). Complexity and structure in visual patterns. Journal of Experimental Psychology:

General, 106:269-301. '

Corbetta, M., Miezin, F. M., Dobmeyer, S., Shulman, G. L., & Petersen, S. E. (1990). Attentional modulation

of neural processing of shape, color, and velocity in humans. Science, 248:1556-1559.

Doll, T. J., & Hanna, T. E. (1989). Enhanced detection with biomodal sonar displays. Human Factors,

31:539-550.

Fodor, J. (1983). The modularity of mind. Cambridge, MA: MIT Press.

Fox, P. T., Mintun, M. A., Raichle, M. E., Miezin, F. M., Allman, J. M., & Van Essen, D. C. (1986). Mapping

human visual cortex with positron emission tomography. Science, 323:806-809.

Garner, W. R. (1970). The stimulus in information processing. American Psychologist, 25:350-358.

Garner, W. R. (1974). The processing of information and structure. Hillsdale, NJ: Erlbaum.

Garner, W. E., & Felfoldy, G. L. (1970). Integrality of stimulus dimensions in various types of information

processing. Cognitive Psychology, 1:225-241.

Klatzky, R. L., Lederman, S., & Reed, C. (1989). Haptic integration of object properties: Texture, hardness,

and planar contour. Journal of Experimental Psychology: Human Perception and Performance, 15:45-57.

Kubovy, M., & Pomerantz, J. R. (Eds.). (1981). Perceptual organization. Hillsdale, NJ: Erlbaum.

Lederman, S. J., & Klatzky, R. L. (1987). Hand movements: A window into haptic object recognition.

Cognitive Psychology, 19:342-368.

Liberman, A. M., & Mattingly, I. G. (1989). A specialization for sound perception. Science, 243:489-494.

Marshall, J. C. (1988). The life blood of language. Nature, 331:560-561.

McGurk, H., & MacDonald, J. (1976). Hearing lips and seeing voices. Nature, 264:746-748.

Naval Studies Board. (1988). Research opportunities in behavioral sciences. Washington, DC: National Acad-

emy Press.

O’Hare, J. J. (1956). Intersensory effects of visual stimuli on the minimum audible threshold. Journal of

General Psychology, 56:167-170.

Pachella, R. G., Somers, P., & Hardzinski, M. (1981). A psychological approach to dimensional integrality. In

D. J. Getty & J. H. Howard, Jr. (Eds.), Auditory and visual pattern recognition (pp. 107-126). Hillsdale, NJ:

Erlbaum.

Petersen, S. E., Fox, P. T., Posner, M. I., Mintun, M., & Raichle, M. E. (1988). Positron emission tomo-

graphic studies of the cortical anatomy of single-word processing. Nature, 331:585-589.

Pollack, I. (1948). Monaural and binaural threshold sensitivity for tones and for white noise. Journal of the

Acoustical Society of America, 20:52-57.

Posner, M. I., Petersen, S. E., Fox, P. T., & Raichle, M. E. (1988). Localization of cognitive operations in the

human brain. Science, 240:1627-1631.

Ryan, T. A. (1940). Interrelations of the sensory systems in perception. Psychological Bulletin, 37:659-698.

Shaw, R., & Bransford, J. (Eds.). (1977). Perceiving, acting, and knowing: Toward an ecological psychology.

Hillsdale, NJ: Erlbaum.

Simmons, J. A., Moss, C. F., & Ferragamo, M. (in press). Convergence of temporal and spectral information

into acoustic images of complex sonar targets perceived by the echolocating bat, eptesicus fuscus. Providence,

RI: Brown University, Department of Psychology.

Summerfield, Q. (1979). Use of visual information for phonetic perception. Phonetica, 36:314-331.

Treisman, A. (1986). Properties, parts, and objects. In K. Boff, L. Kaufman, & J. Thomas (Eds.), Handbook of

perception and human performance: Vol. II. Cognitive processes and performance (pp. 35-1 35-70). New

York: Wiley.

. Treisman, A. (1988). Features and objects: The fourteenth Bartlett memorial lecture. The Quarterly Journal of

Experimental Psychology, 40A:201-237.

Treisman, A., & Gelade, G. (1980). A feature integration theory of attention. Cognitive Psychology, 12:97-

136. 2

75 Years of Scientific Thought

The Washington Academy of Sciences, one of the oldest scientific organiza-

tions in the greater Washington, DC area, has published a book entitled ‘*75

years of scientific thought’’ commemorating the first 75 years of the existence

of the Journal of the Academy.

This compilation, generally aimed at a broad-based scientific readership, con-

tains 25 of the most significant Journal articles, each being of truly enduring

value. Eight of those landmark papers were written by Nobel laureates in-

cluding such preeminent scientific giants as Hans Bethe, Percy Bridgman,

Harold Urey, and Selman Waksman.

This book is the product of an intensive two-year study conducted by a blue-

ribbon multidisciplinary Committee on Scholarly Activities which was chaired

by Dr. Simon W. Strauss, the Academy’s Distinguished Scholar in Residence.

The subject matter, which includes papers on topics.such as Theories of Heat

and Radiation, Chemical Nature of Enzymes, High Pressure in Physics, Cul-

tural Implications of Scientific Research, and Separation of Isotopes, covers

a wide variety of scientific fields, including physics, chemistry, biology, an-

thropology, and general science. The 25 papers provide a classic portrayal of

scientific thought over the past three-quarters of a century. For a complete

listing send a self-addressed stamped envelope to the Academy address shown

below.

1987, 374 pp., author and chronological title indexes, softbound.

Price for Academy members is $15, and for non-members it is $30.

Send orders to the following address:

Washington Academy of Sciences

1101 N. Highland Street

Arlington, VA 22201

DELEGATES TO THE WASHINGTON ACADEMY OF SCIENCES,

REPRESENTING THE LOCAL AFFILIATED SOCIETIES

PASE ical SOCIELY Ol WiaASMINGLON 65.5 05-02 f26 done cea se ot Be od ese emae scene James F. Goff

Aothranolocical Society Of WaShington «2... 20... 5.0 oa eee aie ence ee let eed Edward J. Lehman

PRGOteASOCIeDyiOl WASHINGTON . 2 oc. 5 2 yale Ghee Set ese Ghee dees oe eee Austin B. Williams

Chemical Society of Washington ................ Pee rasa wee ya ‘... Jo-Anne A. Jackson

Eareamompical Society Of WaShingion ...2.....2.2..6...ec0e nce o be feaeeees Manya B. Stoetzel

MISRatICOPTAMMIC SOCIETY: fc... sc ots ce eile emis a,c e See eda Vedas cn Stanley G. Leftwich

eolmrcah society Of Washington 2... 2222. sees ee ee eens cede James V. O’Connor

Micoicalisociety of the District of Columbia. 222.2. 0.2.2. 2c cs cece enc cases es John P. Utz

MainOiamalStOnCal SOCIEDY 2 2+ 22..c60 = 5 hele od oeclslee banks 2508. 08) cence’ 8% Paul H. Oehser

Bersuical Society Ot WaSMINGION .i..6 6.0 oi eek ecco ce eee tease en 5 eter Conrad B. Link

Society of Anierican Foresters, Washington Section .........2......../.4 Forrest Fenstermaker

Sts I EOMESOCISIY: OF PNOINCETS 265.0 --.05.,06 500s. ond ciate oda eneeeeUsccecgiaws Alvin Reiner

Institute of Electrical and Electronics Engineers, Washington Section sree George Abraham

American Society of Mechanical Engineers, Washington Section ................. Michael Chi

etm mnolocical Society of Washington ......... <0. cis. cuss ced case be veeees Kendall G. Powers

American Society for Microbiology, Washington Branch ................... To be determined

Society of American Military Engineers, Washington Post .............. Charles A. Burroughs

American Society of Civil Engineers, National Capital Section ........... Herbert A. Pennock

Society for Experimental Biology and Medicine, DC Section .............. Cyrus R. Creveling

Atmencan Society for Metals, Washington Chapter .....:...:.......0...... Pamela S. Patrick

American Association of Dental Research, Washington Section ............. J. Terrell Hoffeld

American Institute of Aeronautics and Astronautics, National Capital

SSE DA oo ee ies Tene MRE Pn rie hd a ne a Reginald C. Smith

Amencan Meteorological Society, DC Chapter .......:.....000.0c008 cee 0es A. James Wagner

Eos GienecrSOclery Of WaShiNPtOM 2. 2... fc. 2 cence cost cca esuntecdstesccassees Ralph Webb

Acoustical Society of America, Washington Chapter ........................ Richard K. Cook

mimemeanonucicar Society, Washington Section 22 2...2 0.0 ee cee secre ce tees Kamal Araj

Institute of Food Technologists, Washington Section -.......5....0......00.000800s Elvira L. Paz

American Ceramic Society, Baltimore-Washington Section ............... Joseph H. Simmons

Per amne MING AISOGICLY. sh a5 cn 6s ae oicace sae a ode wm ove § OSs ee ok ie ies 6 Vera bias Alayne W. Adams

Paismmeron tistory of Science Club 2.2.5.8 os. yc a eee eae ee es. Albert G. Gluckman

American Association of Physics Teachers, Chesapeake Section ............... Peggy A. Dixon

Optical Society of America, National Capital Section ...................... William C. Graver

American Society of Plant Physiologists, Washington Area Section ..... Walter Shropshire, Jr.

Washington Operations Research/Management Science Council .............. John G. Honig

InstromentSociety of America, Washington Section ......0...20.5...00. cece eee es Carl Zeller

American Institute of Mining, Metallurgical and Petroleum Engineers,

AY SLIVER CTO ASOT 0 Se I a CI oe AED Ue lc moat Ri Ronald Munson

MaMenalCapital ASimMOMOMETS: [.2 co. oe ec cc cee ete e coy Cee cbe vem ece ses Robert H. McCracken

Mathematics Association of America, MD-DC-VA Section ................. Alfred B. Willcox

Bisinct of Columbia Institute of Chemists «2. 2... eeadees ssc eset... Miloslav Rechcigl, Jr.

istrict of Columbia Psychological Association, ........0. 0.2006 elec eee eee eee Jane Flinn

Peasmingiony bamt hechnical-Group) i686. 228s ook | he awe eke Robert F. Brady, Jr.

American Phytopathological Society, Potomac Division ................... Deborah R. Fravel

Society for General Systems Research, Metropolitan Washington

Oe Te ars eg este Nes oe eae ey nn eri eh Meet NV Ronald W. Manderscheid

man Hactors Society, Fotomac' Chapter 2 isi. bo sce es wees ee oes Thomas B. Malone

American Fishenes society, Potomac Chapter (.°< 2 is. 5< 55sec ae yee we ee . Robert J. Sousa

Association for Science, Technology and Innovation ...................... are Ralph I. Cole

Basten: sociologicall Society. (. 0.0 bo occ. 5 Wackinedin cs fae eet ek oe. Ronald W. Manderscheid

- Institute of Electrical and Electronics Engineers, Northern Virginia Section ..... Ralph I. Cole

Association for Computing Machinery, Washington Chapter ............. Charles E. Youman

PVs ston Statistieals SOCIety 245 ins. 5 Sas on can pos Hee Pe has oe ew cet Robert Jernigan

Society of Manufacturing Engineers, Washington, DC Chapter ............... James E. Spates

Society of Industrial Engineers, Chapter 14 ....<.......0.0.:.e.ce02-ceee Shee John Larry Baer

Delegates continue to represent their societies until new appointments are made.

Washington Academy of Sciences 2nd Class Postage Paid

1101 N. Highland St. at Arlington, Va.

Arlington, Va. 22201 _ and additional mailing offices.

Return Requested with Form 3579

1 & ir

1 |

yas 7 |

N t] VOLUME 81

Number 2

Journal of the | The 1091

WASHINGTON

ACADEMY .. SCIENCES

ISSN 0043-0439

Issued Quarterly

at Washington, D.C.

CONTENTS

Conference on Measurement of Individual Differences, jointly sponsored by

the Human Factors Society, Potomac Chapter, and the American

Psychological Association, Division of Experimental and Engineering

Psychologists, held in Arlington, VA, on February 28—March 1, 1991.

SPECIAL EDITOR FOR THIS ISSUE: Robet S. Kennedy.

Articles:

ANNE ANASTASI, “The Gap Between Experimental and Psychometric

MONET OMNALIOINS oe ers oe RN ad atest s MN Pek ey RIOD Mh RNS, SRE Regt

MARSHALL B. JONES, “Serial Averaging in Performance-Test Theory:

PALL e RITA IRC DORE 6 spree ee Lie os eee Te em yee gen oe RUN Me ie Ne

JAMES R. LACKNER and PAUL DIZIO, “Space Adaptation Syndrome:

Multiple Etiological Factors and Individual Differences” .....................

CAROL A. MANNING, “Individual Differences in Air Traffic Control

Specialists ramming PenlOnmMance ea alec eek Be et ar dynes ceake

JOSEPH ZEIDNER and CECIL D. JOHNSON, “Classification Efficiency and

SystemsiDesione iscc2 00. 62 one GRU eesay A ati gel eR pL See Nk ie Leonia Oa aaa

Washington Academy of Sciences

Founded in 1898

EXECUTIVE COMMITTEE

President

Walter E. Boek

President-Elect

Stanley G. Leftwich

Secretary

Edith L. R. Corliss

Treasurer

Norman Doctor

Past President

Armand B. Weiss

Vice President, Membership Affairs

Cyrus R. Creveling

Vice President, Administrative Affairs

Grover C. Sherlin

Vice President, Junior Academy Affairs

Marylin F. Krupsaw

Vice President, Affiliate Affairs

Thomas W. Doeppner

Board of Managers

James W. Harr

Betty Jane Long

John H. Proctor

Thomas N. Pyke

T. Dale Stewart

William B. Taylor

REPRESENTATIVES FROM

AFFILIATED SOCIETIES

Delegates are listed on inside rear cover

of each Journal.

ACADEMY OFFICE

1101 N. Highland Street

Arlington, VA 22201

Phone: (703) 527-4800

EDITORIAL BOARD

Editor:

John J. O'Hare, CAE-Link Corpora-

tion

Associate Editors:

Bruce F. Hill, Mount Vernon College

Milton P. Eisner, Mount Vernon Col-

lege

Albert G. Gluckman, University of

Maryland .

Marc Rothenberg, Smithsonian Insti-

tution

Marc M. Sebrechts, Catholic Univer-

sity of America

Edward J. Wegman, George Mason

University

The Journal

This journal, the official organ of the Washing-

ton Academy of Sciences, publishes original

scientific research, critical reviews, historical

articles, proceedings of scholarly meetings of

its affiliated societies, reports of the Academy,

and other items of interest to Academy

members. The Journal appears four times a

year (March, June, September, and De-

cember). The December issue contains a di-

rectory of the current membership of the

Academy. |

Subscription Rates

Members, fellows, and life members in good

standing receive the Journal without charge.

Subscriptions are available on a calendar year

basis, payable in advance. Payment must be

made in U.S. currency at the following rates:

U:S. and'Canad@ ....2. 3 $25.00

Other countries -.... 45-3 30.00

Single copies, when available La 10.00

Claims for Missing Issues

Claims will not be allowed if recerved more

than 60 days after the day of mailing plus time

normally required for postal delivery and |

claim. No claims will be allowed because of

failure to notify the Academy of a change of

address.

Notification of Change of Address

Address changes should be sent promptly to

the Academy office. Such notifications should

show both old and new addresses and zip-code

numbers, where applicable.

Published quarterly in March, June, September, and December of each year by the

Washington Academy of Sciences, 1101 N. Highland Street, Arlington, VA 22201.

Second-class postage paid at Arlington, VA, and additional mailing offices.

Journal of the Washington Academy of Sciences,

Volume 81, Number 2, Pages 61-73, June 1991

The Gap Between Experimental And

Psychometric Orientations

Anne Anastasi

Fordham University

ABSTRACT

With the increasing specialization of psychological research, there is danger that investiga-

tors may lose contact with other relevant specialties, as illustrated by the widening gap

between experimental and psychometric orientations. Because variability—both within and

between individuals—is an essential feature of human behavior, the experimental psycholo-

gist needs to include a statistical approach in both experimental design and data analysis.

Notable examples of the merging of statistical and experimental methodology include analy-

sis of variance, structural equation modeling, and factor analysis. A brief overview of the

development and current status of these three procedures is provided.

One of the inevitable consequences of the rapid growth of psychology is an

increasing specialization in the training and functioning of psychologists. Spe-

cialization is obviously needed if one is to attain sufficient depth of knowledge

and expertise to make an effective contribution to either research or practice. At

the same time, specialization creates hazards which are becoming increasingly

apparent in psychology. There is the likelihood of losing contact with neighbor-

ing specialties that may be relevant to one’s work. And there is the danger that

the methodological focus becomes too circumscribed to provide an adequate

picture of so complex a phenomenon as human behavior. As a result, one’s data

may be incomplete and one’s conclusions incorrect.

This weakening of contact is what I see in the widening gap between the

psychometric and the experimental orientations in psychological research. As

long ago as 1966, in my address to APA Division 5 (Division of Evaluation and

Measurement), my main thesis was that psychological testing was becoming

dissociated from the mainstream of contemporary psychology (Anastasi, 1967).

Psychologists specializing in psychometrics had been concentrating more and

more on elaborating and refining the techniques of test construction, while

losing sight of the behavior they set out to measure. As a result, outdated inter-

pretations of test performance remained insulated from the impact of the grow-

62 ANASTASI

ing knowledge base provided by behavioral research. I argued then—and I have

continued to argue—that much of the criticism of testing stems from the isola-

tion of psychometrics from other relevant areas of psychology. Because I am

now talking to members of Division 21 and the Human Factors Society, I shall

look at the other side and try to see what the gap may mean for experimental

psychology.

The Psychometric Orientation

First, what do I mean by the psychometric orientation? I define the field of

psychometrics methodologically, to include psychological testing and statistical

analysis, and substantively, as covering the nature and sources of individual

differences (1.e., differential psychology). The psychometric orientation is essen-

tially statistical, because it takes variability as a basic phenomenon to be investi-

gated in its own right. Statistical methodology was originally developed as a

means of dealing with variability. It is needed in analyzing any data on human

behavior, because human behavior presents extensive variability in many

forms.

Not only is there wide variability in the responses of different individuals, but

there is also variability among specific response indicators, such as the items

making up a test. The latter kind of variability is typically assessed by measures

of a test’s internal consistency, as with the Kuder-Richardson reliability or

various split-half reliability coefiicients. There are also random differences in

response over time, which reflect the many variables that may affect an individ-

ual’s performance over short time intervals, including both changes in the indi-

vidual’s physical and psychological condition and changes in the external situa-

tion. The usual retest reliability refers to this kind of random variability. These

random changes should be distinguished from systematic, progressive and cu-

mulative changes, such as would result from learning and other lasting changes

in the person.

It is noteworthy that in the early days of experimental psychology all forms of

random variability were regarded as “errors” insofar as they tended to restrict

the applicability of general findings about human behavior—which was what

most nineteenth century psychologists were looking for. Hence we have in-

herited the error nomenclature, in standard error, error variance, and so forth.

To the psychometrician, however, these are not errors that could be prevented

by improved procedures. They are inalienable facts of behavior to be taken into

account.

Variability will not go away if you ignore it. If ignored, it remains to distort

EXPERIMENTAL-PSYCHOMETRIC ORIENTATIONS 63

your results and probably lead to a wrong conclusion. The psychometric orienta-

tion concerns not only the sorts of variability assessed by sampling errors and

errors of measurement, which can be handled by checking the statistical signifi-

cance of mean differences, finding the confidence interval of a score, or looking

up the significance of a correlation coefficient. It concerns also the variability

among persons, as represented by the total distribution of performance mea-

sures. Thus it is not only the standard error of a mean that we are concerned

with, but the standard deviation of the whole distribution. We may want to

know what range of individual variation would be appropriate to consider for a

particular purpose. Should it be 1.96 times the standard deviation on both sides

of the mean, which would cover approximately the middle 95% of the group? Or

should it be a narrower or a wider range? Such questions deal with real, objec-

tively assessed individual differences in performance. They are as close as you

can get to empirically observed facts.

Another example is provided by the correlation coefficient, which we all take

for granted. The reason we need correlation coefficients is because the relation

between any two variables varies from person to person. If the same relation

between two variables held for all persons, such that each person occupied the

same relative position in both variables, the correlation between the two vari-

ables would be +1.00 and we would not need to compute it. But in reality, one

individual may be high in both variables, another high in one and mediocre in

the other, still another above average in one and below average in the other, and

so on. The size of the correlation coefficient tells us how much individuals differ

in the way the two variables are related. Had we administered a whole battery of

tests, a score profile for each person would reflect this kind of individual differ-

ence. Such score profiles could prove especially useful in assigning persons to

different training programs, different jobs, or different experimental treatments

of any sort.

Some Notable Mergers of Statistical and Experimental Methodology

Intrinsically, there is nothing in either statistical or experimental procedures

that keeps them apart. They are two highly compatible aspects of scientific

method. In this connection, there is a favorite story, which is not apocryphal; it

tells of an experience that has probably occurred many times to young re-

searchers in psychology. The investigator in the story has been busy collecting an

extensive body of data in the effort to test one or more hypotheses. Faced with an

overabundance of numerical data, the investigator decides to consult a well-

known statistician for expert advice on how to analyze the data. The statistician

64 ANASTASI

tries to do the best that he or she can to help, but with a sad shake of the head,

remarks, “I could have been of real help if you had contacted me before you

gathered your data.” This, of course, is the question of experimental design,

which is closely linked to statistical considerations.

Analysis of Variance

A clear example of this linkage is analysis of variance, familiarly known as

ANOVA. R. A. Fisher (1925), who introduced ANOVA in the 1920’s, was Chief

Statistician at the Rothamsted Experimental Station, a British agricultural re-

search center. In Fisher’s own writing and lecturing, considerations of experi-

mental design constituted a major portion of his treatment of ANOVA. When

ANOVA was subsequently adopted by psychological researchers, it was used

chiefly to assign individuals to groups in order to identify the effect of specific

experimental variables. The simplest application involves the comparison of

control and experimental groups, in order to assess the effect of the experimental

variable on performance. A major contribution of ANOVA, however, was to

permit the simultaneous study of the effects of several independent variables,

including both experimentally manipulated and naturally occurring variables.

For instance, we could assess the effect of sex and socioeconomic level on me-

chanical aptitude test scores. With two independent variables, we can measure

not only the total effect of each variable, but also the interaction between the

two. Thus the results might show that in high socioeconomic levels there was no

significant sex difference in mechanical aptitude, while in low socioeconomic

levels the men performed significantly better than the women.

To take an example with experimentally manipulated variables, suppose we

wish to evaluate two training methods for instructing trainees in the perfor-

mance of an occupational test. On the basis of scores on a general aptitude test,!

we divide the trainees into three groups (high, medium, and low), with 40

persons in each group. Of the 40 high-aptitude persons in Group 1|, 20 are taught

by Method A and 20 by Method B. The 40 in the middle-scoring group are

similarly assigned to the two methods, and the same assignment is followed for

the low-aptitude group. If we compare overall mean task performance of the

three groups after training, we would probably find that the high-aptitude group

performs best and the low-aptitude group poorest. But when we compare the

two training programs within each aptitude group, we may find that Method B is

more effective than Method A for the high-aptitude trainees, while the reverse is

true for the two lower-aptitude groups. This would represent a significant inter-

' Such as the College Board’s Scholastic Aptitude Test (SAT) or the AFQT composite of the Armed Services

Vocational Aptitude Battery (ASVAB).

EXPERIMENTAL-PSYCHOMETRIC ORIENTATIONS 65

action between training method and aptitude level, a finding for which there is

considerable empirical evidence (see, e.g., Berliner & Cahen, 1973; Bialek, Tay-

lor, & Hauke, 1973; Cooper, 1974; Cronbach, 1975; Cronbach & Snow, 1977;

Fox, Taylor, & Caylor, 1969).

As we add more independent variables to our experimental design, we can

compute three-way, four-way and progressively higher-order interactions; but as

the number of variables within a single study increases, we need very large

samples in order to demonstrate statistically significant interactions. For practi-

cal purposes, available basic research on the interaction of specific variables can

suggest guidelines in the construction of two or three trial programs, which can

then be empirically tested as whole programs with groups that differ conspicu-

ously in personal characteristics (Bialek, Taylor, & Hauke, 1973).

The concept of interaction among variables provides the theoretical rationale

for tradeoffs in optimizing personnel procedures. Job performance is being

viewed increasingly within a comprehensive system that integrates the effects of

personnel selection and classification, training procedures, and job variables

upon individual performance. Such a system covers multiple aptitude and per-

sonality variables of the worker, features of the training program, nature of job

activities, equipment design, physical characteristics of the work environment,

and such organizational variables as incentives and supervisory techniques

(Campbell, Dunnette, Lawler, & Weick, 1970; Uhlaner, 1972; see also Anastasi,

1979, pp. 6-9, 46-49, 103-104, 460-461).

Structural Equation Modeling

A second example of close linkage between statistical and experimental meth-

odology is of more recent origin. Technically known as structural equation

modeling, it is a close relative of path analysis; and both are informally called

causal modeling.” Unfortunately, there is a confusing variety of terminology in

publications on this procedure. But let us see what structural equation modeling

is designed to do. We all remember learning in elementary statistics that correla-

tion does not indicate causation. If A and B are correlated, it may mean that A

causes B, or B causes A, or both are caused by a third variable C, which 1s

correlated with both. A familiar example is spurious age correlation. In a group

* For strict accuracy, many psychologists avoid the term “cause,” aware that their empirical data demon-

strate only regularity of succession between events, and such observed regularity is not necessarily absolute but

is inferred from frequency of succession. Moreover, the investigator can rarely identify the mechanism

whereby A leads to B or the intervening chain of events that brings about this sequence. In order to avoid

philosophical arguments about causation, psychologists prefer more neutral expressions, such as independent

and dependent variables, or the statements that A determines, influences, or affects B to a specified degree—if

the latter can be estimated. Often the term “causal” may be used loosely, with the assumption that its limita-

tions are understood (see, e.g., James, Mulaik, & Brett, 1982, chap. 1).

66 ANASTASI

of children ranging in age from 5 to 10, we are likely to find a high correlation

between height and knowledge of arithmetic; but we cannot conclude that either

affects the other.

In the effort to disentangle causal relations, psychologists in the 1960s began

to work with cross-lagged experimental designs (see Campbell & Stanley, 1966;

Cook & Campbell, 1976, pp. 284-293). Thus, if you wanted to investigate the

relative influence of attitude and ability on an individual’s performance, you

could, for instance, obtain measures of attitude toward math and ability in math

at two points in time. Then you could compute the cross-lagged correlation

between math attitude at Time | and math performance at Time 2, and com-

pare it with the correlation between math performance at Time | and math

attitude at Time 2. For a few years this seemed a neat way to assess the effects of

two variables on each other, and several published studies reported results ob-

tained by this procedure.

Before long, however, careful logical and statistical analyses revealed serious

weaknesses in the use of cross-lagged correlations. Although the basic cross-

lagged design was excellent, the use of simple, zero-order correlations was likely

to lead to distorted results and incorrect conclusions about causal relations

(Rogosa, 1980). Among the various sources of error in this procedure are the

failure to take into account, first, intercorrelations between both initial and

subsequent variables; second, the reliability of the variables and their stability

over time; and third, the possible contribution of unmeasured variables. Struc-

tural equation modeling provides more sophisticated attempts to avoid such

difficulties.

Essentially, structural equation modeling uses regression equations to predict

the dependent from the independent variables in cross-lagged or other causal

models. In this procedure, partial correlations are used in finding the regression

coefficients, thereby utilizing all intercorrelations among the variables; both

measurement and sampling errors are taken into account; and some provision 1s

made at least to recognize the possibility of additional, unmeasured causal vari-

ables (Bentler, 1988; James, Mulaik, & Brett, 1982; Loehlin, 1987; Rogosa,

1979). Specifically, the first step is to design a model of the hypothesized causal

relations to be tested. It is important that this model be based on thorough

familiarity with existing knowledge about the variables and situation under

investigation. The hypothesized relations should have a sound theoretical ratio-

nale. Special attention should be given to possible unmeasured variables that

may themselves be correlated with measured variables in the study.

Another noteworthy feature of structural equation modeling is that causal

relations are typically computed between constructs, rather than between 1so-

lated measured variables. For instance, to assess a learner’s attitude toward

EXPERIMENTAL-PSYCHOMETRIC ORIENTATIONS 67

math, several indicators could be used, such as measures of interest, goal orienta-

tion, self-concept of math aptitude, and other relevant affective variables. The

common variance among these indicators would then define a construct of the

individual’s attitude toward math, which can itself be related to subsequent

math achievement. The use of constructs provides more stable and reliable

estimates, in which the error and specific variances of the separate indicators

cancel out.

The actual testing of the model is accomplished by solving a set of simulta-

neous linear regression equations. In causal modeling, there are usually more

equations than unknowns, which permits solution for several alternative mod-

els; each model can be compared with the original, empirical correlation matrix

for goodness of fit. In this process, models can be modified and the best fitting

model identified. Usually, the goodness of fit is tested by Chi Square (x7), al-

though in this situation the larger the x? (and hence the higher the probability

that the differences did arise by chance), the closer the fit. This fit is often

expressed as a ratio between the amount of covariance explained by the model

and the total amount of covariance present in the original data.

The actual computations are carried out by available and widely used com-

puter programs.’ But such programs do not preclude the need for thorough

content knowledge; they cannot do the thinking for you. It should also be noted

that the procedure I have described is a highly simplified version of one variant.

There are several alternative approaches, as well as individual modifications and

procedural refinements currently under consideration (see, e.g., Anderson &

Gerbing, 1988; Bentler, 1990; Breckler, 1990; James, 1980; La Du & Tanaka,

1989; Mulaik et al., 1989). While still in a state of development, however,

structural equation modeling is a promising procedure for combining experi-

mental and statistical approaches. It has already been widely applied to prob-

lems in such areas as developmental, personality and social, industrial, and

educational psychology.*

Factor Analysis

A third example of the merging of statistical and psychometric approaches is

provided by the use of factor analysis in research on the organization of human

behavior. The principal object of the technique of factor analysis is to simplify

the description of data by reducing the number of necessary variables, or dimen-

sions. Thus, if we find that five factors are sufficient to account for all the

> Such as LISREL (Hayduk, 1988; Joreskog & Sdrbom, 1986, 1989) and EQS (Bentler, 1985).

* For examples of well designed studies, see Graves & Powell (1988), James & James (1989), Parkerson,

Lomax, Schiller, & Walberg (1984), Shavelsen & Bolus (1982).

68 ANASTASI

common variance in a battery of 20 tests, we can for most purposes substitute 5

scores for the original 20 without losing any essential information. The usual

practice in this context would be to retain from among the original tests those

providing the best measures of each of the factors.

All techniques of factor analysis begin with a complete matrix of intercorrela-

tions among a set of variables, such as tests, and end with a factor matrix, that is,

a table showing the weight or loading of each factor in each test. Several different

methods for analyzing a set of variables into common factors have been derived.

As early as 1901, Pearson pointed the way for this type of analysis, and Spear-

man (1904, 1927) developed a precursor of modern factor analysis. Later, Ho-

telling (1933) and Thurstone (1947) in the United States and Burt (1941) in

England did much to advance the method. Alternative procedures, modifica-

tions, and refinements were developed by many others. The availability of high-

speed computers led to the adoption of more mathematically precise and labori-

ous techniques. Although differing in their initial postulates, most of these

methods yield similar results.°

Today, available computer programs carry out all the necessary steps in the

factor analysis of a battery of tests (or other measured variables).° Nevertheless,

familiarity with certain major steps in the procedure helps to understand pub-

lished reports of factor analytic research in psychology. Factors can be repre-

sented geometrically as reference axes in terms of which the factor loading of

each test can be plotted. The tests with high loading on any one factor will tend

to cluster in one region. It should be noted that the position of the reference axes

is not fixed by the data, but is determined by the method of factor analysis used.

The original correlation table determines only the position of the tests in rela-

tion to each other. The same points can be plotted with the reference axes in any

position. For this reason, factor analysts usually rotate axes until they obtain the

most satisfactory and easily interpretable pattern. This is a legitimate procedure,

somewhat analogous to measuring longitude from, let us say, Chicago rather

than Greenwich.

Thurstone introduced two criteria for the rotation of axes, which are still

proving widely useful. The first, positive manifold, requires the rotation of axes

to such a position as to eliminate negative weights. This condition applies partic-

ularly to aptitude tests, where a negative loading would imply that the higher the

individual stands on the particular factor, the poorer will be his or her perfor-

> For a brief introduction to the concepts and psychometric uses of factor analysis, see Anastasi (1988, pp.

374-390); specific procedures are described in several elementary texts on factor analysis, such as Gorsuch

(1983) and, at a more advanced level, Harman (1976).

° For simplicity, I shall henceforth use the term “‘test’”’ instead of “variable” because most applications of

factor analysis in psychology have dealt with tests, although all statements would apply to any kind of mea-

sured variable.

EXPERIMENTAL-PSYCHOMETRIC ORIENTATIONS 69

mance on the test. The second criterion, simple structure, means essentially that

each test shall have loadings on as few factors as possible. Both of these criteria

are designed to yield factors that can be most readily and unambiguously inter-

preted. If a test has a high loading on a single factor and no significant loading on

any other, we can learn something about the nature of the factor by examining

the content of the test. If, instead, the test has moderate to low loadings on six

factors, it can tell us little about the nature of any one of these factors.

A useful distinction to note is that between correlated and uncorrelated fac-

tors. In the geometric representation of factors, uncorrelated factors are repre-

sented by orthogonal axes, that are at right angles to each other; correlated

factors are represented by oblique axes, the angle between pairs of axes corre-

sponding to the correlation between those factors. When the factors are them-

selves correlated, they can be further analyzed to yield broader, higher-order

factors. Such analyses have led to the concept of a hierarchy of factors, from the

narrowest and most specific variables at the base of the hierarchy to progres-

sively broader factors at successive levels.

A single, general factor, representing what is common to a whole battery of

tests, was first publicized as a “‘g”’ factor by Spearman (1927). The symbol g has

become linked to Spearman’s early theory of intelligence, which proposed that a

single common factor could be identified across all batteries of cognitive tests,

and this factor was regarded as general intelligence. This use of g has survived in

loose, popular discussions of intelligence and has occasioned much confusion

and misunderstanding. Such an adverse effect of factor analytic research illus-

trates one of the consequences of the early weakening of contact between factor

analytic researchers and the mainstream of psychology.

From Primary Mental Abilities to Trait Formation

Factor analysis is of particular interest because it represents what could be

described as an “emerging merger’ between psychometric and experimental

orientations. In its early applications to psychological research, during the first

half of the twentieth century, factor analysis became increasingly dissociated

from developments in other relevant areas of psychology. Its techniques came to

be used more and more by test construction specialists, who were insufficiently

aware of what was happening in other specialties. This period of isolation had

deleterious effects on both the design of factor analytic research and on the

interpretation of results.

The most conspicuous effect was a growing proliferation of factors. The one g

factor identified by Spearman (1927) through the simple statistical techniques

70 ANASTASI

that preceded modern factor analysis is still regarded as of prime importance by

some psychometricians. Soon, however, multiple factor theories were vigor-

ously defended by several researchers, especially in the United States. For some

time, Thurstone’s primary mental abilities dominated the scene. Some dozen

broad group factors were identified by Thurstone and others between the 1930s

and 1950s (Kelley, 1928, 1935; Thurstone, 1938; Thurstone & Thurstone, 1941;

see also, Anastasi, 1988, pp. 381-390; French, 1951). Among the best known of

these factors are verbal comprehension, arithmetic reasoning, spatial orienta-

tion, perceptual speed, and associative memory.

Other factor analysts then began to identify narrower group factors within the

previously identified broad areas. Some, such as Guilford, suggested simple

models for organizing the multitude of rapidly proliferating factors. Guilford’s

structure-of-intellect model, incorporating the findings of more than 20 years of

research, provided a three-dimensional, box-like schema that made room for

120 or more differentiable abilities (Guilford, 1967; Guilford & Hoepfner,

1971). In commenting on this large number of abilities, Guilford argued that

human nature is exceedingly complex and a few factors could not be expected to

describe it adequately.

Controversies among the competing factor theorists were lively and pro-

longed. Each protagonist claimed to be seeking—and at least partially finding—

the basic units of human intelligence, or thinking, or cognition. Gradually the

controversies were dissipated by the proposal of comprehensive hierarchical

models (Burt, 1949; Gustafson, 1989; Humphreys, 1962; Vernon, 1960). At the

same time, there were methodological developments demonstrating that the

different factor analytic solutions are mathematically equivalent and transpos-

able from one to another (Harman, 1976, chap. 15; Schmid & Leiman, 1957).

It is now coming to be widely recognized that individual differences in intellec-

tual functioning can be described at different levels of generality, from a single

common factor within a whole set of variables, through increasingly narrower

factors at successively lower levels. For different practical purposes, one or an-

other level of this hierarchy is most appropriate. The trait hierarchy thus pro-

vides a comprehensive theoretical model that permits practical flexibility in test

development and use for specific purposes. For example, if we want a test to aid

in college admission decisions, broad factors such as verbal comprehension and

numerical reasoning represent the most appropriate level; but to assess the

prerequisite skills and knowledge for assignment to such occupational special-

ties as airline pilot, navigator, or air traffic controller, more narrowly defined

sensorimotor and perceptual factors may be needed. In this connection, it may

be of interest that Thurstone derived his primary mental abilities largely from

data on college students, whereas Guilford began his project while developing

EXPERIMENTAL-PSYCHOMETRIC ORIENTATIONS 71

tests for the Air Force in World War II. Furthermore, Spearman identified his g

factor in studies conducted largely with school children, among whom abilities

are less differentiated than at older ages or at higher educational levels (Anastasi,

1948, 1970, 1983; Spearman, 1927).

The concentration of the early factor analysts on statistical techniques, with

the neglect of psychological content and context, led to a second and more

serious limitation. In interpreting the results of factor analysis, the early re-

searchers treated the traits they identified—at whichever level of the hierarchy

—as fixed, underlying causal entities, which the individual somehow possessed,

which were predetermined by heredity, and which in turn accounted for that

individual’s observable performance. Although still surviving in popular mis-

conceptions about human abilities and about test scores, this early view has been

seriously challenged by a growing accumulation of findings in such areas as

developmental psychology, behavior genetics, the experimental psychology of

learning, and cross-cultural research on the composition of intelligence.’

More and more, researchers are now concerned, not with the discovery of

rigidly fixed underlying human traits, but rather with the process of trait forma-

tion (see Anastasi, 1970, 1983, 1986, 1990, August). This is a continuing pro-

cess, occurring throughout the individual’s life-long learning history. What is it

in an individual’s learning history that led to the behavioral consistencies (or

correlations) through which trait constructs are identified? One proposed mech-

anism involves the contiguity or co-occurrence of learning experiences, as illus-

trated by the development of a broad verbal factor running through all activities

learned in school. Another proposed mechanism of trait formation is differen-

tial transfer of training. The breadth of the transfer effect determines whether

the resulting trait is broad, like verbal comprehension, or narrow, like a special-

ized perceptual skill.

Concluding Comment

When dealing with human behavior, in any form and from any angle, you will

encounter variability—extensive and pervasive variability. If you ignore this

variability, it will come back to haunt you in the form of incorrect conclusions in

basic research and wrong decisions in applied research and practice. Equally

serious are the consequences of becoming totally immersed in the statistics of

variability, while ignoring the psychological content and context of the behavior

itself. The experimental and psychometric approaches are not only intrinsically

’ For more detailed treatment and references, see Anastasi (1983, 1986, 1990, August).

1p: ANASTASI

compatible but also mutually interdependent. Each depends upon the other for

effective functioning in research design, in data analysis, and in the interpreta-

tion of results.

References

Anastasi, A. (1948). The nature of psychological “traits.” Psychological Review, 55, 127-138.

Anastasi, A. (1967). Psychology, PEMD OEE and psychological testing. American Psychologist, 22, 297-

306.

Anastasi, A. (1970). On the formation of psychological traits. American Psychologist, 25, 899-910.

Anastasi, A. (1979). Fields of applied psychology (2nd ed.). New York: McGraw-Hill.

Anastasi, A. (1983). Evolving trait concepts. American Psychologist, 38, 175-184.

Anastasi, A. (1986). Experiential structuring of psychological traits. Developmental Review, 6, 181-202.

Anastasi, A. (1988). Psychological testing (6th ed.). New York: Macmillan.

Anastasi, A. (1990, August). Are there unifying trends in the psychologies of 1990? Invited address delivered at

Annual Convention of the American Psychological Association, Boston, MA.

Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recom-

mended two-step approach. Psychological Bulletin, 103, 411-423.

Bentler, P. M. (1985). Theory and implementation of EQS: A structural equations program. Los Angeles:

BMDP Statistical Software.

Bentler, P. M. (1988). Causal modeling via structural equation modeling. In J. R. Nesselroade & R. B. Cattell

(Eds.), Handbook of multivariate experimental psychology (2nd ed., pp. 319-335). New York: Plenum

Press.

Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238-246.

Berliner, D. C., & Cahen, L. S. (1973). Trait-treatment interactions and learning. In F. N. Kerlinger (Ed.),

Review of research in education (pp. 58-94). Itasca, IL: Peacock.

Bialek, H. M., Taylor, J. E., & Hauke, R. N. (1973). Instructional strategies for training men of high and low

aptitude (HumRRO Tech. Rep. 73-10). Alexandria, VA: Human Resources Research Organization.

Breckler, S. J. (1990). Applications of covariance structure modeling in psychology: Cause for concern?

Psychological Bulletin, 107, 260-273.

Burt, C. (1941). The factors of the mind: An introduction to factor analysis in psychology. New York: Mac-

millan.

Burt, C. (1949). The structure of the mind: A review of the results of factor analysis. British Journal of

Educational Psychology, 19, 110-111, 176-199.

Campbell, D. T., & Stanley, J.C. (1966). Experimental and quasi-experimental designs for eres Chicago:

Rand McNally.

Campbell, J. P., Dunnette, M. D., Lawler, E. E., III, & Weick, K. E., Jr. (1970). Managerial behavior,

performance, and effectiveness. New York: McGraw-Hill.

Cook, T. D., & Campbell, D. T. (1976). The design and conduct of quasi-experiments and true experiments in

field settings. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology (pp. 223-

326). Chicago: Rand-McNally. (Original work republished by Wiley, New York, 1983).

Cooper, R. (1974). High aptitude, low aptitude—training must fit the man. Training, 11(11), 42-43, 58-60.

Cronbach, L. J. (1975). Beyond the two disciplines of scientific psychology. American Psychologist, 30, 116-

Wile

Cronbach, L. J., & Snow, R. E. (1977). Aptitudes and instructional methods: A handbook for research on

interactions. New Y ork: Irvington.

Fisher, R. A. (1925). Statistical methods for research workers. Edinburgh (Scotland): Oliver & Boyd. (4th ed.,

1932).

Fox, W. L., Taylor, J. E., & Caylor, J. S. (1969). Aptitude level and the acquisition of skills and knowledges in a

variety of military training tasks (HumRRO Tech. Rep. 69-6). Alexandria, VA: Human Resources Research

Organization.

French, J. W. (1951). The description of aptitude and achievement tests in terms of rotated factors. Psychomet-

ric Monographs, No. 5.

Gorsuch, R. L. (1983). Factor analysis (2nd ed.). Hillsdale, NJ: Erlbaum.

Graves, L. M., & Powell, G. N. (1988). An investigation of sex discrimination in recruiters’ evaluations of

actual applications. Journal of Applied Psychology, 73, 20-29.

Guilford, J. P. (1967). The nature of human intelligence. New York: McGraw-Hill.

Guilford, J. P., & Hoepfner, R. (1971). The analysis of intelligence. New York: McGraw-Hill.

Gustafson, J-E. (1989). Broad and narrow abilities in research in learning and instruction. In R. Kanfer, P. L.

EXPERIMENTAL-PSYCHOMETRIC ORIENTATIONS 73

Ackerman, & R. Cudack (Eds.), Abilities, motivation, and methodology (pp. 203-237). Hillsdale, NJ: Erl-

baum.

Harman, H. H. (1976). Modern factor analysis (3rd ed.). Chicago: University of Chicago Press.

Hayduk, L. A. (1988). Structural equation modeling with LISREL: Essentials and advances. Baltimore, MD:

Johns Hopkins University Press.

Hotelling, H. (1933). Analysis of a complex of statistical variables into principal components. Journal of

Educational Psychology, 24, 417-441, 498-520.

Humphreys, L. G. (1962). The organization of human abilities. American Psychologist, 17, 475-483.

James, L. A., & James, L. R. (1989). Integrating work environment perceptions: Explorations into the mea-

surement of meaning. Journal of Applied Psychology, 74, 739-751.

James, L. R. (1980). The unmeasured variable problem in path analysis. Journal of Applied Psychology, 65,

415-421.

James, L. R., Mulaik, S. A., & Brett, J. M. (1982). Causal analysis: Assumptions, models, and data. Beverly

Hills, CA: Sage Publications.

Jéreskog, K. G., & Sérbom, D. (1986). LISREL: Analysis of linear structural relationships by maximum

likelihood, instrumental variables, and least squares methods (4th ed.). Mooresville, IN: Scientific Software.

Joreskog, K. G., & Sérbom, D. (1989). LISREL 7 User’s guide. Mooresville, IN: Scientific Software.

Kelley, T. L. (1928). Crossroads in the mind of man: A study of differentiable mental abilities. Stanford, CA:

Stanford University Press.

Kelley, T. L. (1935). Essential traits of mental life. Cambridge, MA: Harvard University Press.

La Du, T. J., & Tanaka, J. S. (1989). Influence of sample size, estimation method, and model specifications on

goodness-of-fit assessments in structural equation models. Journal of Applied Psychology, 74, 625-635.

Loehlin, J. C. (1987). Latent variable models: An introduction to factor, path, and structural analysis. Hills-

dale, NJ: Erlbaum.

Mulaik, S. A., James, L. R., Van Alstine, J., Bennett, N., Lind, S., & Stilwell, C. D. (1989). Evaluation of

goodness-of-fit indices for structural equation models. Psychological Bulletin, 105, 430-445.

Parkerson, J. A., Lomax, R. G., Schiller, D. P., & Walberg, H. J. (1984). Exploring causal models of educa-

tional achievement. Journal of Educational Psychology, 76, 638-646.

Pearson, K. (1901). On lines and planes of closest fit to systems of points in space. Philosophical Magazine

(Series 6), 2, 559-572.

Rogosa, D. (1979). Causal models in longitudinal research: Rationale, formulation, and interpretation. In J. R.

Nesselroade & P. B. Baltes (Eds.), Longitudinal research in the study of behavior development (pp. 263-302).

Orlando, FL: Academic Press. ;

Rogosa, D. (1980). A critique of cross-lagged correlation. Psychological Bulletin, 88, 245-258.

Schmid, J., & Leiman, J. (1957). The development of hierarchical factor solutions. Psychometrika, 22, 53-61.

Shavelson, R. J., & Bolus, R. (1982). Self-concept: The interplay of theory and methods. Journal of Educa-

tional Psychology, 74, 3-17.

Spearman, C. (1904). “General intelligence” objectively determined and measured. American Journal of

Psychology, 15, 201-293.

Spearman, C. (1927). The abilities of man. New York: Macmillan.

Thurstone, L. L. (1938). Primary mental abilities. Psychometric Monographs, No. 1.

Thurstone, L. L. (1947). Multiple factor analysis. Chicago: University of Chicago Press.

Thurstone, L. L., & Thurstone, T. G. (1941). Factorial studies of intelligence. Psychometric Monographs,

No. 2.

_Uhlaner, J. E. (1972). Human performance effectiveness and the systems measurement bed. Journal of

Applied Psychology, 56, 202-210.

Vernon, P. E. (1960). The structure of human abilities (rev. ed.). London: Methuen.

Journal of the Washington Academy of Sciences,

Volume 81, Number 2, Pages 74-88, June 1991

Serial Averaging in Performance-Test

Theory: An Interim Report

Marshall B. Jones

The Pennsylvania State University College of Medicine, Hershey, PA

ABSTRACT

Performance tests sample what a person can do (remember, track, aim, detect, recognize,

and so on); the unit of analysis is a trial, not an item as in knowledge tests; and the person

taking the test usually has at least a rough idea as to how well or poorly he or she is doing. Asa

result, practice effects cannot be ignored in a performance test, even when that test is very

short. The present paper outlines an approach to performance tests that treats them as tasks

to be learned, with a focus on reliability as a function of test length (number of trials adminis-

tered). One conclusion is that low reliability in some performance tests is not corrigible by

increasing test length. Results are presented for the Army Project-A, computer-adminis-

tered tests.

INTRODUCTION

The Theoretical Problem of Performance Testing

The distinction between knowledge and performance testing turns on what

one is trying to measure. A knowledge test samples what a person Knows, a

performance test what he or she can do. Plainly, this distinction is not absolute.

A mathematics test, for example, may involve not only what someone knows

but also what he or she can do with that knowledge. A memory task may be

facilitated if a person has seen an unusual symbol before and knows what it is,

say, the Greek letter omega. Nevertheless, most tests fall lopsidedly into one

category or the other.

In a knowledge test a testee does not usually know whether he or she is right or

wrong; hence, practice effects are limited to auxiliary aspects of the test (test-tak-

ing skills) and, while they exist, are not large (Messick & Jungblut, 1981; Wing,

1980). In a performance test, however, it is usually not possible to prevent a

person from obtaining some idea as to how well or poorly he or she is doing. The

result is that testees tend to do better on a test the more times it is administered

to them (Bittner et al., 1983; Kennedy et al., 1981). In effect, each test adminis-

tration becomes a trial of practice.

PERFORMANCE-TEST THEORY 75

Psychometric theory is based on knowledge tests. The unit of analysis is an

item and the order of administering the items is arbitrary. In performance test-

ing, however, the unit of analysis is a trial and order of administration is not only

nonarbitrary but often the only thing that distinguishes one trial from another.

In a knowledge test it is not unreasonable to suppose that mean performance

and inter-item correlations are independent of order of administration. In a

performance test it is. Typically, performance improves with practice and inter-

trial correlations fall into a definite pattern as a function of order: the super-

diagonal form (Jones, 1962).

The consequences of these differences for theory are drastic. It has long been

known, for example, that inter-trial correlations, unlike inter-item correlations,

may yield spurious results when subjected to conventional factor analysis

(Humphreys, 1960). Also, the familiar formulae for adjusting reliability and

validity for test length assume that average inter-item (inter-trial) correlation, 7,

does not change with test length. In a superdiagonal form, as will be seen below, 7

definitely does change with test length. As a result, the Spearman-Brown and

related formulae (Gulliksen, 1950) have to be reworked and reinterpreted if

their use in performance testing is to be helpful and not misinformative.

The Practical Problem of Performance Testing

During the Second World War performance testing based on electromechani-

cal apparatus (rotary pursuit, complex coordination, two-hand tracking, and the

like) was widely and successfully used in military selection, especially for pilot

training (Melton, 1947). The equipment, however, was heavy, bulky, difficult to

maintain, and more difficult to replace. By the late 1950s all three military

services had abandoned performance testing in favor of paper-and-pencil tests

exclusively. Then in the late 1970s the advent of microcomputers reopened the

possibility of performance testing, this time with equipment that occupied little

space, did not break down frequently, and was easily replaced when it did. At the

- Same time experimental psychology was undergoing a revolution of its own, as

the discipline’s central focus shifted from learning theory to cognition and infor-

mation-processing. The joint effect of these two developments was a new genera-

tion of cognitively oriented, microcomputer-based performance tests. The com-

puter-administered tests in Project A are cases in point (Eaton, Hanser, &

Shields, 1986; Peterson, 1987).

Unfortunately, all has not been clear sailing for this new generation of perfor-

mance tests. The most serious problem has been that many tests have low

reliabilities (Kyllonen, 1985). Predictive validities against real-world criteria are

still sparse, but it seems likely that oftentimes they will also be low. An appro-

priate response to these difficulties involves more than making and trying out

76 JONES

Table 1.—Hypothetical Correlations, with the Average Correlation (r,) and Reliability (R;) up to a Given

Trial

Tnial

Tnial 1 2 3 4 5 6 7

1 — .80 65 50 #5) .20 05

2 — .80 .65 50 135 .20

3 — .80 65 50 ae hs)

4 — .80 65 50

5 — .80 65

6 ‘ — .80

7 if}

ie — .80 45 IQ .65 .60 55

R;* .800 .889 .900 .903 .902 .900 895

* R; (1 = 2) is calculated from 7, using the Spearman-Brown formula; when i = 1, R, =.

new tests. What is needed 1s a theory of performance tests, that is, an approach to

test construction and validation that recognizes and capitalizes upon the dis-

tinctive properties of performance tests.

Approach

Superdiagonal form is one of the best established regularities in human learn-

ing (Jones, 1962, 1969). It refers to the essentially universal tendency for trials of

practice to correlate more strongly the closer they are together in the practice

sequence. Table | presents a hypothetical example. The correlation between

neighboring trials is .80. When there is one intervening trial, the correlation

drops to .65. When two trials intervene, the correlation drops to .50. The weak-

est correlation is between the first and last trials in the sequence, in the example,

.05. In a motor-skills experiment, where each data point typically represents as

much as 20 min. of practice, the superdiagonal pattern is always present and, in

large samples, usually quite regular. In correlations among individual trials of

practice, as in performance testing, the pattern may be very irregular. Almost

always, however, if correlations are averaged over groups of consecutive trials,

the pattern can still be seen.

Table 1 illustrates another point, this one directly relevant to performance

testing. In conventional test theory the Spearman-Brown (S-B) formula (Gullik-

sen, 1950) states that the reliability of a test 7 units in length

a: iR,

oe a | yee,

where R, is the reliability of a test of unit length. When i = 2, R, is taken as the

average correlation among the / units, that is, 7,. The first row at the bottom of

PERFORMANCE-TEST THEORY at

the table shows this average correlation for the first two trials, the first three, out

to all seven trials. As is clear from the table, these averages decrease from the first

to the last trial. Since the correlations decrease along any row to the right, each

new trial adds to the average a column of correlations lower than those already

in it; hence 7, drops a notch.!

Low reliability in a knowledge test is corrigible. It may be laborious to do, but

in principle one can always lengthen the test, while maintaining the same aver-

age inter-item correlation, and thereby improve its reliability. In a performance

test, however, 7, does not remain the same as the test is lengthened; it decreases.

The bottom row in Table 1 gives R; as calculated by the S-B formula for 1 =

1,..., 7. As 7 increases, 7, both decreases and is more strongly amplified by the

S-B formula. The amplification, however, is negatively accelerated while, in this

example, the decrease in 7, proceeds at a constant rate. The upshot is that R;

increases sharply at first, reaches a maximum (at i = 4), and then decreases

gently. In this case, therefore, reliability would not be improved by lengthening

the test. In fact, the test could be shortened to 4 trials with no loss of reliability.

The superdiagonal pattern in Table | is perfectly regular, that is, constant

within any given diagonal and regularly decreasing between diagonals. As we

have seen, however, it nevertheless tends to yield reliabilities that increase to an

optimum and then decrease gently. This tendency may be reinforced by other

considerations. As the number of trials increases, some testees may become

fatigued or lose concentration; and performance in the presence of fatigue and

wavering attention tends to be fitful and erratic. These changes introduce novel

variance not present in earlier trials of practice. The effect is to produce a drop in

correlational level late in practice and, therefore, to bring about a forward opti-

mum earlier than it would have occurred in a perfectly regular pattern.

In practice, reliability as calculated by the S-B formula from a series of acqui-

sition (test) trials is less interesting than temporal stability—that is, the correla-

tion between test and retest over appreciable periods of time (months or years),

where a person’s score 1s his or her average performance over the first / trials at

test or retest. In a single test series a perfectly regular superdiagonal pattern, like

the one in Table 1, suffices to produce a forward optimum (that is, a maximum

prior to the last trial), provided the gradient away from the superdiagonal is steep

and the series is long enough. When testees are tested in two well-separated series

of trials, the conditions are somewhat different. Specifically, the matrix of test-

retest correlations decomposes into three submatrices: the correlations among

test trials, the correlations among retest trials, and the square matrix of correla-

' As given above, the S-B formula assumes that all trials have the same variance. In the section on reliability

this restriction is relaxed, by restating the formula in terms of variances and covariances.

78 JONES

tions between test and retest trials. The first two submatrices both tend to follow

superdiagonal form, and the average correlation between test and retest trials

often decreases as 7 increases. As a result, the average correlation in all three

submatrices decreases as / increases, and temporal stability rises to an optimum

and decreases gently thereafter. Like low reliability, therefore, low temporal

stability is not necessarily corrigible by increasing test length. Moreover, when

temporal stability can be improved by increasing test length, the gain may be

sharply limited. An optimum may be reached after only a modest increase in

test length, and the optimal value may be well short of unity.

Forward optima in reliability and temporal stability have important implica-

tions for the construction of performance tests. If a test shows a forward opti-

mum in stability, the implication 1s that lengthening the test will not improve its

stability. It is true that if the test was lengthened, stability, after decreasing for a

stretch of trials, might start increasing again to a second optimum. To date,

however, I have not seen any such second increase. One does see small depar-

tures from increasing, level, or decreasing curves but not a second increasing

trend in a curve’s general direction. If, however, an optimum once reached will

not be exceeded or, in the worst case, not exceeded by much, then lengthening

the test will not improve its temporal stability.

If a test has not reached an optimum or asymptote with the number of trials

given, it is possible to project where the optimum would fall if a test series was

lengthened. This projection is based primarily on extrapolating the course of

average correlations, either among test (or retest) trials or between test and retest

trials as the test lengthens. Such projections are, of course, no better than the

extrapolations on which they are based. Still, forward averages provide an empir-

ical basis for decisions regarding the length of a performance test. It can tell us

when a series of trials is already long enough, whether it might be shortened

without loss of stability, how much it would have to be lengthened to reach an

optimum, and how much of a gain could be realized by so lengthening it.

Once a test has been constructed, it may be used to predict performance on

numerous external criteria. At this point the issue is no longer test construction

(test length) but test scoring. The usual practice is to average all trials given.

Forward averages, however, may also be correlated with an external criterion.

When they are, the correlations (predictive validities) always rise at first and

sometimes reach an optimum, after which they decrease. It has long been recog-

nized that the differential content of a task, the abilities involved in it, may

change with practice (Ackerman, 1987; Fleishman & Hempel, 1954) and that,

as they do, the relation of the practiced task to an external criterion may also

change. There is, therefore, no reason to be surprised if the average of the first 7

trials sometimes predicts a criterion better than the average of all trials given.

PERFORMANCE-TEST THEORY 79

If, however, a forward optimum in predictive validity exists, then averaging

only those trials up to and including the optimum will yield a higher predictive

validity than the usual practice. Since the differential composition of a test may

change with practice and an external criterion may be most strongly related to

those components of a test that predominate at the beginning (say) or in the

middle of a practice series, stability and validity optima do not necessarily fall on

the same trial. For the same reasons, the optimal forward average for purposes of

prediction may vary from one external criterion to another.

Averaging from the first trial forward is only one way to generate a series of

averages from a series of test trials. Another way is to average from the last trial

backwards. The temporal stability of backward averages can, of course, be cal-

culated and sometimes a maximum occurs before the first trial is reached (a

backward optimum). Backward optima, however, are not informative about

how changes in test length might affect temporal stability. A forward average of,

say, 5 trials retains its meaning (refers to the same trials) regardless of how many

trials are ultimately given. A backward average of 5 trials, however, refers to

trials 6-10 if 10 trials are given and to trials 11-15 ifa total of 15 trials is given. A

backward average changes its meaning when the total number of trials changes.

As a consequence, no conclusions regarding changes in test length can be drawn

from a backward stability optimum.

Backward averages may also be correlated with an external criterion. When

they are, the correlation (predictive validity) rises at first and may reach an

optimum prior to the first trial. In these cases, as in the corresponding cases

involving forward optima, averaging only those trials up to and including the

optimum (following it in the practice series) yields a higher predictive validity

than averaging all trials given. Backward optima are especially helpful in im-

proving a test’s validity when a forward validity optimum also exists.

One final point should be noted. It may happen that predictive validity takes

different values in different subsets of trials. Where this happens, one may

restructure the test to consist exclusively of subsets with high validities, very

much as in conventional item analysis. Once such a restructuring is done, how-

ever, it should be followed (a) by forward averaging to determine the optimal test

length for reliability and temporal stability, and (b) by both forward and back-

ward averaging to determine optimal scoring for predictive validity.

Altogether, serial (forward or backward) averaging has four areas of applica-

tion in performance-test theory: reliability, temporal stability, predictive valid-

ity, and subset analysis. The present paper reports results obtained using the

computer-administered tests in Project A (see below). It is an interim report for

two reasons. First, results are presented for reliability only. Second, validation

results only are presented; a cross-validation is currently underway.

80 JONES

Table 2.—Number of Test Trials and Total Length of Time Required, Including Instructions and Inter-Trial

Intervals, for the 10 Project-A, Computer-Administered Tests

Number of Total Time

Test Tmials (n) (min.)

Simple Reaction Time 10 2

Choice Reaction Time 30 3

Short-Term Memory Test 36 i

Target Tracking | 18 8

Perceptual Speed & Accuracy 36 6

Target Tracking 2 18 7

Number Memory 28 10

Cannon Shoot 36 7

Target Identification 36 4

Target Shoot 30 5

Tests, Testees, and Procedures

The Project-A Tests

Project A is a large, multi-year effort to improve the Armed Forces Vocational

Aptitude Battery (Eaton, et al., 1986; Peterson, 1987). Included in this effort are

10 newly developed, computer-administered performance tests. Brief descrip-

tions of the 10 tests are given below. The tests are administered in the order

described. Table 2 shows the number of trials a person receives on each test and,

approximately, the total length of time each test requires.

Simple reaction time. The testee is instructed to place his or her hands in the

ready position. When the word YELLOW appears in a display box, the testee

strikes the yellow key on the test panel as quickly as he or she can. The depen-

dent measure is average time to respond.

Choice reaction time. This test is much the same as Simple Reaction Time.

The major difference is that the stimulus in the display box is BLUE or WHITE

(rather than YELLOW), and the testee is instructed to strike the corresponding

blue or white key on the test panel. The dependent measure is average time to

respond on trials in which the testee makes the correct response.

Short-term memory. A stimulus set, consisting of 1, 3, or 5 letters or sym-

bols, is presented on the display screen. Following a delay period, the set disap-

pears. When the probe stimulus appears, the testee must decide whether or not it

was part of the stimulus set. The dependent measure is average time to respond

on trials in which the testee makes the correct response.

Target tracking 1. Thisisa pursuit tracking test. The testee’s task is to keep a

crosshair centered within a box that moves along a path consisting exclusively of

vertical and horizontal lines. The dependent measure is the average distance

from the crosshair to the center of the target box.

PERFORMANCE-TEST THEORY 81

Perceptual speed and accuracy. This test measures a testee’s ability to com-

pare rapidly two stimuli presented simultaneously and determine whether they

are the same or different. The stimuli may contain 2, 5, or 9 characters and the

characters may be letters, numbers, or other symbols. The dependent measure is

average time to respond on trials where the testee’s response is correct.

Target tracking 2. This test is the same as Target Tracking 1, except that the

testee uses two sliding resistors instead of a joystick to control the crosshair. The

dependent measure is the same as in Target Tracking 1.

Number memory. The testee is presented with a number on the computer

screen. When the testee presses a button, the number disappears and another

number appears along with an operation term, e.g., “Add 9” or “Multiply by 3”’.

Another press of the button and another number and operation term are pre-

sented. This procedure continues until finally a solution to the problem is pre-

‘sented. The testee must then indicate whether the solution presented is correct

or incorrect. The dependent measure is total time to respond on trials in which

the testee correctly identifies the solution presented as correct or incorrect.

Cannon shoot. ‘The testee’s task is to fire a shell from a stationary cannon so

that it hits a target moving across the cannon’s line of fire. The dependent

measure is a deviation score indicating the difference between time of fire and

optimal fire time (for example, a direct hit yields a deviation score of zero).

Target identification. The testee is presented with a target and three stimu-

lus objects. The objects are pictures of tanks, planes, or helicopters. The target is

the same as one of the three stimulus objects but rotated or reduced in size. The

testee must determine which of the three stimulus objects is the same as the

target object. The dependent measure is average time to respond on trials in

which the testee makes the correct response.

Target shoot. The testee’s task is to move a crosshair over a moving target

and then press a button to fire. The dependent measure is distance from the

crosshair to the center of the target when the testee fires.

The Criterion Task

In addition to the Project-A tests, each person participating in the study was

administered a criterion task. This task was Anti-Aircraft, game #1 in the Atari

Air-Sea Battle cartridge (CX-2624). In this game the individual controls a gun

placed two thirds of the way from left to right at the bottom of a television

screen. Four different kinds of aircraft traverse the screen above the gun, in

different numbers, at different speeds and altitudes, and from left to right or vice

versa. The purpose of the game is to shoot down as many aircraft as possible in a

2-min.-and-16-sec. game. The control devices are a joystick for positioning the

gun and a button for firing the missile. The missile itself was the smaller of two

82 JONES

possible sizes (difficulty position “‘A”’). The dependent measure is number of

aircraft shot down per game. Anti-Aircraft is a complex psychomotor skill with a

high performance-ceiling. No testee comes close to reaching the maximal possi-

ble performance with the amount of testing given.

Participants and Procedures

The participants in the study were 102 central Pennsylvania undergraduate

college students, 50 men and 52 women. Each student was administered the

Project-A tests at the start of the fall semester (September—October) and then

again four months later at the start of the spring semester (January—February).

The Project-A tests were taken in a single sitting that lasted between 45 and 75

min. depending on how quickly the student responded to the tests and the

instructions that preceded them. The entire administration, both test and retest,

instructions as well as the tests themselves, were computer controlled.

In the fall, following the Project-A tests, each student was administered five

sessions of Anti-Aircraft, each session consisting of seven games, for a little more

than 16 min. of playing time. All five sessions were completed within a ten-day

period, with no more than two sessions taking place on a given day. In the spring

semester, again following the Project-A tests, each student was given three ses-

sions of Anti-Aircraft with the same number of games per session and the same

conditions as to distribution as in acquisition.

RESULTS

Comparison with Army Data

Table 3 compares the present results with those collected by Peterson, Hough,

Dunnette, Rosse, Houston, Toquam, and Wing (1990) in overlapping samples

of Army enlisted people ranging in number from 8,892 to 9,269 for different

tests. The college students performed better on all tests, but some of the differ-

ences were sizable whereas others were small. The largest differences were for the

two memory tests, in both cases half a standard deviation (SD) or more.” The

next largest differences were for the two perceptual tests (Perceptual Speed &

Accuracy and Target Identification), approximately .4 SD. The differences for

Choice Reaction and the two tracking tests were approximately .33 SD, while

those for Simple Reaction and the two aiming tests (Cannon Shoot and Target

Shoot) were less than .2 SD. All differences except the last three were significant

* «Standard deviation” refers here to the square root of the pooled within-group variance.

PERFORMANCE-TEST THEORY 83

Table 3.—Comparison of Soldier and Student Results on the Project-A Tests*

Temporal

Test/Measure Testee M SD Reliability Stability

Simple Reaction** Soldier 31.84 14.82 .88 23

(mean dec. time) Student 29.38 4.94 88 50

Choice Reaction Soldier 40.83 Oya 97 .69

(mean dec. time) Student 36.54 6.48 97 si]

Short-Term Memory Soldier 87.72 24.03 .96 .66

(mean dec. time) Student 70.98 17.43 97 .69

Target Tracking | Soldier 2.98 0.49 .98 14

(mean In dist. + 1) Student Lhd 0.43 .98 87

PerceptualS & A Soldier 236.91 63.38 94 .63

(mean dec. time) Student 202.42 47.10 95 73

Target Tracking 2 Soldier 3.70 0.51 98 £035)

(mean In dist. + 1) Student 3.45 0.52 98 91

Number Memory Soldier 160.70 42.63 .88 62

(final resp. time mean) Student 118.39 27.89 91 .69

Cannon Shoot Soldier 43.94 9.57 65 52

(mean abs. time disc.) Student 43.80 8.52 cult ps:

Target Identification Soldier 193.65 (18 ya BS) 97 78

(mean dec. time) Student 163.84 45.08 95 wil

Target Shoot Soldier Da 0.24 74 ey

(mean In dist. + 1) Student 2.14 0.20 a | 570

* All times are in hundredths of a second. Logs (In) are natural logs.

** Simple Reaction in the Army battery had 15 trials; the students were given only 10 trials. Number of trials

in the remaining tests were the same in the soldier as in the student battery. The numbers are shown in Table 2.

at the p = .01 level as determined by f¢ tests. These differences were broadly what

one would expect, 1.e., the more cognitive a test, the larger the difference in favor

of the students.

Variabilities were greater in the soldier than in the student data, except for

Target Tracking 2, but not greatly so, except for Simple Reaction. The variance

of Simple Reaction was nine times as large among the enlisted people as among

the students. Simple Reaction was the first test in the battery, and there may

have been some confusion among the Army testees as to what they were sup-

posed to do. If so, it would explain the high variability of the Simple Reaction

’ Test in the Army data.

The column headed Reliability contains, for the soldier data, odd-even corre-

lations corrected for test length by the Spearman-Brown formula and, for the

student data, Spearman-Brown projections from the average correlation involv-

ing all trials. Thus, both figures make use of all trials administered and both use

the Spearman-Brown formula. The correspondence between the two sets of

figures is startlingly close.

The column headed Temporal Stability contains two-week test-retest correla-

tions for the Army data and four-month test-retest correlations for the student

data. Again, an individual’s score on any given test is the average of his or her

scores on all trials administered. Temporal stability was better among students

84 JONES

than soldiers for all tests except Target Identification and may have been better

even for Target Identification, given that the retest interval was eight times

longer in the college than in the Army sample. There are at least four procedural

differences that may have contributed to better stability among students. First,

of course, was the difference in population: college students versus enlisted

people. Second, the sex ratio in the college sample was essentially 50-50, whereas

males predominated in the Army sample. Third, the tests were administered to

the college students by a single, very experienced person, whereas the Army data

were collected at many places by many people, some of them not experienced

test administrators. Fourth, the students were tested one or two at a time,

whereas the soldiers were tested in groups of as many as two or three dozen at

a time.

In general, the differences between the two arrays of stability results were not

large. The low stability for Simple Reaction in the Army data was probably

related to that test’s high variability. No obvious explanation exists for the low

stability of Target Shoot in the Army sample, except perhaps that it was the last

test in the battery.

Reliability

Figure | presents reliability results for the Simple Reaction Test. The average

correlation (7,) tended to decrease sharply from the 2nd to the 10th trial. A

straight line has been fitted to those nine points and extended out to trial 25. The

correlations R; are Spearman-Brown projections for a test of length, 7, given that

a test of unit length has reliability, R, = 7,. The smooth curve was obtained by

applying the Spearman-Brown formula to corresponding points, 7;, on the re-

gression line. The smooth curve has also been extended to trial 25. Such a curve

reaches a maximum at

ent. Saja ila es

HY SRE: PERRET

where a and / are the intercept and slope, respectively, of the regression line. In

this case i* was estimated to occur at about the 19th trial. The reliability of

Simple Reaction could be improved only slightly by lengthening the test; dou-

bling the number of trials would increase reliability by .02 but still leave it at

.897, well short of unity. More than doubling the number of trials would be

counterproductive. .

Figure | can be improved in two key respects. First, the linear regression line

in Figure | was obtained by weighting the 7; equally. The r,, however, are based

on very different numbers of correlations. For example, r, is based on only one

correlation, whereas 7,,. is based on 45. It would make sense on strictly statistical

PERFORMANCE-TEST THEORY 85

Smooth curve projected from straight line by

Spearman-Brown formula

Correlation (r, or R;)

Straight line calculated from Tr,

using equal weights

_—

aes

5 10 15 20 25

Number of Trials (i)

Fig. 1. Average correlation and Spearman-Brown reliability up to trial i for Simple Reaction Time. The

straight line was calculated weighting all average correlations equally.

grounds to weight the 7, for the number of correlations on which each one is

based. It makes especially good sense when one remembers that the main pur-

pose in fitting the regression line is to predict the course that 7, will follow beyond

the total number of trials administered (n). The 7, usually follow a decreasing,

‘ negatively accelerated course. Therefore, the best prediction of where /, will lie

when i > n is the slope of the 7, curve, not overall, but just before the adminis-

tered sequence reaches its end. Weighting the 7, for the number of correlations

on which each one is based effectively approximates such a slope. The early

points are heavily discounted in favor of the last few points. The resulting line is

almost always shallower than the one obtained by equal weighting of the 7.

Hence, the number of trials for optimal reliability, 7*, is increased (pushed

further out).

The second key improvement concerns how to estimate R,. The estimation

based on 7; assumes that all trials have equal variances. If this is not so (and it

never is), the appropriate estimate becomes

86 JONES

-——

-—.

es [enue an

-_-—

Smooth curve projected from straignt

line by Spearman-Brown formula

Correlation (r; or R,)

Straight line calculated from, weighting

each pooled correlation for the number

of covariances on which it is based

5 10 15 20 a)

Number of Trials (i)

Fig. 2. Pooled correlations (cov/var) and Spearman-Brown reliability up to trial / for Simple Reaction Time.

The straight line was calculated by weighting each pooled correlation for the number of covariances on which

it is based. =

COV;

var,

nee

where cov, and var, are, respectively, the averages of all covariances and vari-

ances up to trial 7. In effect, 7, weights the correlations for the variances involved

in them. Correlations between trials with large variances count for more than

correlations between trials with small variances. This improvement has no sys-

tematic effect on i*. Sometimes it increases i* and sometimes, as in the case of

the Simple Reaction Test, it decreases *.

Figure 2 presents the reliability results for the Simple Reaction Test, using 7;

and a weighted regression line. The net effect is to decrease 7* to 14.8 and to

reduce the optimal reliability to .874.

Table 4 presents reliability results for all ten tests, using 7, and a weighted

regression line. For two tests (Perceptual Speed & Accuracy and Target Tracking

PERFORMANCE-TEST THEORY 87

Table 4.— Reliability Results for the Project-A Tests, Using Pooled Averages (f,) and Weighted Regression

Projected Reliability

Optimal No.

No. of Slope of Trials At Trial n At Trial i*

Test Trials (b X 10°) (i*) (Ri) (Ri+)

Simple Reaction 10 —14.42 .864 14.8 .874

Choice Reaction 30 —1:05 .963 195.3 .988

Short-Term Memory 36 = 16.9 952 159.4 O77

Target Tracking | 18 —2.46 .983 100.3 992

PerceptualS & A 36 +0.16 939 0) 1.000

Target Tracking 2 18 +1.69 .984 eo) 1.000

Number Memory 28 alee 869 96.2 921

Cannon Shoot 36 —1.09 499 29.9 510

Target Identification 36 —0.50 947 310.9 .987

Target Shoot 30 mleoe 701 33.8 .704

2) the regression line had positive slope (b > 0). In those two cases, therefore, /*

was indefinitely large and there was no optimal reliability short of unity. In the

other eight tests the slopes were negative, /* finite, and optimal reliability some

value less than unity. In five of these eight cases, however, /* was very large and,

with the exception of Number Memory, the projected optimal reliability, R:.,

was close to unity. In one case (Cannon Shoot), however, i* <n, i.e., the number

of trials for optimal reliability was less than the number administered. In such a

case reliability cannot be improved by lengthening the test. In fact, the test could

be shortened without reducing optimal reliability. In two tests, Target Shoot and

Simple Reaction Time, i* lay less than five trials ahead of where the adminis-

tered sequence stops. In both cases little would be gained by increasing the

number of trials and optimal reliability was well short of unity.

Comment

At one level, serial averaging is a prosaic data-processing procedure. It is

based, however, on a view of performance testing that departs fundamentally

from classical test theory. The gist of that departure is not to replace one theory

with another but to hybridize classical test theory with the study of individual

differences in skill acquisition and retention. This hybrid is more a matter of

approach and concept than content. Conventional test theory is purely struc-

tural; time has no place in it. The study of skill acquisition and retention,

however, is processual; everything in it is embedded in time and is, therefore,

temporally ordered. Large parts of this structural-processual hybrid have been

taken over from classical theory. Other parts, however, come from the process-

ual component in the hybrid: for example, the treatment in terms of trials, the

88 JONES

centrality of order, or the recognition and use of established regularities such as

superdiagonal form. The overall approach is open, moreover, to further imports

from the study of skill acquisition. Distribution and transfer effects, reminis-

cence, many possible results from cognitive science, may ultimately find a place

in a theory of performance testing hybridized to include performance as well as

test phenomena.

Acknowledgement

This work has been supported by Contract No. MDA 903-86-C-0145 from

the Army Research Institute for the Behavioral and Social Sciences and Grant

No. N00014-90-J-1994 from the Office of Naval Research.

References

Ackerman, P. L. (1987). Individual differences in skill learning: An integration of psychometric and informa-

tion processing perspectives. Psychological Bulletin, 102, 3-27.

Bittner, A. C., Jr., Carter, R. C., Krause, M., & Harbeson, M. M. (1983). Performance Evaluation Tests for

Environmental Research (PETER): Moran and computer batteries. Aviation, Space, and Environmental

Medicine, 54, 923-928.

Eaton, N. K., Hanser, L. M., & Shields, J. (1986). Validating selection tests for job performance. In J. Zeidner

(Ed.), Human productivity enhancement. Vol. 2, Acquisition and development of personnel (pp. 382-438).

New York: Praeger.

Fleishman, E. A., & Hempel, W. E., Jr. (1954). Changes in factor structure of a complex psychomotor test as a

function of practice. Psychometrika, 19, 239-252.

Gulliksen, H. (1950). Theory of mental tests. New York: John Wiley & Sons.

Humphreys, L. G. (1960). Investigation of the simplex. Psychometrika, 20, 173-192.

Jones, M. B. (1962). Practice as a process of simplication. Psychological Review, 69, 274-294.

Jones, M. B. (1969). Differential processes in acquisition. In E. A. Bilodeau (Ed.), Principles of skill acquisition

(pp. 141-170). New York: Academic Press.

Kennedy, R. S., Bittner, A. C., Jr., Carter, R. C., Krause, M., Harbeson, M. M., McCafferty, D. B., Pepper,

R. L., & Wiker, S. F. (1981). Performance Evaluation Tests for Environmental Research (PETER): Col-

lected papers (NBDL-80R008). New Orleans, LA: Naval Biodynamics Laboratory.

Kyllonen, P. C. (1985). Theory-based cognitive assessment (AFHRL-TP-85-30). Brooks Air Force Base, TX:

Air Force Human Resources Laboratory.

Melton, A. W. (Ed.). (1947). Apparatus tests (AAF Aviation Psychology Program Research Report No. 4).

Washington, DC: U.S. Government Printing Office.

Messick, S., & Jungblut, A. (1981). Time and method in coaching for the SAT. Psychological Bulletin, 89,

191-216.

Peterson, N. G. (Ed.). (1987). Development and field test of the trial battery for Project A (Technical Report

739). Alexandria, VA: U.S. Army Research Institute.

Peterson, N. G., Hough, L. M., Dunnette, M. D., Rosse, R. L., Houston, J. S., Toquam, J. L., & Wing, H.

(1990). Project A: Specification of the predictor domain and development of new selection/classification

tests. Personnel Psychology, 43, 247-276.

Wing, H. (1980). Practice effects with traditional test items. Applied Psychological Measurement, 4, 141-155.

Journal of the Washington Academy of Sciences,

Volume 81, Number 2, Pages 89-100, June 1991

Space Adaptation Syndrome:

Multiple Etiological Factors

and Individual Differences

James R. Lackner and Paul DiZio

Brandeis University, Waltham, Massachusetts

ABSTRACT

Space motion sickness is a significant operational concern in the American and Soviet

space programs. Nearly 70% of all astronauts and cosmonauts are affected to some degree

during their first several days of flight. It is now beginning to appear that space motion

sickness like terrestrial motion sickness is the consequence of multiple etiological factors. As

we come to understand basic mechanisms of spatial orientation and sensory-motor adapta-

tion we can begin to predict etiological factors in different motion environments. Individuals

vary greatly in the extent to which they are susceptible to these different factors. However,

individuals seem to be relatively self-consistent in terms of their rates of adaptation to

provocative stimulation and their retention of adaptation. Attempts to relate susceptibility

to motion sickness during the microgravity phases of parabolic flight maneuvers to vestibu-

lar function under 1G and OG test conditions are described.

Introduction

Space motion sickness 1s a significant operational problem in both the Ameri-

can and Soviet space programs. It affects nearly 70 percent of all astronauts and

cosmonauts (Jennings, Davis & Santy, 1988). Its onset can be as early as an hour

after insertion into orbit or microgravity and as long as many hours or even a

day or two. Usually, it is self-limiting and largely abates by Mission Day 3 so that

normal or near normal activities are possible. Astronauts also vary considerably

in terms of how severely they experience space motion sickness. They report

that keeping the head still can prevent or suppress symptoms, but of course this

is usually not practical under operational conditions. In general, the symptoms

are much like those of motion sickness experienced on Earth and can include

drowsiness, nausea, vomiting, apathy and the wide constellation of signs and

symptoms characteristic of motion sickness.

A key concern in the space program has been to identify the primary etiologi-

90 LACKNER AND DIZIO

cal factors in space motion sickness in order to provide the basis for developing

tests of susceptibility and adaptation or training procedures for decreasing sus-

ceptibility. This has proven to be a very difficult task. One of the reasons is that

investigators in the field, at least initially, approached space motion sickness as if

it were a unitary phenomenon arising from a single factor associated with being

in a weightless environment. For example, early on it was thought that the

redistributions of body fluid associated with the absence of hydrostatic pressure

in the circulatory system in OG were responsible (Gibson, 1974; Kerwin, 1974;

Thornton, Hoffler & Rummel, 1974). The fluid shift was thought to alter the

pressures within the organs of equilibrium of the inner ear. This notion was

appealing because of its simplicity, but experimental manipulations of fluid shift

magnitudes on Earth had no discernible effect on motion sickness susceptibility

during provocative stimulation (Graybiel & Lackner, 1977).

One of the features which has characterized motion sickness research over the

years has been the difficulty of making generalizations about an individual’s

susceptibility across different motion exposure conditions, even ones involving

ostensibly similar patterns of stimulation. Our work concerning space motion

sickness and individual differences may enable us to better understand the rea-

sons for this variability. We have come to the conclusion that a given exposure

situation may be provocative for a variety of reasons and that individuals vary in

their susceptibility to different potentially provocative aspects of these situa-

tions. Put differently, individual differences in susceptibility may not be simply

reflecting a greater or lesser amount of “noise” but rather be pointing to the

complexities of the test situation and of spatial orientation mechanisms. This

means that in developing assessment and adaptation procedures “individual

differences” are pointing to the need to assess and train on multiple dimensions.

Our own work represents an attempt to understand why astronauts become

motion sick during exposure to microgravity, and what kinds of procedures can

be employed to prevent or attenuate sickness. In some of our first studies, we

looked at susceptibility to motion sickness in a variety of different “motion

environments” including parabolic flight (Lackner & Graybiel, 1983, 1986,

1987) where periods of weightlessness and high force alternate, and during ex-

posure to provocative vestibular stimulation in the form of repeated impulsive

decelerations from constant velocity rotation (Graybiel & Lackner, 1980;

Lackner & Graybiel, 1979). Half of the individuals also participated in studies in

a slow rotation room, where the stressor is generated by head movements; if the

head is kept still, motion sickness does not result. In each of these different

situations, the individuals had participated on multiple occasions. Thus, it was

possible to measure their initial susceptibility as well as their acquisition and

retention of adaptation to some extent, and to compare these values for the

SPACE ADAPTATION SYNDROME | 91

different test conditions for each individual as well as collectively. The general

conclusion we came to from these studies was that despite great individual

differences in susceptibility, persons who had a high rate of adaptation and a

high level of retention of adaptation in one situation, regardless of initial suscep-

tibility, showed a similar pattern of adaptation rate and retention in other situa-

tions. Moreover, in these experiments, the performance on the impulsive vestib-

ular stimulation task seemed to be most predictive of acquisition and retention

in the other situations (Graybiel & Lackner, 1983).

The observations to be described below represent a variation on this theme.

We sought to develop tests that would allow us to relate susceptibility to motion

sickness in parabolic flight with alterations in vestibulo-ocular function. We

took this approach because perhaps the single most important finding in the

motion sickness literature is that without exception persons who lack labyrinth-

ine function cannot be made motion sick (Graybiel & Johnson, 1963; Money &

Friedberg, 1964). Moreover, there is also strong evidence that individuals with

partial loss, e.g., half-sided destruction of function, are less susceptible to motion

sickness than normal persons after recovery from the acute loss (Graybiel &

Johnson).

We also wanted to use test procedures in which we would be able to quantify

quite precisely the stimulus delivered and the responses obtained for function-

ally relevant situations. In addition, we wanted to have some conditions that

would incorporate voluntary head movements. Our approach was to use sudden

stops from constant angular velocity rotation to deliver step stimuli to the semi-

circular canals and in some trials to have the individual voluntarily make a head

tilt after the trials. On Earth, such post-rotary head tilts will suppress or attenu-

ate the nystagmus elicited by a velocity step, this suppression is often referred to

as dumping (Raphan, Cohen & Matsuo, 1977). It is thought to be related to the

re-orientation of the linear acceleration sensitive otolith organs of the inner ear

to the force of gravity.

- Dumping is thought to represent an abolishment of velocity storage (Raphan,

Cohen & Matsuo, 1977). Figure 1 illustrates the concept of velocity storage. The

post-rotary response to vestibular stimulation considerably outlasts the periph-

eral response of the semicircular canals. For example, the time constant of the

human horizontal semicircular canals is computed on allometric considerations

to be about 8 or 9 s. By contrast, the time constant of post-rotary nystagmus 1s

often 15s or more. This is because there are thought to be two signals driving the

eyes, one directly from the semicircular canals and one a “velocity storage”

signal from a brain stem integrator which receives a copy of the canal signal.

Dumping is thought to be an attenuation of velocity storage.

Interestingly, too, in virtually all terrestrial situations in which dumping oc-

92 LACKNER AND DIZIO

DIRECT PATHWAY asses INDIRECT PATHWAY #-##

0 EYE VELOCITY s»««»—a 0 - 40 SEC

OmMmMa~Qme

Fig. 1. Simulation of the slow phase velocity of vestibular nystagmus elicited by a sudden stop from 60°/s

constant velocity rotation, according to the model of Raphan, Cohen, and Matsuo (1977). In this model, eye

velocity is dependent on two factors: 1) semicircular canal activity which is conveyed by the direct pathway,

and 2) the activity of a brain stem integrator which integrates a copy of the canal signal, this “velocity storage”’

is then conveyed by the indirect pathway. The combined direct and indirect pathway activity determines eye

velocity.

curs (1i.e., in which velocity storage is continuously suppressed or in which it is

dumped by a head movement), motion sickness also is common. In fact, we do

not know of situations involving dumping which are not provocative. However,

people can become motion sick in situations where dumping is either empiri-

cally lacking or theoretically not expected; this will be discussed later with regard

to implications for multiple etiological mechanisms. In these experiments we

attempted to relate individual differences in motion sickness susceptibility dur-

ing head movements 1n parabolic flight with individual differences in peripheral

vestibular mechanisms, velocity storage, and dumping.

Methods

The experiments were carried out during parabolic flight maneuvers (Figure

2) in which the gravitoinertial resultant force varied from 0G to 1.8G, with the

high and low force phases each lasting about 25 s (DiZio & Lackner, 1988). Our

observations were carried out during these periods and during straight and level

flight at 1G. Fifteen individuals who ranged in age from 18 to 45 took part. Each

had passed an FAA Class II flight physical and was without known sensory

SPACE ADAPTATION SYNDROME . 93

33 PUSH- OVER

oo] OUT PULL-UP PULL-OUT PULL-UP PULL-OUT PULL-UP

2.06 2.06 2.06 206 2.06 2.06

PEAK PEAK PEAK PEAK PEAK PEAK

SEC. -40 -30 -20 -0 0 0/20 30 40, 50 6| 70 80/90 100 WO [20 130 40 50 160 }70 180 190 200] 210 220 230 240

Pp ! | ! 3

(ACCELEROMETER !

RECORDING)

ALTITUDE IN THOUSAND FEET

2 ce 2 -—-—~

|. SUBGRAVITY PERIOD

2. SUPRAGRAVITY PERIOD

3. WEIGHTLESS PERIOD

Fig. 2. Schematic illustration of the flight profile of the KC-135 aircraft during parabolic maneuvers and the

associated alterations in gravitoinertial force level.

or motor abnormalities (except several wore spectacles). They were not medi-

cated during the experiments.

On any given flight day, the participants received only six trials, which were

balanced across 0G, 1G and 1.8G, head movements and no head movements.

On a given flight day, 40 parabolas would be flown and a person would be tested

in the first, second, third, or fourth set of ten. During a given flight week there

~ would be four consecutive days of 40 parabola flights. A particular individual

would receive the vestibulo-ocular tests on, at most, two of these days. Each

person participated in multiple flight weeks, usually separated by several

months or more.

Six of the individuals also participated in later observations in which they

were tested early and late in flight with sudden stops during the course of a single

40-parabola flight to determine whether adaptation or habituation would occur.

Motion sickness susceptibility was assessed during an individual’s first two

days of a parabolic flight week, some persons were assessed on multiple occa-

sions. On the first of these days, the individuals sat in an aircraft seat restrained

by a lap belt, they were permitted normal vision and were free to move their

94 LACKNER AND DIZIO

Table 1.—Diagnostic Categorization of Different Levels of Severity of Acute Motion Sickness

Pathognomonic Major Minor Minimal AQS*

Category 16 Points 8 Points 4 Points 2 Points 1 Point

Nausea syndrome Vomiting or Nauseaf II, II] NauseaI Epigastric Epigastric awareness

retching discomfort

Skin Pallor III Pallor II‘ Pallor I Flushing/Subjective

warmth = II

Cold sweating Ill II I

Increased salivation Ill II I

Drowsiness ~ Hi II I

Pain Headache = II

Central nervous Dizziness

system

Eyes closed = II

Eyes open III

Levels of Severity Identified by Total Points Scored

Frank Sickness Severe Malaise Moderate Malaise A Moderate Malaise B Slight Malaise

(S) (M III) (M ITA) (M IIB) (M I)

>16 points 8-15 points 5-7 points 3-4 points 1-2 points

* AQS = Additional qualifying symptoms. +III = severe or marked, IJ = moderate, I = siight.

head and eyes. On the second day, they were free to move about the aircraft

except for brief participation in tests not involving rotation or provocative stim-

ulation. Throughout these two flight periods, the individuals were carefully

monitored for symptoms of motion sickness after every 10 parabolas using the

symptom identification system (Table 1) developed by Graybiel and his col-

leagues (Graybiel, Wood, Miller & Cramer, 1968). We averaged the number of

motion sickness points accumulated on the four sets of parabolas on a given day,

and the scores for an individual’s two days were averaged for a final score.

For the vestibular tests, the blindfolded participants were rotated in a servo-

controlled chair about the z-axis of the body, at a constant velocity of 60°/s for at

least 60 s. They were brought to a sudden stop within 3 s after a transition into

the desired G level. In trials involving head movements after the stop, the indi-

vidual voluntarily moved the head backwards, the head movement being me-

chanically confined to the pitch plane and to an amplitude of 40°. Appropriate

instrumentation was used to record G level, rotational velocity, eye movements,

and head movements. Appropriate procedures (DiZio & Lackner, 1988) were

used to compute the step gain and dominant time constant of decay of vestibular

nystagmus.

Results

Table 2 shows a rank ordering of the individuals according to their susceptibil-

ity to motion sickness along with the step gain of their vestibulo-ocular reflex to

SPACE ADAPTATION SYNDROME . 95

Table 2.—Motion Sickness Susceptibility in Parabolic Flight, Step Gain of the Vestibulo-Ocular Reflex

Following Sudden Stop Stimulation (SS), and Modification of Step Gain De Gravitoinertial Force Level (G)

Early and Late in Flight

Change in Step Change in

Motion Gain Relative to Reduction of Step

Sickness Step Gain 1G, SS Baseline Gain Early Vs. Late

Dames em SUSccPUbIity:. Me Me Thc co a ee ee ee

ID Score SSG SS10G" GSS, .L8SG@ SS. 0G. SS, 1:8G". °SS0G. 2 SS. 1-8G

l 0.0 .265 258 .192 —.007 Or OS +.024

2 0.6 .430 .420 2503 —.010 123 = 401 O22

3 1.0 DES 313 1327. .058 .067 —.008 —.058

4 Pei 593 .666 542 .073 —.051 +.039 —.043

5 6.8 332 328 305 —.004 027 +.023 203i

6 9.9 .430 358 395 072 = 085)

7 11.8 .738 774 .786 .036 .048

8 12.9 500 45181 .463 .031 —,.037

9 13.0 395 361 .480 —.034 085

10 15.3 234 .202 2g 032 Old

bit 19.6 540 560 528 .020 =.012 —.068 +.024

12 20.1 .468 35 362 —.153 —.106

13 22.4 293 283 292 —.010 021

14 25.5 351 .406 362 055 O11

15 27.9 398 443 .466 045 .068

Xi 415 414 417 001 .002 —.006 —.018

SD .138 158 158 .058 .064 .037 .034

sudden stops in the different background force levels. In comparing the step gain

across individuals as a function of background force level no significant differ-

ences were observed. This means that the peripheral end organs, the semicircu-

lar canals, were not affected by variations in force level. The gain averaged .415

and ranged from .234 to .738, and had a high test-retest correlation (r = .8911, p

< .001). There was no significant rank correlation between step gain and motion

sickness susceptibility (rho = .0393).

The time constants of slow phase velocity decay for sudden stops without

head movements ranged from 11.7 to 26.4 with an average of 16.6 s (Table 3)

and also had a high test-retest reliability (r = .8444, p < .001). The time con-

stants were not significantly correlated with step gain and showed a non-signifi-

cant rank correlation (rho = .4964, p = .08) with motion sickness susceptibility

(Figure 3A). Relative to the 1G baseline values, the vestibular-ocular reflex time

constants decreased both in OG and 1.8G, ¢t = 8.65, p < .0001, and ¢ = 3.76, p<

.0005, respectively. These decreases were highly correlated with each other (7 =

.6706, p < .01) but neither was correlated with susceptibility to motion sickness

(Figure 3B).

Ten individuals had also participated in sudden stops followed by head move-

ments. The baseline, 1G, no head movement trials had an average time constant

of 16.7 s. This value was reduced to 11.3 s when head movements were made in

96 LACKNER AND DIZIO

A b

rho=0.4964, p=0.08 rho=0.3085, p>0.1

15 15

12 12

x a

= 9 ag

4 =

z 6 x 6

3 < 3

@) @)

0) 6) 6 9 12 VS) @) 3 6 is) 12 15

SUSCEPTIBILITY RANK SUSCEPTIBILITY RANK

mo—O0:18998, p<0:0i1 rho=0.4587, p>0.1

10 we 6

8 Lu

= . 4

< l

s “ai

fs :

=< 2

<i 2 =

0 ale ea

@) 2 4 6 8 10 @) Z 4 6

SUSCEPTIBILITY RANK SUSCEPTIBILITY RANK

Fig. 3. Graphs of motion sickness susceptibility rankings (lower rank means lower susceptibility) during

head movements in parabolic flight versus rankings based on various vestibular tests: A. Rankings of the

dominant time constant of slow phase velocity decay following sudden stops in 1G with the head kept still

(lower TAU, rank = shorter time constant). B. Rankings of the reduction in dominant time constant in 0G

relative to 1G with no head movements (lower A TAUg, rank = smaller reduction). C. Rankings based on

reduction in time constant due to post-rotary head movements in 1G (lower A TAU, rank = smaller

reduction). D. Rankings of the difference, early versus late in flight, in the reduction of the time constant in 0G

relative to 1G (A TAUpg EARLY-LATE; lower rank = smaller difference).

1G (Table 3). The decreases in time constants for individuals and their motion

sickness susceptibilities (Figure 3C) were positively rank correlated (rho = .8598,

p< 01).

For those six participants who were tested repeatedly during a flight there was

SPACE ADAPTATION SYNDROME 97

Table 3.—Motion Sickness Susceptibility in Parabolic Flight, Time Constant (s) of Decay of Nystagmic

Slow Phase Velocity Following Sudden Stop Stimulation (SS), and Modification of the Time Constant by Head

Movements (HM) and by Gravitoinertial Force Level (G) Early and Late in Flight

Change in

Reduction

of Time

Constant

Change in Time Constant Early Vs.

Motion Time Constant Relative to SS, 1G Baseline Late

Sickness [Ace eo en a ea

Participant Susceptibility SS, SS+HM, SS, SS) =SS—-HIM; SS, SS, SS, SS,

ID Score 1G 1G 0G 1.8G 1G 0G 18G 0G 1.8G

1 0.0 Lf 12-2 126, W326 =3).5) SO Oe — eS teeO Mm

2 0.6 itabed/ 9.1 8.5 8.5 we SS SB Ole 3e5

3 1.0 16.7 13.6 hg eS ecctil MS a —3.1 So Oe oe eee

4 De, 16.0 LOH HOS =§.9: =525.+0.2 -+0:4

S 6.8 12.0 8.5 LOS ee, ENS) Sh SO. STL 0.0

6 9.9 14.7 8.4 roe dea (Ss 653 —=6.5" —1e4

7 11.8 17.4 ee 12.86 Sl5i8 = o3 =4.6,Se—1k6

8 12.9 19.1 12.4 15.4 20.9 Sh 7/ —3.7 +1.8

9 13.0 16.3 10.5 bHO} PLS.3 —5.8 =43: =10

10 5e3 1327, Face Ns I Ses ee]

11 19.6 26.4 15.6 1G: Weg —10.8 —10.2 -8.5 +0.4 +0.3

12 20.1 1257 9508 T1Or3 = 6) PR te

13 22.4 Otol eG 20.3 35) tS

14 DISS 71 10.4 eS ae ba is Olt = eens eis

15 27.9 18.6 D2 Ais 9 0 Shia

¥ 16.6 1s ese eg, —5.4 —53 -25 -06 —0.6

SD 3.79 2.33 322. 2 S102 2.44 2355 (29 levee 93 1.59

a general habituation of time constants. For example, the 1G values decreased

from 16.1 to 13.3 s from the beginning to the end of a flight, however, the G

force influence remained. In OG, early in flight, the average reduction relative to

1G was 4.1 s, later in flight it was 3.5 s, a non-significant change; for 1.8G trials

the average reductions were 2.0 s early and 1.4 s late, also a non-significant

difference. Motion sickness susceptibility was not rank correlated with the early

versus late changes in the effect of G level on time constants (Figure 3D). A

similar habituation of step gain was also seen (Tables 2 and 3). The 1G values

ranged from .481 early to .406 late, but there were no significant mean changes

in the 0G and 1.8G gains early or late relative to the early and late 1G gains.

Motion sickness ranking was also not correlated with G related gain changes,

early versus late in the flight.

Discussion

Over the years there have been many attempts to correlate the gain of the

vestibulo-ocular reflex or the time constant of slow phase velocity decay with

98 LACKNER AND DIZIO

susceptibility to motion sickness (deWit, 1954; Bles, deJong, and Oosterveld,

1984). In general, the results have been disappointing because of low correla-

tions or because of limited replicability. The results of the present study insofar

as they use like measures are quite similar. For example, we did not find the step

gain of the vestibulo-ocular reflex to be correlated with susceptibility to motion

sickness, regardless whether step gain was assessed in 1G, 0G, or 1.8G. The

aforementioned studies failed to find a relationship between sea sickness and

step gain of the 1G, vestibulo-ocular reflex. We interpret our findings of a lack of

G-related differences in step gain and a lack of correlation step gain with suscep-

tibility to mean that (1) there was no general response decline caused by the

participants’ states of motion sickness and (2) that individual differences in

peripheral vestibular function (within the “normal range’’) are not predictive of

motion sickness susceptibility in parabolic flight.

Earlier studies have also failed to find a relationship between time constants of

nystagmus decay and susceptibility, for sudden stops in 1G. In the present study,

we have found a positive correlation of .49 between the rank orderings of time

constants for sudden stops and motion sickness susceptibility. Additionally

there was a trend for longer time constants and increased susceptibility to be

associated. Here it should be noted that earlier studies may have obscured a

possible relationship by comparing average values for “‘susceptibles” and “‘non-

susceptibles,” rather than taking advantage of individual variability on the dif-

ferent parameters. Interestingly, our correlation was achieved using a much

smaller set of individuals than many of the earlier studies.

The most robust correlation that we observed was the close correlation be-

tween the amount of dumping elicited by head movements in 1G and suscepti-

bility: a positive correlation at the p = .01 significance level, with a sample of

only ten individuals. This observation is extremely interesting because coupled

with the positive correlation between duration of time constants for sudden

stops and susceptibilities, it suggests that the greater the velocity storage and the

greater the amount of dumping elicited by head movements in 1G, the greater

the susceptibility to motion sickness in parabolic flight. And, it may be, that

these correlations will hold for other provocative situations as well.

We know for example that many other situations on Earth which involve

dumping are extremely provocative. For example, off-vertical rotation and bar-

beque spit rotation both involve modulation of velocity storage and are highly

provocative. A stable visual input can also suppress velocity storage or dump

velocity storage that already may have developed. For example, in the sudden

stop vestibulo-visual interaction test (Graybiel & Lackner, 1980) participants

are exposed to repeated decelerations from constant velocity. The first part of

the test is conducted with the eyes closed (the first 20 stops), the next parts with

SPACE ADAPTATION SYNDROME 99

the eyes open. The eyes-closed portion of the tests would involve velocity storage

and is relatively unprovocative for most individuals; however, the parts with full

vision to suppress and dump velocity storage are highly provocative.

The present observations may also be useful in understanding other situations

in which motion sickness develops. For example, if head movements are made

during rotation in a slow rotation room, there will be cross-coupled stimulation

of the semicircular canals (one set of semicircular canals will gain angular mo-

mentum and another set will lose it as the head is tilted with respect to the plane

of rotation). Such head movements are more provocative if full vision is allowed

compared with being blindfolded. This visual input would suppress velocity

storage. It should now be possible to see whether an individual’s susceptibility in —

different test situations correlates with the degree of dumping he or she exhibits

in that situation. If so, it would provide for the first time a purposeful way to look

at susceptibility differences across test situations and relate them to the activity

of known physiological mechanisms that can be behaviorally assessed.

We raised the point earlier that people can get motion sick in situations where

velocity storage does not occur or is not dumped. We would not expect the

measures of an individual’s capacity for velocity storage and sensitivity to

dumping to predict motion sickness in such situations. By extension, we would

not expect an individual’s susceptibility in a situation where charging and

dumping of velocity storage are prevalent to be correlated with susceptibility ina

situation where they were not evoked. For example, susceptibility during post-

rotary head tilts where velocity storage is first charged and then dumped would

not predict susceptibility during vertical oscillation where neither velocity stor-

age nor dumping are evoked. Vertical oscillation is a major component of the

stimulus for sea sickness.

These considerations are consistent with our view that patterns of individual

differences in susceptibility to motion sickness across situations will only be

comprehensible when we (1) understand basic spatial orientation mechanisms,

(2) can measure how individuals differ in the operation of such mechanisms,

and (3) can discriminate which control mechanisms are etiologically relevant in

particular provocative situations. |

An important aspect of future experiments concerning vestibulo-ocular adap-

tation to head movements during rotation will be to look at changes in the

vestibulo-ocular response to cross-coupling stimulation, to impulsive decelera-

tions, and to impulsive decelerations followed by head movements to produce

dumping. The pattern of variation in these situations should provide insight into

factors that elicit motion sickness, changes associated with adaptation of vesti-

bulo-ocular and motion sickness responses, and factors that determine whether

adaptation will transfer from one test situation to another.

100 LACKNER AND DIZIO

Acknowledgement

This work was supported by NASA Contract NAS9-15147 and NASA grants

NAG9-295 and NAG9-515.

References

Bles, W., deJong, H. A. A., & Oosterveld, W. J. (1984). Prediction of sea sickness susceptibility. In AGARD

Conference Proceedings, No. 372, Motion Sickness: Mechanisms, Prediction, Prevention and Treatment (pp.

27.1-27.6). Neuilly-sur-Seine, France: Advisory Group for Aerospace Research and Development.

deWit, G. (1954). Seasickness (motion sickness): A labyrinthological study. Acta Otolaryngologica,

116(Suppl.), 24.

DiZio, P., & Lackner, J. R. (1988). The effects of gravitoinertial force level on oculomotor and perceptual

responses to sudden stop stimulation. Experimental Brain Research, 40:485-495.

Gibson, E. (1974). Skylab 4 crew observations. In Proceedings of the Skylab Life Sciences Symposium, Vol. 1,

NASA TM X-58154, JSC-09275 (pp. 47-54). Houston, TX: Johnson Space Center.

Graybiel, A., & Johnson, W. H. (1963). A comparison of symptomatology experienced by healthy persons and

subjects with loss of labyrinthine function when exposed to unusual patterns of centripetal force in a

counterrotating room. Annals of Otology, Rhinology and Laryngology, 72:1-17.

Graybiel, A., & Lackner, J. R. (1977). Comparison of susceptibility to motion sickness during rotation at 30

rpm in the Earth-horizontal, 10° head-up, and 10° head-down positions. Aviation, Space and Environmen-

tal Medicine, 48:7-11.

Graybiel, A., & Lackner, J. R. (1980). A sudden-stop vestibulo-visual test for rapid assessment of motion

sickness manifestations. Aviation, Space and Environmental Medicine, 51:21-23.

Graybiel, A., & Lackner, J. R. (1983). Motion sickness: Acquisition and retention of adaptation effects

compared in three motion environments. Aviation, Space and Environmental Medicine, 54:307-311.

Graybiel, A., Wood, C. D., Miller, E. F., & Cramer, D. B. (1968). Diagnostic criteria for grading the severity of

acute motion sickness. Aerospace Medicine, 39:453-455.

Jennings, R. T., Davis, J. R., & Santy, P. A. (1988). Comparison of aerobic fitness and space motion sickness

during the Shuttle program. Aviation, Space & Environmental Medicine, 59:448-451.

Kerwin, J. P. (1974). Skylab 2 crew observations and summary. In Proceedings of the Skylab Life Sciences

Symposium, Vol. 1, NASA TM X-58154, JSC-09275, (pp. 55-59). Houston, TX: Johnson Space Center.

Lackner, J. R., & Graybiel, A. (1979). Some influences of vision on susceptibility to motion sickness. Aviation,

Space and Environmental Medicine, 50:1122-1125.

Lackner, J. R., & Graybiel, A. (1983). Some etiological factors in space motion sickness. Aviation, Space and

Environmental Medicine, 54:675-68 1.

Lackner, J. R., & Graybiel, A. (1986). Head movements made in non-terrestrial force environments elicit

symptoms of motion sickness: Implications for the etiology of space motion sickness. Aviation, Space and

Environmental Medicine, 57:443-448.

Lackner, J. R., & Graybiel, A. (1987). Head movements in low and high gravitoinertial force environments

elicit motion sickness: Implications for space motion sickness. Aviation, Space and Environmental Medi-

cine, 58(Suppl.), A212—A217.

Money, K. E., & Friedberg, J. (1964). The role of the semicircular canals in causation of motion sickness and

nystagmus in the dog. Canadian Journal of Physiology and Pharmacology, 42:793-801.

Raphan, T., Cohen, B., & Matsuo, V. (1977). A velocity storage mechanism responsible for optokinetic

nystagmus (OKN), optokinetic afternystagmus (OKAN), and vestibular nystagmus. In R. Baker & A.

Berthoz (Eds.), Control of gaze by brainstem neurons (pp. 37-47). Amsterdam: North-Holland Biomedical

Press.

Thornton, W. E., Hoffler, G. W., & Rummel, J. A. (1974). Anthropometric changes and fluid shifts. In

Proceedings of the Skylab Life Sciences Symposium, Vol. 2, NASA T™ X-58154, JSC-09275, (pp. 637-

658). Houston, TX: Johnson Space Center.

Journal of the Washington Academy of Sciences,

Volume 81, Number 2, Pages 101-109, June 1991

Individual Differences in Air Traffic

Control Specialist Training Performance

Carol A. Manning

FAA Civil Aeromedical Institute, Oklahoma City, OK

ABSTRACT

This presentation provides an historical perspective of the Federal Aviation Administra-

tion’s (FAA) investigations into the role of individual differences in predicting performance

in FAA Air Traffic Control Specialist (ATCS) training programs. Most of the research in this

area has been done to identify tests that could be used to select air traffic controllers.

Introduction

Before discussing ATCS selection procedures, it is necessary to briefly de-

scribe the job of the air traffic controller. There are three types of air traffic

controllers. En route controllers work at one of about 22 Air Route Traffic

Control Centers (ARTCCs) in the United States. Their primary duty is to con-

trol trafic moving between airports, although some may clear or approve traffic

to take off or land at uncontrolled airports (those not manned by other con-

trollers). En route controllers work high altitude sectors in which the aircraft are

moving at high altitudes and at high speeds, or low altitude sectors, where

aircraft are beginning to converge on an airport, and are in the process of begin-

ning to slow down and move closer together.

Terminal controllers control traffic during arrivals and departures from spe-

cific airports. Controllers in the tower cab environment control aircraft as they

move on the ground and as they take off or land; their control extends to about a

5-mile radius of the airport. Terminal radar controllers, or “approach con-

trollers,” control the traffic within about 5 and 40 miles of the airport, providing

coverage between the en route and tower cab controller. In the United States,

there are some four hundred terminal facilities, both tower cabs and radar

approach facilities. Within those categories of facilities there are different facility

levels, indicating different numbers and types of aircraft controlled and there are

101

102 MANNING

also a few other distinctions between types of facilities that provide different

services to aircraft.

Flight Service Station (FSS) controllers file flight plans (primarily for general

aviation pilots), interpret weather information so they can brief pilots on

weather conditions, and provide services to orient lost aircraft. Because FSS

controllers do not formulate clearances to separate aircraft, they are usually

considered independently from the other controllers with regard to part of the

selection process and with regard to their career progression.

The question with regard to the en route and terminal controllers is whether

they perform different jobs. All controllers formulate clearances to ensure air-

craft separation; all communicate with pilots and coordinate activities with

other controllers. On the other hand, their job functions differ because: 1) the

aircraft they control are moving at different speeds and are in different stages of

converging on an airport; 2) they may control different numbers of aircraft; and

3) they may or may not use radar as a tool. Radar allows the controller to have a

two-dimensional representation of a part of the airspace as a cue for what is

going on in the three-dimensional real world. Most en route controllers use

radar unless they work in areas where radar coverage is not available (e.g., over

the ocean). Terminal approach controllers use radar (but employ different

equipment than the center controller) to sequence aircraft into the vicinity of an

airport. Tower cab controllers at only the largest facilities have a small radar

scope called BRITE to aid in identifying positions of aircraft. However, these

controllers do not use the radar to ensure that separation is maintained between

aircraft; they also do not receive radar training. Other tower cab controllers do

not have access to radar at all. In the absence of radar (which happens only

infrequently), controllers use nonradar procedures, which means they keep

track of an aircraft’s movement through their airspace without a visual represen-

tation. They are aided in keeping track of this information by paper flight

progress strips on which they indicate the aircraft’s flight plan and times when

the aircraft is expected to be at a certain location. Tower cab controllers may

sometimes make notes on a pad of paper about aircraft being controlled and

what they have been approved/cleared to do.

Clearly, air traffic control is a very complex job environment. Before 1964,

the only selection criterion for an ATCS was having prior military experience

performing the job of air traffic controller. Research had been conducted since

the 1940s to identify tests to select controllers, but most of the tests used initially

were job samples specifically developed to measure skills used on the job. Such

tests as memory for flight information, coding flight data, receiving oral mes-

sages, coordinating clearances, devising clearances, and issuing oral communica-

tions were developed as predictors, not as criteria. The development of job

ATC SPECIALIST TRAINING . 103

AIRCRAFT ALTITUDE SPEED ROUTE

10 7000 480 AGKHC

20 7000 480 BGJE

30 7000 240 AGJE

40 6500 240 CHKJF

50 6500 240 DIKGB

60 8000 480 DIKJE

70 8000 480 FJKID

SAMPLE QUESTION

WHICH AIRCRAFT WILL CONFLICT?

04812

Seg AKIE on MILEAGE SCALE

C. 20 AND 30

D. NONE OF THESE

Fig. 1. Multiplex controller aptitude test (MCAT) item.

sample tests continued—the first Civil Service Commission (CSC) battery con-

tained a test called Air Traffic Problems. The current Office of Personnel Man-

agement (OPM) selection battery implemented in 1981 contains a test called the

Multiplex Controller Aptitude test (MCAT); an example problem is shown in

Figure i. While these tests may be predictive of performance, it is not clear from

looking at the test what they really measure--the MCAT clearly does not mea-

sure a single aptitude construct.

Studies on Performance Prediction

Harris (1987) administered a set of marker tests to Academy entrants to

- examine their relationship to the MCAT; the MCAT had high correlations with

marker tests that measure Integrative Processes, General Reasoning, Spatial

Orientation, Logical Reasoning and Spatial Scanning. Low correlations were

found between MCAT and verbal comprehension.

These findings are consistent with findings from other studies conducted

throughout the 1960s. During the 1960s a number of aptitude and temperament

tests were administered to a group of controller trainees, and studies were con-

ducted to follow up on their performance for several years afterward (Brokaw,

1984). The tests administered included eight tests of computational and arithme-

tic reasoning, six tests of perceptual and abstract reasoning, 4 verbal tests, two

tests of perceptual speed and accuracy and four temperament tests. The tests

104 MANNING

Table 1.—Biographical Data as Predictors

e High school grades (math, physical sciences)

e Self report—likelihood of remaining in FAA ATC work

e Self ratings of future performance

e Number of times aptitude battery taken

e Prior ATC experience

e Age at Academy entry

;b+1+++

used were from the Differential Aptitude Test (DAT), California Test of Mental

Maturity, Air Force tests, and California Test Bureau temperament tests. The

temperament tests had virtually no relationship with the criteria measures and

verbal tests had correlations below .2. Of the other tests, those contributing most

to prediction of criterion measure were Air Traffic Problems, Arithmetic Rea-

soning, Symbolic Reasoning, Code Translation and prior experience.

Just as Arithmetic Reasoning and Abstract Reasoning have consistently been

found to be predictors of performance in ATC training, so has prior ATC experi-

ence. While not sufficient to use as the sole selection criterion, prior experience

contributes to the prediction of performance. In more recent years, it has been

determined that control of air trafic using IFR (Instrument Flight Rules) proce-

dures is significantly related to performance at the Academy, but prior control of

air trafic using VFR (Visual Flight Rules) procedures is not (VanDeventer and

Baxter, 1984). For a number of years, even after implementation of the initial

aptitude battery, simply having had prior ATC experience qualified an appli-

cant for extra credit points to be added to the rating based on their OPM test

scores; now applicants can only gain extra credit points if they demonstrate job

knowledge via a paper and pencil test of occupational knowledge.

Other biographical factors have also been found predictive of training perfor-

mance (Table 1). A biographical questionnaire is administered to Academy

students during their orientation period (first 3 days of class). It has been found

in a number of studies looking at thousands of students who entered the Acad-

emy between 1980 and 1987 (e.g., Collins, Manning, and Taylor, 1984; Collins,

Nye, and Manning, 1990) that prior ATC IFR experience, grades in high school

math courses, and self-reports of both expectations of future performance and

likelihood of remaining in ATC work are positively correlated with training

success, while the number of repetitions of the aptitude battery and age at entry

are negatively correlated.

The age factor is especially interesting (Table 2). Cobb and Nelson (1974)

looked at the same group of trainees at two different points in their career—at

completion of Academy training, and at completion of field training. Age was

negatively correlated with attrition in both early and later training. This finding

was substantiated by later studies; older students have consistently been found

ATC SPECIALIST TRAINING 105

Table 2.—Age as a Predictor of Attrition

1. 1969 Academy entrants (Cobb & Nelson, 1974)

% Academy % Field % Retention

Age Attrition Attrition after 4 yrs N

<30 iss 18.9 66.0 1684

>30 35:5 24.7 39.8 665

2. Subsequent studies of Academy attrition (Kegg, 1988; VanDeventer, 1984)

% Attrition from Academy

1977-82 1982-88

Age Entrants Age Entrants

<24 26% <24 34%

25-30 39% 25-29 45%

>30 48% >30 53%

si Field training attrition for 1981-85 Academy graduates (Manning, 1988)

% Attrition from field training

Age En route Terminal

<25 28% 13%

26-29 40% 16%

>30 52% 26%

more likely than younger students to attrite (fail or withdraw) from Academy

training and from later field training as well, in spite of the requirement limiting

the maximum entry age to 30 for new hires.

A number of tests and biographical factors have been identified that have

been found consistently over the years to be predictive of success in training.

What has not yet been discussed is the criteria measures used in these studies. In

the early studies of the 1950s, the criteria were usually “lecture grades and

instructor ratings” (Brokaw, 1984); follow up studies used supervisor ratings as

criteria. Studies conducted during the 1960s used instructor and supervisor

ratings, course grade composites, as well as attrition, career progression, and

disciplinary actions. In these studies supervisor ratings were considered the

‘‘real”’ measures of job performance and the other criteria used were evaluated

on how effectively they predicted the supervisor ratings.

More measures that might be used as criteria have become available in recent

years. Some of these measures might be obtained from the Academy nonradar

screen program. The Academy screening program has not yet been addressed

during this presentation. In one sense the Academy screen is a training program.

Students spend 5 weeks learning aviation and ATC concepts and 4 additional

weeks being tested on their ability to apply those concepts in a set of simulated

106 MANNING

BLOCK TESTS

PHASE TESTS; erie am 12% MANDATORY

LAB AVERAGE 18% INSTRUCTOR

(best 5 of 6) 60% = ASSESSMENT

TECHNICAL

30% ASSESSMENT

CONTROLLER | 20% | (errors)

SKILLS TEST

Fig. 2. Nonradar phase weighting system.

laboratory problems which comprise 60% of the final grade (see Manning, Kegg,

and Collins, 1989, and Della Rocco, Manning, and Wing, 1990, for details). The

Civil Aeromedical Institute obtains all test (and item) scores and the laboratory

scores as well as an indication of the number of errors made in various categories

on each laboratory problem. However, there is a problem involved in using

these scores as criteria measures, although they are commonly used because of

convenience. First, the Academy program is a selection procedure; if candidates

don’t pass, they are fired. The implication of that is that the program occurs at

the beginning of a student’s career and thus does not measure how well students

perform on the job. In fact, the correlation between the score in the Screen and

status in later training (corrected for restriction in range but not for attenuation)

is only about .44 (Manning, Della Rocco, and Bryant, 1989). Furthermore, the

Screen measures performance on the laboratory problems under timed condi-

tions; it is clear that not all candidates have learned to perform the problems

when they are tested; thus, we are not measuring asymptotic performance, but

are instead measuring performance that occurs somewhere on the learn-

ing curve.

Second, the laboratory problems are graded by instructors who are former air

traffic controllers. It can be seen in Figure 2, the lab grade is comprised of a

technical assessment (TA; based on the number of various types of errors made)

and an instructor assessment (IA), which is another subjective instructor rating.

The IA suffers from the same problems encountered with most subjective rating

systems. Furthermore, the TA score contains error as well; technical grading 1s

ATC SPECIALIST TRAINING 107

done by instructor counts of the number of errors made. Despite institutional

procedures designed to standardize instructor grading, it.is not possible to en-

sure that observation of performance occurring under varying conditions is

conducted in a standardized fashion. In fact, the TA and IA are correlated with

each other in the range of .8 to .9, and the IA has a slightly higher correlation

than the TA with field training status. However, it is not clear what contributes

to the measurement of the criterion.

Third, the laboratory problems in the Screen are based on nonradar scenarios,

not radar. Nonradar control problems have been found to predict retention in

later field training that is based on radar principles, but the screen and field

training do not involve the same job tasks. The rationale for using a nonradar-

based program for the screen is that if the student can control traffic using

nonradar, they'll be able to use radar; that is, if they can keep the information

straight in their head without radar, then they can surely do it if they have a

two-dimensional cue to help them. However, you can easily find controllers

who knew someone who wasn’t successful in radar training after passing non-ra-

dar, but that hypothesis has never been empirically tested.

While using performance on nonradar ATC problems may be an appropriate

predictor of future performance, using such measures as criteria representing

job performance is probably not at all appropriate, for the reasons described

above. However, if the Screen is considered an inappropriate criterion, what

other measures are available to use as appropriate criteria? Several measures of

performance in laboratory simulation and on-the-job training are available.

Information is collected for every controller trainee regarding every phase of

field training they complete. This includes the dates they took the course, the

number of OJT (on-the-job training) hours required, the grade, and a rating of

their performance as compared with the performance of all others previously

rated made by an instructor or supervisor or other person knowledgeable of the

student’s performance.

The measures of field training performance have a variety of problems which

limit their utility as criteria. The available information isn’t detailed enough to

describe specific strengths and weaknesses in training performance. Further-

more, much of the information may be contaminated. For example, elapsed

time by training may be affected by operational policy because the trainee may

have to work on a sector on which he or she has been certified instead of training

on a new sector (a sector is a piece of airspace for which the controller is responsi-

ble; training occurs on one sector at a time). The complexity (the number and

types of aircraft and their configurations in the airspace) of individual sectors

differs in relation to each other, and students are not assigned to sectors in any

particular order (e.g., they don’t always train on the easy one first). Furthermore,

108 MANNING

students may have more than one OJT instructor. During an evaluation of

facility training in which the author participated, training records showed that

one student had 25 OJT instructors on a sector, and another had 9 instructors in

one day.

Facility philosophies differ regarding training. Some facilities consider train-

ing to be a high priority activity and require all trainees to obtain a certain

amount of training per week. Other facilities do not consider training to be as

important, and trainees at those facilities may not receive training as often.

There are differences between options as well; radar training may occur after

nonradar at terminal facilities, while nonradar and radar associate training

usually occur before, or at the same time as, radar training in the en route

option. Finally, air traffic control is considered to be almost an art or craft rather

than a standardized job. Controllers utilize different techniques to control air-

craft, and as long as their individual techniques maintain aircraft separation and

promote efficiency (and conform to the generally defined operating proce-

dures), the methods or techniques used are left to the discretion of the controller.

Because of the factors described above, it is clear that there 1s much error

included in the measurement of training performance. In spite of the errors of

measurement involved, performance in the Screen program still has a correla-

tion (adjusted for restriction in range) of .44 with field training status. On the

other hand, it is possible that the correlation between the two measures reflects

the degree to which the students get along with their instructors, both at the

Academy and when they reach field training.

Summary

Clearly, measurement of individual differences in the performance of air

traffic control specialists is complicated by the complexity of the occupation, the

resulting complexity that must be incorporated in their training, and a number

of factors which introduce error into the measurement of training and occupa-

tional performance. Considerable research has been conducted which identifies

aptitude and biographical factors which have consistently been found predictive

of success in training. However, the message that should be conveyed by this

discussion is that we need to be very careful about using performance in the

selection/training screen program or measures of performance or success in field

training as performance criteria. We must better understand how controllers

perform the job, in order to obtain better measures of job performance. When

those performance measures have been developed, it will be necessary to review

past findings and conduct additional studies to determine to what extent the

ATC SPECIALIST TRAINING 109

predictors of performance identified during years of prior research are still pre-

dictive of new performance-based criteria.

References

Brokaw, L. D. (1984). Early research on controller selection, 1941-1963. In S. B. Sells, J. T. Dailey, & E. W.

Pickrel (Eds.), Selection of air traffic controllers. Report No. FAA-AM-84-2. Washington, DC: FAA Office

of Aviation Medicine.

Cobb, B. B., & Nelson, P. L. (1974). Aircraft-pilot and other pre-employment experience as factors in the

selection of air traffic controller trainees. Report No. FAA-AM-74-8. Washington, DC: FAA Office of

Aviation Medicine.

Collins, W. E., Manning, C. A., & Taylor, D. K. (1984). A comparison of prestrike and poststrike trainees:

Biographic factors associated with Academy training success. In A. D. VanDeventer, W. E. Collins, C. A.

Manning, D. K. Taylor, & N. E. Baxter. Studies of poststrike air traffic control specialist trainees: I. Age,

selection test performance, and biographic factors related to Academy training success. Report No. AM-84-6.

Washington, DC: FAA Office of Aviation Medicine.

Collins, W. E., Nye, L. G., & Manning, C. A. (1990). Studies of poststrike air traffic control trainees: III.

Changes in demographic characteristics of Academy entrants and biodemographic predictors of success in

air traffic controller selection and Academy screening. Report No. FAA-AM-90-4. Washington, DC: FAA

Office of Aviation Medicine.

Della Rocco, P. S., Manning, C. A., & Wing, H. (1990). Selection of air traffic controllers for automated

systems: Applications from today’s research. Report No. FAA-AM-90-13. Washington, DC: FAA Office of

Aviation Medicine.

Harris, P. A. (1987). A construct validity study of the Federal Aviation Administration Multiplex Controller

Aptitude Test. Washington, DC: U.S. Office of Personnel Management.

Kegg, P.S. (1988). Relationship between age and performance in the FAA ATCS Academy screening program.

Unpublished manuscript.

Manning, C. A. (1988). Relationship between age and performance in ATCS field training. Unpublished

manuscript.

Manning, C. A., Della Rocco, P. S., & Bryant, K. (1989). Prediction of success in FAA air traffic control field

training as a function of selection and screening test performance. Report FAA-AM-89-6. Washington, DC:

FAA Office of Aviation Medicine.

Manning, C. A., Kegg, P. S., & Collins, W. E. (1989). Selection and screening programs for air traffic control.

In R. S. Jensen (Ed.), Aviation psychology. Surrey, UK: Gower Publishing Company Limited.

VanDeventer, A. D., & Baxter, N. E. (1984). Age and performance in air traffic control specialist training. In

A. D. VanDeventer, W. E. Collins, C. A. Manning, D. K. Taylor, & N. E. Baxter. Studies of poststrike air

traffic control specialist trainees: I. Age, selection test performance, and biographic factors related to Acad-

emy training success. Report No. AM-84-6. Washington, DC: FAA Office of Aviation Medicine.

Journal of the Washington Academy of Sciences,

Volume 81, Number 2, Pages 110-128, June 1991

Classification Efficiency

and Systems Design

Joseph Zeidner and Cecil D. Johnson

The George Washington University, Washington, DC

ABSTRACT

The broad objectives of our investigation were: to provide a description of some major

validity findings in selection/classification and to give results of a utility study of the Armed

Services Vocational Aptitude Battery (ASVAB); to outline several theoretical and practical

controversies involving classification efficiency; and to detail future manpower, personnel

and training considerations in systems design.

Introduction

The traditional interest of individual differences measurement researchers

has emphasized predictor and criterion measures, as well as predicted job perfor-

mance and the expression of benefits and costs in utility terms such as productiv-

ity gains in dollars. A critical first step in developing a classification-efhicient

job-matching technology is the adoption of a theoretical framework. Three

viable competing theories currently exist to explain predicted performance. An

earlier theory, situational specificity, was influential a decade or more ago, but

today serves mainly as a strawman, advancing a g-based theory as recently

represented by validity generalization proponents and others. A second theory,

specific aptitude, is derived from the tradition of Thurstone’s primary factors.

Both theories are directed at incremental predictive validity as affected by the

structure of the predictor space. The third theory, differential assignment theory

or DAT, has a dual focus: (1) on incremental predictive validity—for single job

selection, and; (2) on mean predicted performance (PP) for each job computed

after assignment from a common pool—for personnel classification processes.

DAT is directed at the structure of the joint predictor-criterion space.

General cognitive ability theory suggests that one general cognitive ability

factor underlies all valid specific cognitive abilities. The underlying variable, g,

causes specific aptitudes to have validity in predicting job performance. If it is

110

CLASSIFICATION AND EFFICIENCY 111

true that there is a single factor (g) underlying specific aptitudes, and specific

aptitudes do not provide any greater prediction than g alone, assuming predic-

tive validity is an adequate index of classification efficiency, then the efficient

classification of individuals to jobs based on specific aptitudes, or on general

cognitive aptitude, is not a pertinent issue. However, if there are several factors

that differentially predict performance in various jobs (even though the specific

factor validities do not exceed the validities of g for each job), then classification

is a relevant issue and DAT can make a contribution to both science and the

work place.

Specific aptitude theory on the other hand, suggests that job performance is

best predicted by one or more specific aptitudes required by the job, rather than

by general cognitive ability. For example, performance as an editor would better

predicted by verbal and perceptual speed abilities than by g. According to this

theory, g has only an indirect relation to job performance since it is mediated by

specific aptitudes, 1.e., results from the correlations among specific aptitudes.

This theory strongly contributes to the concept of situational specificity to ex-

plain subtle differences in job requirements in different settings.

We propose the third theory, DAT, postulating that several factors differen-

tially predict performance in various jobs, providing a coherent framework for

job classification. DAT stresses the existence of real differences among predicted

performance measures and emphatically denies that classification efficiency is a

function of mean predictive validity alone. Instead DAT proposes that classifica-

tion efficiency is approximated by the mean intercorrelation among predictive

performance measures, the number of jobs to which individuals are assigned,

the selection ratio, and mean predicted validity. DAT states that the joint pre-

dictor-criterion space is multidimensional and factors other than an unidimen-

sional g factor contribute a non-trivial amount of classification efficiency.

The key to success in manning a new system with sufficient numbers of

trained personnel is early identification of the manpower requirements. Man-

power, personnel and training (MPT) planning is an integral part of the design

and acquisition process. The MPT programs of the military services provide

methods for systematic analysis throughout the development process and, when

applied from the outset, result in the manpower issues being addressed with

enough lead time to allow MPT to adequately respond to requirements. A more

advanced goal of MPT systems is to interact early in the consideration of design

alternatives and evaluate trade-offs between MPT and designs and thus go

beyond merely supporting MPT requirements of new systems.

A recently completed National Research Council study was concerned with

the next generation of aircraft carriers that will become operational around the

year 2040. The sponsors of the study were interested in improving manpower

112 ZEIDNER AND JOHNSON

Table 1.—Needs to Improve Manpower Utilization on Future Aircraft Carriers

Weapons systems are increasingly complex and numerous.

MPT cost to support Navy weapon system is 50% of the total life cycle cost of a system.

Youth population will decrease during the next decade; Improves slightly but will not reach levels of 1970s

in the early years of next century.

Youth composition will have large number with academic deficits; Navy now recruits 11% in CAT IV

(lowest aptitude ratings).

utilization for the reasons listed in Table 1. Note that the MPT cost to support a

system is 50 percent of the total life cycle cost of a system. Also note the decline

in the youth population during the next several decades. This indicates that it

may be more difficult and expensive to recruit for the military services.

Modern carriers are manned by 6000 personnel; the number of authorized

maintenance ratings alone on a carrier is greater than 3000. Thus a focus on the

possible reduction in the number of maintenance personnel appears to be a key

approach to manpower reduction. Such a reduction would not only permit

alternative use of space and considerable monetary savings, but if accompanied

by the application of improved selection and classification techniques would

also result in large productivity gains. Among the approaches to bring about

potential improvements in manpower utilization in maintenance systems 1s

condition-based maintenance. The objective of condition-based maintenance is

to identify maintenance tasks by detecting degraded or potentially degraded

performance by electronic monitoring, rather than responding to failures with

corrective maintenance and scheduling “‘open and inspect” tasks of preventive

maintenance. The implementation of condition-based maintenance requires a

new philosophy of systems and equipment designs, repair procedures and main-

tenance training. In contrast, the approach detailed in this report concerns

advanced selection/classification techniques and measuring their utility.

Selection

The Army General Classification Test (AGCT) was quite successful in select-

ing men for specialist training during World War II, as evidenced by the magni-

tude of the many hundreds of validity coefficients (r) obtained, a few of which

are shown in Table 2 (PRS staff, 1945). Since most of the samples for these

studies had been pre-selected on the AGCT or on some highly correlated factor,

the obtained relationships were, in general, quite restricted and thus consider-

ably underestimated the operational or true effectiveness of the tests. Also the

standard error of r varied greatly as a function of sample size. Validity findings

shown in Table 2 strongly contributed to the concept of situational specificity,

CLASSIFICATION AND EFFICIENCY | 113

Table 2.—Training Validities for the Army General Classification Test (World War II)

Tested Population N* r

ADMINISTRATIVE CLERICAL TRAINEES

AAF 2947 .40

AAF 123 44

Armored 119 35

MAAC 199 62

OFFICER CANDIDATES

Infantry 103 .30

Ordnance 190 41

Signal Corps 213 36

Tank Destroyers ae 44

Transportation 314 38

WAAC 787 46

Infantry 201 ale

Ordnance 190 09

Combat Arms (13) 5186 28

_ Source: Personnel Research Section (1945).

1.e., validities of the same tests for the same jobs, but in different settings, varied

because of subtle differences in job requirements. Note, for example, that the

four AGCT validities given for clerical courses ranged from .35 to .62 and that

validities for the Officer Candidate courses ranged from .09 to .46, using grades

as the criteria. Such results served to reinforce the perception of the need for

empirical validation of tests for each new application.

Ghiselli made it his life’s work to aggregate selection validity data from the

1920s through the 1960s. He analyzed an enormous amount of validity infor-

mation, with many of the results looking like the findings shown in Table 2. He

classified predictors into 20 different test types and jobs into 21 different job

families and reported the grand average validity across all tests and jobs as .39

for training and .22 for job performance (uncorrected for attenuation and range

restrictions). Ghiselli (1959) expressed his dismay at the lack of validity general-

ization: “A confirmed pessimist at best, even I was surprised at the variation in

. findings concerning a particular test applied to workers on a particular job. We

never anticipated them to be worlds apart” (pp. 397-398). Hunter & Hunter

(1984) reanalyzed Ghiselli’s data, correcting for criterion unreliability and for

restriction in range, and reported an increase of r to .48 for job performance.

Such findings led Guion (1976) to state that the inability to generalize validity

findings from one setting to another was the major hurdle separating selection as

technology from selection as science.

Charles Mosier (1951) first introduced the concept of validity generalization

and later Schmidt and Hunter (1977) developed a procedure for correcting for

statistical artifacts to arrive at wider generalizations of validity across jobs and

work settings than had been recognized earlier.

114 ZEIDNER AND JOHNSON

Table 3.—Average Training Validities of ASVAB Composites for Four Job Families by Military Service

JOB FAMILY

Military [OT eh iG MTR Hea ences

Service M&C B&C E&E HS & T Total

Army 49 45 49 50 .48

Air Force .70 14 AE] 74 714

Navy 50 49 SS) 58} 5 1

Marines 58 58 5 61 58

Total 56 2) Dy) sy) 58

Source: Hunter, Crosson, & Friedman (1985). M & C = Mechanical and Crafts; B & C = Business and

Clerical; E & E = Electronics and Electrical; HS & T = Health, Social and Technology. Number of jobs = 180;

Total N = 103,700.

Table 3 shows the average corrected (for range restriction) validities for four

military occupational composites in each service against training success criteria

(final course grades) reported by Hunter, Crosson, and Friedman (1985). It

should be noted, however, that for the Army validities, skill qualification test

scores obtained about a year after administration of the Armed Services Voca-

tional Aptitude Battery (ASVAB) were used rather than training grades. Validi-

ties obtained for the Army sample more properly should be considered as job

proficiency validities using a job knowledge criterion. The different nature of the

criteria and differences in time between ASVAB and criterion testing reduces

the interpretability of between-service comparisons. Mean validities were:

Army, .48; Air Force, .74; Navy, .51; and Marines, .58. The grand mean, across

services and 190 jobs for a sample size of 103,700, was .58. The higher level of

Air Force validities against course grades, might be attributed to the more tech-

nical content or higher complexity of jobs in that service or to methodological

differences in the criteria or to increased criterion reliability (with increased

course length).

Hunter et al. (1985) drew a very significant conclusion after analyzing the

occupational composite validities for each job family in each service. If different

aptitudes best predict different job families, the validity of each occupational

composite should be highest for its own associated job family and lower for the

other job families. Such a result would be indicative of differential validity. The

results, unfortunately, indicated that each occupational composite is almost as

valid for other job families as for its own. The conclusion reached, then, is that

the ASVAB operational composites provide high predictive validity but little

differential validity across job families. To translate differential validity findings

into a precise measure of classification efficiency, however, would require a

simulation study. The results should not be construed as evidence that the

ASVAB as a battery does not possess potential allocation efficiency (PAE) nor

CLASSIFICATION AND EFFICIENCY 115

Table 4.—Mean Validities of GATB for Training and Job Proficiency at Five Levels of Job Complexity

Training Outcomes Job Proficiency Measures

Job Complexity Multiple Multiple

(decreasing) GVN_ SPO" KEM R GVN SPQ KFM R

Setting up 65 53 .09 .65 56 ys .30 Sy)

Synthesizing/coordinating 50 .26 a3 .50 58 so B) 21 58

Analyzing/compiling

computing Sy 44 ail 9 I 40 oy 58)

Comparing/copying 54 a J) 40 59 40 35 43 50

Feeding/off bearing — os — — 8) .24 48 250)

Weighted Mean 55D Al .26 7 45 oi), oT 153

' Source: Hunter (1983). Test categories: GVN = General Mental Ability; SPQ = Perceptual Ability; KFM

= Psychomotor Skill. Number of jobs = 515. Total N = 39,500.

that full least-square (FLS) test composites of the ASVAB would not demon-

strate PAE. Hunter et al. results merely show the inadequacy of the current

ASVAB composites.

Table 4 shows the mean validities of the General Aptitude Test Battery

(GATB) for training as reported by Hunter (1983). Jobs are clustered into job

families on the basis of complexity rather than on task similarity. Hunter’s

method of ordering job complexity is based on Fine’s (1955) functional job

analysis dimension scheme for rating people, data and things. Of the 515 valida-

tion studies which were accomplished, 90 used criteria of training success and

425 used criteria of job proficiency. The average sample size of the studies was

75. The 515 jobs were considered representative of the entire work force job

spectrum. Validities were corrected for range restriction and attenuation. A

_ reliability of .80 was assumed for the training criterion of job knowledge and a

reliability .60 was assumed for the job proficiency criterion of ratings.

From Table 4 it can be seen that the validity of cognitive ability (GVN) for job

proficiency decreases from .56 to .23 with decreases in job complexity, and

conversely, psychomotor ability (K FM) validity increases with decreasing job

complexity. Results of cognitive ability for training, however, show a fairly high

validity regardless of job complexity, although the validity of .65 for the highest

complexity job level was much higher than the validity for the lower job com-

plexity levels, with validities ranging from .50 to .57. Psychomotor ability results

for training again show an increase of validity as a function of decreasing job

complexity. Hunter (1983) states that the findings for training are consistent

with the need for good psychomotor abilities for hands-on training situations.

116 ZEIDNER AND JOHNSON

Table 5.—Project A— Measurement Methods and Performance Dimensions

Measurement Method Performance Dimension Overall Performance

Hands-on Tests (Specific Skills)

Written MOS Tests (Specific Skills)

Supervisor Ratings (Specific Skills)

MOS-Specific

Knowledge and Skill

Hands-on Tests (Common Skills) Basic Soldiering

Written Tests (Common Skills) Knowledge and Skill

Ratings of Leadership/Effort

Awards and Certificates Leadership and Effort JOB PERFORMANCE

Combat Effectiveness

Ratings of Discipline

Avoiding Article 15 Personal Discipline

Being Promoted on Time

Ratings of Fitness

Physical Readiness bites Appeatatiee

Source: Wise, Campbell, McHenry, & Hanser (1986). MOS = Military Occupational Specialty.

On the other hand, cognitive ability increases in validity as a predictor of job

proficiency as job demands become increasingly complex. Similarly, jobs that

have low cognitive demands have a significant psychomotor demand. Of course,

there are jobs that are exceptions to this inverse relationship. Taken as a whole,

job complexity shows a strong effect on validity.

The complementary patterns of cognitive and psychomotor abilities lend

themselves to various ability combinations or multiple correlations. Table 4

shows that the average multiple correlation for a training criterion 1s .57 as

compared to an average of .55 using cognitive ability alone.

Importantly, Table 4 provides a clear indication of differential validity re-

quired for classification efficiency. For example, the GVN composite (general

mental ability) provides high prediction for high complexity job proficiency

criteria and low prediction for low complexity criteria. On the other hand, the

KFM composites (psychomotor) has the opposite pattern. This is strong evi-

dence of differential validity and for potential allocation efficiency.

Table 5 summarizes the measurement methods and performance dimensions

characterizing the common latent structure across nine different jobs in the

Army’s Project A, a large-scale longitudinal validation study. The latent struc-

ture model which was specified included the five job performance constructs

shown in Table 6. A confirmatory analysis showed that the overall model fitted

extremely well (Wise, Campbell, McHenry, & Hanser, 1986). The latent perfor-

mance structure appears to be composed of very distinct performance compo-

CLASSIFICATION AND EFFICIENCY 117

Table 6.—Validity of Cognitive, Non-Cognitive, and Combined Predictor Composites—Project A

Job Performance Criterion Factor

Technical Generali Personal Physical

Predictor Skill Skill Leadership Discipline Fitness

Cognitive composites (K = 11) 65 .69 roe aT 78)

Non-cognitive composite

(K = 13) 44 44 38 535 38

Combined cognitive and

non-cognitive composites

(K = 24) 67 0 44 oi) 42

Validity gain of combination 04 05 iM 21 22

Source: McHenry (1987). N = 4039. K = number of measures.

nents, and this suggests that different constructs could be predicted by different

types of tests. The pattern of validity gains shown in Table 6 confirms that

non-cognitive measures, for example, predict motivationally-based criteria bet-

ter than they predict specific job proficiency criteria.

However, later analyses indicated little potential for differential validity using

any criterion other than the specific job proficiency criterion. For generalization

across jobs, within each criterion factor, one equation fitted the data for the

other four performance components (Wise, McHenry, and Campbell, 1990).

In Table 6 the impact of considering a composite of 24 cognitive and non-

cognitive measures for each of the five job performance criteria is clear. Those

results highlight the magnitude of validities obtainable by considering all tests

types together. The increments over ASVAB validities for the total combined

composite of 24 cognitive and non-cognitive tests range from .04 for Military

Occupational Specialty (MOS) Specific Technical Skills to .22 for Military Bear-

ing/Physical Fitness. Of equal interest is the actual level of validities reached

against the five job performance criteria factors, with multiple correlations rang-

ing from .17 to .70. It may be inferred that if the five criterion components were

‘combined they would most likely contribute to increased selection efficiency

but not to increased classification efficiency.

Classification

Traditionally, in selection and placement, only a single job is involved, and

can be accomplished with one or more predictors. The outcome is determined

by an individual’s position along a single predicted performance continuum.

Classification decisions provide the basis for assigning a selected pool of individ-

uals to more than one job. As in selection, these assignments can be made on the

118 ZEIDNER AND JOHNSON

basis of a single predictor continuum adjusted to predict performance by reflect-

ing job validities and/or values. When the predictors are adjusted in such a

manner that the mean adjusted predictor scores and the mean criterion scores

have the same rank order across jobs, a hierarchical layering effect that makes a

positive contribution to the benefits obtainable from classification is evident. A

hierarchical layering effect due to either a variation across jobs of the validities of

job-specific test composites, or to the value assigned to each job and reflected in

predictor score means and/or variances, assures that the assignment process 1s,

at least in part, influenced by hierarchical classification.

Classification that does not capitalize on hierarchical layering effects is re-

ferred to as “allocation” (Johnson and Zeidner, 1991). While hierarchical classi-

fication can be unidimensional, e.g., based entirely on a single predictor, alloca-

tion requires multiple predictors measuring more than one dimension in the

joint predictor-criterion space. Validity is determined individually against each

job’s performance criterion; the set of job criteria should also be multidimen-

sional. Thus a classification battery requires a separate assignment variable

(criterion specific composite) for each criterion, if allocation is to be maximized.

Brogden (1959) directly linked measurement of classification efficiency (CE)

to mean predicted performance (MPP) and thus to utility. His allocation equa-

tion expresses MPP as a function of predictive validity, intercorrelations among

the least-square estimates (LSEs) of job performance, and the number of job

families. Often this equation is misunderstood, e.g., the intercorrelations in the

equation pertain to the full LSEs of job performance and not to the intercorrela-

tions among predictor composites. Also, the equation makes it clear that predic-

tive validity is only one term in the equation and thus classification efficiency

can not be adequately described by predictive validity alone.

An example of this misunderstanding is revealed in this remark by Murphy

and Davidshofer (1988), ‘““The [ASVAB] composites do not provide informa-

tion about different sets of aptitudes, and that, in fact, any one of the composites

could be substituted for another with little loss of information. For most of its

major goals, the military would do just as well with a short test of general

intelligence” (p. 208). The quote implies that the intercorrelations of aptitude

areas are so high that the use of multiple predictor composites are ineffectual.

Research shows that in ASVAB there is a mean r of .95 among LSEs of job

performance. But this high value still was capable of producing considerable

classification efficiency (Nord and Schmidt, 1991). Additionally, the comment

also implies that the predictive validity is the appropriate measure of classifica-

tion efficiency (CE) and that the current operational composites best reflect the

potential for differential validity.

Brogden’s (1949) historic equation for estimating costs and benefits of a selec-

CLASSIFICATION AND EFFICIENCY . 119

Table 7.—Brogden’s Cost-Benefit Equation

AO Nia SDPZ.. — NC

the total utility gains of a selection program;

the mean number of years selectees remain on the job;

the correlation between the predictor and job performance in the applicant population (the validity

of the predictor);

SD, = the standard deviation of job performance in dollars;

the average score of those selected in applicant population predictor standard score; and

the cost of selection per individual.

Howl

tion program as a means of demonstrating that testing can save money is shown

in Table 7. How much is saved depends on the predictive efficiency of the

selection device, the selection ratio and two important recently applied situa-

tional variables—the variance of dollar-valued performance to the organization

and costs associated with testing and recruiting.

Despite the availability of Brogden’s equation since 1949, utility analysis had

not received widespread attention until recent years when practical and rational

means of estimating job performance in dollars were developed, such as the

global estimates procedure described by Hunter and Schmidt (1983).

Utility of the ASVAB

In fiscal year 1987, 315,000 enlistees entered the All Volunteer Force; and of

these, 130,000 or 41 percent were recruited into the Army. The Services rely

heavily on aptitude information (ASVAB), since most recruits have little or no

work experience. The Services’ selection and assignment systems are dependent

on an interrelated set of complex factors including policies, goals, recruiting

resources, recruiter incentives, formal and informal enlistment standards, the

- willingness of young people to enlist, and the efficiency of the job assignment

system in person-job matching.

Nord and Schmitz (1991) conducted an empirical analysis of productivity

gains attributable to simultaneous changes in job entry standards (minimum

cutting scores), assignment policies and assignment procedures to provide deci-

sionmakers with realistic information in making rational choices for allocating

scarce resources among alternative strategies. Taken together, the need for real-

ism and the need to consider opportunity costs imply that, in order for this

utility analysis to be useful, it must be context-specific and credible. A total of

thirty-three different policies were analyzed. Eleven different job assignment

policies and procedures were first simulated under 1984 enlistment entry stan-

120 ZEIDNER AND JOHNSON

Table 8.—Average Validities of Army Composite Used for Assignment to Job Families

Mean Validity

Job Family 1981? 1984” 1987° 19879

Clerical/Administrative 3 49 .60 59

Combat 56 44 4 AO

Electronic Repair 59 45 By)? .65

Field Artillery 63 45 39 = )3)

General Maintenance 16 .40 54 .60

Mechanical Maintenance ‘ey 45 .62 o>

Operators/Food awe 61 50 61 .60

Surveillance/Communication 25 47 5) By)

Skilled Technical 5 “Oi. 54 Bp)

* Maier, & Grafton (1981).

> McLaughlin, Rossmeissl, Wise, Brandt, & Wang (1984).

“McHenry (1987); Eaton (1987).

4 Zeidner (1987).

dards, then under the assumption that those standards were raised by five stan-

dard-score points for all Army jobs, and finally under the assumption of a

ten-point across-the-board increment in standards. All thirty-three policies were

simulated using a same random sample of 4,280 accessions from the 120,281

Army enlistments in 1984.

The utility analysis relied heavily on the work of previous researchers in this

area but extended previous work in several key respects, especially the use of

empirically based simulations, rather than theoretically derived values, to esti-

mate gains under alternative policies, and also the incorporation of realistic

labor market considerations.

Table 8 shows the average job performance validities for nine occupational

clusters of job families (all validities include corrections for range restriction).

The aptitude area composites are constructed from tests on the ASVAB. The

single composite validities used in the present analysis were obtained by com-

bining the validity results of three previous studies and the number of MOS in

each job family included in each study. The resulting weighted average, tabled in

the last column of Figure 8, fell within the range of previous estimates for each

job family and close to the unweighted average for the previous studies. The

main effect of this averaging was to dampen large and inconsistent variation in

validities across job families. Use of optimal full-least-squares (FLS) prediction

equations as the assignment strategy raised MPP .334 of a standard deviation

unit over random unit selection, and .143 of a standard deviation unit over the

current selection and assignment process. Thus, the FLS equations provided 1.7

times the increment of the current policy over random selection and assignment

in predicting MPP. As expected, increasing selectivity, i.e., raising job standards

CLASSIFICATION AND EFFICIENCY 121

Table 9.—Gross and Net Values and Predicted Performance Under Different Policies

Job Selection and Mean Predicted Gross Net

Classification Policy Performance Value Value

Random Selection and Assign .000 =i EY —152.4

Current Select/Random Assign .189 —21.4 —50.4

Current System shO7 0.0 0.0

EPAS aN 32.5 56.5

Optimal FLS 334 228.8 260.0

All values are relative to the CURRENT allocation, under CURRENT selection standards (in millions of

dollars per year). Gross present value is estimated value of performance gains without accounting for changes

in training and recruiting costs. Net present value is equal to Gross value minus these changes.

by 5 points, increased MPP within each of the sets of policies. The use of a

5-point increase in standards using the FLS assignment policy increases MPP

from 0.334 to 0.370 over random selection and assignment.

However, simply to know the impact of a policy on performance is not sufh-

cient. Increasing the job standards involves increasing the applicant pool, which,

in turn, increases recruiting costs. Therefore, performance gains are evaluated

via a benefit-cost model using utility analysis and an alternative economic analy-

sis based on opportunity costs.

Table 9 provides estimates of the gross value of mean predicted performance

change under different policies, as well as the net value under one recruiting cost

assumption. The gross values shown are the estimated present values of each

policy prior to accounting for recruiting and training costs produced by changes

in selection ratios and attrition rates. The net values are estimates produced after

changes in training and recruiting costs are accounted for. The gross value of the

performance gains produced by the current system is about $350 million an-

nually. However, when the large reduction in recruiting costs that could be

realized by moving to a 50 percent high-quality accession pool are taken into

account, the annual savings provided by the current system drop to $152

“million.

Even under the handicap of predictor score optimization, rather than pre-

dicted performance optimization, a computer-based efficient performance allo-

cation system, the Enlisted Personnel and Assignment System (EPAS), provides

significant gains over the current system, with estimated net gains under current

selection standards of $56 million annually.

The final point to be made with respect to these results is that the potential

gains from the use of the FLS predictors are extremely large, yielding estimated

net gains of $260 million under current standards, even when current Armed

Forces Qualification Test (AFQT) quality goals are enforced. Note that this gain

will be over $310 million larger than a current selection system with random

122 ZEIDNER AND JOHNSON

Table 10.—Estimated Cost of Achieving Equivalent Performance by Increasing AFQT CAT I-IIIA Acces-

sions*

Mean Required Avg

Job Selection and Predicted I-IIIA Cost Opportunity

Classification Policy Performance (%) ($K) Cost ($M)

Random Selection and Assign. .000 .46 4,649 —295.6

Current Sel/Random Assign. .189 58 8,142 —20.0

Current System 197 59 8,371 0.0

EPAS 22) .63 9,195 81.6

Optimal FLS 334 78 133155 573.0

* Using the current system.

assignment and $412 million over random selection and assignment (without

use of testing).

The most serious limitations of the foregoing net present-value (NPV)

method is the centrality of the assumption about the dollar value of a standard

deviation in performance. While there is persuasive empirical evidence that 40%

of salary is a conservative estimate, this rule-of-thumb approach is often per-

ceived as subjective, and therefore unreliable. This problem is exacerbated when

the rule is applied to public sector activities where no clear valuation of output is

possible. An alternative to the NPV approach that, in some circumstances, may

provide more useful information for the decisionmaker, is to focus attention on

the cost of obtaining a given level of performance using existing procedures

instead of attempting to directly measure the net value of the gains achieved

under different procedures, 1.e., to focus on opportunity cost of retaining the

existing system. Using an opportunity-cost approach in this context we ask,

‘‘What would it cost to achieve the levels of performance produced under each

evaluated policy if the mechanism used to achieve these gains was to simply

increase the number of high quality recruits and assign them using the current

system?”

Table 10 shows the opportunity costs. In general, the opportunity cost esti-

mates parallel those arrived at under the NPV method in terms of relative

magnitude, but are considerably higher in absolute magnitude. The estimated

cost of achieving the performance gains provided by EPAS under current selec-

tion standards through the recruitment of additional mental categories I-II[A

soldiers is $81 million, compared with the NPV estimated gains of $56.5 mil-

lion. (Mental category I is the highest category and category IV is the lowest

acceptable level for entry into the service.) The increased cost of using the

current system, combined with the recruiting of higher quality recruits to

achieve the performance provided by the optimal FLS option under current

standards, would exceed $570 million.

CLASSIFICATION AND EFFICIENCY 123

On the basis of our utility analysis we can suggest changes in military opera-

tional classification systems that are based solely on our simulation results. The

changes depend entirely on better utilization of information contained in the

present ASVAB. Only technical changes in assignment policy and procedures

are needed to obtain the productivity gains estimated. Specifically, we suggest

the use of: (1) mean predicted performance, 1.e., FLS composites, as the objec-

tive function for the assignment algorithm, rather than aptitude area composite

scores; (2) the FLS prediction equations as the assignment variables rather than

equally weighted and reduced numbers of tests comprising aptitude area com-

posites; and (3) an efficient optimal allocation algorithm. The assumption is

made that the preponderance of recruits could be persuaded to accept the jobs in

which they can perform best.

Effective Selection and Assignment in the Development of New Systems

Emphasis needs to be placed on more effective selection and assignment of

personnel in the development of new systems. The first step in this process 1s the

determination of qualitative personnel requirements of each new job in terms of

multidimensional measures of predicted performance. We have already pro-

gressed almost as far as we can go in the achievement of utility within the

constraints of a system which measures quality of input in terms of general

cognitive intelligence. The next step is the assessment of personnel resources

employing the same measures used to determine requirements. By matching

resources and requirements we obtain the constraints that must be applied in the

design of a system that makes the optimal use of personnel resources. We con-

tend that an optimal distribution of the applicant population to jobs in such a

way as to maximize predicted performance on the job requires multidimen-

sional predictors for selection, classification and assignment. Classification efh-

-clency must be considered in developing the selection and classification batter-

les, in the determination of sets of job families, and the determination of

corresponding test composites for making assignments to jobs.

Capabilities for meeting the manpower quality requirements of a new hu-

man-machine system under development can be increased at least three ways. If

we can afford the increased recruiting costs we can decrease the selection ratio,

that is reject more applicants. Or, we can reduce the complexities of jobs by

shredding out each specialty into two or more specialties. These approaches

provide means of trading off the required quantity (and cost) of applicants

against the quality of personnel that can be placed on the job. Second, design

and human-factor engineers can reduce qualitative requirements by simplifying

124 ZEIDNER AND JOHNSON

the tasks of the human component or by improvement of training aids. Third,

apply an economical means of increasing quality of assigned personnel through

the use of classification-efficient measures for both selection and assignment.

Applied psychologists working in the field of human-machine interface have

been heavily concerned with individual differences in the conduct of research

and development for new systems. Implementation of MANPRINT in the

Army, HARDMAN in the Navy, and IMPACT in the Air Force, assure that

MPT components of new systems are not ignored. Many of the aspects of MPT

can be improved by the substitution of classification-efficient multidimensional

test-batteries.

There are several primary factors for increasing utility, one of which is related

to classification eficiency (CE). CE is heavily dependent on the increase of

mean predicted performance on the job, 1.e., MPP, which can be best realized

through increasing classification efficiency in both selection (using multidimen-

sional selection) and classification. The other factors are reduced attrition and

recruiting costs, and trade offs between all three.

Utility is closely associated with the value of mean predicted performance on

the job in several ways. A higher MPP permits more effective hardware designs,

reduces training costs and provides higher skill levels on the job. Skills are

retained longer and productivity is generally increased—raising system per-

formance. ;

Effective selection and classification of personnel to multiple jobs requires the

integration of procedures and predictors that are classification efficient. The

development of such a personnel system draws upon the skills of operations

researchers, management, and computer specialists, as well as measurement

psychologists who provide the classification efficient measures. There is also a

role for the human factors specialist in assuring that human skills most predic-

tive of total system performance are identified and described to the measure-

ment specialists.

We define our classification efficiency (CE) index as the MPP standard score

computed after personnel have been assigned to the job from a common pool.

The potential classification efficiency (PCE) index is this same index except that

assignment to the job must be accomplished using an optimal assignment pro-

cedure in which MPP is maximized and each predicted performance (PP) score

is expressed as a least square estimate of the performance criteria for each job.

Each PP score is computed using the full classification battery.

When we undertake research relating to the classification efficiency of instru-

ments, test batteries, assignment (test) composites, or sets of job families, we use

PCE as the basis for making comparisons across alternative approaches (condi-

tions). In recent years, most measurement specialists have emphasized the 1m-

CLASSIFICATION AND EFFICIENCY 125

Table 11.—Common Erroneous Assumptions in Classification

Use of 3 or 4 tests in an assignment battery provides as much CE as 10 or more tests.

Maximizing SE will maximize CE without special consideration of CE in choosing predictors.

Since predictive validity of unit-weighted composites in cross samples may be superior to LSE weights, LSEs

are assumed to be less effective than unit-weighted composites for classification.

A small number of job families and associated test composites used as assignment variables can provide as

much CE as a larger number of families.

New predictors for use in classification batteries can be effectively selected solely on the basis of their

contributions to predictive validity.

provement of predictive validity in classification situations where PCE would be

more appropriate. A number of these specialists in the forefront of the validity

generalization movement have concluded that since a single general cognitive

ability measure dominates most personnel test batteries, the MPP obtainable

from the use of differential test composites in an independent sample can only

trivially exceed the MPP obtainable from a single efficient measure of g. The

validity generalization movement has been highly instrumental in the general

rejection of the earlier theory of situational specificity—a theory which held that

a tailored test composite is required for predicting performance on each job

situation. As noted in the introductory section, we propose a third theory, differ-

ential assignment theory (DAT) which accepts many of the tenets of the validity

generalization movement without adopting their pessimism regarding the util-

ity obtainable from personnel classification, and emphasizes the importance of

measuring classification efhciency by means of either PCE or CE (Johnson and

Zeidner, 1991).

Validity generalization proponents often accept some classification assump-

tions (Table 11) as facts. Belief in these tenets leads to discounting the utility

obtainable from personnel classification. We have conducted two model sam-

pling experiments and have four more in progress which clearly prove these

assumptions to be erroneous. For example, more PCE is obtained from a battery

of 10 best tests than from a battery containing 5 best tests. More potential

classification efficiency can be obtained when test selection indices that measure

classification efficiency are used-as compared with the use of indices that reflect

predictive validity. Also, optimal assignments to 18 jobs provide higher PCE in

independent samples than when assignment is made to 9 jobs. Differential

assignment theory is based on three major assumptions. The most important of

these is the optimistic premise that the joint predictor-criterion space, as com-

monly encountered, provides a non-trivial degree of multidimensionality and

provides a classification efficient structure, 1.e., rotatable into simple structure.

We have evidence that this can be present even when the first principal compo-

nent provides 80% of the total factor contributions. The second assumption is

126 ZEIDNER AND JOHNSON

that a utility model making use of MPP as the starting point is appropriate for

the measurement of classification efiiciency. The third assumption is that the

state-of-the-art in computer technology has made it practical to make common

use of computationally complex algorithms.

DAT has been used to generate more than 20 principles. The first of these

selected principles relates to the contention of some investigators that classifica-

tion efficiency is automatically maximized under circumstances that maximize

selection eficiency. DAT clearly indicates otherwise and we have accumulated

considerable simulation results supporting this principle. We also have convinc-

ing simulation results that support these principles: (a) more predictors can

contribute to maximization of CE as compared to SE; (b) predictive validity

does not measure CE; and (c) the equation provided by Brogden (1959), de-

scribed earlier, is a robust and useful model.

Differential assignment theory requires as a research paradigm the simulation

of the classification process in order to compute either CE or PCE. Such a

simulation can be based on empirical scores from a data bank or may use

synthetic scores generated by a model sampling process. The latter involves

generating random normal deviate scores and the transformation of these scores

to scores with expected means and covariances equal to those of a designated

population. Model sampling provides synthetic scores with the same statistical

properties as would be provided by a random sample of empirical scores drawn

from a multi-variate normally distributed population.

Model sampling experiments can incorporate safeguards against the presence

of sampling and other biases - safeguards that are difficult to provide in simula-

tion experiments based on empirical data. A typical model sampling paradigm is

provided in Table 12. It commences, in the upper left hand corner, with a

designated population. The parameters of this population are used to generate

synthetic scores, thus providing an analysis sample and 20 or more independent

cross-validation samples.

Summary

The potential gains for various approaches to improve classification efficiency

have been estimated. A study by Sorenson (1965) and the results of on-going

model sampling experiments at The George Washington University suggest that

the substitute of FLS composites for the existing Army aptitude area composites

would provide a 100% gain in an optimal assignment process. Based partly on

DAT psychometric predictions, and partly on model sampling results, we be-

lieve increasing the number of Army job families from 9 to 20 would provide a

CLASSIFICATION AND EFFICIENCY : 127

Table 12.—Typical Model Sampling Research

Population

Sample designated

as population

Computation of

weights to be used

for evaluation

variables”

Generation of entity

samples using

parameters of

designated population

Analysis Sample!

Cross- Validation

Samples®

Computation of

weights to be used

for assignment

variables

MPP RESULTS‘

'Job validation sample sizes equal to those used in Project A first-term concurrent validation study.

*Evaluation weights computed from Project A empirical sample designated as the population. *Sample size of

assigned entities number from 200-300; in the aggregate, N numbers in the thousands for each strategy.

“Predicted performance is computed using the same evaluation variable and same weights for each job across

all experimental conditions.

further gain of 100%. If the differential validity of current predictors was in-

creased, the expected gain in MPP could approximate 100%. The use of appro-

priate multiple-criterion components common to all job families to varying

degrees of importance should provide for the detection of further gains that are

not estimated here. The total gains from the implementation of all these im-

provements should easily exceed several hundred millions of dollars annually

and/or greatly increase the quality of personnel performance in the military

Services.

References

Brogden, H. E. (1949). When testing pays off. Personnel Psychology, 2:171-183.

Brogden, H. E. (1959). Efficiency of classification as a function of number of jobs, percent rejected, and the

128 ZEIDNER AND JOHNSON

validity and intercorrelation of job performance estimates. Educational and Psychological Measurement,

19:181-190.

Eaton, N. K. (1987, March). The Army’s Project A, improving selection, classification, and utilization of Army

enlisted personnel. Briefing for HQDA Personnel Proponent General Officer Steering Committee Confer-

ence, Alexandria, VA.

Fine, S. A. (1955). A structure of worker functions. Personnel and Guidance Journal, 34:66-73.

Ghiselli, E. E. (1959). The generalization of validity. Personnel Psychology, 12:397-402.

Guion, R. M. (1976). Recruiting, selection and job placement. In M. D. Dunnette (Ed.), Handbook of indus-

trial-organizational psychology. Chicago: Rand McNally.

Hunter, J. E. (1983). Validity generalization for 12,000 jobs: An application of job classification and validity

generalization analysis to the General Aptitude Test Battery (GATB) (USES Test Res. Rep. No. 45). Wash-

ington, DC: U. S. Employment Service, U. S. Department of Labor.

Hunter, J. E., Crosson, J. J., & Friedman, D. H. (1985). The validity of the Armed Services Vocational

Aptitude Battery for civilian and military job performance. Rockville, MD: Research Applications.

Hunter, J. E., & Hunter, R. F. (1984). Validity and utility of alternative predictors of job performance.

Psychological Bulletin, 96:72-98.

Hunter, J. E., & Schmidt, F. L. (1983). Quantifying the effects of psychological interventions on employee job

performance and work-force productivity. American Psychologist, 38:473-478.

Johnson, C. D., & Zeidner, J. (1991). The economic benefits of predicting job performance: Vol. 2. Classifica-

tion efficiency. New York: Praeger.

Maier, M. H., & Grafton, F. C. (1981, May). Aptitude composites for ASVAB 8, 9 and 10 (Res. Rep. 1308).

Alexandnia, VA: U. S. Army Research Institute.

McHenry, J. J. (1987, April). Project A validity results: The relationship between predictor and criterion

domains. Paper presented at the annual conference of the Society for Industrial and Organizational Psychol-

ogy, Atlanta, GA.

McLaughlin, D. H., Rossmeissl, P. G., Wise, L. L., Brandt, D. A., & Wang, M. M. (1984, October). Valida-

tion of current and alternative ASVAB composites, based on training and SOT information of FY81 and

FY82 enlisted accessions (Tech. Rep. 651). Alexandria, VA: U. S. Army Research Institute.

Mosier, C. I. (1951). Problems and designs of cross validation. Educational and Psychological Measurement,

11:5-11.

Murphy, K. R., & Davidshofer, C. O. (1988). Psychological testing principles and applications. Englewood

Cliffs, NJ: Prentice-Hall.

Nord, R., & Schmitz, E. (1991). Estimating performance and utility effects of alternative selection and classifi-

cation policies. In J. Zeidner & C. D. Johnson, The economic benefits of predicting job performance: Vol. 3

The gains of alternative policies (pp. 73-154). New York: Praeger.

Personnel Research Section, Classification and Replacement Branch, Adjutant General’s Office. (1945). The

Army General Classification Test. Psychological Bulletin, 42:760-768.

Schmidt, F. L., & Hunter, J. E. (1977). Development ofa general solution to the problem of validity generaliza-

tion. Fournal of Applied Psychology, 62:529-540.

Sorenson, R. C. (1965, November). Optimal allocation of enlisted men - full regression equations vs. aptitude

area scores (Tech. Res. Note 163). Washington, DC: U. S. Army Personnel Research Office. (AD 625 224)

Wise, L. L., Campbell, J. P., McHenry, J. J., & Hanser, L. R. (1986, August). A latent structure model of job

performance factors. Paper presented at the annual meeting of the American Psychological Association,

Washington, DC.

Wise, L. L., McHenry, J. J., & Campbell, J. P. (1990). Identifying optimal predictor composite and testing for

generalizability across jobs and performance factors. Personnel Psychology, 43:355-366.

Zeidner, J. (1987, April). The validity of selection and classification procedures for predicting job performance

(IDA Paper P-1977). Alexandria, VA: Institute for Defense Analyses.

ee sa

$$ +

_ eee ee

DELEGATES TO THE WASHINGTON ACADEMY OF SCIENCES,

REPRESENTING THE LOCAL AFFILIATED SOCIETIES

Biilesapiaicdl Society Of Washington: - 22... 4.8. ....e-s080sees cee s deena .. Thomas R. Lettieri

Pee polorical SOciety Of-WashingpoOn |... 22.6606 ie eee cee oa cole e es epee wees Belford Lawson III

Pearieal society Ol Washingtone: 4... 2.42.22. .2.4 50 a0l ices anced hats Soses Kristian Fauchald

Chemical Society of Washington ............... Boe et ee chee BUG ge mate Elise A. B. Brown

Emomoelocical Society of Washington ..... 22... 5.0.06... s0 ene ee ween _F. Christian Thompson

MII COCTAPNIC SOCICLY 2.55748 os eo) eco oe ws ein eve bocce ce Stanley G. Leftwich

mere SOCICLY OL WaShINGtOny oes sed Sle sede ol eae eo James V. O’Connor

Miedicalsociety of the District of Columbia .......5.0 52.00... 0600.00 ec seek ee ene John P. Utz

Pranic society of Washineton, DC). .c5. 20. lo oc cee cae cals Thomas G. Manning

PEE aOMCICLY, OF WVASMINSLON. =o jas sce. oo. os secoc' es cae ve gine s Tec eee o's bale Muriel Poston

Secicryor American Foresters, Washington Section ..2...22.....0.0......05805. Eldon W. Ross

eee OA CIELYVIOL ENOMEETS 22202523. .552205 os soos dec Sa esses de cealvccsses Alvin Reiner

Institute of Electrical and Electronics Engineers, Washington Section ........ George Abraham

American Society of Mechanical Engineers, Washington Section ......... Clayton W. Robson

Pelmmimnpuoorical society of Washington... . 2)... 552. cele cc ke ce cece eens Kendall G. Powers

American Society for Microbiology, Washington Branch .................. Herman Schneider

Society of American Military Engineers, Washington Post .................... James Donahue

American Society of Civil Engineers, National Capital Section .............. John N. Hummel

Society for Experimental Biology and Medicine, DC Section .............. Cyrus R. Creveling

Rat imomational, Washineton Chapter... oo. 0.06. ccc ccc wise et ceensees Pamela S. Patrick

American Association of Dental Research, Washington Section ............ . J. Terrell Hoffeld

American Institute of Aeronautics and Astronautics, National Capital

SSG LED 2 bol, 5 is te RS RRS ee mi gee PM reels ti RE rte a ree ene Reginald C. Smith

madenican Meteorological Society, DC Chapter ..... 00... ie secs oe A. James Wagner

bosieseemee society Of Washington... .... 0.0.5. 2228. idee else cas Seeeacees st To be determined

Acoustical Society of America, Washington Chapter .....................0.. Richard K. Cook

Punetican Nucicar Society, Washington Section.............2.......0.5 seca seceaees Kamal Araj

Institute of Food Technologists, Washington Section ...................-. George W. Irving, Jr.

American Ceramic Society, Baltimore-Washington Section .................. Curtis A. Martin

2p ED TPD TETRA D 7] NO SLE pete ee eit a a a a Paul Natishan

Mavmdatoaastonyvot science:Club:. 20. 5.025 000505 bode eek e b in das Albert G. Gluckman

American Association of Physics Teachers, Chesapeake Section ............. Robert A. Morse

Optical Society of America, National Capital Section ...................... William R. Graver

American Society of Plant Physiologists, Washington Area Section ............. Steven J. Britz

Washington Operations Research/Management Science Council .............. John G. Honig

Imstrumient Society Of America, Washington Section ............0..5...02.20- Donald M. Paul

American Institute of Mining, Metallurgical and Petroleum Engineers,

Ber Poni E ES CETOI 8 een ees Oe mesa eter orden a aed David M. Sutphin

PHMOnA@apItalASITOMOMETS: ©!) fo kG dee at oe cee oan cow enee wees s Robert H. McCracken

Mathematics Association of America, MD-DC-VA Section ..................... Alice Schafer

Seusicict of Columbia Institute: of-Chemists .2.5 00.0000 .05. anced sce ste William E. Hanford

Mstract of Columbia Psycholopical Association . 5.2.2 06. 020. woe ec ee hee Sue Bogner

mi asmmeroneaint Nechnolory Group... <..424. 0 sews ain toe Jas Oe eee ok: Lloyd M. Smith

American Phytopathological Society, Potomac Division .................... Kenneth L. Deahl

Society for General Systems Research, Metropolitan Washington

CREY OY PST Ry teeless SS 8 pa Et = goo ne Tate AARETE ay ony RE Ace a John H. Proctor

Fiuman Hactors Sseciety, Potomac Chapter 3522. 24068 ee oe cee hese Thomas B. Malone

mmerican’ Fisheries Society, Potomac Chapter .. 3.06.0. 2 5. oa ee sk David A. Van Vorhees

Association for Science, Technology and Innovation ................:..... =. Ralph. Cole

astern Sociological SOcieby —..e0 sco ow sa ck eicipee ew Se i Seeman Ronald W. Manderscheid

Institute of Electrical and Electronics Engineers, Northern Virginia

| SCHON G ys te eerie ae are ee ete a cee UE LG Oc Blanchard D. Smith

Association for Computing Machinery, Washington Chapter ............. Charles E. Youman

Sy ASiaNSLON, StalisticalSOCICLY 2.0. he nee ee Een kes Nancy Flournoy

Society of Manufacturing Engineers, Washington, DC Chapter ....... NOTE Aten James E. Spates

Institute of Industrial Engineers, National Capital Chapter ................... James S. Powell

Delegates continue to represent their societies until new appointments are made.

Washington Academy of Sciences 2nd Class Postage Paid

1101 N. Highland St. at Arlington, Va.

Arlington, Va. 22201 and additional mailing offices.

Return Requested with Form 3579

VOLUME 81

Number 3

od Our nal of the ' September, 1991

WASHINGTON

ACADEMY... SCIENCES

ISSN 0043-0439

Issued Quarterly

at Washington, D.C.

CONTENTS

Articles:

NICHOLAS L. CHANDLER, KENNETH LUCAS, and JOHN J. O'HARE,

mIDIStAnce learning im Witlicany* LrAUMIIN 1... cy, cisicse cic seyeeshol sl sites oleleie eee ere eee os

ALLAN KROOPNICK, “What a Difference a Sign Makes” .................

ARTHUR J. REPAK, “Effect of Cyst Age, Media, pH, Temperature, and

Time, on Encystment of Blepharisma stoltei Isquith” ..................:.:...

PERCY A: WELLS, “Penicillin Production Saga Recalled’? 3220.05.26. 000.....

Washington Academy of Sciences

Founded in 1898

EXECUTIVE COMMITTEE

President

Walter E. Boek

President-Elect

Stanley G. Leftwich

Secretary

Edith L. R. Corliss

Treasurer

Norman Doctor

Past President

Armand B. Weiss

Vice President, Membership Affairs

Cyrus R. Creveling

Vice President, Administrative Affairs

Grover C. Sherlin

Vice President, Junior Academy Affairs

Marylin F. Krupsaw

Vice President, Affiliate Affairs

_ Thomas W. Doeppner

Board of Managers

James W. Harr

Betty Jane Long

John H. Proctor

Thomas N. Pyke

T. Dale Stewart

William B. Taylor

REPRESENTATIVES FROM

AFFILIATED SOCIETIES

Delegates are listed on inside rear cover

of each Journal.

ACADEMY OFFICE

1101 N. Highland Street

Arlington, VA 22201

Phone: (703) 527-4800

EDITORIAL BOARD

Editor:

John J. O’Hare, CAE-Link Corpora-

tion

Associate Editors:

Bruce F. Hill, Mount Vernon College

Milton P. Eisner, Mount Vernon Col-

lege

Albert G. Gluckman, University of

Maryland

Marc Rothenberg, Smithsonian Insti-

tution

Marc M. Sebrechts, Catholic Univer-

sity of America

Edward J. Wegman, George Mason

University

The Journal

This journal, the official organ of the Washing-

ton Academy of Sciences, publishes original

scientific research, critical reviews, historical

articles, proceedings of scholarly meetings of

its affiliated societies, reports of the Academy,

and other items of interest to Academy

members. The Journal appears four times a

year (March, June, September, and De-

cember). The December issue contains a di-

rectory of the current membership of the

Academy.

Subscription Rates

Members, fellows, and life members in good

standing receive the Journal without charge.

Subscriptions are available on a calendar year

basis, payable in advance. Payment must be

made in U.S. currency at the following rates:

U.S: and Canada: .:2.5.520 00 eee $25.00

Other countries)... cc eee 30.00

Single copies, when available ....... 10.00

Claims for Missing Issues

Claims will not be allowed if received more

than 60 days after the day of mailing plus time

normally required for postal delivery and —

claim. No claims will be allowed because of

failure to notify the Academy of a change of

address.

Notification of Change of Address

Address changes should be sent promptly to

the Academy office. Such notifications should

show both old and new addresses and zip-code

numbers, where applicable.

Published quarterly in March, June, September, and December of each year by the

Washington Academy of Sciences, 1101 N. Highland Street, Arlington, VA 22201.

Second-class postage paid at Arlington, VA, and additional mailing offices.

Journal of the Washington Academy of Sciences,

Volume 81, Number 3, Pages 129-147, September 1991

Distance Learning in Military Training

Nicholas L. Chandler, Kenneth Lucas, and John J. O’Hare

CAE-Link Corporation, Alexandria, VA

ABSTRACT

Instruction by telecommunications-based distance learning systems connected to remote

U.S. military sites has demonstrated that students attain test scores equivalent to those

obtained with traditional classroom instruction. Student attitudes toward remote methods

of instruction indicate that critical to greater acceptance of those techniques is the enhance-

ment of instructor-student interactions during and after training-sessions. Installation and

operational costs for fully interactive, audio-visual instructional systems are high but declin-

ing. Costs for less robust systems are lower, and it is estimated that some approaches can be

amortized over several months.

Introduction

The duration of a technological generation has been estimated to be about five

years (Gopher & Kimchi, 1989), and that brief span demands that more fre-

quent training be provided if the technical currency of the working population is

to be maintained. Within some organizations, particularly the military, the

maintenance of technical skill-levels in an era of declining force-size has addi-

tional value as a force multiplier. Costs for school-house training for the update

of technical skills are high, and there is a need to determine the cost-benefits of

alternate modes of training, especially for personnel who are distributed over a

wide geographic area. This is an especially acute problem for the military where

. the number of people to be trained, costs, and resource demands are consider-

able, and there is a requirement to bring training directly to individuals, rather

than the reverse. It has been estimated (Fletcher, 1990) that on an average day in

FY90, about 208,000 active duty personnel and 47,000 National Guard and

Reserve personnel received formal training, at an annual cost of approximately

$18.3 billion, and with a resource requirement of 176,000 military and civilian

personnel to conduct and support those instructional programs.

There has been a progression in the instructional approaches used by U.S.

military training organizations to provide efficient and effective training pro-

129

130 CHANDLER, LUCAS, & O’HARE

grams. Computer-based training is one technique that has met some local needs.

A variation of computer-based training made use of video-disc systems

(Fletcher, 1990) whereby students interacted with knowledge bases and were

aided by computer-based systems which tailored the pace and level of instruc-

tion to individual ability and preference. But to serve the training needs of

widely dispersed students who require long-term programs, the military has

evaluated several telecommunications-based distance learning systems. Future

training systems are expected to use artificial intelligence techniques to individ-

ualize instruction with interactive remediation and retesting and to assess the

learning style of each student (Griffin & Hodgins, 1991).

Distance learning technologies. Distance learning has been described

(Wells, 1990) as a teaching approach that has these components: (a) instructor,

students, and an implicit contract between them; (b) physical separation be-

tween instructor and students, as well as between students and the instructional

institution; (c) two-way communication between instructor and students; and

(d) use of course materials specifically designed for distance study. Subsumed

under that term would be the various methods that are made available for

communication between instructor and student, such as teletraining and tele-

conferencing.

Teleconferencing is a generic, industrial term that incorporates four major

forms of telecommunications technology used for transmissions between indi-

viduals and groups: video, audio, audiographic, and computer. The capabilities

of those four media-forms offer different possibilities for the implementation of

instructional strategies: .

Video. Television systems are used to deliver training materials and they can

be linked to remote sites by satellite, land line, or microwave communication

systems. Video teleconferencing may be presented in an interactive, digital form

which will allow two-way video and audio messages to be exchanged between

instructor and students at each site on the training network. An analog form

(usually called business TV) provides one-way video imagery combined with

two-way audio interaction between instructor and the training participants;

however, the interaction is essentially audio (Chandler & Harris, 1991). Where

the element of student-instructor and student-student interaction is deemed

essential, interactive video systems have been adopted (Bailey, Sheppe, Hodak,

Kruger, & Smith, 1989). On the other hand, analog video has been demon-

strated to be adequate within industrial firms, e.g., IBM, Sears, and General

Electric, for company-wide meetings directed at information transfer. More

than 50 networks and 11,000 reception sites have been developed by industry

(Johansen, 1988). In most cases, these systems are used for both business-infor-

mation purposes and for corporate distance learning.

DISTANCE LEARNING 131

Audio. Conferencing with audio systems is achieved with an interactive con-

ference call over low-cost commercial telephone lines that connect the instruc-

tor with the students. Instructional costs are lower with these systems and they

have proven to be adequate where there is no need to transfer video imagery and

graphic data.

Audiographic. In these systems the instructor employs a telephone confer-

ence call to send timing signals to student computers that key stored, digitized

images which become accessible to the students for instructional purposes.

Training costs are reduced when the type of instructional program makes it

feasible to forward the requisite video materials prior to the instructional session

rather than transmitting them during the session.

Computer. Student workstations are linked to the instructor’s computer and

the students are able to access the instructor’s files, download that information

to their own computers, and exchange information with the instructor and the

other students about the instructional program. Those exchanges may occur in

real time between the instructor and fellow students, but in most cases, the

communications are asynchronous, 1.e., at times preferred by the individual

student (Johansen, 1988). Studies within an experimental meeting room (Stefik,

Foster, Bobrow, Kahn, Lanning, & Suchman, 1987) have established the feasi-

bility of collaborative work by 2-6 persons via asynchronous communication

and discussion. The successful operation of another computer-mediated com-

munication system (Malone, Grant, Tarbak, Brobst, & Cohen, 1987) demon-

strated that an intelligent information-sharing system can adequately support

problem-solving activities by groups.

Until recently, the term teleconferencing was used to describe any of these

four telecommunications media without regard to the purpose of the confer-

ence. Now, the term fe/etraining is being adopted to refer to teleconferences

specifically directed toward distance learning. Therefore, teletraining may be

defined as the use of teleconferencing technology to convey educational and

’ training information to participants who are geographically dispersed.

Training Effectiveness. Distance learning for adults has been fostered by a

number of educational institutions with diverse programs (Guthrie-Morse &

Julian, 1989; Pugh, Parchman, & Simpson, 1991; Saffo, 1990) which used tele-

training for instructional purposes. Training programs that employ teletraining

are also widely used by business firms for professional training. Surveys con-

ducted on the policies and experiences with those approaches at a variety of

institutions uniformly conclude that its educational/training outcomes are

equivalent to traditional residential-training, and find that students have only a

slight preference for conventional training methods (Pugh, Parchman, & Simp-

son, 1991). Despite large investments in these techniques, there has been a

132 CHANDLER, LUCAS, & O’HARE

shortfall on formal, controlled studies to determine their compatibility with

student expectations and instructor capabilities, functionality in courses of dif-

ferent types, and the match with the cognitive-processing abilities of students

(Bailey et al., 1989). Experimental evidence is insufficient for analysts to tease

out from case studies on distance learning that which is most relevant to the

assessment of the principal questions that have been posed on its effectiveness

and utility.

In their recent review of the research literature on human performance,

Gopher and Kimchi (1989) identified the display of, and interaction with, com-

plex visual information as major issues. Among the critical questions in commu-

nication systems were the methods to represent properly the world one wishes to

depict, and the physical attributes of that representation. These questions have

led research investigators to theoretical studies of how perceptual events are

encoded, and the answers are often sought within the framework of feature-ana-

lytic or bottom-up processes (Treisman, 1986), global or top-down processes

(Navon, 1977), and the in-between boundary processes (Kinchla & Wolfle,

1979). These notions as well as theories of data-driven and concept-driven pro-

cessing (Rumelhart, 1977) could provide a rich basis for guidance in the design

of training programs that depend on the presentation of complex information.

Such guidance is essential for the selection of instructional methods like video-

based teletraining and teleconferencing which put strong emphasis on the visual

medium (reducing the textual), mute the affective content, and enlarge the

cognitive components of instruction. However, those major factors in training

effectiveness have not been considered in the design and evaluation of field

studies with military distance-learning systems.

Cost factors. Instructional designers have envisioned the development of

cost-effective systems wherein students would be trained at locations of conve-

nience yet receive instruction that was comparable in effectiveness to that pro-

vided in conventional classroom-settings (U.S. Congress, 1989). Costs for most

training systems are expressed in terms of student throughput, 1.e., the cost-per-

student. Early on, television emerged as a primary medium for instruction and

educators exploited the power of that technology for instructional purposes,

both directly and with videotapes, for broadcast to distributed locations. To

offset the high costs of television studios, satellite bandwidth, and a professional

production staff, large numbers of students had to be involved in each television

broadcast. As the number of students increased, the opportunities for student-

to-teacher and student-to-student interaction decreased. This lack of interaction

in those using television techniques was found to be incompatible with a stu-

dent’s need to communicate with the instructor and their fellow students (Pugh,

DISTANCE LEARNING | 133

Parchman, & Simpson, 1991, p. 29). Therefore, while these systems were cost

efficient, they were often found not to be training effective.

Operating costs for teletraining systems are fairly easy to quantify. The cost

factors include: facilities construction/modification needed to accommodate

the technology employed; hardware/software; transmission or telecommunica-

tions; maintenance; and personnel for production and support staff. Of the

teletraining technologies discussed here, only analog video teleconferencing

(Business TV) requires special facilities, although digital video teleconferencing

may require some site preparation to accommodate the telecommunications

equipment. This technology also requires a dedicated staff to operate equipment

at the originating site. The other technologies all require some level of training to

operate, but all, including the digital video teletraining, can be operated by an

instructor or student as a collateral duty.

The telecommunications costs associated with both types of video teletraining

systems remain the most complex issue involved in determining the cost of a

teletraining system. There are different transmission media, such as satellite,

microwave, and fiber optics; there are private and public networks; and there are

dedicated, as well as pay-as-you-go or ad hoc services. Given these variables, it 1s

difficult to give exact costs. Generally speaking, the greater amount of band-

width used, the higher the cost. Analog television uses the equivalent of a 90

megabits per second digital signal. When that signal is compressed many times

its original size, as it is with digital video, to 512 kilobits per second, the cost 1s

lower. However, while an analog broadcast signal may cost more per hour, that

cost is constant regardless of the number of sites in the network. If the cost 1s

$2,000. per hr., it is the same $2,000, whether it is downlinked to 5 sites or 100

sites. While digital signals may only cost $200. per hr., that cost is repeated for

each site on the network, as each site is both a downlink and an uplink. If there

are 10 sites in a digital network, the cost is $200. x 10, or $2,000. per hr.

The more important components for distance learning are those that control

access, provide storage, and manipulate information, such as micro-computers,

display devices, optical memories, facsimile machines, and scanners for graphic

images. In the future it is anticipated (Simpson, 1990) that multimedia worksta-

tions that access computer data-bases will be joined with high-definition televi-

sion, facsimile, and telephone devices, to create a versatile linkage between

student and instructor that will be rich in information, and essential for some

training purposes.

The introduction of analog video teletraining systems has been slowed by

investment costs. A complete uplink facility with studio and its associated elec-

tronics, has been estimated to cost between $1,000,000-—3,000,000. Downlink

134 CHANDLER, LUCAS, & O’HARE

sites are less expensive, costing on the average of $6,000.—10,000. Satellite trans-

mission is generally distance insensitive, but its cost varies with the bandwidth

that is used: (a) C-band (6 GHz on uplink and 4 GHz for downlinks) costs

between $200.-500./hr.; and (b) Ku-band (14 GHz on the uplink and 12 GHz

on downlinks) runs between $200.-600./hr. Under development are Ka-band

systems (30 GHz uplink and 20 GHz downlinks) that require very small dishes

(1 m) for which costs have not yet been determined. A satellite downlink with

receiver, transmitter, and dish, costs between $800.-18,000., depending on

bandwidth that is used, and the informational features that are required (voice,

data, and/or video). When a steerable rather than a fixed dish is employed, the

cost is multiplied by about three. Finally, operational and programming costs

demand a substantial, long-term financial commitment.

Instructor factor. The attitude of the instructor toward these systems has to

be weighed since the teacher is a key element in the success of distance training.

Romiszowski (1981) has reflected on the issues that bedevil designers when new

techniques are to be introduced for the improvement of instruction, and his

analyses have focussed on the level of the instruction, the instructor, and the

adequacy of the methods being replaced. He has argued that there is a need for

different instructional strategies for the exposition of facts, discovery of princi-

ples, development of algorithmic procedures, and generalization to concepts. In

his view, at the level of skill development, instructional media should provide an

opportunity for a student to practice the behavior to be learned, and to use the

appropriate sensory channels for communicating the requisite information.

Among the instructor factors, Romiszowski cautions that efforts have to be

made to prepare the teacher to accept and use the new medium, to determine the

essentials that are to be taught, to reduce demands on the instructor for signifi-

cant changes in teaching practice, and to allay any fears or negative attitudes that

new gadgetry might engender. It is anticipated (U.S. Congress, 1989) that

teacher preparation for distance instruction (teletechniques) will require in-

creased attention to: (a) the creation of rapport with students; (b) the encourage-

ment of participation; (c) refinement of presentation style, such as volume and

tone of voice; and (d) provisions for regular feedback to the student. Most

teachers have mastered these skills in face-to-face instruction to a level of auto-

maticity but for distance learning, they would have to be exercised more

consciously and aggressively.

Effectiveness factors. The Office of Technology Assessment (U.S. Congress,

1989) has thoroughly examined the effectiveness of video distance-learning for

elementary and secondary-school students, and outlined the technological, eco-

nomic, institutional, and policy barriers to its further development. The general

conclusion was that such methods work with motivated older students and

DISTANCE LEARNING 135

professionals, however doubts were expressed about their suitability for young

and/or academically weak students. All of these considerations need to be as-

sessed carefully before recommendations for adoption of distance-learning ap-

proaches are made. Caution has to be exercised in the interpretation of data on

the relative effectiveness of distance-learning approaches because there are seri-

ous questions on their equivalence to traditional instruction (Wells, 1990). The

materials that are developed for remote learning are selected because they are

most appropriate for the learning task, the instructor teaches differently, and the

array of technology—particularly computer-mediated—exceeds that employed

in the conventional teaching situation. The analysis reported here will focus on

the experience of military trainers with three forms of distance learning: digital

and analog video teletraining, and asynchronous computer teletraining.

Effectiveness and Acceptance of Distance Learning in the Military Environment

Four military distance-training systems, two in the U.S. Navy and another

two supported by the U.S. Army, have been identified that serve adult learners

and have been the object of evaluative studies. They will be examined for mea-

sures of performance effectiveness relative to conventional training and for

assessments of student acceptance of that mode of instruction. Military training

requirements represent a highly promising area for the employment of distance-

learning techniques for a number of reasons. The number of students and vari-

ety of courses served by the military training establishment is very large, so that

instructional costs are significant. Attendance at training programs involves

costs for travel, per diem, learning resources, and time away from regular duties,

for both students and instructors. Another important factor that influences

training opportunities is the narrow temporal window that is available for train-

ing, especially in the case of the Reserve forces.

Navy Fleet Combat Training Center-Atlantic. The digital video teletraining

center at NFCTC-Atlantic has been in operation since early 1989 at Dam Neck,

VA, and currently links six different sites, each capable of originating and re-

ceiving audio-visual signals, along the East Coast, by Ku-band satellite stations.

There are two remote sites on Naval installations within the Dam Neck area,

and the others are at: Newport, RI; Norfolk, VA; Charleston, SC; and Mayport,

FL. The network exchanges compressed, digital, encrypted, two-way video and

audio information (Chandler & Lucas, 1989; Pugh, Parchman, & Simpson,

1991), at a transmission rate of 512 kilobits per second. KG-81 encryption

devices allow training to be conducted up to the security level of secret.

Classrooms are equipped with fixed cameras that are focussed on the instructor,

136 CHANDLER, LUCAS, & O’7HARE

ELECTRONIC SCHOOLHOUSE NETWORK

GTE- Ren,

ey 325 KG 81

Norfolk, VA

a 2 re 1

KG 8i a Bre

Mayport, FL

Dam Neck, VA

Fleet Combat Training Center, Atlantic

Mu ipoint interactive

Video Teleconilerence Network

Fig. 1. Teletraining network originating at Naval Fleet Combat Training Center-Atlantic (Chandler &

Lucas, 1989).

the students, and downward for graphic presentations; each room also has a

small number of sound-activated microphones and the instructor wears a clip-

on microphone. Whenever a student at a remote site queries the instructor, all

sites on the network are switched by that audio signal to a video of that

classroom. The majority of courses delivered with this system are lecture-based

and completed within one week.

Formal experiments were conducted with five courses that originated at Dam

Neck and transmitted to five of the remote sites (Fig. 1). The courses were

concerned with diverse topics such as planning, supervision, gun battery align-

ment, and ammunition administration. An equal number of students (178)

participated in the courses at Dam Neck and at the remote sites (Griffin &

Hodgins, 1991). The courses were transmitted over a three-week period. Data

collected on student evaluations at the originating and remote sites allowed for a

DISTANCE LEARNING | 137

comparison of the effectiveness and acceptability of this teletraining system with

the standard class-room setting to which these students were accustomed (Ru-

pinski & Stoloff, 1990). Results of those assessments showed no significant

performance differences in coursework between the students at the originating

site and the various remote sites; teletraining was equally effective as residential

training in meeting the objectives of all of the courses. No significant differences

were reported by the students on: (a) general attitude toward the instructors;

however, expressions of disappointment regarding their non-availability to the

remote sites after the class sessions did appear; (b) quality of the audio-visual

systems was regarded as similar at all locations; (c) tests/homework were not

rated as any more burdensome or difficult; and (d) course contents were rated at

similar levels. Sizeable differences, however, were noted in the responses of the

students at the remote sites with regard to: (a) the degree of interaction with the

instructor; it was cited as insufficient by 20% of the students; (b) the number of

opportunities to talk and ask questions; this factor troubled 11% of the students;

and (c) preference for teletraining vs. residential training methods; 12% of the

students opted for the traditional mode of instruction. When invited to suggest

improvements for the courses, the students focussed on the technical problems

that had occurred in the operation of the equipment, and the decreased level of

interaction with the instructor. Bailey et al. (1989) have compiled a set of guide-

lines, based on experiment, experience, and expert opinion, that provides design

solutions for most of the criticisms that have been made of the teletraining

approach.

Navy Fleet Combat Training Center—Pacific. During a survey (Simpson,

1990) of current communication systems for delivery of training programs,

alternative designs for a distance-learning system were considered and led to the

specification of optional systems for use in Navy instructional environments.

The most serious training problem to be met by such systems was expected to be

the ability to meet the requirements of individuals aboard ships at sea, dockside,

and at the 235 Naval Reserve training sites in the continental U.S. One of the

distance-learning configurations that was proposed was a teletraining system for

group instruction, similar to that in use at the NFCTC-Atlantic. However, for a

test system installed at NFCTC-Pacific in San Diego, a land line with a trans-

mission rate of 1.54 megabits per second was employed instead of a satellite to

link two Naval sites on the West Coast. As with the East Coast network, digital

signals were encoded with a VideoTelcom (VTC) System 300 Codec, and en-

crypted with KG-81 devices, to provide secure two-way audio and video trans-

missions (Simpson, Pugh, & Parchman, 1990). Careful attention was given to

the design of the originating (Fig. 2) and remote (Fig. 3) classroom facilities

which imposed few demands on the instructor. The equipment was operated by

138 CHANDLER, LUCAS, & O’HARE

VTC System 300 Audio Mixer FAX Phone

Yy A a al LLL,

Yy Gare ee ey

Ideo from S.D Video from T.I.

to T.I. to S.D.

Video from ELMO

MIC MIC

(i &

Student Student

Table Table

MIC MIC

& if

Siudent Student

Table Table

MIC MIC

29] @

Student Student

Table Table

MIC MIC

al Cc

Student Student

Table Table

Camera

b” NANA

RM A234

INN AAA AAA OA

ML,

Fig. 2. Equipment configuration of teletraining system originating at Naval Fleet Combat Training Center-

Pacific (Simpson, Pugh, & Parchman, 1990).

VLLLLLLLLLLLLL LLL LLL

DISTANCE LEARNING

he te — ote =

Newbridge 3600

Equipment Rack

Remote Classroom

Bidg. 461 Rm. 242 & 243

RM 242 Treasure Island

RH Hoh hhh ooo ooo oy

SES HHH HH ooo» yg >-

ULLLULLLL LILIA UL/LULLLLL LALIT LALLTUULLLL LALLA LA LLILLALLLLLLLLL LULL TLASILA

MAA AANA

VTC System 300

Monltors

RM 243

VTC ovate 300

Phone

beens hitcere

Student Student

Table Table

S a

MIC MIC

Student Student

Table Table

Student Z Student

Table Table

@ oa

MIC MIC

Student Student

Table Table

ra] fe]

MIC MIC

Camera Video from T.lI. toS.D.

\ ; Video from S.D. to T.I. /

35” Monitor

artition

Partition

MIC

SNSNSAASANASSAANNAAANADN NS

SOAAAAAAAAAANNANNNANAAAAANANAN ST

ULL

LLM LMA LALLA LLL LLL LLL LLL LLL LLL LLL

139

Fig. 3. Equipment configuration of teletraining system at remote site networked to Naval Fleet Combat

Training Center—Pacific (Simpson, Pugh, & Parchman, 1990).

140 CHANDLER, LUCAS, & O7HARE

Support personnel as the instructor presented the course material in the accus-

tomed manner. In this equipment configuration, the instructor could hear the

comments of students who spoke into their desk microphones (MIC), and could

see, on 12” video monitors, the students at both locations. The students at both

locations could see the instructor and the hardcopy graphic-materials picked up

by the Elmo easel camera on a 35” video monitor; they could hear, but not see,

the students at the other location.

Three lecture courses, rich in terms of the training processes required and with

information on the shape and movement of solid objects were experimentally

evaluated (Simpson, Pugh, & Parchman, 1990) over a 2-month period with

active duty and reservist personnel undergoing required training. A facilitator at

the remote site administered and scored the tests that were required, and trans-

mitted the results to the instructor. The courses originated in San Diego with the

remote site at San Francisco (Treasure Island). Twenty-one students partici-

pated at the originating site while 27 students enrolled at the remote location.

Three regular instructors were recruited for this field study and were assigned to

one or more of the courses to be taught. The instructors were then brought into a

video-training studio to practice teaching in that environment but no measure-

ments were made at that time of instructor attitudes or their skill in that teaching

situation. Test scores were compared on performance outcomes for two of the

courses, where the maximum possible score was 100 and the passing score was

70, but they revealed no significant difference between the two groups. Perfor-

mance on one course at the originating site (MW = 86.5; SD = 2.9) varied slightly

from the scores obtained at the remote site (WV = 86.8; SD = 6.2), while the

performance-levels on the other course at the originating course (M = 86.0; SD

= 4.4) did not differ greatly from the scores obtained at the remote site (M@ =

82.6; SD = 4.9). However, some informative outcomes appeared in student-at-

titude ratings of the courses, their interactions with the instructors, and their

reports of technical problems with the training system. Student attitude mea-

surements showed: (a) no differences in the appraisal of the instructor; however,

students at the remote site rated access to the instructor outside of class hours, as

more important; (b) slight difference in the assessment of the adequacy of the

physical system; an exception was that students at the remote site rated, as less

satisfactory, the loudness and quality of the audio transmissions; (c) assessments

of tests and homework assignments were similar at the two locations; (d) no

significant difference was seen in the quality of course contents; and (e) the

frequency and opportunity for instructor-student interactions during the 1n-

structional periods were reported as similar. :

When asked whether they would prefer teletraining to the conventional

method of instruction, about a third of the students at the remote site were

DISTANCE LEARNING 141

Table 1.—Student Preference (%) for Teletraining vs. Conventional Training Methods (Simpson, Pugh, &

Parchman, 1990)

Originating Remote

Video teletraining where instructor is on TV 11 15

Traditional methods of instruction where instructor is

physically present in the classroom a7 31

Indifferent between video teletraining and traditional methods

of instruction 52 54

firmly committed to the conventional system (Table 1). This percentage was at

about the same degree of preference reported by students at the originating site.

The majority of students at the remote sites either preferred teletraining (15%) or

were indifferent (54%) to the instructional system that was used. When the

responses of the students at the remote site are contrasted with individuals at the

originating site, negative attitudes toward teletraining were attributed to: (a)

fewer opportunities to talk with the instructor (11% vs. 6%); and (b) the sufh-

ciency of the number of opportunities to obtain remedial instruction outside the

regular class hours (13% vs. 21%). Regular instructors, qualified to teach the

courses, had been given informal training and practiced instructional delivery

with the teletraining system and the use of its equipment, in order to enhance

student-instructor interaction. In post-course questionnaires completed by two

of the three instructors, satisfaction was expressed with the quality and usability

of the equipment, effectiveness of the system for the remote sites, and the high

level of class-participation by both student groups.

Army Research Institute. The Army Research Institute has supported re-

search studies on the merits of asynchronous computer teletraining systems as

media for distance learning. Training on computer-based teletraining systems

imposes fewer equipment and telecommunications demands and thus can be

operated at much lower cost for instruction than the two Navy teletraining

‘ programs. Correspondence courses are the usual options for the Reserve forces

who are unable to attend conventional training programs for a variety of rea-

sons. And not to be overlooked, to the advantage of asynchronous training

approaches, is the fact that courses by that method are available on a 24-hour

demand basis via computer terminals. Though they agreed that such an ap-

proach 1s suitable for training programs that support independent study, Simp-

son & Pugh (1990) questioned whether that technique would function well as a

primary instructional delivery-system. Nonetheless, the Army Research Insti-

tute has pursued studies on the feasibility of computer teletraining since 1986

and has accumulated a sizeable body of data on performance, cost, and student

attitude toward computer teletraining as an instructional mode. The lessons

142 CHANDLER, LUCAS, & O’7HARE

learned from this research that can assist training developers to design, conduct,

and evaluate computer teletraining courses have been summarized in a series of

reports (Hahn, Ashworth, Phelps, Wells, Richards, & Daveline, 1990; Hahn,

Harbour, Wells, Schurman, & Daveline, 1989; Harbour, Daveline, Wells,

Schurman, & Hahn, 1990; Wells, 1990).

In that computer teletraining system, the structure and content of the residen-

tial-course materials were adapted for the home or office environment to pro-

vide paper-based readings and problems, computer-based training, and video

tapes. The computer teletraining links permitted discussions by the students

with the instructor and with each other. Each student was provided with a

personal computer loaded with software on course management, communica-

tion, computer-based training and tests, word-processing, and a spreadsheet. In

addition, the students were given a hot-line phone number in case any hard-

ware/software problems arose. Identical equipment and training on teaching

techniques with the computer teletraining system were provided for the instruc-

tional staff.

In a typical research project to assess the computer teletraining system, a

group of 14 Army Reserve students completed a portion of the Army Engineer

Officer Advanced course (Hahn, Ashworth, Phelps, & Byers, 1990). That course

work required 66 hours of instruction and covered Army doctrine, engineering

topics, leadership, and presentation skills. Performance comparisons were made

with all students who took the same courses in residence at the U.S. Army

Engineer School over a period of about 32 weeks. Teletraining permitted the

student to work with prepackaged learning materials, the instructional staff, and

other members of the student group. When time delays would prove to be

impractical for a given task, all students accessed the system at the same time

and worked together. The instructional staff included a full-time course man-

ager and three part-time instructors. The instructors were given a 40-hr. training

course on system operation and teaching techniques appropriate to that system.

Results showed no significant differences in performance scores on the test,

homework, and practical exercises taken by the two groups. Comparison of

skill-changes, before and after taking the course showed a marginal difference in

favor of the students who were in the teletraining program. Survey and interview

data on student acceptance of teletraining were collected after course comple-

tion, and on-line comments were analyzed to assess acceptability of this instruc-

tional approach. |

Practically all students who completed the courses preferred teletraining to a

correspondence course (Hahn, Ashworth, Phelps, Well, et al., 1990). Very large

differences on two measures were reported when residential training was con-

trasted with independent study: (a) time to complete the course was shorter (2

DISTANCE LEARNING ; 143

Cost ($K)

1 2 | 4 5 6 7 8 9 10

Iteration

= AGG In-House —— ACC Contracted —# Resident

Fig. 4. Cost estimates ($K) for three types of training modes (residential, in-house and contractor-furnished

computer teletraining) as a function of the number (N) of course iterations (Hahn, Ashworth, Phelps, & Byers,

1990).

weeks vs. 31 weeks); and (b) percentage of students not completing the course

was smaller (5% vs. 36%). Independent study emerges as a highly unfavorable

instructional mode. However, those results are similar to other forms of non-res-

idential training, such as correspondence courses. Dropout rates of 65% are not

uncommon in Army correspondence courses (Wells, 1990). Cost data for 10

course-iterations were estimated for converting a course to teletraining delivery,

based on the following formula, where cost includes equipment purchase, in-

structor training, and recurring expenses:

_ (class size X % throughput)

op cost X 100%

Figure 4 shows that costs for this teletraining system could be amortized over

four iterations when training was conducted by outside contractors. In-house

developed teletraining had costs similar to resident-training initially, and then

decreased immediately during subsequent iterations; the estimated dollar sav-

ings in 1990 for in-house training rose by the tenth iteration to about 48%. This

Savings 1s substantial when it is appreciated that the Reserve component makes

up about one-half of total Army strength.

The Naval Supply System Command has supported the design of a prototype

144 CHANDLER, LUCAS, & O’7HARE

audiographics teletraining system whose costs were estimated at only $10,000

per site for hardware and software (Young, 1991). In that distance-learning

system, a telephone conference call was combined with video imagery loaded

into each student’s computer. It was reported that satisfactory comparisons were

realized in favor of teletraining vs. residential training by students from the

Naval Supply environment.

Army Satellite Education Network. When networks are employed for train-

ing, they may be characterized by the formation of a highly-skilled staff at a

central location, which is another method for cost-control in the development of

training programs. The U.S. Army Logistics Management College operates such

an educational network, the ASEN, that originates at Ft. Lee, VA, uplinks to a

satellite communication system with 58 downlinks throughout the continental

U.S. to provide instruction on acquisition management, logistics support, and

logistics operations (Brockwell, 1989). Unencrypted one-way video and two-

way audio are transmitted to each instructional site which is equipped with a

46” rear-projection TV monitor. Diskettes with digitized images are mailed to

the remote sites but their presentation can be controlled by the instructor at the

central location over narrow bandwidth phone-lines. This system provides an

analog video teletraining system with operating costs that are estimated (Pugh,

Parchman, & Simpson, 1991) to be almost 30% less than residential training. It

is expected that the analog format for video transmissions will be replaced by a

digital system; this latter technology requires a smaller bandwidth and thus,

costs will be lower. The Army has circulated a draft standard for video telecon-

ferencing systems that will assure the interoperability of such systems for the

secure transmission of video, audio, graphics, and data (MIL-STD-188-131,

1989).

A typical class session on ASEN is transmitted to 7-8 remote sites with a

maximum of 30 students at each of those locations; and the duration of a course

ranges from | day to 3 weeks. Presentations can be taped at the remote site and

re-run at dates and times preferred by students. Class-performance data showed

that examination scores did not differ significantly from equivalent resident

courses; but student ratings indicated that there was a slight preference for

residential training over analog video teletraining instruction.

Conclusions

Experience on four distance-learning systems supports the feasibility of tele-

training systems in terms of performance and acceptance by adult learners.

Field surveys indicate that those systems would be more effective if several

DISTANCE LEARNING 145

features were enhanced: (a) the downtime due to communication conditions

was reduced; (b) greater attention was paid to the maintenance of high-quality

audio and video; (c) additional effort was focussed on improvement of student-

instructor interactivity; and (d) arrangements for instructor accessibility after

class sessions were mandated. The first two issues are technical problems that

are tractable but the latter two features require increased attention to instructor

preparation or support. Some instructional designers in the Navy (Grifhin &

Hodgins, 1991) suggest that interactive courseware and distance-learning sys-

tems which utilize available artificial-intelligence (AI) algorithms might be com-

bined to provide the individualized interaction that is desired by students. They

envision the introduction of procedures that react to incorrect student answers,

individualize remediation, and provide training content that is tailored to the

style of the learner. They also imagine a schoolhouse equiped with virtual-image

technology which will allow student and instructor to communicate as if they

were face-to-face, via holographic and 3-D stereoscopic simulations. The intro-

duction of AI and virtual-image technologies will complement distance-learn-

ing techniques. However, AI training technologists envision future training envi-

ronments other than the traditional classroom model found in the military

training systems that have been examined in the present study. An important

research question will be how to meld the best features of the group approach

inherent in the conventional classroom with the individualized, active learning

that is fostered in Al-based tutoring systems.

Rapid improvements in communications technology at reasonable cost-lev-

els are needed to enhance the physical system, such as compressed digital televi-

sion which allows for acceptable dynamic video imagery over narrow communi-

cation bandwidths. The substantial difference in cost for a two-way vs. a

one-way video system is an important factor in the acquisition of these systems

that could be reduced by computer networking. However, the overall costs vary

widely and are determined in large part by these factors: (a) the number of

‘remote sites; (b) the level of instructional demand at those sites; (c) the complex-

ity of the technical content of the courses; and (d) the initial equipment costs,

which can be substantial (U.S. Congress, 1989). Nonetheless, dollar costs have

been declining for all of these systems in recent years, and those downward

trends are expected to continue (Bailey et al., 1989).

Most of the literature on distance learning has been on case studies with scant

empirical evidence (Wells, 1990) and the data are not adequate, so far, to isolate,

measure, and determine the features of course content, instructional strategy,

and instructor preparation, that are most salient for those training approaches.

Unusual demands are placed on the teleteacher to function as a dynamic televi-

sion personality, as well as a tightly organized, well-paced instructor. Those

146 CHANDLER, LUCAS, & O7HARE

factors have been reported (Bailey et al., 1989) as sufficiently different from the

requirements of traditional instruction that careful assessment of the distance-

learning environment is required to assure effective training performance.

Experimental studies to define optimal features could be conducted economi-

cally with hardwired connections between classrooms in the same or nearby

facilities to the originating training site. Multimedia workstations for individual

students is one of the options for the future; and systems that serve large student-

groups might consist of computers, high-definition TV displays, facsimile ma-

chines, and push-to-talk telephones. Practical advantages of distance learning

are expected from; (a) reduced demand for the construction of residential

schools; (b) reduced need for travel for training purposes; (c) increased accessibil-

ity to training programs by personnel at remote locations; (d) savings from

reduced number of instructor hours that are applied to the design of training

materials; and (e) systematic refinement and re-use of course materials will

promote higher-quality instruction. No doubt the value of distance learning for

a specific instructional configuration will be determined, in part, by assessments

of its cost-effectiveness, but the available data for such estimations are presently

too sparse. Nonetheless, it is likely that military training will be able to make a

strong case for the adoption of distance-learning approaches to meet its critical

requirements.

References

1. Bailey, S. S., Sheppe, M. L., Hodak, G. W., Kruger, R. L., & Smith, R. F. (1989, December). Video

teletraining and video teleconferencing: A review of the literature (Tech. Rep. 89-036). Orlando; FL: Naval

Training Systems Center.

2. Brockwell, J. E. (1989). The satellite education program (SEP) at the Army Logistics Management Col-

lege: Its beginning. Ft. Lee, VA: U.S. Army Logistics Management College.

3. Chandler, N. L., & Harris, P. A. (1991, July) Training technology survey. Washington, DC: Office of

Personnel Management.

4. Chandler, N., & Lucas, K. (1989). The electronic schoolhouse. Teleconference Magazine, 8(2):1-4.

5. Fletcher, J. D. (1990, July). Effectiveness and cost of interactive videodisc instruction in defense training

and education (IDA paper P-2372). Alexandria, VA: Institute for Defense Analyses.

6. Gopher, D., & Kimchi, R. (1989). Engineering psychology. In M. R. Rosenzweig & L. W. Porter (Eds.),

Annual Review of Psychology (pp. 431-455). Palo Alto, CA: Annual Reviews, Inc.

7. Griffin, G. R., & Hodgins, M. M. (1991). VTT in the Navy: Training now and for the future. 7. H. E.

Journal, 18(12):65-67.

8. Guthrie-Morse, B., & Julian, C. A. (1989). A small college’s tool for effectiveness: Telecommunication.

AACJC Journal, Oct/Nov, 1-6.

9. Hahn, H. A., Ashworth, R. L., Jr., Phelps, R. H., & Byers, J. C. (1990). Performance, throughput, and

cost of in-home training for the Army Reserve: Using asynchronous computer conferencing as an alterna-

tive to resident training. In Proceedings of the Human Factors Society 34th Annual Meeting (pp. 1417-

1421). Santa Monica, CA: Human Factors Society. ;

10. Hahn, H. A., Ashworth, R. L., Jr., Phelps, R. H., Wells, R. A., Richards, R. E., & Daveline, K. A. (1990).

Distributed training for the Reserve component: Remote delivery using asynchronous computer conferenc-

ing (Final Rep.). Idaho Falls, ID: Idaho National Engineering Laboratory.

11. Hahn, H. A., Harbour, J. L., Wells, R. A., Schurman, D. L., & Daveline, K. A. (1989, November).

Distributed training for the Reserve component: Course conversion and implementation guidelines for

computer conferencing (Res. Rep.). Idaho Falls, ID: Idaho National Engineering Laboratory.

12. Harbour, J. L., Daveline, K. A., Wells, R. A., Schurman, D. L., & Hahn, H. A. (1990, April). Distributed

13.

14.

15.

16.

17.

18.

19.

20.

21.

_ ~ CRM 90-36). Alexandria, VA: Center for Naval Analyses.

24.

25.

26.

Di:

28.

29.

30.

DISTANCE LEARNING . 147

training for the Reserve component: Instructor handbook for computer conferencing (Res. Rep.). Idaho

Falls, ID: Idaho National Engineering Laboratory.

Johansen, R. (1988). Groupware: Computer support for business teams. New York: Free Press.

Kinchla, R. A., & Wolfle, J. M. (1979). The order of visual processing: “top down’, “bottom up’, or

“middle out”. percaien & Psychophysics, 25:225-231.

Malone, T. W., Grant, K. R., Turbak, F. A., Brobst, S. A., & Cohen, M. D. (1987). Intelligent information-

sharing systems. Communications of the ACM, 30:390-402.

MIL-STD-188-131 (1991, July). Interoperability and performance standard for video teleconferencing

(Draft). Ft. Monmouth, NJ: Joint Tactical Command, Control and Communications Agency.

Navon, D. (1977). Forest before trees: The precedence of global features in visual perception. Cognitive

Psychology, 9:353-383.

Pugh, H. L., Parchman, S. W., & Simpson, H. (1991, March). Field survey of videoteletraining systems in

public education, industry, and the military (Tech. Rep. TR-91-7). San Diego, CA: Navy Personnel Re-

search and Development Center.

Romiszowski, A. J. (1981). Designing instructional systems: Decision making in course planning and

curriculum design. London: Kogan Page.

Rumelhart, D. E. (1977). Toward an interactive model of reading. In S. Dornic (Ed.), Attention and

Performance VI. Hillsdale, NJ: Erlbaum.

Rupinski, T. E., & Stoloff, P. H. (1990, May). An evaluation of Navy video teletraining VTT (Tech. Rep.

Saffo, P. (1990). Same-time, same-place groupware. Personal Computing, 14(3):57-58.

Simpson, H. (1990, June). The evolution of communication technology: Implications for remote-site train-

ing in the Navy (Rep. TN-90-22). San Diego, CA: Navy Personnel Research and Development Center.

Simpson, H., & Pugh, H. L. (1990, September). A computer-based instructional support network: Design,

development, and evaluation (Tech. Rep. TR-90-6). San Diego, CA: Navy Personnel Research and Devel-

opment Center.

Simpson, H., Pugh, H. L., & Parchman, S. W. (1990, September). 4 two-point videoteletraining system:

Design, development, and evaluation (Tech. Rep. TR-90-5). San Diego, CA: Navy Personnel Research and

Development Center.

Stefik, M., Foster, G., Bobrow, D. G., Kahn, K., Lanning, S., & Suchman, L. (1987). Beyond the chalk-

board: Computer support for collaboration and problem solving in meetings. Communications of the

ACM, 30:32-47.

Treisman, A. (1986). Properties, parts, and objects. In K. B. Boff, L. Kaufman, & J. P. Thomas (Eds.),

Handbook of perception and human performance, Vol. II: Cognitive processes and performance (pp.

35-1-35-70). New York: Wiley.

U.S. Congress, Office of Technology Assessment. (1989, November). Linking for learning: A new course

for education (Rep. OTA-SET-430). Washington, DC: U.S. Government Printing Office.

Wells, R. A. (1990). Computer-mediated communications for distance education and training: Literature

review and international resources (Res. Rep.). Boise, ID: Boise State University.

Young, R. (1991, Summer). Audiographic teletraining meets evolving needs of Navy’s distance learners.

FETA Newsletter, 4-5.

Journal of the Washington Academy of Sciences,

Volume 81, Number 3, Pages 148-150, September 1991

What a Difference a Sign Makes

Allan Kroopnick

Social Security Administration, Baltimore, MD

ABSTRACT

It is formally demonstrated that contrary to the general case, the uniqueness of a non-lin-

ear differential equation may not be lost when the Lipschitz condition is violated.

Mathematics majors taking an introductory course in differential equations

are usually exposed to the well-known result that solutions to such equations are

unique provided that a Lipschitz condition is satisfied. A Lipschitz condition

means that for a bounded domain D contained in R? and a continuous function

f(t, x), we have |f(t, x) — f(t, y)| <M|x —y| where M > 0(Ross, 1964, pp. 334

ff.). M is usually called the Lipschitz constant and its existence is guaranteed if

|0/dx f(t, x)| is bounded on D. If f is independent of t, this reduces to having

\f’(x)| bounded on D. For example, consider the function f(t, x) = tx? on the

unit square [0, 1] < [0, 1]. Now we have |f(t, x) — f(t, y)| = |tx? — ty?| = |t(x

— y)(x + y)| <2|x — y| on the unit square. In this instance, 2 is the Lipschitz

constant.

The method of proof for uniqueness usually involves a fixed-point theorem or

the method of successive approximations. In this note, two examples of a non-

linear differential equation with a unique solution are given when the Lipschitz

condition is violated. Furthermore, it is shown that when the sign is changed, an

infinite number of solutions occur for each differential equation, illustrating the

subtleties involved with uniqueness theorems.

Consider the differential equation, x’ = f(x) = —x!/?, with initial condition,

x(0) = 0. Note that f’(0) is infinite at x = 0, so the Lipschitz condition is violated

at that point since f’(0) is infinite. Also, the trivial solution, x = 0, satisfies this

equation with the given initial condition and is the only solution. In order to

demonstrate this, multiply the equation by x(t) and integrate from 0 to t, to

obtain:

148

WHAT A DIFFERENCE A SIGN MAKES | 149

x(1)2/2+ | x(s)'ds = 0 | (1)

All terms on the left-hand side of (1) are non-negative while the right-hand side

is 0. Hence, the trivial solution is the only solution satisfying this equation with

the initial condition, x(0) = 0.

When the sign of f(x) is changed, then the resulting equation does not have a

unique solution (Bellman, 1953, p. 69). In fact, it has a non-denumerable

number of solutions given by:

K(1) = 0 olor a-=4'= b fa’ =< 0. Db =0)

Mi (a ye don it <a

x1) ={t —bj7,+ for t>b (2)

Here, a is an arbitrary negative number, and b is an arbitrary positive number.

Note, that x(t) = 0 is a solution as well.

Next, consider the second-order differential equation:

x x2 =O (3)

It has initial conditions of x(0) = 0 and x'(0) = 0. The solution to this equation

is also the trivial solution, x(t) = 0. In order to demonstrate that this solution is

unique, multiply equation (3) by x’(t) and integrate from 0 to t thereby obtain-

ing:

M2, exh: /3 = (4)

That equation is satisfied by only the trivial solution, x(t) = 0. However, if the

sign of x!/? in (3) is changed then the resulting equation becomes:

x’ —x!/2=9 (5)

* Equation (5) has an infinite number of solutions satisfying the initial condi-

tions, x(0) = 0 and x’(0) = 0. They are given by:

x(t)=0 a<t<b (a<0,b>0)

Mite ee 144d)

x(t) — (F— hb) / 144 9G Sb) (6)

The above solutions satisfy the relationship:

wie? —2x(t)7/7 73 — 0, (ae)

which is analogous to equation (4). Again, by changing the sign 1n equation (4),

a non-denumerable number of solutions to the given equation is obtained.

150 ALLAN KROOPNICK

Thus, uniqueness may occur even when the Lipschitz condition is grossly vio-

lated, although in general this is not the case.

References

1. Bellman, R. (1953). Stability theory of differential equations. New York: McGraw-Hill.

2. Ross, S. (1964). Ordinary differential equations. Waltham: Blaisdell.

Journal of the Washington Academy of Sciences,

Volume 81, Number 3, Pages 151-156, September 1991

Effect of Cyst Age, Media, pH,

Temperature, and Time, on Excystment

of Blepharisma stoltei Isquith

Arthur J. Repak

Quinnipiac College, Hamden, CT

ABSTRACT

The effects of various media on excystment processes of Blepharisma stoltei were studied

using four different media: Cerophyll, wheat germ infusion, distilled water, and Brandwein’s

buffer solution without light over 72 hr. Excystment optimally occurred at 26° C in wheat

germ infusion and to a lesser extent in Cerophyll but not in distilled water or Brandwein’s

solution. The effective pH range for excystment was 7.2—7.8 with an optimal value at 7.7.

The presence of suitable bacteria or their metabolic products appears to contribute toward

the process of excystment. Other factors involved are the age of the cysts since their creation

and time. Excystment optimally occurs within 24—48 hr and continues to a lesser extent in

the following days depending upon the pH and nutritional value of the medium in question.

Introduction

The existence of the resting cyst of Blepharisma lateritum, a reddish to pink

ciliated protozoan, was first noted by Cienkowsky [1]. A resting cyst is a stage in

the life cycle of a protozoan, where the organism secretes a wall about its self

several layers thick. Cysts may be produced under inclement conditions or for

reproductive purposes. Later investigations [2-11] were devoted to studies of

various morphological and cytochemical aspects of cystation in a variety of

strains of Blepharisma. The conditions leading to the formation of cysts have

been examined in various ciliates [12-15]. Some of the factors that affect excyst-

ment in Blepharisma have been only partially studied by Giese [16] using B.

americanum. Past investigations used a variety of different media to grow this

ciliate [16-19]. Growth and excystment were noted in polyxenic cultures using

Cerophyll and wheat germ infusion but did not occur in distilled water or

Brandwein’s solution [18]. The purpose of this study was to investigate the

effects of cyst age, different culture media, temperature, time and pH on the

excystment of Blepharisma stoltei.

151

152 ARTHUR J. REPAK

Materials and Methods

Blepharisma stoltei (Federsee, Germany strain) Isquith 1967 was maintained

in finger bowls containing bacterialized infusions of 0.3% weight/volume (w/v)

Cerophyll. The cultures were held at room temperature without light and sub-

cultured weekly. Blepharisma stoltei, in Cerophyll, formed cysts on a regular

basis within three weeks.

To determine the effects of different media on excystment, 50 cysts, washed 3

times in two-day old Cerophyll, were placed into wells of a Falcon multiwell

sterile, plastic, tissue culture-dish containing 10 ml of (1) wheat germ infusion

(WGI) [1.e., 0.3% Cerophyll (w/v) plus one distilled water washed wheat kernel];

(2) 0.3% (w/v) Cerophyll; (3) Brandwein’s buffer solution (BWS) [20]; or (4)

distilled water. Duplicate sets were used. The cysts were placed in a moist

chamber at 26° C for 24, 48, 72 and 96 hr. in the dark. Ten trials were used in

each medium. Each day, the number of completely excysted trophozoites were

counted and removed. The ratio of the number of excysted forms per total

number of viable cysts was used to compute the percent excystment per me-

dium.

The effect of pH on excystment was determined by using 5 ml. of fresh

Cerophyll media contained in wells of multiwell tissue culture plates. Thirty

cysts, handpicked through the use of micropipet and a stereo-dissecting micro-

scope, were introduced into each well. The plates were incubated in the dark at

26° C and checked at 24, 48 and 72 hr. Alloquots of fresh Cerophyll were

adjusted to pH 4, 5, 6, 7, 8, 9, 10, 11. Unaltered Cerophyll (pH 5.8) and distilled

water (pH 5.8) served as controls. The experiments were performed twice. Cysts

were also introduced into fresh culture media and the pH checked daily.

The effects of temperature on excystment in each medium were also deter-

mined (using the same procedures described for the study the effects of different

media) at 5, 14, 20, 26 and 34° C following 24, 48 and 72 hr. in the dark. There

were ten trials in each medium at each temperature.

Cysts used were a month old except in one instance where cysts were 5 days

old.

The significance of the resulting data was statistically analyzed using GML

Analysis of Variance (ANOVA).

Results

No excystment occurred in distilled water or BWS. WGI induced a signif-

cantly (p = .05) larger percentage of excystment (M = 9.7%, SE = .59) than

did Cerophyll (M = 1.8%"; SE = .63). Best results were obtained for excyst-

CYST PROCESSES IN Blepharisma . 153

Excystment (% per day)

4 20 26 34

Temperature C C)

O WHEAT GERM INFUSION + CEROPHYLL © DIST.WATER 4 BRANDWEIN’S SOLN.

Fig. 1. The effects of temperature on the percentage of excystment of Blepharisma stoltei in four different

media.

ment at 26° C for both media (Fig. 1). No excystment was observed at 20° C for

either medium. At 34° C, the differences between the two media were not

statistically significant but they were at 26° C (p = .05).

Under preliminary conditions the greatest amount of excystment occurred

when pH was 6.3-7.2. Under the conditions of this experiment the pH of Cero-

phyll was adjusted to eight different values on day 0 and the cysts were put into

the medium (Table I). After 24 hr. the greatest amount of excystment was found

at pH 9. Excystment was seen between pH 5-10 but the percent of excystment

was not statistically different from the average. In all adjusted media (pH 2-11)

after 24 hr. the pH values tended to swing toward equilibrium settling near 7.3.

Extremely low (<5.0) or high (<10.0) pH values were detrimental to excyst-

ment.

Fresh Cerophyll with wheat germ kernels had an initial pH of 5.8. Following

incubation for 24 hr., the pH was 7.2. After 48 hr., the pH of Cerophyll was

determined to be 7.6 + .1. During the following 2 to 3-week period, the pH of

most cultures appeared to increase slowly until encystment occurred and re-

mained at pH 8.0 + .2.

In general, time had a highly significant effect on the % excystment. A media x

154 ARTHUR J. REPAK

Table 1.—Effect of pH! Levels of Cerophyll on Excystment of Blepharisma stoltei

Mean % Excysted pH of Medium

Adjusted abelian” 7 Riis seaplane SAR ER Shing |

Medium Levels Day | Day 2 Day 0 Day 2

@inliAr 0.00 28:38 5.8 7.65

Ctrl B? 0.00 0.00 5.8 5.80

Nt 0.00 0.00 4.0 4.60

2 0.00 3:33 Su} pes

3 1.85 10.93 6.0 7.45

4 We 6.90 7.0 Tego

5 2.87 22.67 8.0 7.70

6 10.00 26.67 9.0 7.65

i 0.00 el 10.0 7.80

8 0.00 0.00 11.0 7.85

M 2.06 22 ZG

SD 3.18 9.27 1.02

Min 0.00 3.33 4.60

Max 10.00 23:39 7.85

' The pH was adjusted before the cysts were put into the media.

* Control A contained Cerophyll as a culture medium with the pH unaltered.

3 Control B contained distilled water with B. sto/tei with the pH unaltered.

time first-order interaction was highly significant (p = .05) demonstrating the

lack of independence between the two variables.

Finally, percent excystment using one-week old cysts was significantly differ-

ent (3.4 times greater) from that of cysts taken three weeks after encystment.

Discussion .

Excystment of Blepharisma occurs in cultures which facilitate bacterial

growth. Bacteria probably influence the initiation of excystment and the buffer-

ing and stabilization of the media pH. The growth of bacteria, like other cells, is

time and temperature-dependent, and specific bacteria can only grow under

certain ranges of temperatures. The influence of bacteria on excystment has

been demonstrated in Didinium nasutum [12, 14]. Excystment of Nassula or-

nata was studied by Beers [21] who introduced cysts into 0.1% proteose-pep-

tone, previously inoculated with bacteria taken from the surrounding environ-

ment. McLoughlin [6] found that a whole milk and horse-dung infusion easily

induced excystment in his cultures of Blepharisma. Strickland and Haagen-

Smit [22] used a simple mixture of 0.3 M ethanol and 10°-* M potassium phos-

phate to induce excystment in Colpoda duodenaria. Giese [16] tried this simple

concoction on Blepharisma with limited success.

Osmolality, pH changes, and differences in quantities of potassium and phos-

phate were indicated as major factors in the excystment of Vorticella micros-

CYST PROCESSES IN Blepharisma 155

toma [23]. Jefferies [24] noted that fructose diphosphate, several amino acids

and citric acid may act as excysting agents for Pleurotricha lanceolata. Demar-

Gervais and Génermont [25] found that Fabrea salina excysted at 37° C. They

also indicated that during excystment a chemical signal is produced promoting

excystment in other cysts. Experimentally, Demar-Gervais and Génermont also

demonstrated that trypsin promotes excystment in Fabrea salina. The undeter-

mined chemical inducer is believed to be a proteolytic enzyme and induces an

unknown step of the excystment process. Singh [13] indicated that a bacterial

environment is necessary for excystment of many kinds of protozoa.

One more point to consider is the number of organisms that excyst in labora-

tory cultures. More organisms form cysts (encyst) than are released from the cyst

(excyst) under the culture conditions used in this study. Considering that Ble-

pharisma cultures are far from sterile and also that newly encysted forms do

_ better than older ones, there is the possibility that a parasitic fungus is responsi-

ble for the low numbers of viable cysts. Foissner and Foissner [26, 27] reported

parasitism of cysts of the hypotrich Kahliella simplex by a Ciliomyces spectabi-

lis, amember of the Lagenidiaceous fungi. Although this author has not person-

ally observed a fungus in the cysts of Blepharisma, many cysts were found to be

empty during excystment experiments.

The majority of reported species of Blepharisma are inhabitants of fresh

waters. These ciliates are subject to a variety of natural and unnatural environ-

mental fluctuations depending upon the seasons and human-oriented activities,

e.g., acid rain and pollution. Like many protozoa, Blepharisma meet changes in

food content, pH and temperature, by forming numerous cysts. When more

favorable conditions return, some of the surviving inactive cells redifferentiate,

emerge from the cyst and resume a vegetative life. This study has attempted to

define some of the limitations upon excystment faced by this ciliate.

References

jot

. Cienkowsky, L. (1855). Uber Cystbildung bei Infusorien. Zeit. Wiss. Zool., 6:301-306.

2. Penard, E. (1922). Etudes sur les infusoires d’eau douce. Geneva: Georg & Cie.

3. Stolte, H. A. (1922). Verlauf, Ursachen, und Bedeutungder Encystierung bei Blepharisma undulans.

Verh. Deutsch. Zool. Ges., 27:79-81.

4. Stolte, H. A. (1924). Morphologische und Physiologische Untersuchungen an Blepharisma undulans

Stein. Arch. Protistenk., 48:245-300.

5. Suzuki, S. (1954). Morphogenesis in the regeneration of Blepharisma undulans Stein with special refer-

ence to macronuclear variation. J. Sci. Hiroshima Univ. Ser. B. Div. 1, 15:205-220.

6. McLoughlin, D. C. (1955). A study of cystic phenomena and of macronuclear morphogenesis in strains of

the heterotrichous ciliate Blepharisma undulans Stein. Unpublished doctoral dissertation, University of

Illinois. Dissert. Abstr., 15:902.

7. Chunosoff, L., & Hirshfield, H. I. (1964). A cytochemical investigation of cysts of a species of Blephar-

isma. J. Protozool., 11(suppl.):90.

8. Isquith, I. R., Repak, A. J., & Hirshfield, H. I. (1965). Blepharisma seculum, sp. nov., a member of the

subgenus (Compactum). J. Protozool., 12:615-618.

156

ARTHUR J. REPAK

. Repak, A. J. (1966). Cortical studies of Blepharisma and associated phenomena. Unpublished doctoral

dissertation, New York University.

. Repak, A. J., & Pfister, R. M. (1967). Electron microscopical observations on the extracellular structure

of the resting cyst of Blepharisma stoltei. Trans. Amer. Microsc. Soc., 86:417-421.

. Repak, A. J. (1968). The encystment and excystment of the heterotrichous ciliate Blepharisma stoltei

Isquith. J. Protozool., 15:407-412.

. Beers, C. D. (1946). Excystment in Didinium nasutum with special reference to the role of bacteria. J.

Exptl. Zool., 103:201-231.

. Singh, B. N. (1965). Encystation and excystation in protozoa. Progress in protozoology, Intern. Congress

Series, 91:39. New York: Excerpta Medica Fdtn.

. Butzel, H. M., Jr., & Horwitz, H. (1965). Excystment of Didinium nasutum. J. Protozool., 12:413-416.

. van Wagtendonk, W. J. (1955). Encystment and excystment in protozoa. In S. H. Hutner and A. Lwoff

(Eds.), Biochemistry and physiology of protozoa: Vol. II (pp. 85-90). New York: Academic Press.

. Giese, A. C. (1973). Blepharisma: The biology of a light sensitive protozoan. Palo Alto, CA: Stanford

University Press.

. Hilden, S. A., & Giese, A. C. (1969). Effect of salt concentration on regeneration rate in Blepharisma

acclimated to high salt levels. J. Protozool., 16:419-422.

. Smith, S. G. (1965). Nutrition and axenic culture of Blepharisma intermedium. Unpublished doctoral

dissertation, Stanford University.

. Christie, S. L., & Hirshfield, H. I. (1969). Morphological variations between axenic and bacteria-fed

Blepharisma. J. Protozool., 16(suppl.):20.

. Brandwein, A. V. (1935). The culturing of fresh water protozoa and other small invertebrates. Amer. Nat.,

69:628-632.

. Beers, C. D. (1966). The excystment process in the ciliate Nassula ornata Ehrb. J. Protozool., 13:79-83.

. Strickland, A. G. R., & Haagen-Smit, A. J. (1947). Chemical substratum inducing excystment of the

resting cysts of Colpoda duodenaria. J. Cell. Comp. Physiol., 30:381-390.

. Finley, H. E., & Lewis, A. C. (1960). Observations on excystment and encystment of Vorticella micros-

toma. J. Protozool., 7:347-351.

. Jefferies, W. B. (1959). A survey of certain chemicals as excysting agents for Pleurotricha lanceolata. J.

Protozool., 6(suppl.):15.

. Demar-Gervais, C., & Géenermont, J. (1971). Données expérimentales sur le mécanisme de l’éclosion des

kystes de Fabrea salina. Protistologica, 7:421-—433.

. Foissner, I., & Foissner, W. (1986). Ciliomyces spectabilis, nov.gen., nov.spec., a Zoosporic fungus which

parasitizes cysts of the ciliate Kahliella simplex. I. Infection, vegetative growth and sexual reproduction.

Z. Parasitenkd., 72:29-41.

. Foissner, I., & Foissner, W. (1986). Ciliomyces spectabilis, nov.gen., nov.spec., a Zoosporic fungus which

parasitizes cysts of the ciliate Kahliella simplex. II. Asexual reproduction, life cycle and systematic ac-

count. Z. Parasitenkd., 72:43-45.

Journal of the Washington Academy of Sciences,

Volume 81, Number 3, Pages 157-161, September 1991

Penicillin Production Saga Recalled

Percy A. Wells’

U.S. Department of Agriculture (retired), Abington, PA

ABSTRACT

The epic journey of Nobel laureate Sir Howard Walter Florey and his associate, Dr.

Norman Heatley, to the U.S. in 1941, seeking help in making enough penicillin to confirm

their promising clinical evaluation studies, is described. The trip ended with spectacular

success even though Florey on returning to England, believed he had failed in that mission.

The Northern Regional Research Laboratory (now the National Center for Agricultural

Utilization Research) of the U.S. Department of Agriculture was deemed to be the best place

where the production problem could be solved, and indeed, two major research discoveries

at that laboratory provided the basis for successful large-scale manufacture of the drug.

This short note summarizes some key events in U.S. Department of Agricul-

ture (USDA) involvement in the development of penicillin and extends the

remarks of Moberg (1991) concerning an honor that came recently to Norman

Heatley. It is most heartening that Oxford University has conferred on Heatley

an honorary degree of Doctor of Medicine. The honor is well deserved for

Heatley played a crucial and successful role in the penicillin studies at Oxford,

headed by Howard Walter Florey. Although Heatley, like others (Bickel, 1972),

was not able to produce penicillin in large amounts, he succeeded against over-

whelming odds in making enough of the drug to establish its great potential in

treating bacterial diseases.

Further studies needed to confirm this early work were seriously stymied by

the very low yields of penicillin produced by the mold Penicillium notatum. It

was at this point, in June 1941, as related by Moberg, that Dr. Florey together

with Dr. Heatley came to the U.S. seeking help in making enough of the drug to

complete their promising laboratory and clinical work. And thus began one of

the great science sagas of modern times.

Shortly after they arrived in the U.S. I became a link in a chain of circum-

stances which quickly took these men to the one place in the world where the

' Former Director, USDA Eastern Regional Research Laboratory, Wyndmoor, PA.

157

158 PERCY A. WELLS

basis for a major break in the problem of penicillin production already existed

although no one knew it at the time. Dr. Heatley was directly involved and I

think the story needs recalling. In the end a great new penicillin industry was

created so their trip to the U.S. was a huge success even though at the time Dr.

Florey returned to Oxford he felt that his trip was a failure.

In early July 1941, Florey and Heatley concluded their harrowing war-time

trip from England to New York City. There was barely enough time to report

their recent findings to the supporting Rockefeller Foundation officials before

moving on to New Haven over the July 4 holiday which enabled Dr. Florey to

visit his two children who were staying with the Fultons for the duration of

World War II. John Fulton and Howard Florey were great friends dating back to

Rhodes Scholar days at Oxford and when Dr. Fulton heard Florey’s story noth-

ing could stop him. In a few days he had his English visitors in Washington, DC

where they contacted Dr. Ross Harrison, Chairman of the National Research

Council Board, who immediately put them in touch with Dr. Charles Thom, a

most noted mycologist in the U.S. Department of Agriculture (USDA). Dr.

Thom seemingly knew everything that was going on in his field of expertise and

without any preliminaries brought his English guests to see Mr. H. T. Herrick,

Assistant Chief of USDA’s Bureau of Agricultural and Industrial Chemistry. But

he found me there instead. Much against my will I was spending the entire

month of July in Washington, DC, backing up Mr. Herrick while he traveled.

There we were: Dr. Florey, Dr. Heatley, and Dr. Thom, along with me. These

two Englishmen were our guests that hot afternoon of July 9, 1941. It was sticky

hot and we had no air-conditioning. Dr. Florey was the spokesman. He quickly

explained their need for a large supply of penicillin to complete their clinical

studies. I had never heard of penicillin before that moment but I never doubted

his statement about this miraculous substance, quite probably because I wanted

it to be true. Before he finished my mind was made up. Mold fermentation

research had been my special field of interest from 1930-39 at our Bureau’s

Color Laboratory near the Pentagon site across the Potomac River in Virginia.

My research associates of those years had been transferred to the newly estab-

lished Northern Regional Research Laboratory (NRRL) of USDA in Peoria, IL

in late 1940 when those facilities became available. As I wrote later (Wells,

1975), it was in my view the one place in the world where the job could best be

done and it was there that the penicillin production problem must be attacked,

and that was my proposal. Dr. Florey accepted at once! My immediate telegram

to Dr. Orville E. May, Director of the NRRL, was written while they were with

me and the next morning I was able to tell Dr. Florey that all was in readiness for

their visit. Florey and Heatley arrived in Peoria at noon on July 14, 1941, justa

couple days after meeting with me in Washington and that same afternoon,

PENICILLIN PRODUCTION 159

research on the improved production of penicillin began. Dr. May insisted that

work should first be undertaken to improve the penicillin yields. Dr. Robert

Coghill, Head of the NRRL Fermentation Division, was placed in overall charge

of the project and the immediate Laboratory phase was assigned to Dr. Andrew

Jackson Moyer, a former associate of mine at the Color Laboratory in Virginia.

Dr. Heatley stayed on at the NRRL for several months to work with Moyer and

others to teach his penicillin assay method and for other reasons.

Poor Heatley! His assignment to work with Andy Moyer was an extremely

difficult one. Moyer was totally, one hundred per cent, anti-British for he be-

lieved as did many others at that time that Winston Churchill was dragging the

U.S. into World War II. To this day many Americans believe that was true. So it

is apparent that Heatley was placed in a most uncomfortable situation. Those of

us who had worked with Moyer for years knew that his bark was worse than his

‘bite and somehow, we and Heatley found satisfactory accommodation. In spite

of these personal asides all of us recognized Dr. Moyer’s expertise in nutritional

mycology.

I must back up a moment here to explain that in 1937 during studies on the

sorbose fermentation at the Color Laboratory, due to a shortage of funds, we

sought and found a low cost substitute for yeast extract used in our medium.

This substance was known in industry as corn steep liquor, a byproduct in the

wet milling process for the manufacture of corn starch. Dr. Moyer was a close

witness to this work and used it in at least one mold fermentation study.

Within a short time after the penicillin problem was assigned to him he

demonstrated that the addition of corn steep liquor to the British medium in the

optimum amount greatly increased the penicillin yield. It was a miraculous

thing and it provided a tremendous boost to the penicillin project when, on Dec.

17, 1941, Dr. Coghill was able to report this finding at an industrial conference

in New York. Up to this point the fermentation industry people, although

sympathetic to Florey’s needs, were unable to commit themselves to making

~ penicillin because of the poor yields. With this new information, the interest of

both industry and government agencies was assured. Production of penicillin by

means of a surface fermentation was undertaken (Bickel, 1972).

But even with this great improvement in penicillin yields there was in railroad

parlance a SLOW SIGN. The surface fermentation was inherently slow and the

overall yields of product needed improvement. The USDA researchers at the

Peoria laboratory knew there was a better way. Beginning in 1930 at the Color

Laboratory, research on submerged mold culture fermentation was undertaken

and by 1937, successful large-scale fermentation equipment had been built and

two fermentation processes had been developed on a pilot-plant scale to make

gluconic acid and I-sorbose (Wells, Lockwood, Stubbs, Roe, Porges, & Gastrock,

160 PERCY A. WELLS

1939; Wells, Lynch, Herrick, & May, 1937; Wells, Moyer, Stubbs, Herrick, &

May, 1937).

There was nothing really new about submerged fermentations. Far from it.

Man was into this thing in prehistoric times when he made his first intoxicating

beverages by means of the alcoholic fermentation. But it was new for mold

fermentations until it was done in 1937 by USDA fermentologists. So an organ-

ism that would grow in a submerged environment and produce large amounts of

penicillin was sought and in July 1943 it happened. Through the prolonged

efforts of Dr. Ken Raper at the NRRL, a culture of Penicillium chrysogenum

was obtained right there in the fruit market of Peoria that met these qualifica-

tions. Later this mold was tailored by X-ray and other means in other research

institutions to give very high yields of penicillin (Kauffman, 1975; Raper, 1978).

These two major research developments: (1) the use of corn steep liquor in the

medium; and (2) the use of the tailored mold Penicillium chrysogenum were

quickly adopted by industry which then was able to make its giant contributions

to the effort. Throughout this period of development the NRRL was able and

did relay new research information to industry by means of conferences so there

was no delay in transferring and using this vital information. Thus technology

transfer was at its best decades before this modern term came into use.

From the time of my meeting with Florey and Heatley on July 9, 1941, there

was a continued sense of urgency about penicillin. Thus it came about that

massive amounts of this miracle drug were available for our armies on D-Day,

June 6, 1944, a bit less than three years from our first encounter with the

problem. It was a fantastic achievement in which everyone who participated can

take pride. It ranks in the view of many, right along with the war-time develop-

ments of radar and the atom bomb. In addition to other prestigious awards the

Northern Regional Research Laboratory was named an historic site by the

American Institute for the History of Pharmacy and the Illinois Pharmacists

Association on Sept. 5, 1980, for its contributions to the large-scale production

of penicillin. The bronze plaque presented that day reads as follows:

On this site the Northern Regional Research Laboratory, USDA, made key

contributions to the development of large-scale penicillin production (1941-

1946). These included the introduction of submerged culture fermentation, the

use of precursors to produce more effective penicillin, and the discovery of a

mold strain more productive of penicillin.

References

1. Bickel, L. (1972). Rise up to life New York: Charles Scribner’s Sons.

2. Kauffman, G. B. (1975). The penicillin project: From Petri dish to fermentation vat. Chemistry, 51:11-17.

3. Moberg, C. (1991). Penicillin’s forgotten man: Norman Heatley. Science, 253:734-735.

4. Raper, K. B. (1978). The penicillin saga remembered. ASM News, 44:645-651.

PENICILLIN PRODUCTION 161

. Wells, P. A. (1975). Some aspects of the early history of penicillin in the United States. J. Washington

Academy of Sciences, 65:96-101.

. Wells, P. A., Lockwood, L. B., Stubbs, J. J., Roe, E. T., Porges, N., & Gastrock, E. A. (1939). Sorbose from

sorbitol. Industrial and Engineering Chemistry, 34:1518-1521.

. Wells, P. A., Lynch, D. F. J., Herrick, H. T., & May, O. E. (1937). Translating mold fermentation research

to pilot plant operation. Chemical and Metallurgical Engineering, 44:188-190.

. Wells, P. A., Moyer, A. J., Stubbs, J. J., Herrick, H. T., & May, O. E. (1937). Gluconic acid production.

Industrial and Engineering Chemistry, 29:653-656.

75 Years of Scientific Thought

The Washington Academy of Sciences, one of the oldest scientific organizations

in the greater Washington, DC area, has published a book entitled “75 years of

scientific thought’”’> commemorating the first 75 years of the existence of the

Journal of the Academy.

This compilation, generally aimed at a broad-based scientific readership, con-

tains 25 of the most significant Journal articles, each being of truly enduring

value. Eight of those landmark papers were written by Nobel laureates including

such preeminent scientific giants as Hans Bethe, Percy Bridgman, Harold Urey,

and Selman Waksman.

This book is the product of an intensive two-year study conducted by a blue-rib-

bon multidisciplinary Committee on Scholarly Activities which was chaired by

Dr. Simon W. Strauss, the Academy’s Distinguished Scholar in Residence.

The subject matter, which includes papers on topics such as Theories of Heat

and Radiation, Chemical Nature of Enzymes, High Pressure in Physics, Cul-

tural Implications of Scientific Research, and Separation of Isotopes, covers a

wide variety of scientific fields, including physics, chemistry, biology, anthropol-

ogy, and general science. The 25 papers provide a classic portrayal of scientific

thought over the past three-quarters of a century. For a complete listing send a

self-addressed stamped envelope to the Academy address shown below.

1987, 374 pp., author and chronological title indexes, softbound.

Price for Academy members is $15, and for non-members it is $30.

Send orders to the following address:

Washington Academy of Sciences

1101 N. Highland Street

Arlington, VA 22201

162

ae Pe

: pecaes

DELEGATES TO THE WASHINGTON ACADEMY OF SCIENCES,

REPRESENTING THE LOCAL AFFILIATED SOCIETIES

Potlesaphical Society of Washington’). ...)2.4. 40.6 foe ee ew oO ... Thomas R. Lettien

Paicapolorical society Of Washimeton «. 22.522. 2.0 2 sel oe os eso ee eae. Belford Lawson III

Pmtarieal SOCIely Oli WASHINGTON 2. 8 i5.).) 65. cae cee sche docs eats cease Pe eec Kristian Fauchald

Chemical Society of Washington ......... ful Aree ot Ste Ae RE ae tet SAS A 2 Elise A. B. Brown

Paromoiorical society of Washington .... 02... 42. 2...0.6.4-25060.5- F. Christian Thompson

aC COPTADINIG SOCIETY 4/0.) 5-08 sree he bess Sheen halo Sicwareuhen Stanley G. Leftwich

eee al SOCICLyOL WaSMINStOM 05.6 akc o 2 ieee ck de Sis vie ve lee sb eeces James V. O’Connor

Meee society ol the District of Columbia’ 0.650828 be cc hes ec de cs loka John P. Utz

Eeaecieansoctcey.ot Washineton:; DC i..55..5.88) oo.cn bec kw ew be Deak Thomas G. Manning

Beesieal Society Ol WashiMetOn fe. ss .2.c.e sco iae 6 sawn cde see obese $e da deals Muriel Poston

Sovicu on Amerncan Foresters, Washington Section .....2.: 2.0... Bo. eel ee es Eldon W. Ross

Me aie IMT SOCIeTyiOl ENOINECELS 2755). S oc aie eons eos calles ea bos gle be wielowie seo cbe Alvin Reiner

Institute of Electrical and Electronics Engineers, Washington Section ........ George Abraham

American Society of Mechanical Engineers, Washington Section ......... Clayton W. Robson

Melmnimological society of Washington ... 5... 022.0656. ceeds ce ect ee. Kendall G. Powers

American Society for Microbiology, Washington Branch .................. Herman Schneider

Society of American Military Engineers, Washington Post .................... James Donahue

American Society of Civil Engineers, National Capital Section .............. John N. Hummel

Society for Experimental Biology and Medicine, DC Section .............. Cyrus R. Creveling

American Association of Dental Research, Washington Section ............. J. Terrell Hoffeld

American Institute of Aeronautics and Astronautics, National Capital

CEES TE 2 Ske SEC pan epee Aner ote ite ae Garni Mn Ae ma eee eee Reginald C. Smith

Amencan Meteorological Society, DC Chapter ......0....... 2.0.6. 500.00005 A. James Wagner

Reta seieHeS SOCIELY Of Washington: . fo oo). 6. oe wane din sore ores aee-nnys os en cise To be determined

Acoustical Society of America, Washington Chapter ........................ Richard K. Cook

Pimenicam Nuclear Society, Washington Section .. 2... ..,..co. cc ccc eee ewan oe Kamal Araj

Institute of Food Technologists, Washington Section .................... George W. Irving, Jr.

American Ceramic Society, Baltimore-Washington Section .................. Curtis A. Martin

Be reset ETNA TP SOCIOL a 35635) ape at tle OU NINE bed cla fGias rote Mate ee atee bee Paul Natishan

Masimetandustory-of science Club)... 5.0.c.:. ce. coal cigs oe leap oe also’ Albert G. Gluckman

American Association of Physics Teachers, Chesapeake Section ............. Robert A. Morse

Mpiicalsociety of America, National Capital Section”........0:5.00.00.00..0..% William R. Graver

American Society of Plant Physiologists, Washington Area Section ............. Steven J. Britz

Washington Operations Research/Management Science Council .............. John G. Honig

Instrument Society of America, Washington Section ......................05. Donald M. Paul

American Institute of Mining, Metallurgical and Petroleum Engineers,

RiasininetOneScCHOM Ct csc 50 a Sek tis Sea Meg ey lel David M. Sutphin

Gc Oapital ASIFONOMICTS. «).5/80 5. dea iad oS nus 6iclae ew ete a cars Sie baie ous Robert H. McCracken

Mathematics Association of America, MD-DC-VA Section ..................... Alice Schafer

Bisinct-ot Columbia Institute of Chemists, .-., 20.56. ec eee aaa eset William E. Hanford

Pustrict of Columbia’ Psychological Association: { >. 2.0.2.6. Sho. 6 oP oon ces cies Sue Bogner

Masameton Pamt Technolopy Group 2.6) S25 ee Ase link eta os Ses oe ok es Lloyd M. Smith

American Phytopathological Society, Potomac Division .................... Kenneth L. Deahl

Society for General Systems Research, Metropolitan Washington

CUETO CSTR ENE se OE a ERO LOR hh ASHE Ee Sa PRS eR TOR 9 Me ee OE John H. Proctor

Human Factors society. Potomac Chapter .: i270 use l booek ol Thomas B. Malone

American Fisheries Society, Potomac Chapter...s..04 0.0... sis sends Sa David A. Van Vorhees

Association for Science, Technology and Innovation ................:..... ..... Ralph I. Cole

Eastern: Sociological Society sc oe eae eee eS i Gael Ah Ronald W. Manderscheid

Institute of Electrical and Electronics Engineers, Northern Virginia

SSCL GS Tuner eT la aah ON REL ey MUU I ee ei SAP ae ean, ae ar eee Blanchard D. Smith

Association for Computing Machinery, Washington Chapter ............. Charles E. Youman -

Washinston StatisticaliS@ciety: 0 i. nesta ee we eee WO Oe fe Nancy Flournoy

Society of Manufacturing Engineers, Washington, DC Chapter ......... Sheep! James E. Spates

Institute of Industrial Engineers, National Capital Chapter ................... James S. Powell

Delegates continue to represent their societies until new appointments are made.

Washington Academy of Sciences 2nd Class Postage Paid

1101 N. Highland St. at Arlington, Va.

Arlington, Va. 22201 and additional mailing offices.

Return Requested with Form 3579

we (7

VOLUME 81

Number 4

“i our nal of the | December, 1991

WASHINGTON

ACADEMY .. SCIENCES

ISSN 0043-0439

Issued Quarterly

at Washington, D.C.

EMTHSON A

NOV 021998

LIBRARIES

CONTENTS

Article:

VALERY F. VENDA and YURI V. VENDA, “Transformation Dynamics in

Complex Systems”

Academy Reports:

C. R. CREVELING, “The 1991 Washington Academy of Sciences Awards

Program for Scientific Achievement”

“The Bylaws of the Washington Academy of Sciences”

“1991 Washington Academy of Sciences Membership Directory” ............ 203

Washington Academy of Sciences

Founded in 1898

EXECUTIVE COMMITTEE

President

Walter E. Boek

President-Elect

Stanley G. Leftwich

Secretary

Edith L. R. Corliss

Treasurer

Norman Doctor

Past President

Armand B. Weiss

Vice President, Membership Affairs

Cyrus R. Creveling

Vice President, Administrative Affairs

Grover C. Sherlin

Vice President, Junior Academy Affairs

Marylin F. Krupsaw

Vice President, Affiliate Affairs

Thomas W. Doeppner

Board of Managers

James W. Harr

Betty Jane Long

John H. Proctor

Thomas N. Pyke

T. Dale Stewart

William B. Taylor

REPRESENTATIVES FROM

AFFILIATED SOCIETIES

Delegates are listed on inside rear cover

of each Journal.

ACADEMY OFFICE

1101 N. Highland Street

Arlington, VA 22201

Phone: (703) 527-4800

EDITORIAL BOARD

Editor:

John J. O’Hare, CAE-Link Corpora-

tion

Associate Editors:

Bruce F. Hill, Mount Vernon College

Milton P. Eisner, Mount Vernon Col-

lege

Albert G. Gluckman, University of

Maryland

Marc Rothenberg, Smithsonian Insti-

tution

Marc M. Sebrechts, Catholic Univer-

sity of America

Edward J. Wegman, George Mason

University

The Journal

This journal, the official organ of the Washing-

ton Academy of Sciences, publishes original

scientific research, critical reviews, historical

articles, proceedings of scholarly meetings of

its afhliated societies, reports of the Academy,

and other items of interest to Academy

members. The Journal appears four times a

year (March, June, September, and De-

cember). The December issue contains a di-

rectory of the current membership of the

Academy.

Subscription Rates

Members, fellows, and life members in good

standing receive the Journal without charge.

Subscriptions are available on a calendar year

basis, payable in advance. Payment must be

made in U.S. currency at the following rates:

U.S: and Canada... 2)... eee $25.00

Other countries: .......242. 455 4ee eee 30.00

Single copies, when available ....... 10.00

Claims for Missing Issues

Claims will not be allowed if received more

than 60 days after the day of mailing plus time

normally required for postal delivery and

claim. No claims will be allowed because of

failure to notify the Academy of a change of

address.

Notification of Change of Address

Address changes should be sent promptly to

the Academy office. Such notifications should

show both old and new addresses and zip-code

numbers, where applicable.

Published quarterly in March, June, September, and December of each year by the

Washington Academy of Sciences, 1101 N. Highland Street, Arlington, VA 22201.

Second-class postage paid at Arlington, VA, and additional mailing offices.

Journal of the Washington Academy of Sciences,

Volume 81, Number 4, Pages 163-184, December 1991

Transformation Dynamics

in Complex Systems

Valery F. Venda

The University of Manitoba, Winnipeg, Canada

and

Yuri V. Venda’

Moscow State University, Russia

ABSTRACT

For more than one hundred years, researchers of psychological theories of learning, eco-

nomics, human factors, education, and management, have described the processes of devel-

opment in individuals, firms, technologies, and nations, as simply monotonic. That assump-

tion was adopted even though it did not always predict and adequately describe those

development processes. When the structures in industry, professional education and train-

ing, and society, changed slowly, monotonic models were more or less acceptable. Now, the

success of dynamic systems, like advanced manufacturing, retraining, and progress in the

former Soviet Union, eastern Europe, and other countries, depends on a process of deep

transformation for the very structure of those systems. Old theory is not sufficient in those

cases, transformation dynamics theory which studies changes in complex system structures

is required to help in the prediction of the dynamics of efficiency, to minimize inevitable

losses, and to speed up the attainment of the highest levels of productivity, quality, reliabil-

ity, and safety. A newly discovered law of transformations is described as a fundamental

basis for that new transformation theory.

Traditional Monotonic Models

Monotonic dynamics models came from traditional psychological learning

theory. Studies on learning have been a central problem in world psychology.

Systematic experimental research on learning processes was initiated by H.

' While he was critically analyzing and generalizing 20 years of experimental data on transformation learn-

ing processes, Y. Venda discovered the Law of Transformations. He was tragically killed, at the age of 22, on

August 9, 1991, at a summer camp during a tornado. This paper is dedicated to his memory.

163

164 VENDA & VENDA

Ebbinghaus (1885). His numerous experiments led him to identify the existence

of monotonic (actually exponential) regularity in learning dynamics. E. L.

Thorndike (1898) and C. L. Hull (1943) confirmed that regularity. Many subse-

quent experimental studies and mathematical models of learning and develop-

ment processes confirmed the same monotonic shape of the dynamic curves

(Atkinson & Crothers, 1964; Bower, 1961; Bush & Mosteller, 1955; Estes, 1950,

1959; Greeno, 1967; Luce, 1959).

Monotonic learning curves were generally accepted in behavioral science. For

example in a handbook on experimental and design techniques in engineering

psychology, Chapanis (1959) recommended monotonic exponential approxi-

mations for any kind of experimental data on individual training. On this basis

Zarakovski, Korolyou, B. M., Medvedev, V. I., & Shlaen, P. Y. (1977) proposed

to compute learning curves based on three or even two empirical points. As a

result most authors in the last 50 years considered dynamics in development as

very simple, 1.e., monotonic and exponential. The experiments by Bryan and

Harter (1899) on training of operators using telegraphic code, where interme-

diate plateaus had been found, were ignored because no theory or regularity was

given for these unusual phenomena. A short explanation, with a hypothesis

regarding plateaus as indicators of changes of the observer’s strategies was given

by Woodworth (1938). Hence only two types of learning dynamics curves were

usually considered during the early years: monotonic exponential, and mono-

tonic with intermediate plateaus. Monotonic exponential dynamics of develop-

ment was regarded as classic and universal.

From Models of Monotonic Statics to Monotonic Dynamics

Monotonic exponential models were widely used not only in analysis of indi-

vidual learning processes, but in many other areas of science and practice. For

example, many economists tried to predict dynamics of the financial state of

companies and industries as monotones exponential processes. They found that

long term predictions up to 30 years or more could be more or less modeled with

that type of progress curve. However attempts to use traditional monotones

curves for prediction of shorter periods were unsuccessful. Yanch (1974) named

middle-term prediction periods as difficult periods. One could imagine how

many companies fell into bankruptcy with this primitive modeling and predic-

tion of dynamics by monotones processes. Monotones learning theory has made

some people unhappy. Teachers and coaches expect steadily increasing aca-

demic and athletic success. Any declines lead to such dramatic results as re-

moval of non-performing people. Wrong theory led to wrong decisions. The

TRANSFORMATION DYNAMICS 165

Q orl, 0 1), SR neta Q _ _ Qmax pus we

I T

Fig. 1. Illustration of the Fitts’ Law (left side) and monotones dynamics as its consequence (right side). Q =

efficiency of performance; I = information volume perceived; T = time.

‘same trouble was also common in science and industry. Temporary declines of

efficiency are inevitable and play a positive role in the development of the

individual, firm, technology, science and every complex human-machine-en-

vironment system. Why has monotonic development dynamics hypnotized peo-

ple for such a long time? There is an experimental basis for the phenomenon. A

theoretical basis for monotones exponential theory of dynamics was provided

by Fitts’ law (1954), illustrated on the left side of Fig. 1. The monotones influ-

ence of the volume of information perceived by an observer’s efficiency of

information processing was stated by Fitts. The static characteristic of that influ-

ence is monotones. It is easy to understand (monotones theories are very attrac-

tive because they are easily comprehended) that dynamics also will be like

monotones. Indeed a learning process is always based on gradual increasing of

information displayed and perceived by the individual. According to Fitts’ law

the efficiency of the performance studied should also increase gradually. Zara-

kovski et al. (1977) have drawn this simple logical conclusion as shown on the

right side of Fig. 1. After reaching the single maximal level of efficiency (Qmax)

_ determined by Fitts’ law, further learning will not affect efficiency. The same

authors adopted an over-simplification of learning theory leading to approxi-

mations of the monotones learning curve with only three, and then just two,

experimental data points. That simplification is far from real learning and devel-

opment dynamics in individuals and human-machine-environment systems,

and monotones models are wrong and useless in many important practical

cases.

From Fitts’ Law of Monotones to a Law of Transformations

Sheridan has noted that Fitts’ model accords with the experimental data in a

number of relatively simple studies, but cautions that “like so many elegant

166 VENDA & VENDA

F i

Fig. 2. The law by Yerkes-Dodson and F. W. Taylor (left side) and its consequence as a bell-like shape of

performance efficiency dynamics (right side). Q = efficiency; F = ergonomic factor; T = time.

models for human behavior, Fitts’ model breaks down for more complex manip-

ulations” (1992, p. 123). In the very beginning of this century Yerkes and Dod-

son (Woodworth, 1938) and F. W. Taylor (Freivalds, 1987; Konz, 1990) found

that every human or animal performance has an optimal condition for maximal

performance efficiency. That means that if efficiency of information processing

(understanding of texts, decision making, diagnosis) is a bell-shaped function of

the information volume perceived, it can not be constant when information

volume surpasses some certain, optimal level for maximal efficiency. Further

increasing of information volume will lead to decreases in performance efh-

ciency. Hence, performance efhciency dynamics for gradual information vol-

ume increases will also have a bell-like shape (Fig. 2). Several series of experi-

ments (Venda, 1990) were conducted that examined hypotheses about

non-monotonic behavior in the efficiency of human performance:

1. Observers were asked to identify words of 8 letters’ length. Different, randomly (for

static characteristics), and gradually (for dynamic), increasing numbers of letters

were displayed: from 0 to 90. The probability of correct answers changed as shown in

Fig. 2. This bell-like shape of the Q function meant that Fitts’ law was not correct for

volumes of information greater than optimal. In other experiments (Venda, 1989-

91) the optimal volume was equal to the length of the word, i.e., 8 specific letters. If

more than 8 letters were displayed, the exact word became masked. Masking was

greater with further increases in the number of letters. After 15—17 letters the proba-

bility of a correct answer was practically equal to zero.

2. Observers were asked to read words with a length of 8 letters (Venda, 1990). All

letters were displayed, but with special fonts making reading more or less difficult, so

that the observers were reading by separate letters, by syllables and by whole words.

Eye movements were recorded. In another series of experiments, text by separate

letters, syllables and words was displayed on the computer screen. In addition, text

moving on the screen with different speeds was displayed for reading by letters,

syllables and words. The static characteristic curves for eye movements during three

strategies of reading (letter, syllable, or word) as a result of the experiments are shown

TRANSFORMATION DYNAMICS 167

(iio ge AAS Se Me ee ae BE Qe sive renee ES IC ee en eee ae

oo) ES Ye

---—-— 0) iar nn a Sb to Se

le toile Sa/ _—Sa to Sb

= LS

Fe opt’ Fb opt: 10 Fa opt: 22 a0 6k - - 3 T, days

Fig. 3. Transformation dynamics in information perception: characteristic curves of reading strategies: Sa

= reading by letter; Sb = by syllable; Sc = by word (left side); and transformations of those strategies (right

side). Q = efficiency (probability of successful trial divided by time spent); F = ergonomic factor (number of

eye movements during perception); T = time (Venda, 1990).

in Fig. 3 (left side). The mght side of Fig. 3 shows learning curves with changing

strategies from Sa to Sb and then into Sc or directly from Sa to Sc. These characteris-

tic curves are static because the curves display the influence of a randomly changed

factor, not a process in time. The characteristic curves were obtained by selecting

data (Q and F) for different performance strategies. For example, analysis of eye

movements during information perception enables us to classify each strategy of

reading.

Fundamentals of Transformation Dynamics Theory

Several important findings led to a fundamental theory of transformation

dynamics:

The same human performance could be accomplished with different strategies, and

every strategy has its specific characteristic curve, 1.e., a correlation between perfor-

mance efficiency (or other criteria) and ergonomic (psycho-physiological) factors of

performance. The same items could be produced by using different technologies and

management structures.

. If an ergonomic, economic, technological or management factor is increased gradu-

ally, monotonically, and the performance has only one strategy, the dynamics of

efficiency (learning curve) will have bell-like shape.

. If an ergonomic or some other factor is increased gradually, monotonically, but

different strategies are used in practice, efficiency changes to a wavy-like process,

with monotones, exponential phases of development for every concrete structure-

strategy, and efficiency decreasing when a previous strategy is transformed into a new

one.

So, in addition to monotones exponential learning curves discovered by Ebbinghaus

(1885) and learning curves with intermediate plateaus discovered by Bryan and

Harter (1899) wavy-like learning processes were found. Woodworth (1938) suggested

that the phenomenon of plateau arises because observers change their cognitive

strategies while executing the task. The same explanation can be used in our case, but

168 VENDA & VENDA

the difference between strategies (distance between their optimal F values) needs to

be big in order to have a wave.

5. Transformation of one strategy to another depends on the level of efficiency (Q) and

value of factor (F) common to both strategies. In Fig. 3, transformation states are

illustrated by the crossing points of the characteristic curves for the strategies Sa, Sb,

and Sc.

6. Psycho-physiological structures which are the bases of the respective strategies of

performance have some common parts. When a transformation is starting, that part

which is specific to the structure is eliminated. Only the common part remains.

Obviously the efficiency of this part is lower than that of the whole structure. Subse-

quently, on the base of this common part a whole new structure is synthesized and

the efficiency rises to a higher level.

7. Sometimes, performance structures include many different levels of the human or-

ganism. Barabash (1982) of the Novosibirsk Branch of the Russian Academy of

Science showed that the training of astronauts affects not only the psycho-physiologi-

cal level but also biological and cell levels of their organism. Transformations of

structures can be, at times, very fast (in our experiments it could take only minutes)

but sometimes, very slow. It is well-known that the training of athletes takes many

years. We suggest that transformation strategies are important not only in the direc-

tion of ever-increasing complexity, efficiency and achievement, but also in the re-

verse case of reduction in complexity, efficiency and achievement, such as in factory

downsizing, de-automation to meet reduced demands, and retirement of highly mo-

tivated and skilled staff. It is important not to slide down too fast on the left side of

the curve. Previously challenged individuals have met untimely illness, depression or

even death, once challenge is removed (e.g., retirees from executive levels, or former

athletes).

8. Examples have been found of such wavy-like processes in physics, chemistry, engi-

neering, non-linear control theory, metallurgy, optics, electricity, social processes, as

well as in economics (Venda, 1990). Those transformation dynamics processes have

similar mechanisms and should be studied with the single methodology of transfor-

mation dynamics (Y. Venda has proposed the name, Transformatics, for this future

science).

Law of Transformations

Y. Venda worded the Law of Transformations as follows: Transformations of

structures-strategies of any system go through states common to the previous and

following structures-strategies. By system he meant a complex unit with con-

stant components and energy-material resources. By structure of the system he

meant the regularity of mutual adaptation processes between inner components

of the system. The regularity could be displayed as a technological scheme

(technological structure), scheme of organization hierarchy (management struc-

ture), sequences of operations (algorithm structure), structure of the control

system (dynamic links between parameters with transferring equations), and the

tree structure of a work operation. By strategy of the system he meant the

regularity of mutual adaptation of the system with its enviroment (the charac-

TRANSFORMATION DYNAMICS 169

teristic curves at the left sides of Figures 2—5 display different strategies). Hence,

strategy depends on interaction between internal structures and external condi-

tions. It is very important always to study the pairs: the structures and respective

strategies of the system.

Transformation Dynamics in Decision Making

Y. Venda was interested in testing the Law of Transformations in many

different systems and conditions. He proposed to study transformations in learn-

ing as a long-term psychological and decision-making process. He started to

analyze my old experimental protocols, and was especially interested in natural,

field, and industrial experiments.

Special emergency experiments were devised at Moscow’s fossil power plant

#21. At that time complex equipment for the control room of the plant was

designed and implemented (Venda, 1982). Moscow Power Plant Headquarters

gave special permission for the conduct of emergency experiments for objective

studies of efficiency, reliability and safety of control room equipment, and analy-

sis of the performance of human operators under normal and emergency situa-

tions. Ten emergency situations were created during night time (between 2 and

5 AM). Suddenly and secretly, without the knowledge of the operators, impor-

tant technological equipment like working feed water pumps, air and dust fans,

and fuel lines, were turned off. Operators were supposed to recognize, diagnose,

locate and eliminate those emergencies with information obtained from annun-

ciator-flashing labels, mnemonic schemes, and computer displays. Five opera-

tors participated in the experiments, each operator two times. All commands,

comments, inquiries, operations, motions, and eye movements, were recorded

with movie camera, computer, and telemetric psycho-physiological instru-

ments.

Three cognitive strategies were found among operators: Sa—perception of

information by separate elements; Sb—perception of information simulta-

neously by small chunks (2-11 functionally connected elements); and Sc—per-

ception of information simultaneously by big chunks (20-50 functionally con-

nected information elements). Characteristic curves of these strategies and their

transformations during decision making processes are shown in Fig. 4. When

different strategies are used simultaneously, in parallel, to solve especially com-

plicated, multi-factor tasks, a special group of different specialists with the aid of

a Hybrid Intelligence system can be implemented. The system would use indi-

vidually adapted information displays (Venda, 1990).

170 VENDA & VENDA

a a ee me a ae a ee ae ee ee ee we ee ee el oe ee ee

Fc opt: Fb opt: Fa opt: F - 3 5 7 T, min

Fig. 4. Transformation in human operator performance under emergency conditions at a power station:

characteristic curves of operator information-perception strategies (Sa = by separate elements; Sb = by small

chunks; Sc = by whole technological units) (Venda, 1982).

Transformation Dynamics in System Safety

It was found during the emergency experiments that it 1s especially important

to teach operators not only to perform main control strategies but to quickly and

easily transform one strategy to another, e.g., to transform a strategy and psy-

cho-physiological state adequate for a normal control situation into one which 1s

adequate for a suddenly occurring emergency. Fig. 5 (right side) shows processes

of transformations of a normal strategy. Sn, into an émergency (alarm) strategy,

Sa, and back into Sn. The Factor F is the relative level of psycho-physiological

strain as a compiex parameter of electrical brain activity (in alpha, beta, and

delta intervals), average number of eye movements per minute; or angle of

movement and fixation duration of the eyes (Venda, 1982).

To consider the multilevel character of human structure it 1s important to

explain these phenomena:

1. For individuals (as well as other complex systems), structures and strategies in their

explicit representations in the processes of mutual adaptation, are plural. Human

Fo opt Fna_ Fa opt F Tab-ATsn tosa—p T aeq— ATsa to a T

Fig. 5. Transformation dynamics of operator’s strategies at the start and end of emergency situations. Sn =

strategy in normal situation; Sa = strategy in alarm situation; Q = efficiency; T = time.

TRANSFORMATION DYNAMICS 171

performance can be achieved with several different structures and strategies. Effec-

tive and safe performance is based on adequate structure and strategy.

2. Human structures and strategies are discrete. There are essential intervals between

values of every factor of mutual adaptation efficiency and complexity that is optimal

for different structures-strategies. The same individual can respond very differently

to the same information in normal and emergency situations.

3. A human operator well trained to work separately in normal and in emergency

conditions may fail in the transformation from normal to an emergency strategy,

spending too much time and lowering reliability. So, training in transformations of

_Strategies would be very important in the training of nuclear and fossil power-plant

operators, aircraft pilots, and air traffic controllers.

4. When the time needed by a human operator for transformation of normal strategy

into an emergency one is longer than the time for critical changes of the control

object, automatic shut-off of the safety system would be activated, because the hu-

man operator can not follow control processes. Inability of human operators to

transform their strategies synchronously with changes in the control-object dynamic

- structure, in the absence of active shut-off system, was one of the main causes of

Chernobyl Nuclear Power Plant (NPP) catastrophe in 1986 (Venda, 1990).

Rasmussen’s methodology of ecological interface design (1986) and Bel-

tracchi’s model based information systems with combination of technological

and physical structures (1984) are perfect examples of organization of effective

transformations of human operator strategies under quick changing conditions

from normal to emergency and in the reverse direction at a NPP. Beltracchi

proposed using a thermodynamic model of the heat engine Rankine cycle as an

external mental model of a nuclear power plant for human operator-computer

interface. No new type of model-based display for the human-machine control

processes was found necessary; it was shown that with a model based upon the

Rankine cycle as an interface, it is much easier for a human to employ a first

monitoring strategy to evaluate plant performance. Rasmussen and Vicente

(1987) and Beltracchi (1987, 1988) proposed models of analysis with knowl-

edge-based, rule-based, and skill-based behaviors. Transformation methodol-

ogy allows the study, predicts safety and efficiency changes, and specially orga-

nizes transformations between those types of behavior strategies when it is

necessary for optimizing the control processes. Yufik, Sheridan, and Venda

(1992) provide general theoretical and methodological bases for knowledge mea-

surement in mutual human-machine adaptation. The methodology of transfor-

mation dynamics, hybrid intelligence and mutual adaptation in human-ma-

chine-environment systems have been successfully used in ergonomic design

and in the improvement of many experimental and industrial complex systems

in the former USSR (Venda, 1982, 1990).

Transformation Dynamics in Information Systems

The higher level of efficiency of strategy Sc over Sa or Sb, and Sb over Sa, in

the emergency experiments (Fig. 4) could be explained with simple information

172 VENDA & VENDA

measures. Higher efhiciency means (in this case) lower response time and corre-

sponding task complexity. The natural question is: How does a change in strat-

egy make it possible? Let us analyze the following example. Suppose an observer

is taught to identify, 1.e., diagnose, 16 states of an object which is described by

binary values on a total of 30 dimensions. Each binary value may be thought of

as representing either the normal or pathological state for that dimension.

There are two important consequences for practice: a) strategy Sc with more

narrow range of factor F values than Sa: (Fc max — Fe min) < (Fa max —

Fa min) usually has higher maximal efficiency: Qc max > Qa max. We say in

this case that Sc is more specialized and Sa more universal. Obviously, reading

by words is more effective in optimal environmental conditions (light, size and

style of the fonts) than reading by letters. But the same deviation from optimal

conditions will cause greater decreases in efhiciency of Sc than of Sa.

In the experiments on reading (Fig. 3) as well as in the emergency experiments

(Fig. 4), the strategies used by operators were usually in the sequence Sa-Sb-Sc.

Sometimes, after short and unsuccessful trials with strategies Sc and Sb, opera-

tors made reverse transformations back to Sb and Sa, and later, to Sb and Sc.

That means that decision making processes include using and assessing different

cognitive strategies, using various methods of combining (chunking) of infor-

mation elements as well as direct and reverse transformations of strategies.

Microanalysis of Performance Transformations

In another series of experiments we used a power plant training center where

similar emergency situations were modeled. Engineering students (25 in num-

ber) participated as observers with 10 trials by each. They were supposed to

make decisions, carry out operations, and give commands, like operators at the

power plant. The experiments showed the same dynamics of human perfor-

mance during learning processes as those observed in the previous experiments.

The main difference was that in the learning processes, reverse transformations,

as well as transformations from Sa directly to Sc, were found in only two stu-

dents out of the 25. Wavy-like processes were studied in teaching students spe-

cial skills in speed-reading with use of a metronome, tachistoscope and pacer

(fast moving text on the display screen), teaching power-plant operators to per-

ceive information from mnemonic schemes with different information struc-

tures and to maintain tracking control with one to six simultaneously perceived

dynamic signals (Venda, 1986).

What is especially interesting in transformation dynamics? We found that

when analysis of performance becomes more and more detailed, transformation

TRANSFORMATION DYNAMICS 173

waves can be seen at any phase of performance and development. For example,

transformations occur in long-term individual professional development and

the waves are seen during periods of changing of occupation, functions and

positions of a person in a company. Besides those waves, macroanalysis will

show smooth, monotones processes. More detailed analysis of individual perfor-

mance during one year, then one month, one day and even during one act of

decision making, will reveal more waves in human performance and develop-

ment. More detailed analysis shows more transformations; less detailed analysis

masks many transformation processes and a whole individual career may look

like a smooth monotones process. Everybody knows that there are many ups

and downs in a career.

Appropriate detailization in the analysis of human performance and develop-

ment is important for studying transformation processes. Macro and micro-

- analysis are relative definitions. In practice, if transformations are not found

with macroanalysis, changing the methodology to more detail (microanalysis)

will do so.

Human life is wavy-like. It is inevitable from the point of view of our transfor-

mation theory that deep temporary decreases in health, performance efficiency

during changes in occupation, life style, sport activity, aging, need to be studied

attentively. Adapting to a high level of business or sport activity normally takes

much time and includes several big waves. If somebody tries to stop those

activities quickly it could be dangerous. Many early deaths of sportsmen have

occurred among those who stopped training without necessary reversal waves of

transformations back to low load. Every training, sport, or profession leads to

deep changes in the organism, shaping step-by-step the appropriate psychologi-

cal, physiological and biological structures. Transformations of the structures

takes time and effort, losses in those periods are inevitable. These are some of the

implications of natural law of transformations.

Transformations in Biomechanics and Professional Training

As with any other complex system, the human body allows for a plurality of

its structures and respective strategies of behavior. In changing the structures/

strategies, their transformations are based on human evolution and on individ-

ual development of abilities for mutual adaptation with environment, machines

and other people. Such plurality of mental and physical structures/strategies is

particularly important when employees are requested to adapt to new technolo-

gies, workplaces or to develop work skills.

While the mind and body adapt to the new activity, new strategies and new

174 VENDA & VENDA

muscle groups are incorporated in the task, and error rate and effort are reduced

through a natural process of optimization. By recognizing the general law of

transformations, training programs can be made more effective and training

time reduced. The major benefits of recognizing and using transformation

theory are not confined to simple cost efficiencies; the reaction of the physiologi-

cal system to external stimulus is such that proper adaptation to change can

reduce the incidence of stress-related death (such as heart attack) as well as

absenteeism in the work place. Since adaptation is essentially the process of

adopting new operational strategies, and new strategies are adopted through a

behavior compatible both with the old and the new, it is apparent that when the

degree of change requested of an individual or of a system is such that very little

is in common, drastic failure of the system 1s imminent.

When a task is re-organized or a new job or level of job is attained, there is an

associated stress to which the physiology must adapt. In industry, the purpose of

re-organization is to increase productivity and often is predicated on an increase

in performance of individuals. Figures 3 and 4 can be used as a qualitative

portrayal of the performance benefits of three possible strategies (Sa, Sb, Sc) and

the expected performance Q for each. If a task is currently employing strategy

Sa, then the individual would shift to strategy Sc to elicit a higher productivity.

However, the law of transformations requires that the new strategy will be

adopted and adapted through a performance level no greater than that which is

common to both strategies. Figure 4 indicates that such a performance efh-

ciency during direct transformation of Sa into Sc is extremely low. The intersec-

tion of the strategy curves 1s at a performance level one-fifth the normal perfor-

mance using the current strategy. A corporation attempting to introduce new

technology without having its personnel comprehend it, may be in exactly the

position described above. The possibility of financial disaster is great under

those circumstances, particularly if the performance measure 1s goods produced.

However, the focus here is on the complexity level F, associated with perfor-

mance (Venda, 1990). F could be as any human factor, for example, level of

work stimulation. Note that F increases with each new strategy and thus the

individual worker must adapt to a new task environment, the biomechanical

motion requirements of that environment, and the new biochemical environ-

ment that the exogenous and endogenous stimulation produces.

To adopt new work strategies, planning is needed to reach the most lucrative

strategy. Refer to Figure 4, where the use of the intermediate strategy Sb, could

reach the Sc strategy without the loss in productivity associated with a direct

attempt to go from Sa to Sc. In the move from Sa to Sb, the value of performance

Q drops to the value of the intercept of those two curves, so performance efh-

ciency is higher than the intercept of curves Sa and Sc, and productivity is

TRANSFORMATION DYNAMICS 175

maintained. The productivity saved is the difference in the performances inte-

grated over the time period of the projected strategy changes. The curves of

Figure 4 also have a second meaning—a more human meaning. As stimulation

level rises, the emotional and psychological stress levels are kept lower if per-

sonal productivity can be maintained. Therefore, by approaching the goal

through intermediate strategies, stress-related occupational diseases are mini-

mized and the skills of the work-force are retained.

The eiement of time is important in the adaptation process. The denial of

time can result in the stimulation level rising to the point of over-stimulation but

forms of biomechanical structure adaptation also require time and a pro-

grammed approach (Venda & Thornton-Trump, 1992). While biomechanical

adaptations are often not considered in industrial situations, new executives

notice that clothes begin to feel tight as they gain weight from days and nights at

a desk. The point is, physiology adapts to the level of physical activity such that it

becomes dangerous to exert oneself at levels that once were a normal part of a

job. It is from this observation and from athletic training programs that infor-

mation can be drawn on the biomechanical adaptation process and its confor-

mance to the law of transformation.

The energy supply system for human motion is multi-partite. A runner may

draw energy from creatine phosphate (CP) as a result of energy release from a

phosphate bond as a result of the decomposition reaction. Energy may be gained

from glucolysis, as glucose is lysed to lactic acid. A third source of energy is the

aerobic oxidation of proteins, fats, and carbohydrates which at the same time

replenish the reserves of adenosine triphosphate (ATP). What is important here

is the rapidity with which the energy can be accessed and the level at which the

activity can be maintained. In training, the runner forces physiological processes

to increase the rate of energy release due to decomposition of creatine phosphate

as well as the rate at which glycolysis takes place. Each individual has a different

_ potential for the upper limit of this activity and so not all can be good sprinters,

but the physiology responds according to the qualitative behavior shown in

Figure 6.

Due to the time period of the events a sprinter takes part in, the primary

energy reserves used are from creatine phosphate (CP) as shown as curve Q1 of

Figure 6, and glycolysis shown as curve Q2. The adaptation of runners to sprint-

ing may be monitored by measuring the energy reserves used from these two

sources after they have performed the event for which they are being trained.

Since CP has three times the total energy of the glycolysis system, the rate of

release of CP is of extreme importance in adapting to sprint events. For long-dis-

tance runners, the third energy source, oxidation of fats, proteins and carbohy-

drates, becomes important. Since the oxidation source has four and a half times

176 VENDA & VENDA

Q,.% 100 200

100

50 Q2

0 5 10 15 20

Fig. 6. Locomotor energy sources for a runner. Q! = creatine phosphate consumption; Q2 = glycolysis; Q3

= oxidation of fats, proteins, and carbohydrates.

less energy than glycolysis, it is the last source of energy to be switched on during

locomotor activity. Indeed, sedentary people may never operate at a high rate.

In this latter case, physical exhaustion sees an early onset and the stimulation of

the organism moves into low performance efficiency level on the characteristic

curves (Fig. 3 and 4). In the case of a runner in training, if the exercises under-

taken are changed to utilize new muscle groups, the energy release targets are

changed, representing a change in strategy for the physiology thus an erosion of

performance in sprinting would be seen prior to the hoped-for increase in perfor-

mance. Such changes have been observed by coaches.

A more common situation in which the law of transformation can be seen to

apply is in biomechanical adaptation to a knee injury. In such a case, the phase

relationship of the locomotor muscle firings is changed for both the injured and

the uninjured limb. The amplitude and duration of the firing of the muscle

groups also changes. The reasons for such changes are primarily the attempt of

the two nearest uninjured joints to alter activity to compensate for the injured

joint, but also may be a pain-avoidance mechanism. The results of the motion

restriction and of the phase changes in muscle group activity are to cause the

speed of normal locomotion to decrease and to change the floor reaction force

TRANSFORMATION DYNAMICS . 177

record such that the Fourier components of that record are changed in magni-

tude and in phase relationship (Thornton-Trump & Suzuki, 1991).

The law of transformation is seen to be valid for many biomechanical and

psycho-physiological processes and can be applied in the design of training

programs when changes in work-place are such that new sequences of move-

ment, levels of movement, or new groups of tasks, are required of an individual.

Parameters can be developed and measured to assess the adaptation level of

individuals and to determine training programs appropriate to their ability to

adapt.

Productivity can be maintained at a higher level during a transition in strate-

gies when intermediate strategies are used. By designing transitions recognizing

the law of transformation, both corporate and individual health can be best

served. Of primary significance in the changing of strategies is a recognition of

the fact that new strategies can only be adopted through the elements common

to both the old and the new strategy. It is through the recognition of this, the law

of transformation, that a more reliable basis for the estimation of costs involved

in the adoption of new work skills, can be established.

Using Transformation theory leads to many practical observations. For exam-

ple:

1. The longer time that an individual remains at a stable level (plateau) of strategy Si,

the longer will be the time needed for transformation of Si into Si + 1;

2. Individual learning capacity, creativity, adaptability, mutual adaptation with new

machines, and environment are dependent on an ability to transform to new strate-

gies, especially distant ones (with big differences between Fi-opt.);

3. It is necessary to teach human operators, pilots, and sportsmen, not only different

effective strategies (Such as normal and emergency operative conditions) but also

how to transform those strategies for appropriate mutual adaptation with the environ-

ment. Effective transformations in both directions, forward and reverse, are needed

in many types of human performances;

4. In the learning, training and retraining processes, no exams, tests or competitions

should be organized during the periods of transformations for structures, strategies,

or skills.

Transformations in Manufacturing Technologies

The main ergonomic requirement and methodological principle in the design

of productive, reliable and safe technology is mutual, multi-level adaptation

between all components of the human-machine-environment system (Venda,

1982). This requirement can be satisfied relatively easily in constant technology

conditions. But companies with constant technology, management, and social

178 VENDA & VENDA

relations, quickly become noncompetitive. Only companies with the quick abil-

ity to change technologies and products are able to survive in difficult times.

Industrial ergonomics is a science for mutual multi-level adaptation and Trans-

formation in human-machine-environment systems (HMES).

Ergonomic studies of human-machine-environment interaction in dynamic

advanced manufacturing facilities means that the features of machines (work-

stations, assembly lines, computers, telecommunications, control rooms) and

industrial environments (working space, shifts, light, noise, micro climate) are

being analyzed and designed in connection with dynamic psychological, physio-

logical, and biomechanical characteristics of human beings.

Study of mutual adaptation and transformation is of particular importance in

the implementation of new technologies, advanced manufacturing, manage-

ment, flexible team work, shifting of work places and worker’s functions, profes-

sional training and retraining. The law of transformation can be a new basis for

multi-level mutual adaptation in dynamic industrial HMES.

There is no general theory or methodology for the analysis and synthesis of the

structural dynamics of advanced manufacturing HMES in contemporary hu-

man factors and ergonomics. Therefore, analysis and design of such systems will

start almost from the beginning, with each design having its own individual

character, making such studies very expensive and time consuming and slowing

the pace of progress. This is especially visible in the time and effort spent by

ergonomists during implementations of computer integrated manufacturing

(CIM), Just In Time (JIT), Total Quality Management (TQM), Team Owned

Processes, and other innovations in dynamic manufacturing.

The principles of Mutual Adaptation and Transformation are fundamental

and generally applicable to all kinds of systems and to every component of the

manufacturing system, i.e., human beings, machines, working environment,

and to the system as a whole, in its mutual adaptation to other manufacturing,

trade, supply, communication, control, and management systems.

Transformation Dynamics and Mutual Adaptation in Ergonomics Development

The general structure and goals of future research are displayed in Fig. 7

which shows the various areas of research and development and how they relate

to each other. The main aim of that research program would be to work out the

theory, methodology and practical methods of human factors/ergonomics in

dynamic manufacturing systems. That aim can be achieved by conducting in

parallel a wide range of theoretical, experimental and applied research.

The theory of mutual adaptation in HMES would be oriented toward synthe-

TRANSFORMATION DYNAMICS 179

Theoretical Experimental Applied

Transformation Theory of

Professional Training and

Retraining

Adaptation of Humans to

New Technologies and

Workstations

Ergonomic Analysis of

Workstations and

Assembly Lines

Information Interaction

in Human Factors

Design Analysis’

Theory of Individual

Adaptation to the

User

Human Work Strategies

and Information

Structure

Ergonomic Parameters of

Mutual Adaptation

Theory of Mutual

Adaptation

Assessment and

Improvement of JIT

environment in C!M

Transformation Theores

and Quadragram Models

Dynamics of Human Work

Stratedgies

Transfer of New

Technologies to Industry

Theory, Methodology and

Practical Methods of the

Mutual Adaptation and

Transformation Dynamics

Fig. 7. A structure of research on transformation dynamics and mutual adaptation for Human-Machine-

Environment Systems.

sis of two main and traditionally separate directions in psychological and hu-

man factors/ergonomics studies: 1. Adaptation of humans to new machines,

environment, functions and tasks by using the methods of learning, education,

‘professional and physical training; 2. Adaptation of machines and technologi-

cal, educational environments to people by using the methods of ergonomic

analysis and design. Our proposed methodology of mutual adaptation helps to

combine possibilities of both previous tendencies, to make HMES more produc-

tive, effective, reliable and safe. In the books by Venda (1982, 1990) and the

paper by Yufik, Sheridan, and Venda (1992) the principle of mutual adaptation

and transformation theory are used for improving the methodology of analysis

of decision making processes, design of information display systems, and de- _

creasing of intellectual complexity of human operator functions. The theory of

individual adaptation of machines to the user describes in what way machines

need to be designed to fit various types of machine users (Lomov & Venda,

180 VENDA & VENDA

1977; Venda, 1990). The theory of mutual adaptation of humans with machines

and the environment describes how the best trade-off for adjusting the human to

the machine and environment, and reworking the machine and the environ-

ment to the human, can be derived. The theory is being implemented in human

factors design, usability testing, interface design and innovative hardware and

software solutions, for terminal and user-interface design. The Quadragram

Models of the HMES structural transformations (Venda, 1988) describe how a

human changes cognitive strategies as learning progresses, and shows how this

learning may be sped up.

The experiments on adaptation of humans to new technologies and worksta-

tions are based on the transformation theory of learning, training and retraining

connected with the analysis of human working skills, strategies of information

perception, thinking and decision making, and change of the strategies when a

new technology or workstation is implemented. Mutual adaptation of human

cognitive strategies and adequate information structures is extremely important

for optimal human-computer interaction. The influence of information on hu-

man decisions is studied at facilities and with laboratory simulation of human

performance in CIM. In addition to the previous stage of experimental studies,

an opposite direction of adaptation, the adaptation of a machine to the human

individual, will be studied as a second stage. Instead of changing human knowl-

edge, skills, and work strategies, at this stage the local optimum of human-ma-

chine system efficiency will be found experimentally by using wide changes of

machine characteristics in design, and operative adaptation of information dis-

plays, workstations, control rooms, and assembly lines.

In the studies of ergonomic parameters of mutual adaptation which relate to

HMES in dynamic manufacturing, criteria and factors of human performance

efficiency, complexity, reliability and safety are measured. The third stage of

experimental studies on mutual adaptation in human-machine systems should

include searching for the global optimum of the systems with coordinated adap-

tation of human to machine and machine to human. Mutual individual adapta-

tion in human-computer interactive systems is based on recording and com-

puter analysis of observer self reports on psychological factors in decision

making. Practical methods of searching for the factors of human operator work

complexity and efficiency were described by Venda (1982). These methods

allow study of the dynamics of human work strategies when criteria and factors

may be changed quantitatively and qualitatively, as in the transformation pe-

riods. Fast and effective implementation of new technologies in the manufac-

turing environment can be organized on the base of mutual adaptation of

workers and industrial facilities with dynamic ergonomic criteria, and factors of

HMES efficiency. Industrial ergonomic experiments on the transferring of new

TRANSFORMATION DYNAMICS | 181

technologies have been conducted at the assembly plants of the Northern Tele-

com Canada, Ltd. Ergonomic analysis, industrial design and improvement of

work stations, should facilitate human-centered processes in mutual multi-level

adaptation of HMES. Research underway at Northern Telecom concentrates on

information display at assembly lines and workstations in CIM and JIT environ-

ments. Their objective is to increase the productivity and reliability of individ-

uals, teams, and systems, working under these conditions through the use of the

principle of transformation and mutual adaptation. Ergonomic analysis and

improvement of the assembly workstation will involve the evaluation of current

methods in designing modular quickset workstations that are product specific.

The present method of manufacturing is a long flow-line of workstations.

Current problems with the straight line process are: decreased communications,

fewer opportunities to solve immediate problems and difficulties with smooth

KAN-BAN operation (Venda, Strong, Hawaleshka, & Rychlicki, 1992).

Transformation theory is very effective in prediction of dynamics and opti-

mal planning in the process of changing of old technology to the new. For

example, if a worker uses professional strategy Sn (Fig. 5) with productivity

Qn max, and then technology is changed so the human factor of work complex-

ity F increases from Fn opt to Fa opt very quickly and the worker retains strategy

Sn, productivity (quality, efficiency) will decrease to zero (right side of Fig. 5).

Hence, transformations of technologies should be synchronized with transfor-

mations of the worker’s strategies.

Transformation theory is very important also for the problem of job rotation

and prevention of repetitive strain injures. An experience at Northern Telecom

confirmed that a well defined analysis of job rotation schedules is very impor-

tant for dynamic manufacturing, where models for dynamic human-machine-

environment interaction of the Quadrigram type were worked out (Venda,

1990).

The methods of Transformation dynamics are effective for economic analysis

and planning of restructuring of technologies, management and facilities. Dur-

ing a recession and heavy competition, this is of great importance. D. Strong has

proposed (Venda, Strong, Hawaleschka, & Rychlicki, 1992) to use transforma-

tion dynamics models to compare economic features of different tactics of a

manufacturer, who can: |. Build a new plant somewhere else, in addition to

operating the old one, using new workers; and when the new plant functions

properly, terminate the workers at the old plant, and sell the old plant and land.

2. Enter a change carefully in one part of his plant, using volunteers to work in

this changed area; if the introduction is successful, introduce the change in other

areas which require the same type of improvement. 3. Introduce changes on

entire facility. Of the three approaches, the third is certainly the most humane

182 VENDA & VENDA

and attractive, but its implementation could be done successfully only on the

basis of transformation dynamics and mutual adaptation between all major

components of the facility as a HMES. There is another important ergonomic

problem, 1.e., organizing collective decision-making at every step of transforma-

tion to find the best decisions and help everybody to consider the decisions as

their own, for the most effective and synchronized implementation of the deci-

sions by every worker, engineer, and manager. This is the problem addressed by

the Hybrid Intelligence System (Venda, 1990).

Conclusions

The methodology of transformation dynamics can be effectively used in

many spheres of HMES: ergonomic analysis and design, especially in training

and re-training; mutual individual adaptation of human-computer dialogue;

and synthesis of hybrid intelligence systems for collective decision-making

under the most complicated situations. The law of transformation may be used

as a very general theoretical basis for study, prediction and improvement of

structure changes in humans, machines, and technologies.

Some of the more practical properties of the transformation dynamics pro-

cesses in advanced dynamic manufacturing human-machine-environment sys-

tems could include:

1. The longer the time that a manufacturing facility (firm, human, operator) remains

on the plateau of the Si th structure, the longer will be the time for a transforma-

tional plateau for the transition from the Si th to the Si + | th structure: Si goes to Si

= JL.

_ 2. The longer the time that a manufacturing facility (firm, human operator) remains

on the plateau of the i-th structure, the shorter will be the time for retransforma-

tional plateaus with the back transition from novel strategies Si + 1 back to the Si: Si

+ | goes to Si.

3. Learning capacity and potential for creativity (innovation) of a system, 1.e., its

mobility, is determined by its ability to execute direct (forward) and reverse (back-

ward) transitions from one distinct strategy to another, and to adopt the most

effective strategy in the mutual adaptation of the system with its dynamic environ-

ment.

4. Learning efficiency with respect to a range of rapid changes of conditions, at which

a system can perform (survive), depends on the batch of mastered strategies and on

the rates of their action during the transformational period.

5. An increase in learning time may result in a deterioration of efficiency criteria; the

learning process should not be stopped during transformational shifts.

6. Individuals’ motivation during learning depends on their personal assessment of

prospects and on the degree of mutual adaptation with the environment.

7. Prediction of transformation dynamics is especially difficult. It is much more effec-

tive if a collective decision-making system, i.e., hybrid intelligence system, is used.

8. The process of learning or professional training ought to be so planned as to exclude

TRANSFORMATION DYNAMICS 183

all kinds of examinations, tests, competitions or responsible assignments, during

transformational periods.

9. Cognitive strategies, amenable to transformation are called associated ones, and the

method of thinking predicated thereon is associative. A process of thinking, espe-

cially a creative one, is based on the transformations of thoughts, images and the

like.

10. Quantitative estimates of the fields of events, images and decision making, are the

target of studies of various transformations.

11. In many practical cases the initial direct (forward) and the subsequent reverse

(back) transformations differ in that the former are clearly of an exploratory,

searching character and are performed by the trial-and-error method.

12. Direct and reverse transformation of structures and strategies of the complex sys-

tem proceed under the same conditions and state of the system.

The law and theory of transformations are general and applicable to any system.

This last consequence of the law in application to the former USSR, that con-

- temporary reverse transformation of socialism to capitalism should be in its

main aspects as difficult and as similar to the direct transformation which oc-

curred in 1917. These transformation processes have been described in detail

(Venda, 1989).

Acknowledgments

I appreciate very much the fruitful discussions and help in editing this paper

by my coileagues at the University of Manitoba and Northern Telecom: Dr.

Doug Strong, Dr. A. B. Thornton-Trump, Prof. Ostap Hawaleshka, Brian

Rychlicki, and Joe I. Wong, Dr. Thomas B. Sheridan of MIT, Dr. Yan M. Yufik

of the Institute of Medical Cybernetics in Washington, DC, and Mr. Leo Bel-

tracchi of the U.S. Nuclear Regulatory Commission.

These studies were supported by the Natural Science and Engineering Re-

search Council of Canada (NSERC), Northern Telecom, and Bell Northern

Research.

References

1. Atkinson, R. C., & Crothers, E. J. (1964). A comparison of paired associate learning models having

different acquisition and retention axioms. Journal of Mathematical Psychology, 1, 285-315.

2. Barabash, P.S. (1982). Physiological adaptation and training of the humans to hypodynamics. In Proceed-

ings of the I. M. Sechenov Institute of the Evolutionary Physiology and Biochemistry (pp. 146-168).

Leningrad.

3. Beltracchi, L. (1984). A process/engineered safeguards iconic display. In Proceedings of the Symposium

on New Technology in Nuclear Power Plant Instrumentation and Control (pp. 241-250). Washington, DC.

4. Beltracchi, L. (1987). A model-based display. In Proceedings of the American Nuclear Society Topical

Meeting on Artificial Intelligence and Other Innovative Computer Applications in the Nuclear Industry

(pp. 68-80). Snowbird, UT.

5. Beltracchi, L. (1988). Alarm coding of a model-based display. In Proceedings of the IEEE Fourth Confer-

ence on Human Factors in the Power Plants (pp. 146-154). Monterey, CA.

6. Bower, G. H. (1961). Application of a model to paired-associate learning. Psychometrika, 26, 255-280.

184

20.

21.

22.

23.

24.

25.

26.

41)

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

VENDA & VENDA

. Bush, R. R., & Mosteller, F. (1955). Stochastic models for learning. New Y ork: Wiley.

. Bryan, W. L., & Harter, N. (1899). Studies on the telegraphic language: The acquisition of a hierarchy of

habits. Psychological Review, 6, 345-375.

. Chapanis, A. (1959). Research techniques in human engineering. Baltimore, MD: The Johns Hopkins

University Press.

. Ebbinghaus, H. (1885). Uber das Gedéchtnis: Untersuchungen zur experimentellen Psychologie. Leipzig:

Duncker und Humblot.

. Estes, W. K. (1950). Toward a statistical theory of learning. Psychological Review, 57, 94-107.

. Estes, W. K. (1959). Component and pattern models with Markovian interpretations. In R. R. Bush &

W. K. Estes (Eds.), Studies in mathematical learning theory (pp. 26-53). Stanford, CA: Stanford Univer-

sity Press.

. Fitts, P. M. (1954). The information capacity of the human motor system in controlling the amplitude of

movement. Journal of Experimental Psychology, 47, 381-391.

. Freivalds, A. (1987). The ergonomics of tools. In Jnternational Review of Ergonomics, vol I (pp. 12-48).

London: Taylor & Francis.

. Greeno, J. G. (1967). Paired-associate learning with short-term retention: Mathematical analysis and data

regarding identification of parameters. Journal of Mathematical Psychology, 4, 430-472.

. Hasset, J., & White, K. M. (1989). Psychology in prospective. New York: Harper and Row.

. Hull, C. L. (1943). Principles of behavior: An introduction to behavior theory. New York: Appleton-Cen-

tury-Crofts.

. Konz, S. A. (1990). Work design: Industrial ergonomics. Worthington, OH: Publishing Horizons.

. Lomoy, B. F., & Venda, V. F. (1977). Human factors: Problems of adapting systems for the interaction of

information to the individual: The theory of hybrid intelligence. In A. S. Neal & R. F. Palesek (Eds.),

Proceedings of the Human Factors Society 21st annual meeting (pp. 1-9). Santa Monica, CA: Human

Factors Society.

Rasmussen, J. (1986). Information processing and human-machine interaction: An approach to cognitive

engineering. New York: North-Holland.

Rasmussen, J., & Vicente, K. (1987). Cognitive control of human activities and errors: Implications for

ecological interface design. In Proceedings of the Fourth International Conference on Event Perception and

Action (pp. 120-148). Trieste, Italy.

Sheridan, T. B. (1992). Telerobotics: Automation and human supervisory control. Cambridge, MA: MIT

Press.

Thorndike, E. L. (1898). Animal intelligence: An experimental study of the associative process in animals.

Psychological Monographs, 2, No. 8.

Thornton-Trump, A. B., & Suzuki, K. (1991). Fourier analysis of reaction force data (Res. Rep). Winnipeg,

Canada: The University of Manitoba.

Venda, V. F. (1982). Engineering psychology and synthesis of information (2nd ed.). Moscow: Mashino-

stroenie.

Venda, V. F. (1986). On transformation learning theory. Behavioral Science, 31(1), 1-11.

Venda, V. F. (1988). The quadragrams of mutual adaptation as a new model of human activity. In

Proceedings of the Xth Congress of the International Ergonomics Association (pp. 462-470). Sydney,

Australia.

Venda, V. F. (1989). The waves of progress. Moscow: Znanie Publishers.

Venda, V. F. (1990). Hybrid intelligence systems: Evolution, psychology, and ergonomics, Moscow:

Mashinostroenie.

Venda, V. F., Strong, D., Hawaleshka, O., & Rychlicki, B. (1992). Human factors and transformations of

manufacturing technologies. In Advances in industrial ergonomics and safety-IV (pp. 93-99). London:

Taylor & Francis.

Venda, V. F., & Thornton-Trump, A. B. (1992). Applications of transformation theory in biomechanics.

In Advances in industrial ergonomics and safety-IV (pp. 87-92). London: Taylor & Francis.

Venda, Y. V. (1989-1991). The law of transformations and its consequences. Unpublished manuscripts,

Moscow State University, Russia.

Venda, Y., & Venda, V. F. (1992). An introduction to transformation dynamics: The law and theory of

transformations. In Advances in industrial ergonomics and safety-IV (pp. 79-86). London: Taylor &

Francis. :

Woodworth, R. S. (1938). Experimental psychology. New York: Holt.

Yanch, E. (1974). Forecasting of scientific and technological progress. Moscow: Progress Publishers.

Yufik, Y. M., Sheridan, T. B., & Venda, V. F. (1992). Knowledge measurement, cognitve complexity and

cybernetics of mutual human-machine adaptation. In C. V. Negoita (Ed.), Cybernetics and applied sys-

tems (pp. 187-238). New York: Marcell Dekker.

Zarakovski, G. M., Korolyov, B. M., Medvedev, V.I., & Shlaen, P. Y. (1977). Introduction to ergonomics.

Moscow: Sovetskoe Radio Publishers.

Journal of the Washington Academy of Sciences,

Volume 81, Number 4, Pages 185-188, December 1991

The 1991 Washington Academy

of Sciences Awards Program

for Scientific Achievement

C. R. Creveling

National Institute of Diabetes, Digestive, and Kidney Diseases

Bethesda, MD

The Washington Academy of Sciences was founded in 1898 as an afhliation

of scientific societies under the sponsorship of the Washington Philosophical

Society to conduct, endow, and assist investigation in any department of

science. At present the Academy is afhliated with 52 scientific societies. In

keeping with the purposes of the Academy, each year, the Committee on

Awards for Scientific Achievement accepts nominations and recognizes scien-

tists and science teachers in the Washington metropolitan area who have made

outstanding contributions to science that are of merit and distinction. Awards

are made for outstanding contributions in the Mathematical and Computer

Sciences, the Behavioral and Social Sciences, the Engineering Sciences, the Bio-

logical Sciences, and the Physical Sciences. Further, in keeping with the goals of

the Academy which include the promotion of excellence in the teaching of

science, the Academy also presents awards for the Teaching of Science. These

awards include the Leo Schubert Award for excellence in the teaching of science

- in college and the Bernice Lamberton Award for excellence in teaching science

in high school. ;

Persons selected for recognition are chosen by panels of experts in each of

the fields. Nominations are made by Academy members or the public.

In 1991, the Awards were presented at a gala reception held at the Mary

Woodward Lasker Center for Health Research and Education, on the grounds

of the National Institutes of Health, in Bethesda Maryland, on Thursday,

April 18th.

185

186 CREVELING

The 1991 awardees were:

Dr. Harold Liebowitz Distinguished Career in Science

Dr. Robert E. Fay, III Mathematics and Computer Sciences

Dr. Andrew F. Brimmer Behavioral and Social Sciences

Dr. David E. Ramaker Physical Sciences

Dr. Robert J. Lutz Engineering Sciences

Dr. Miles Herkenham Biological Sciences

Prof. Glen E. Gordon Teaching of Science in College

Distinguished Career in Science

The award was granted to DR. HAROLD LIEBOWITZ, Dean Emeritus of

the School of Engineering and Applied Science at The George Washington

University and the L. Stanley Crane Professor of Engineering. Dr. Liebowitz led

the School of Engineering and Applied Science with grace and distinction for

over 20 years. Under his leadership the school entered into a period of unprece-

dented growth in both qualitative and quantitative aspects as reflected in the

number of students, faculty, and in research accomplishments. Dr. Liebowitz

actively promoted a very successful partnership between the School and the

National Aeronautics and Space Administration as exemplified by the Joint

Institute for the Advancement of Flight Sciences. In addition Dr. Liebowitz was

selected for his scholarly research achievements in fracture mechanics. The

Academy recognized Dr. Liebowitz as scholar, engineer, author, educator, and

consultant to industry, consultant to the U.S. government and to foreign govern-

ments. Dr. Liebowitz was nominated by Prof. Walter K. Kahn, and selected by

the Awards Chair, Dr. C. R. Creveling.

Mathematics and Computer Sciences

The award in the Mathematical and Computer Sciences was granted to DR.

ROBERT E. FAY, III of the Director’s Office of the U.S. Bureau of the Census,

for his outstanding contributions to the development of major methodological

improvements in survey statistics. Dr. Fay is recognized both nationally and

internationally as an insightful innovator in mathematics. His creative applica-

tions of sound statistical theory led to important applications in survey sample

design, nonsampling error, small area estimation and contingency table analy-

sis. Dr. Fay was nominated by Dr. Edward J. Wegman, President of the Wash-

ington Statistical Society, and selected by the Mathematics and Computer

Science Committee under the leadership of Dr. Abolghassem Ghaffari.

WAS AWARDS PROGRAM 187

Behavioral and Social Sciences

The award in the Behavioral and Social Sciences was granted to DR. AN-

DREW F. BRIMMER, for his contributions in many areas of in general eco-

nomics, in money, banking, and monetary policy, in international finance and

balance of payments and especially for his contributions to the economic devel-

opment of the black community. For his technical and scientific scholarship in

interdependent macroeconomic and social policy. As a result of his contribu-

tions, Afroamericans are more successfully integrated into American society,

the international financial market is more integrated, and systemic risks in

capital markets more fully accounted for. Dr. Brimmer’s contributions have led

to the construction of a fairer and more efhcient American society in a global

world. Dr. Brimmer was nominated by Dr. Robert H. Aten and selected by the

Committee on Behavioral and Social Sciences under the direction of Dr. Cora-

' lee Farlee.

Physical Sciences

DR. DAVID E. RAMAKER received the award in the physical sciences for

his many and important theoretical contributions to quantitatively understand-

ing the role of many-body electron phenomena in electron spectroscopies (X-

ray, photoelectron, X-ray adsorption) and in stimulated desorption (electron

and photon desorption). These contributions has played a major role in the

rapid progress of surface science and its technological applications. In particular,

he showed that significant new electronic structure information can be obtained

from interpretation and understanding of the very complex ““many-body” states

mapped in the experimental electron spectral line shapes, and that this informa-

tion could be obtained in a relatively straight-forward and simple manner. Fur-

thermore, he has shown that these very complex ““many-body” states are the

primary actors in the electron and photon desorption process, and that Auger-

electron spectroscopy can be used to map the systems, including gas phase

hydrocarbons, condensed molecular gases, solids (silicon, silicon dioxide, dia-

mond, graphite, carbides, and the high temperature superconductors) and

atomic and molecular adsorbates on surface (chemisorbed ethylene and carbi-

dic carbon). Dr. Ramaker was nominated by Dr. James S. Murday of the Naval

Research Laboratory, and selected by the Committee on the Physical Sciences

chaired by Dr. Richard K. Cook.

Engineering Sciences

The award in the Engineering Sciences was granted to DR. ROBERT J.

LUTZ, of the National Institutes of Health, for his creative application of engi-

188 CREVELING

neering science and practice in biomedical research and his scholarly contribu-

tions to the study of fluid mechanics. Dr. Lutz made significant contributions to

the development of practical vascular models in persons and in the development

of physiological and pharmacokinetic models of drug and toxin distribution.

Dr. Lutz was nominated by Dr. Robert L. Dedrick, and selected by the Commit-

tee on Engineering Sciences under the direction of Marianne P. Vaishnav.

Biological Sciences

The award in the Biological Sciences was granted to DR. MILES HERKEN-

HAM, Chief of the Section on Functional Neuroanatomy, National Institute of

Mental Health, for his pioneering development of high resolution, autoradio-

graphic techniques for the localization of receptors in the central nervous sys-

tem. Dr. Herkenham has made major contributions toward understanding the

nature of affective disorders in man and toward an understanding of the mecha-

nisms of action in the brain of therapeutic agents and drugs of abuse. Dr. Her-

kenham was nominated by Dr. Kenner C. Rice of the Laboratory of Medicinal

Chemistry, National Institute of Diabetes and Digestive and Kidney Diseases,

and was selected by the Committee on Biological Sciences under the direction of

Dr. C. R. Creveling.

Leo Schubert Award for Teaching of Science in College

The Leo Schubert award for Teaching of Science in College was granted to

PROFESSOR GLEN E. GORDON of the University of Maryland, for his devel-

opment and dynamic teaching of environmental chemistry. Professor Gordon

has provided both science and non-science students with a knowledgable basis

for making decisions on questions ranging from nuclear weapons to fuel econ-

omy in cars. He has provided science students a basis for making critical evalua-

tions of environmental and economic factors in our society and the world and

an appreciation of the risks to which people may be subjected. Professor Gordon

was nominated by Dr. Alice C. Mignerey, and selected by the Leo Schubert

Award Committee under the direction of Marylin F. Krupsaw.

Bernice Lamberton Award for Teaching of Science in High School

No nominations were received for this award.

After a reception for the awardees, a lecture was delivered by the winner of the

Award in the Biological Sciences, Dr. Miles Herkenham, entitled Understand-

ing drug and neurotransmitter actions in the brain.

Journal of the Washington Academy of Sciences,

Volume 81, Number 4, Pages 189-202, December 1991

The Bylaws of the Washington Academy

of Sciences’

ARTICLE I. OBJECTIVES

Section 1. The objectives of the Washington Academy of Sciences (hereinafter

called the Academy) shall be: (a) to stimulate interest in the sciences, both pure

and applied; and (b) to promote their advancement and the development of

their philosophical aspects by the Academy membership and through coopera-

tive action by the Affiliated Societies.

Section 2. These objectives may be attained by, but are not limited to: (a)

publishing a periodical, occasional scientific monographs and such other books

or pamphlets as may be deemed desirable; (b) conducting public lectures of

broad scope and interest in the fields of science; (c) sponsoring a Washington

Junior Academy of Sciences (WJAS); (d) promoting science education and a

professional interest in science among people of high school and college age; (e)

accepting or making grants of funds to aid special research projects; (f ) conven-

ing symposia, both formal and informal, on any aspects of science; (g) calling

scientific conferences; (h) organizing or assisting in scientific assemblies or bod-

ies; (1) cooperating with other academies and scientific organizations; (j) award-

ing prizes and citations for special merit in science; (k) maintaining an office and

staff to aid in carrying out the objectives of the Academy.

ARTICLE Il. MEMBERSHIP

- Section 1. The Academy shall be comprised of individuals and A ffiliated Societ-

ies. Throughout this document when reference is made to individuals, the use of

“‘he” shall be interpreted as “‘he or she.”

' The revised Bylaws of the Washington Academy of Sciences dated May 1982 were replaced by a March

1984 edition. The 1984 version was found to have many imperfections. The effort to correct resulted in a April

1, 1988 version, followed quickly by a April 29, 1988 version which took away the vote of representatives of

Afhliated Societies. The subsequent May 1989 version returned the vote of the affiliates but other problems

came to the forefront. A proposed version dated May 24, 1990 was mailed to the membership for consider-

ation with a cut-off date of August 9, 1990. The revisions were approved by majority vote of the Membership.

189

190 WASHINGTON ACADEMY OF SCIENCES

Section 2. Members shall be individuals who are interested in and will support

the objectives of the Academy and who are otherwise acceptable to at least

two-thirds of the Committee on Membership. A letter or application form re-

questing membership and signed by the applicant may suffice for action by

Committee; approval by the Committee constitutes election to membership.

Section 3. Fellows shall be individuals who by reason of original research or

other outstanding service to the sciences, mathematics, or engineering are

deemed worthy of the honor of election to Academy fellowship.

Section 3(a). Nominations of fellows shall be presented to the Committee on

Membership on a form approved by the Committee. The form shall be signed by

the sponsor (a Fellow who has knowledge of the nominee’s field), and shall be

endorsed by at least one other Fellow. An explanatory letter from the sponsor

and supporting material shall accompany the completed nomination form.

Section 3(b). Election to fellowship shall be by vote of the Board of Managers

upon recommendation of the Committee on Membership. Final action on nom-

inations shall be deferred at least one week after presentation to the Board of

Managers, where two-thirds of the vote cast shall be necessary to elect. The

election process shall be completed upon submission of the processed nomina-

tion forms to the Vice President for Membership A ffairs.

Section 3(c). Each individual (not already a Fellow) who has been chosen to be

the recipient of an Academy Award for Scientific Achievement shall be consid-

ered nominated for immediate election of fellowship. The election process shall

be completed upon submission of the standard nomination forms to the Vice

President for Membership Affairs, thus obviating the procedures of Sections 3(a)

and 3(b) of this Article.

Section 3(d). Any fellow of the Academy may recommend in wniting that an

individual of national eminence be considered for immediate election to fellow-

ship by the Board of Managers, without the necessity of compliance with the

procedures of Sections 3(a) and 3(b) of this Article. Following approval by the

Board of Managers, the election process shall be completed upon submission of

the standard nomination forms to the Vice President for Membership Affairs.

Section 4. Patrons. Members or fellows who have given to the Academy not less

than one thousand dollars, or its equivalent in property or tangible assets, shall

be eligible for election by the Board of Managers as Patrons of the Academy (for

life). Following approval by the Board of Managers, the election process shall be

completed when suitable documentation has been submitted to the Vice Presi-

dent for Membership Affairs.

ACADEMY BYLAWS 191

Section 5. Life Members or Life Fellows shall be those individuals who have

made a single payment in accordance with Article II, Section 11(a) in lieu of

annual dues.

Section 6. Members or fellows in good standing who are retired and are no

longer engaged in regular gainful employment may be placed in emeritus status.

Individuals in emeritus status shall be designated Emeritus Member or Emeritus

Fellow as appropriate. Upon request to the Vice President for Membership

Affairs for transfer to this status, they shall be relieved of further payment of

dues, beginning with the following January first; shall retain rights to hold office

and attend meetings; shall receive notices of meetings without charge; and at

their request, shall be entitled to receive the Academy periodical at cost. This

transfer shall be completed when the Treasurer and the Vice President for Ad-

-munistrative Affairs have been so notified.

Section 7. Members or fellows living more than 50 miles from the White

House, Washington, DC shall be classed as Nonresident Members or Nonresi-

dent Fellows.

Section 8. An election to any dues-paying class of membership shall be void if

the candidate does not within three months thereafter pay his dues or satisfacto-

rily explain his failure to do so.

Section 9. Former members or fellows who resigned in good standing may be

reinstated upon application to the Vice President for Membership Affairs for

approval by the Board of Managers. No reconsideration of the applicant’s quali-

fications need be made by the Membership Committee in these cases.

Section 10. Affiliated Societies. Bona fide scientific societies may apply for

affiliation with the Academy by furnishing to the Secretary of the Academy an

- outline of their objectives, organizational structure and requirements for mem-

bership in their society. The Secretary will transmit the application to the Policy

and Planning Committee for review and recommendation for action by the

Board of Managers.

Section 10(a). Each Affiliated Society shall select one of its members who is also

a member or fellow of the Academy to serve as its representative to the Board of

Managers. The representative shall serve until replaced by his society.

Section 10(b). Each Affiliated Society shall cooperate with the Academy in

sponsoring joint meetings of general scientific interest.

192 WASHINGTON ACADEMY OF SCIENCES

Section 11. Dues. Annual dues for each member class shall be fixed by the

Board of Managers. No dues shall be paid by Emeritus members, Emeritus

fellows, Life members, Life fellows, Patrons, or Affiliated Societies.

Section 11(a). Members and fellows in good standing may be relieved of further

payment of dues by making a single payment that has a value equal to ten years

of dues current at the time of payment. (see Article II, Section 5) Such persons

are to be identified as Life Members or Life Fellows as appropriate. Income from

this source shall be identified as the Life Membership Endowment Fund

(LMEF) and shall be monitored in perpetuity by three 7rustees who are resident

Life Members or resident Life Fellows. The Trustees shall direct the investment

of the Fund (LMEF) in a conservative action and turn over to the Treasurer all

interest from such investments. Trustees shall serve for the duration of life or

until the change to nonresident status or the onset of permanent disability or

resignation.

Section 11(b). Individuals whose dues are 1n arrears for one year (counting from

the ““dues payable date” on the latest dues payment bill) shall neither be entitled

to receive Academy publications nor to vote in Academy elections.

Section I11(c). Individuals whose dues are in arrears for twenty-four (24) months

(counting from the “dues payable date” on the latest dues payment bill) shall be

dropped from the rolls of the Academy, upon notice to the Board of Managers,

unless otherwise directed. Those who have been dropped from membership for

nonpayment of dues may be reinstated upon approval of the Board of Managers

and upon payment of back dues for two years together with dues for the-year of

reinstatement.

ARTICLE YW. ELECTED OFFICERS and BOARD MEMBERS

Section 1. Officers of the Washington Academy of Sciences shall be President,

President-Elect, Vice President for Administrative Affairs, Vice President for

Membership Affairs, Vice President for Affiliate Affairs, Vice President for

WJAS Affairs, Secretary, and Treasurer. All shall be chosen from resident fel-

lows of the Academy.

Section 2. The President shall appoint all committees and such nonelective

officers as are needed unless otherwise directed by the Board of Managers or

provided in the bylaws. He (or his substitute: the President-Elect, the Vice Presi-

dent for Administrative Affairs, the Vice President for Membership Affairs, the

Vice President for Affiliate Affairs, the Vice President for WJAS Affairs, the

ACADEMY BYLAWS 193

Secretary, or the Treasurer, in that order) shall preside at all meetings of the

Academy, the Board of Managers and the Executive Committee.

Section 3. The President-Elect shall succeed to the office of President following

one term as President-Elect. The President-Elect shall serve as Chair of the

Program Planning Committee to arrange speakers and meeting places for the

following year (the year in which the President-Elect succeeds to President).

Other duties may be assigned by the Board of Managers.

Section 4. The Vice President for Membership Affairs shall have general respon-

sibilities for committees related to membership: the Membership Committee,

the Membership Promotion Committee, and the Committee on Awards for

Scientific Achievement. Other duties may be assigned by the Board of Man-

agers.

Section 5. The Vice President for Administrative Affairs shall have general re-

sponsibility for operation of the Business Office of the Academy and the Journal

of the Washington Academy of Sciences, and such other duties as assigned by

the Board of Managers. He shall oversee the activities of the Editorial Advisory

Committee, the Ways and Means Committee, and the Office Manager.

Section 6. The Vice President for Affiliate Affairs shall serve as Chair of the

Afhliated Society Representatives. He shall carry out such other duties as as-

signed by the Board of Managers.

Section 7. The Vice President for WJAS Affairs shall have general responsibility

for the committees relating to organizing and maintaining the Junior Academy

(WJAS). He shall interface with the Joint Board on Science and Engineering

Education, and shall carry out such other duties as assigned by the Board of

Managers.

Section 8. The Secretary shall record and distribute the minutes of the meetings

_ of the Board of Managers and such other meetings as the Board of Managers

may direct. He shall conduct all correspondence relating thereto, except as

otherwise provided, and shall be the custodian of the Corporate Seal of the

Academy.

Section 9. In cooperation with the Vice Presidents for the functional areas

described in Sections 4, 5, 6, and 7, above, the Treasurer shall be responsible for

preparing the Budget of the Academy and submitting it to the Board of Man-

agers for approval. The Treasurer shall also be responsible for distributing to the

Board of Managers in a timely manner records of funds received and expended.

The Treasurer shall be responsible for maintaining records of funds deposited in

banks or other savings instruments. The Treasurer and/or other designated

194 WASHINGTON ACADEMY OF SCIENCES

persons shall sign checks for disbursements of funds as directed by the Board of

Managers. The Treasurer shall prepare annual reports as required by the Inter-

nal Revenue Service, the U.S. Postal Service, etc. He shall deposit and disburse

funds of the Washington Junior Academy of Sciences.

Section 10. The President and the Treasurer, as directed by the Board of Man-

agers, shall jointly assign securities belonging to the Academy and endorse finan-

cial and legal papers necessary for the uses of the Academy, except those relating

to current expenditures authorized by the Board of Managers and those under

cognizance of the Life Membership Endowment Fund Trustees. In case of dis-

ability or absence of the President or Treasurer, the Board of Managers may

designate the President-Elect or another elected officer as Acting President and/

or another elected officer of the Academy as Acting Treasurer, who shall per-

form the duties of these offices during such disability or absence.

Section 11. Two members or fellows of the Academy shall be elected each year

to serve a three-year term as Members of the Board of Managers. To initiate

staggered terms or to fill vacancies, additional Members of the Board of Man-

agers may be selected in the annual election.

Section 12. The newly elected officers and Members of the Board of Managers

shall take office at the close of the annual meeting, when the President-Elect of

the previous year becomes President.

ARTICLE IV. APPOINTED OFFICERS

Section 1. An Office Manager shall be appointed by the Board of Managers.

The Office Manager shall be responsible for the routine business operation of

the Academy. The Board of Managers shall determine the type of business

activity (volunteer workers or contract workers) and the amount of funds to be

allocated to the business office.

Section 2. An Editor for the Journal of the Washington Academy of Sciences

shall be appointed by the Board of Managers. The Editor shall be responsible to

the Vice President for Administrative Affairs for administrative policy and re-

lated activities.

Section 3. An Archivist may be appointed by the President. If appointed he shall

maintain the permanent records of the Academy, including important records

which are no longer in current use by the Secretary, Treasurer or other officer,

and such other documents and material as the Board of Managers may direct.

ACADEMY BYLAWS 195

ARTICLE V. BOARD OF MANAGERS

Section 1. The activities of the Academy shall be guided by the Board of Man-

agers, consisting of the President, the President-Elect, the immediate Past Presi-

dent, the four Vice Presidents, the Secretary, Treasurer, the six Members of the

Board of Managers, and one representative from each of the Afhliated Societies.

The elected officers of the Academy shall hold like offices on the Board of

Managers.

Section 2. The Board of Managers shall set the dues for individual members and

the minimum contribution for Patrons. For prolonged, diligent and well-docu-

mented service in the administrative work of the Academy the Board of Man-

agers may recognize such service of a member or fellow by citation including

‘dues paid for Life.

Section 3. The Board of Managers shall transact all business of the Academy

not otherwise provided for. A quorum of the Board shall be one third of the

membership of the Board of Managers. To be eligible to vote the officer or

member of the Board of Managers must be in good standing, casting one vote

only regardless of the number of offices or Afhliated Societies that he may

represent.

Section 4. The Board of Managers may provide for such standing and special

committees as it deems necessary.

Section 5. The Board of Managers shall have power to fill all vacancies in its

elected membership until the next election. This does not apply to the offices of

the President and Treasurer or to representatives of Afhliated Societies.

ARTICLE VI. COMMITTEES

Section 1. An Executive Committee shall have cognizance of Academy finances

by reviewing the Treasurer’s monthly reports of budgeted expenses and antici-

pated income, and by reviewing the status of several internal accounts: the Life

Membership Endowment Fund, the I.R.S. Form 990 accounts, the Postal Ac-

counts, the WJAS Account, etc.

Section 2. The Executive Committee shall meet one-half hour prior to the

scheduled meeting of the Board of Managers to anticipate and obviate budge-

tary imbalances. The results of such actions shall be reported to the Board of

Managers following the Treasurer’s report.

196 WASHINGTON ACADEMY OF SCIENCES

Section 3. The Executive Committee shall consist of the incumbent elected

officers of the Board of Managers plus two non-elected members of the Board of

Managers chosen by the Board of Managers.

Section 4. Committees under the cognizance of the President are the Executive

Committee, the Nominating Committee, the Policy and Planning Committee,

the Audit Committee, and such other committees as shall be determined by the

Board of Managers.

Section 5. Committees under the cognizance of the President-Elect are the

Program Planning Committee and such other committees as shall be deter-

mined by the Board of Managers.

Section 6. Committees under the cognizance of the Vice President for Member-

ship Affairs are the Membership Committee, the Membership Promotion Com-

mittee, the Committee on Awards for Scientific Achievement, and such other

committees as shall be determined by the Board of Managers.

Section 7. Committees under the cognizance of the Vice President for Admin-

istrative Affairs are the Editorial Advisory Committee, the Ways and Means

Committee, and such other committees as shall be determined by the Board of

Managers.

Section 8. Committees under the cognizance of the Vice President for WJAS

Affairs are the Committee on the Encouragement of Science Talent, Committee

on Grants-in-Aid for Scientific Research, and such other committees as shall be

determined by the Board of Managers.

Section 9. The President shall appoint from the Academy membership such

committees as are authorized by the Board of Managers and such special com-

mittees as necessary to carry out its functions. Committee appointments shall be

staggered as to term whenever it is determined by the Board of Managers to be in

the interest of continuity of committee affairs.

Section 10. The President, with the approval of the Board of Managers, shall

appoint a Nominating Committee of six fellows of the Academy, (see Article VI,

Section 4) at least one of whom shall be a Past-President of the Academy, and at

least three of whom shall have served as representatives of Afhliated Societies for

at least one year.

Section 11. The Nominating Committee shall prepare a slate listing one or

more persons for each of the offices of President-Elect, the four Vice Presidents,

Secretary, Treasurer, and four or more persons for the two Members of the

Board of Managers whose terms expire after three years and at least two persons

ACADEMY BYLAWS 197

for each vacant unexpired term of such position (see Article III, Section 11). The

slate shall be presented for approval at the meeting in December. Not later than

December 15, the Vice President for Administrative Affairs shall forward to

each Academy member and fellow an announcement of the election, the Com-

mittee’s nomination for the offices to be filled, and a list of incumbents. Addi-

tional candidates for such offices may be proposed by any member or fellow in

good standing by letter received by the Vice President for Administrative Affairs

not later than January 3. The letter shall include the concurrence of the nomi-

nees and the names of 25 members or fellows making the proposal. Upon

verification by the nominating committee the names shall be entered on the

ballot. The Board shall remind members and fellows of the foregoing option

with the distribution of the preliminary slate.

Section 12. Not later than February 15, the Vice President for Administrative

Affairs shall prepare and mail ballots to members and fellows. Independent

nominations shall be included on the ballot, and the names of the nominees

shall be arranged in alphabetical order. When more than two candidates are

nominated for the same office, the voting shall be by preferential ballot in a

manner prescribed by the Board of Managers. The ballot shall contain a notice

to the effect that votes not received by the Vice President for Administrative

Affairs before the first Thursday of March, and votes of individuals whose dues

are in arrears for one year or more, will not be counted. The Committee of

Tellers shall count the votes and report the results at the April Meeting of the

Board of Managers.

ARTICLE VI. MEETINGS OF THE ACADEMY

- Section 1. The annual meeting of the Academy shall be held each year in May.

It shall be held on the third Thursday of the month unless otherwise directed by

the Board of Managers. At this meeting, the reports of the President-Elect and

the several Vice Presidents, the Secretary, the Treasurer, and the Committee of

Tellers shall be presented.

Section 2. Meetings of the Board of Managers shall be held as called by the

President, or in his absence by the Secretary, or within ten days after a written

request by six members of the Board of Managers has been sent to all members

of the Board of Managers. Regular meetings of the Board of Managers shall be

set preferably for a fixed place, hour, day of week, and sequence of months

excepting July and August.

198 WASHINGTON ACADEMY OF SCIENCES

Section 3. Other meetings may be held at such time and place as the Board of

Managers may determine.

Section 4. The rules contained in “Robert’s Rules of Order Revised”’ shall

govern the Academy in all cases to which they are applicable, and in which they

are not inconsistent with the bylaws or special rules of order of the Academy.

ARTICLE VUI. REMOVAL FROM OFFICE

Section 1. Members of the Board of Managers and the Executive Committee

shall assure that all business of the Academy is conducted in the highest spirit of

ethics and integrity. This includes the absence of a conflict of interest, which is

defined as the acceptance of positions or contracts with the Academy which

would result or give the appearance of resulting in a profit or other material

advantage to an officer of the Academy. NOTE: Article V, Section 2 is consid-

ered a service citation by the Academy and as such is an exception.

Section 2. If any member of the Board of Managers or the Executive Commit-

tee is found by a vote of two-thirds of the Board of Managers to have violated the

spirit of ethics and integrity or the conflict of interest requirements, he or she

shall be removed from office.

Section 3. The position vacated by such removal shall be filled temporarily by

appointment by the Board of Managers until the next scheduled election or

regular appointment to the affected position.

Section 4. When for approved Academy obligations, circumstances necessitate

payment by persons other than the Academy officers who sign checks, reim-

bursement to such persons shall be made only when appropriate documentation

is submitted to the Treasurer of the Academy.

ARTICLE IX. COOPERATION

Section 1. The term “Affiliated Societies” in their order of seniority (see Article

II, Section 10) shall be held to cover the:

Philosophical Society of Washington;

Anthropological Society of Washington;

Biological Society of Washington;

Chemical Society of Washington;

Entomological Society of Washington;

National Geographic Society;

ACADEMY BYLAWS 199

Geological Society of Washington;

Medical Society of the District of Columbia;

Columbia Historical Society;

Botanical Society of Washington;

Society of American Foresters, Washington Section;

Washington Society of Engineers;

Institute of Electrical and Electronics Engineers, Washington Section;

American Society of Mechanical Engineers, Washington Section;

Helminthological Society of Washington;

American Society for Microbiology, Washington Branch;

Society of American Military Engineers, Washington Post;

American Society of Civil Engineers, National Capital Section;

_ Society for Experimental Biology and Medicine, District of Columbia Section;

American Society for Metals, Washington Chapter;

American Association for Dental Research, Washington Section;

American Institute of Aeronautics and Astronautics, National Capital Section;

American Meteorological Society, District of Columbia Chapter;

Insecticide Society of Washington, now Pest Science Society of Washington;

Acoustical Society of America, Washington Chapter;

American Nuclear Society, Washington Section;

Institute of Food Technologists, Washington Section;

American Ceramic Society, Baltimore-Washington Section;

Electrochemical Society, National Capital Section;

Washington History of Science Club;

American Association of Physics Teachers, Chesapeake Section;

Optical Society of America, National Capital Section;

American Society of Plant Physiologists, Washington Area Section;

Washington Operations Research Council, now Washington Operations Re-

search and Management Science Council;

Instrument Society of America, Washington Section;

American Institute of Mining; Metallurgical, and Petroleum Engineers, Wash-

ington Section;

National Capital Astronomers;

Maryland-District of Columbia-Virginia Section of the Mathematical Associa-

tion of America;

District of Columbia Institute of Chemists;

District of Columbia Psychological Association;

Washington Paint Technical Group;

American Phytopathological Society, Potomac Division;

Society for General Systems Research, Metropolitan Washington Chapter;

200 WASHINGTON ACADEMY OF SCIENCES

Human Factors Society, Potomac Chapter;

American Fisheries Society, Potomac Chapter;

Association for Science, Technology and Innovation;

Eastern Sociological Society;

Institute of Electrical and Electronics Engineers, Northern Virginia Chapter;

Association for Computing Machinery, Washington Chapter:

Washington Statistical Society;

Institute of Industrial Engineers;

Society of Manufacturing Engineers;

and such others as may be hereafter recommended by the Board of Managers

and elected by two-thirds of the members of the Academy voting, the vote being

taken by correspondence. A society may be released from afhliation on recom-

mendation of the Board of Managers, and the concurrence of two-thirds of the

members of the Academy voting.

Section 2. The Academy may assist the afhliated scientific societies of Washing-

ton in any matter of common interest, as in joint meetings, or in the publication

of a joint directory; provided it shall not have power to incur for or in the name

of one or more of these societies any expense or liability not previously autho-

rized by said society and societies, nor shall it without action of the Board of

Managers be responsible for any expenses incurred by one or more of the Affili-

ated Societies.

Section 3. No Affiliated Society shall be committed by the Academy to any

action in conflict with the charter, constitution, or bylaws, of said society, or its

parent society.

Section 4. The Academy may establish and assist a Washington Junior Acad-

emy of Sciences for the encouragement of interest in science among students in

the Washington area of high school and college age.

ARTICLE X. AWARDS AND GRANTS-IN-AID

Section 1. The Academy may award medals and prizes or otherwise express its

recognition and commendation of scientific work of high merit and distinction

in the Washington area. Such recognition shall be given only on approval by the

Board of Managers of a recommendation by the Committee on Awards for

Scientific Achievement. |

Section 2. The Academy may receive or make grants to aid scientific research in

the Washington area. Grants shall be received or made only on approval by the

ACADEMY BYLAWS 201

Board of Managers of a recommendation by the Committee on Grants-in-Aid

for Scientific Research.

ARTICLE XI. AMENDMENTS

Section 1. Amendments to these bylaws shall be proposed by the Board of

Managers and submitted to the members of the Academy in the form of a mail

ballot accompanied by a statement of the reasons for the proposed amendment.

A two-thirds majority of those members voting is required for adoption. At least

two weeks shall be allowed for the ballots to be returned.

Section 2. Any Afhliated Society or any group of ten or more members may

' propose an amendment to the Board of Managers in writing. The action of the

Board of Managers in accepting or rejecting this proposal to amend the bylaws

shall be by a vote on roll call, and the complete roll call shall be entered in the

minutes of the meeting.

ARTICLE XII. DISTRIBUTION OF FUNDS ON DISSOLUTION

In the event of a liquidation, dissolution, termination or winding up of the

Washington Academy of Sciences (whether voluntary, involuntary, or by oper-

ation of law) the total assets of the Washington Academy of Sciences shall be

distributed by the Board of Managers, provided that none of the property or

assets of the Washington Academy of Sciences shall be made available in any

way to any individual, corporation or other organization, except to one or more

corporations, or other organizations which qualify as exempt from federal in-

come tax under Section 501(c)(3) of the U.S. Internal Revenue Code of 1954, as

may be from time to time amended.

ARTICLE XUI. PURPOSE

The Washington Academy of Sciences is organized exclusively for charitable,

educational, and scientific purposes, including, for such purposes, the making of

distributions to organizations that qualify as exempt organizations under Sec-

tion 501(c)(3) of the U.S. Internal Revenue Code (or the corresponding provi-

sion of any future United States Internal Revenue Law.).

202 WASHINGTON ACADEMY OF SCIENCES

ARTICLE XIV. CONTROL OF FUNDS, ACTIVITIES

No part of the net earnings of the Washington Academy of Sciences shall inure

to the benefit of, or be distributable to its members, trustees, officers, or other

private persons, except that the Washington Academy of Sciences shall be au-

thorized and empowered to pay reasonable compensation for services rendered,

and to make payments and distributions in furtherance of the purposes set forth

in Article XII hereof. No substantial part of the activities of the Washington

Academy of Sciences shall be the carrying on of propaganda, or otherwise at-

tempting to influence legislation, and the Washington Academy of Sciences

shall not participate in, or intervene in (including the publishing or distribution

of statements) any political campaign on behalf of any candidate for public

office. Notwithstanding any other provision of these Articles, the Washington

Academy of Sciences shall not carry on any other activities not permitted to be

carried on (a) by an association exempt from Federal income tax under Section

501(c)(3) of the Internal Revenue Code of 1954 (or the corresponding provision

of any future United States Internal Revenue Law) or (b) by an association,

contributions to which are deductible under Section 170(c)(2) of the U.S. Inter-

nal Revenue Code of 1954 (or the corresponding provision of any future United

States Internal Revenue Law).

Journal of the Washington Academy of Sciences,

Volume 81, Number 4, Pages 203-218, December 1991

1991 Washington Academy of Sciences Membership Directory

M = Member; F = Fellow; LF = Life Fellow; LM = Life Member; EM = Emeritus

Member; EF = Emeritus Fellow; NRF = Non-Resident Fellow.

ABDULNUR, SUHEIL F. (Dr) 5715 Glenwood Rd., Bethesda, MD 20817 (F)

ABELSON, PHILIP H. (Dr) 4244 50th St., NW, Washington, DC 20016 (F)

ABRAHAM, GEORGE (Dr) 3107 Westover Dr., SE, Washington, DC 20020 (LF)

ABSOLON, KAREL B. (Dr) 11225 Huntover Dr., Rockville, MD 20852 (F)

ACHTER, MEYER R. (Dr) 417 Sth St., SE, Washington, DC 20003 (EF)

ADAMS, ALAYNE A. (Dr) 8436 Rushing Creek Ct., Springfield, VA 22153 (F)

ADAMS, CAROLINE L. (Dr) 242 N. Granada St., Arlington, VA 22203 (EM)

AFFRONTI, LEWIS F. (Dr) George Washington University School of Medicine, Microbiology, 2300

Eye St., NW, Washington, DC 20037 (F)

ALDRIDGE, MARY H. (Dr) 7904 Hackamore Dr., Potomac, MD 20854-3825 (EF)

ALEXANDER, BENJAMIN H. (Dr) P. O. Box 41126 NE, Washington, DC 20018 (F)

ALICATA, J. E. (Dr) 1434 Punahou St., Apt. #736, Honolulu, HI 96822 (EF)

ALLEN, J. FRANCES (Dr) P. O. Box 284 (Meeker Hollow Rd.), Roxbury, NY 12474-0284 (EF)

ANDRUS, EDWARD D. (Mr) 2497 Patricia Ct., Falls Church, VA 22043 (M)

ARGAUER, ROBERT J. (Dr) 4208 Everett St., Kensington, MD 20895 (F)

ARONSON, CASPER J. (Mr) 3401 Oberon St., Kensington, MD 20895 (EM)

ARSEM, COLLINS (Mr) 10821 Admirals Way, Potomac, MD 20854 (M)

ARVESON, PAUL T. (Mr) 10205 Folk St., Silver Spring, MD 20902 (F)

AXELROD, JULIUS (Dr) LCB-M.H.IRP-NIMH, Room 3A15A, Bldg. 36, National Institute of Men-

tal Health, Bethesda, MD 20892 (EF)

AXILROD, BENJAMIN M. (Dr) 9216 Edgewood Dr., Gaithersburg, MD 20877 (EF)

BAILEY, R. CLIFTON (Dr) 6507 Divine St., McLean, VA 22101 (LF)

BAKER, ARTHUR A. (Dr) 5201 Westwood Dr., Bethesda, MD 20816 (EF)

BAKER, LEONARD (Dr) 4924 Sentinel Drive, Bethesda, MD 20816 (F)

BAKER, LOUIS C. W. (Dr) Georgetown University, Dept. of Chemistry, Washington, DC 20057 (F)

BALLARD, LOWELL D. (Mr) 7823 Mineral Springs Dr., Gaithersburg, MD 20877 (F)

~BARBOUR, LARRY L. (Mr) Rural Route 1, Box 492, Great Meadows, NJ 07838 (M)

BARTFELD, CHARLES I. (Dr) 6007 Kirby Rd., Bethesda, MD 20817 (M)

BATAVIA, ANDREW I. (Mr) 700 Seventh St., S.W., Apt #813, Washington, DC 20024 (LF)

BAUMANN, ROBERT C. (Mr) 9308 Woodberry St., Seabrook, MD 20706 (F)

BEACH, LOUIS A. (Dr) 1200 Waynewood Blvd., Alexandria, VA 22308 (F)

BECKER, DONALD A. (Mr) 13115 Dauphine St., Silver Spring, MD 20906 (F)

BECKER, EDWIN D. (Dr) Bldg 2, Room 122, N.I.H., Bethesda, MD 20892 (F)

BECKMANN, ROBERT B. (Dr) 10218 Democracy Ln., Potomac, MD 20854 (F)

BEKEY, IVAN (Mr) 4624 Quarter Charge Dr., Annandale, VA 22003 (F)

BENDER, MAURICE (Dr) 16518 NE Second PI., Bellevue, WA 98008 (EF)

BENESCH, WILLIAM M. (Dr) 4444 Linnean Ave., NW, Washington, DC 20008 (LF)

BENJAMIN, CHESTER R. (Dr) 315 Timberwood Ave., Silver Spring, MD 20901 (EF)

203

204 WASHINGTON ACADEMY OF SCIENCES

BENNETT, JOHN A. (Mr) 7405 Denton Rd., Bethesda, MD 20814 (F)

BENSON, WILLIAM M. (Dr) 636 Massachusetts Ave., NE, Washington, DC 20002 (F)

BERGMANN, OTTO (Dr) George Washington Univ., Dept. of Physics, Washington, DC 20052 (F)

BERKSON, HAROLD (Dr) 12001 Whippoorwill Ln., Rockville, MD 20852 (M)

BERNETT, MARIANNE K. (Mrs) 5337 Taney Ave., Alexandria, VA 22304 (EM)

BERNSTEIN, BERNARD (Mr) 7420 Westlake Terr., Apt. #608, Bethesda, MD 20817 (M)

BESTUL, ALDEN B. (Dr) 9400 Overlea Dr., Rockville, MD 20850 (F)

BETTS, ALLEN W. (Mr) 2510 South Ivanhoe PI., Denver, CO 80222 (M)

BHAGAT, SATINDAR M. (Prof) 112 Marine Terr., Silver Spring, MD 20904 (F)

BICKLEY, WILLIAM E. (Dr) 6516 Fortieth Ave., University Park, Hyattsville, MD 20782 (F)

BISHOP, WILLIAM P. (Dr) Desert Research Institute, 2505 Chandler Dr., Suite #1, Las Vegas, NV

89120 (NRF)

BLACKMON, RICHARD F. (Mr) 2000 N. Adams St., Apt. #102, Arlington, VA 22201 (M)

BLACKSTEN, HARRY RIC (Mr) 4413 N. 18th St., Arlington, VA 22207 (M)

BLANCHARD, DAVID L. (Dr) 1015 McCeney Ave., Silver Spring, MD 20901 (LF)

BLANK, CHARLES A. (Dr) 255 Massachusetts Ave., Apt. #607, Boston, MA 02115 (NRF)

BLOCH, CAROLYN C. (Mrs) P. O. Box 1889, Rockville, MD 20849 (M)

BLUNT, ROBERT F. (Dr) 5411 Moorland Ln., Bethesda, MD 20814 (F)

BOEK, HEATHER (Dr) Corning Incorporated, SP-DV-2-1, Corning, NY 14831 (M)

BOEK, JEAN K. (Dr) National Graduate University, 1101 N. Highland St., Arlington, VA 22201 (LF)

BOEK, WALTER E. (Dr) 5011 Lowell St., Washington, DC 20016 (F)

BOGNER, M. SUE (Dr) 9322 Friars Rd., Bethesda, MD 20817 (LF)

BONEAU, C. ALAN (Dr) 5305 Waneta Rd., Bethesda, MD 20816 (F)

BOTBOL, JOSEPH MOSES (Dr) 60 Curtis St., Falmouth, MA 02540 (F)

BOURGEOIS, LOUIS D. (Dr) 8701 Bradmoor Dr., Bethesda, MD 20817 (EF)

BOURGEOIS, MARIE J. (Dr) 8701 Bradmoor Dr., Bethesda, MD 20817 (F)

BOWMAN, THOMAS E. (Dr) Smithsonian Institution, Invertebrate Zoology, NHB Mail Stop 163,

Washington, DC 20560 (F)

BOYD, WENDELL J. (Mr) 6307 Balfour Dr., Hyattsville, MD 20782 (M)

BRADSHAW, SARA L. (Ms) 5405 Duke St., Apt. #312, Alexandria, VA 22304 (M)

BRANCATO, EMANUEL L. (Dr) 7370 Hallmark Rd., Clarksville, MD 21029 (EF)

BRANDEWIE, DONALD F. (Mr) 6811 Field Master Dr., Springfield, VA 22153 (EF)

BRENNER, ABNER (Dr) 7204 Pomander Ln., Chevy Chase, MD 20815 (F)

BRIER, GLENN W. (Mr) 1729 N. Harrison St., Arlington, VA 22205 (LF)

BRISKMAN, ROBERT D. (Mr) 6728 Newbold Dr., Bethesda, MD 20817 (F)

BROADHURST, MARTIN G. (Dr) 116 Ridge Rd., Box 163, Washington Grove, MD 20880 (F)

BROWN, ELISE A. B. (Dr) 6811 Nesbitt Pl., McLean, VA 22101 (LF)

BRYAN, MILTON M. (Mr) 3322 N. Glebe Rd., Arlington, VA 22207 (M)

BURAS, EDMUND M., JR. (Mr) 824 Burnt Mills Ave., Silver Spring, MD 20901 (EF)

BUSCH, WILLIAM S. (Mr) 1035 Sun Valley Dr., Annapolis, MD 21401 (M)

BUTTERMORE, DONALD O. (Mr) 34 West Berkeley St., Uniontown, PA 15401 (LF)

CAHNMAN, HUGO N. (Mr) CASSO-SOLAR Corp., P. C. Box 163, Pomona, NY 10970 (M)

CAMPBELL, LOWELL E. (Mr) 14000 Pond View Rd., Silver Spring, MD 20905 (F)

CANNON, EDWARD W. (Dr) 18023 134th Ave., Sun City West, AZ 85375 (NRF)

CANTELO, WILLIAM W. (Dr) 11702 Wayneridge St., Fuiton, MD 20759 (F)

CARROLL, WILLIAM R. (Dr) 4802 Broad Brook Dr., Bethesda, MD 20814 (EF)

CERRONI, MATTHEW J. (Mr) 12538 Browns Ferry Rd., Herndon, VA 22070 (M)

CHAMBERS, RANDALL M. (Dr) 2704 Winstead Circle, Wichita, KS 67226 (NRF)

CHAPLIN, HARVEY R., JR. (Dr) 1561 Forest Villa Ln., McLean, VA 22101 (F)

MEMBERSHIP DIRECTORY

CHAPMAN, ROBERT D. (Dr) 10976 Swansfield Rd., Columbia, MD 21044 (F)

CHEEK, CONRAD H. (Dr) 4334 H St., SE, Washington, DC 20019 (F):

CHEZEM, CURTIS G. (Dr) 3378 Wisteria St., Eugene, OR 97404 (NRF)

CHI, MICHAEL (Dr) 201 International Dr., Apt. #631, Cape Canaveral, FL 32920 (NRF)

CHRISTIANSEN, MERYL N. (Dr) 610 T-Bird Dr., Front Royal, VA 22630 (NRF)

CIVEROLO, EDWIN L. (Dr) 12340 Shadetree Ln., Laurel, MD 20708 (F)

CLAIRE, CHARLES N. (Mr) 4403 14th St., NW, Washington, DC 20011 (EF)

CLARK, GEORGE E., JR. (Mr) 4022 N. Stafford St., Arlington, VA 22207 (F)

CLEVEN, GALE W. (Dr) 2411 Old Forge Ln., Apt. #103, Las Vegas, NV 89121 (EF)

CLIFF, RODGER A. (Dr) 2331 Cheshire Way, Redwood City, CA 94061 (M)

CLINE, THOMAS LYTTON (Dr) 13708 Sherwood Forest Dr., Silver Spring, MD 20904 (F)

COATES, JOSEPH F. (Mr) 3738 Kanawha St., NW, Washington, DC 20015 (F)

COFFEY, TIMOTHY P. (Dr) Naval Research Laboratory, Code 1001, Washington, DC 20375 (F)

COLE, RALPH I. (Mr) 3705 S. George Mason Dr., Apt. #1515S, Falls Church, VA 22041 (F)

COLWELL, RITA R. (Dr) Maryland Biotechnology Institute, 1123 Microbiology Building, University

_of Maryland, College Park, MD 20742 (LF)

COMPTON, W. DALE (Dr) Ford Motor Company, P. O. Box 1603, Dearborn, MI 48121 (F)

CONDELL, WILLIAM J., JR. (Dr) 4511 Gretna St., Bethesda, MD 20814 (F)

CONNELLY, EDWARD McD. (Mr) 1625 Autumnwood Dr., Reston, VA 22094 (F)

COOK, RICHARD K. (Dr) 4111 Bel Pre Rd., Rockville, MD 20853 (F)

COOPER, KENNETH W. (Dr) 4497 Picacho Dr., Riverside, CA 92507 (EF)

CORLISS, EDITH L. R. (Mrs) 2955 Albemarle St., NW, Washington, DC 20008 (LF)

CORMACK, JOHN G. (Mr) 10263 Gainsborough Rd., Potomac, MD 20854 (M)

COSTRELL, LOUIS (Mr) 15115 Interlachen Dr., Apt. #621, Silver Spring, MD 20906 (F)

COTHERN, C. RICHARD (Dr) 4732 Merivale Rd., Chevy Chase, MD 20815 (F)

COTTERILL, CARL H. (Mr) 6030 Corland Ct., McLean, VA 22101 (F)

CREVELING, CYRUS R. (Dr) 4516 Amherst Ln., Bethesda, MD 20814 (F)

CRUM, JOHN K. (Dr) 1155 16th St., NW, Washington, DC 20036 (F)

CULBERT, DOROTHY K. (Mrs) 6254 Seven Oaks Ave., Baton Rouge, LA 70806 (EF)

CURRIE, CHARLES L., S. J. (Rev) Rector, Jesuit Community, St. Joseph’s University, Philadelphia,

PA 19131 (NRF)

CUTKOSKY, ROBERT DALE (Mr) 19150 Roman Way, Gaithersburg, MD 20879 (F)

D’ANTONIO, WILLIAM V. (Dr) 3701 Connecticut Ave., NW, Apt. #818, Washington, DC 20008 (F)

DAVIS, ANDREW V. (Mr) 4201 Massachusetts Ave., NW, Apt. #332, Washington, DC 20016 (M)

DAVIS, CHARLES M., JR. (Dr) 8458 Portland Pl., McLean, VA 22102 (M)

DAVIS, MARION MACLEAN (Dr) Crosslands, Apt. #100, Kennett Square, PA 19348 (LF)

‘DAVIS, MILES (Dr) 1214 Bolton St., Baltimore, MD 21217-4111 (F)

DAVIS, ROBERT E. (Dr) 1793 Rochester St., Crofton, MD 21114 (F)

DAVISON, MARGARET C. (Mrs) 2928 N. 26th St.,- Arlington, VA 22207 (M)

DAVISSON, JAMES W. (Dr) 400 Cedar Ridge Rd., Oxon Hill, MD 20745 (EF)

DAWSON, VICTOR C. D. (Dr) 9406 Curran Rd., Silver Spring, MD 20901 (F)

DEAHL, KENNETH L. (Dr) USDA-ARS-BARC WEST, Bldg. 004, Room 215, Beltsville, MD 20705

(F)

DEAL, GEORGE E. (Dr) 6245 Park Rd., McLean, VA 22101 (F)

DeBERRY, MARIAN B. (Mrs) 3608 17th St., NE, Washington, DC 20018 (EM)

DEDRICK, ROBERT L. (Dr) 1633 Warner Ave., McLean, VA 22101 (F)

DeLANEY, WAYNE R. (Mr) 602 Oak St., Farmville, VA 23901-1118 (M)

DEMING, W. EDWARDS (Dr) 4924 Butterworth Pl., NW, Washington, DC 20016 (F)

DEMUTH, HAL P. (Cdr) 118 Wolfe St., Winchester, VA 22601 (NRF)

DENNIS, BERNARD K. (Mr) 915 Country Club Dr., Vienna, VA 22180 (EF)

206 WASHINGTON ACADEMY OF SCIENCES

DESLATTES, RICHARD D., JR. (Dr) 610 Aster Blvd., Rockville, MD 20850 (F)

DEUTSCH, STANLEY (Dr) 7109 Lavarock Ln., Bethesda, MD 20817 (EF)

DEVEY, GILBERT B. (Mr) 2801 New Mexico Ave., NW, Apt. #617 Washington, DC 20007 (M)

DEVIN, CHARLES, JR. (Dr) 629 Blossom Dr., Rockville, MD 20850 (M)

DeVOE, JAMES R. (Mr) 11708 Parkridge Dr., Gaithersburg, MD 20878 (F)

deWIT, ROLAND (Dr) 11812 Tifton Dr., Rockville, MD 20854 (F)

DICKSON, GEORGE (Mr) 415 Russell Ave., Apt. #1116, Gaithersburg, MD 20877 (F)

DIMOCK, DAVID A. (Mr) 4291 Molesworth Terr., Mt. Airy, MD 21771 (EM)

DOCTOR, NORMAN (Mr) 6 Tegner Ct., Rockville, MD 20850 (F)

DOEPPNER, THOMAS W. (Col) 8323 Orange Ct., Alexandria, VA 22309 (LF)

DONAHUE, JAMES H. (Capt) 3080 N. Oakland St., Arlington, VA 22309 (M)

DONALDSON, EVA G. (Ms) 3941 Ames St., NE, Washington, DC 20019 (F)

DONALDSON, JOHANNA B. (Mrs) 3020 N. Edison St., Arlington, VA 22207 (F)

DONNERT, HERMANN J. (Dr) Kansas State University, Dept. of Nuclear Engineering, Ward Hall,

Manhattan, KS 66506-7039 (F)

DOOLING, ROBERT J. (Dr) 4812 Mori Dr., Rockville, MD 20852 (F)

DOUGLAS, THOMAS B. (Dr) 3031 Sedgwick St., NW, Washington, DC 20008 (EF)

DRAEGER, HAROLD R. (Dr) 1201 N. 4th St., Tucson, AZ 85705 (EF)

DUBEY, SATYA D. (Dr) 7712 Groton Rd., West Bethesda, MD 20817 (EF)

DUFFEY, DICK (Dr) University of Maryland, Chem-Nuclear Engineering Dept., College Park, MD

20742 (LF)

DUKE, JAMES A. (Mr) 8210 Murphy Rd., Fulton, MD 20759 (LF)

DUNCOMBE, RAYNOR L. (Dr) 1804 Vance Circle, Austin, TX 78701 (NRF)

DUNKUM, WILLIAM W. (Dr) 1561 Pensacola St., Apt. #2306, Honolulu, HI 96822 (EF)

DuPONT, JOHN ELEUTHERE (Mr) P. O. Box 358, Newtown Square, PA 19073 (NRF)

DURIE, EDYTHE G. (Mrs) 4408 Braeburn Dr., Fairfax, VA 22032 (EF)

ECKLIN, JOHN W. (Mr) 6143 K Edsall Road, Alexandria, VA 22304 (M)

EDINGER, STANLEY E. (Dr) 5901 Montrose Rd., Apt. #404-N, Rockville, MD 20852 (F)

EDMUND, NORMAN W. (Mr) 407 NE Third Ave., Ft. Lauderdale, FL 33301 (M)

EISENHART, CHURCHILL (Dr) 9629 Elrod Rd., Kensington, MD 20895 (EF)

ELASSAL, ATEF A. (Dr) 1538 Red Rock Ct., Vienna, VA 22182 (F)

EL-BISI, HAMED M. (Dr) 258 Bishops Forest Dr., Waltham, MA 02154 (M)

ELISBERG, F. MARILYN (Mrs) 4008 Queen Mary Dr., Olney, MD 20832 (F)

ELLIOTT, F. E. (Dr) 7507 Grange Hall Dr., Fort Washington, MD 20744 (EF)

EMERSON, K. C. (Dr) 560 Boulder Dr., Sanibel, FL 33957 (F)

ENDO, BURTON Y. (Dr) 1010 Jigger Ct., Annapolis, MD 21401 (F)

ENGLAR, ROBERT JOHN (Mr) 3269 Catkin Ct., Marietta, GA 30066 (F)

ETTER, PAUL C. (Mr) 16609 Bethayres Rd., Rockville, MD 20855-2043 (F)

ETZIONI, AMITAI (Dr) 2700 Virginia Ave., NW, Apt. #1002, Washington, DC 20037 (F)

EVERSTINE, GORDON C. (Dr) 12020 Golden Twig Ct., Gaithersburg, MD 20878 (F)

EWERS, JOHN C. (Mr) 4432 N. 26th Rd., Arlington, VA 22207 (EF)

FARLEE, CORALEE (Dr) 389 O St., SW, Washington, DC 20024 (F)

FARMER, ROBERT F., III (Dr) c/o Akzo Chem, | Livingston Ave., Dobbs Ferry, NY 10522-3401

(NRF)

FAUCHALD, CHRISTIAN (Dr) National Museum of Natural History, Smithsonian Institution, Wash-

ington, DC 20560 (F)

FAULKNER, JOSEPH A. (Mr) 2 Bay Dr., Lewes, DE 19958 (NRF)

FAUST, WILLIAM R. (Dr) 5907 Walnut St., Temple Hills, MD 20748 (EF)

MEMBERSHIP DIRECTORY 207

FEARN, JAMES E. (Dr) 4446 Alabama Ave., SE, Washington, DC 20019 (F)

FEINGOLD, S. NORMAN (Dr) 9707 Singleton Dr., Bethesda, MD 20817 (F)

FERRELL, RICHARD A. (Dr) University of Maryland, Dept. of Physics, College Park, MD 20742 (F)

FINKELSTEIN, ROBERT (Mr) 10001 Crestleigh Ln., Potomac, MD 20854 (M)

FINN, EDWARD J. (Dr) 7500 Lynn Dr., Chevy Chase, MD 20815 (F)

FISHER, JOEL L. (Dr) 4033 Olley Ln., Fairfax, VA 22030 (M)

FLINN, DAVID R. (Dr) 9714 Wildflower Circle, Tuscaloosa, AL 35405 (NRF)

FLORIN, ROLAND E. (Dr) 7407 Cedar Ave., Takoma Park, MD 20912 (EF)

FLOURNOY, NANCY (Ms) 1829 E. Capitol St., Washington, DC 20003 (M)

FOCKLER, HERBERT H. (Mr) 10710 Lorain Ave., Silver Spring, MD 20901 (EM)

FONER, SAMUEL N. (Dr) 11500 Summit West Blvd., Apt. #15 B, Temple Terrace, FL 33617 (NRF)

FOOTE, RICHARD H. (Dr) Box 166, Lake of the Woods, Locust Grove, VA 22508 (NRF)

FORZIATI, ALPHONSE F. (Dr) 15525 Prince Frederick Way, Silver Spring, MD 20906 (F)

FORZIATI, FLORENCE H. (Dr) 15525 Prince Frederick Way, Silver Spring, MD 20906 (F)

FOSTER, AUREL O. (Dr) 4613 Drexell Rd., College Park, MD 20740 (EF)

FOURNIER, ROBERT O. (Dr) 108 Paloma Rd., Portola Valley, CA 94028 (F)

FOWLER, WALTER B. (Mr) 9404 Underwood St., Seabrook, MD 20706 (M)

FOX, DAVID W. (Dr) University of Minnesota, 136 Lind Hall, 207 Church St., SE, Minneapolis, MN

55455 (F)

FOX, WILLIAM B. (Dr) 1813 Edgehill Dr., Alexandria, VA 22307 (F)

FRANKLIN, JUDE E. (Dr) 7616 Carteret Rd., Bethesda, MD 20817-2021 (F)

FRAVEL, DEBORAH R. (Dr) Soilborne Diseases Laboratory, Room 275, Bldg. 011A, BARC-West,

Beltsville, MD 20705 (F)

FREEMAN, ANDREW F. (Mr) 5012 N. 33rd St., Arlington, VA 22207 (EM)

FRIEDMAN, MOSHE (Dr) Naval Research Laboratory, Code 4732, Washington, DC 20375-5000 (F)

FRIESS, SEYMOUR L. (Dr) 6522 Lone Oak Ct., Bethesda, MD 20817 (F)

FRUSH, HARRIET L. (Dr) 4912 New Hampshire Ave., NW, Apt. #104, Washington, DC 20011 (EF)

FURUKAWA, GEORGE T. (Dr) 1712 Evelyn Dr., Rockville, MD 20852 (F)

GAGE, WILLIAM W. (Dr) 10 Trafalgar St., Rochester, NY 14619 (F)

GALASSO, GEORGE J. (Dr) 636 Crocus Dr., Rockville, MD 20850 (F)

GALLER, SIDNEY R. (Dr) 6242 Woodcrest Ave., Baltimore, MD 21209 (EF)

GANEFF, IWAN (Mr) 5944 W. Wrightwood Ave., Chicago, IL 60639 (M)

GARVIN, DAVID (Dr) 18700 Walker’s Choice Rd., Apt. #807, Gaithersburg, MD 20879 (F)

GAUNAURD, GUILLERMO C. (Dr) 4807 Macon Rd., Rockville, MD 20852 (F)

GENTRY, JAMES W. (Prof) University of Maryland, Chem-Nuclear Engineering Dept., College Park,

MD 20742 (F)

‘“GHAFFARI, ABOLGHASSEM (Dr) 7532 Royal Dominion Dr., West Bethesda, MD 20817 (LF)

GHOSE, RABINDRA NATH (Dr) 8167 Mulholland Terr., Los Angeles, CA 90046 (NRF)

GILLASPIE, A. GRAVES, JR. (Dr) 141 -Cloister Dr., Peachtree City, GA 30269 (NRF)

GIST, LEWIS A. (Dr) 1336 Locust Rd., NW, Washington, DC 20012 (EF)

GLASER, HAROLD (Dr) 1346 Bonita St., Berkeley, CA 94709 (EF)

GLASGOW, AUGUSTUS R., JR. (Dr) 4116 Hamilton St., Hyattsville, MD 20781 (EF)

GLOVER, ROLFE E., III (Prof) 7006 Forest Hill Dr., Hyattsville, MD 20782 (F)

GLUCKMAN, ALBERT G. (Mr) 11235 Oakleaf Dr., Apt. #1619, Silver Spring, MD 20901 (F)

GLUCKSTERN, ROBERT L. (Dr) 10903 Wickshire Way, Rockville, MD 20852 (F)

GOFF, JAMES F. (Dr) 3405 34th Pl., NW, Washington, DC 20016 (F)

GOLDEN, A. MORGAN (Mr) 9110 Drake PI., College Park, MD 20740 (F)

GOLUMBIC, CALVIN (Dr) 6000 Highboro Dr., Bethesda, MD 20817 (EM)

GONET, FRANK (Dr) 4007 N. Woodstock St., Arlington, VA 22207 (EF)

GOODE, ROBERT J. (Mr) 2402 Kegwood Ln., Bowie, MD 20715 (F)

208 WASHINGTON ACADEMY OF SCIENCES

GORDON, RUTH E. (Dr) American Type Culture Collection, 12301 Parklawn Dr., Rockville, MD

20852 (EF)

GRAVER, WILLIAM R. (Dr) 6137 N. Ninth Rd., Arlington, VA 22205 (M)

GRAY, IRVING (Dr) 5450 Whitley Park Terr., Apt. #802, Bethesda, MD 20814-2060 (EF)

GREENOUGH, M. L. (Mr) Greenough Data Associates, 616 Aster Blvd., Rockville, MD 20850 (F)

GREER, SANDRA C. (Dr) University of Maryland, Chemistry Dept., College Park, MD 20742 (F)

GRISAMORE, NELSON T. (Prof) 9536 E. Bexhill Dr., Kensington MD 20895 (EF)

GROSS, DONALD (Mr) 3530 N. Rockingham St., Arlington, VA 22213 (F)

GROSSLING, BERNARDO F. (Dr) 10903 Amherst Ave., Apt. #241, Silver Spring, MD 20902 (F)

GRUNTFEST, IRVING (Dr) 140 Lake Carol Dr., West Palm Beach, FL 33411-2132 (EF)

HACSKAYLO, EDWARD (Dr) P. O. Box 189, Port Republic, MD 20676 (F)

HAENNI, EDWARD O. (Dr) 7907 Glenbrook Rd., Bethesda, MD 20814 (F)

HAGN, GEORGE H. (Mr) 4208 Sleepy Hollow Rd., Annandale, VA 22003 (LF)

HAINES, KENNETH (Mr) 3542 N. Delaware St., Arlington, VA 22207 (F)

HALL, E. RAYMOND (Dr) 1637 West Ninth St., Lawrence, KS 66044 (EF)

HAMER, WALTER J. (Dr) 407 Russell Ave., Apt. #305, Gaithersburg, MD 20877-2889 (EF)

HAMMER, GUY S., III (Mr) 8902 Ewing Dr., Bethesda, MD 20817 (F)

HAMMER, JEAN H. (Mrs) 8902 Ewing Dr., Bethesda, MD 20817 (M)

HAND, CADET S., JR. (Prof) Star Route, Bodega Bay, CA 94923 (EF)

HANEL, RUDOLPH A. (Dr) 31 Brinkwood Rd., Brookeville, MD 20833 (F)

HANFORD, WILLIAM E. (Mr) 5613 Overlea Rd., Bethesda, MD 20816 (M)

HANIG, JOSEPH P. (Dr) 822 Eden Ct., Alexandria, VA 22308 (F)

HANSEN, LOUIS S. (Dr) University of California, Oral Pathology, Room S-524, OM&D, San Fran-

cisco, CA 94143-0424 (NRF)

HANSEN, MORRIS H. (Mr) 13532 Glen Mill Rd., Rockville, MD 20850 (LF)

HARR, JAMES W. (Mr) 9503 Nordic Dr., Lanham, MD 20706 (M)

HARRINGTON, FRANCIS D. (Dr) 4600 Ocean Beach Blvd., Apt. #204, Cocoa Beach, FL 32931

(NRF)

HARRINGTON, MARSHALL C. (Dr) 10450 Lottsford Rd., Apt. #2207, Mitchellville, MD 20721

(EF) ;

HARTLEY, JANET WILSON (Dr) N.I.H., NIAID, Laboratory of Immunopathology, Bethesda, MD

20892 (F)

HARTMANN, GREGORY K. (Dr) 10701 Keswick St., Apt. #317, Garrett Park, MD 20896 (EF)

HARTZLER, MARY P. (Ms) 1250 S. Washington St., Apt. #203, Alexandria, VA 22314 (M)

HASKINS, CARYL P. (Dr) 1545 18th St., NW, Suite 810, Washington, DC 20037 (EF)

HASS, GEORG H. (Mr) 7728 Lee Ave., Alexandria, VA 22308 (F)

HAUPTMAN, HERBERT A. (Dr) The Medical Foundation of Buffalo, Inc., 33 High St., Buffalo, NY

14203-1196 (F)

HAYDEN, GEORGE A. (Dr) 1312 Juniper St., NW, Washington, DC 20012 (EM)

HAYNES, ELIZABETH D. (Mrs) 4149 N. 25th St., Arlington, VA 22207 (M)

HEADLEY, ANNE RENOUF (Dr) The Metropolitan Square, 655 15th St., NW, Suite #330, Washing-

ton, DC 20005 (F)

HEIFFER, MELVIN H. (Dr) 11107 Whisperwood Ln., Rockville, MD 20852 (F)

HENDERSON, EDWARD P. (Dr) 4600 Connecticut Ave., NW, Washington, DC 20008 (EF)

HENNEBERRY, THOMAS J. (Dr) 1409 E. Northshore Dr., Tempe, AZ 85283 (NRF)

HERMACH, FRANCIS L. (Mr) 2201 Colston Dr., Apt. #311, Silver Spring, MD 20910 (F)

HERMAN, ROBERT (Dr) 8434 Antero Dr., Austin, TX 78759 (NRF)

HERSEY, JOHN B. (Mr) 923 Harriman St., Great Falls, VA 22066 (M)

HEYER, W. RONALD (Dr) Amphibian and Reptile, N.H.B., Smithsonian Institution, Washington,

DC 20560 (F)

—

MEMBERSHIP DIRECTORY 3 209

HIBBS, EUTHYMIA (Dr) 7302 Durbin Terr., Bethesda, MD 20817 (M)

HILLABRANT, WALTER J. (Dr) 1927 38th St., NW, Washington, DC 20007 (M)

HILSENRATH, JOSEPH (Mr) 9603 Brunett Ave., Silver Spring, MD 20901 (F)

HOBBS, ROBERT B. (Dr) 7715 Old Chester Rd., Bethesda, MD 20817 (F)

HOFFELD, J. TERRELL (Dr) 11307 Ashley Dr., Rockville, MD 20852-2403 (M)

HOGE, HAROLD J. (Dr) 65 Grove St., Apt. #146, Wellesley, MA 02181 (EF)

HOLLINGSHEAD, ARIEL (Dr) 3637 Van Ness St., Washington, DC 20008 (F)

HOLSHOUSER, WILLIAM L. (Mr) P. O. Box 1475, Banner Elk, NC 28604 (F)

HONIG, JOHN G. (Dr) 7701 Glenmore Spring Way, Bethesda, MD 20817 (F)

HOOVER, LARRY A. (Mr) P. O. Box 491, Gastonia, NC 28053-0491 (M)

HOPP, THEODORE H. (Mr) Bldg 220, Room A127, National Institute of Standards and Technology,

Gaithersburg, MD 20899 (M)

HORNSTEIN, IRWIN (Dr) 5920 Bryn Mawr Rd., College Park, MD 20740 (EF)

HOROWITZ, EMANUEL (Dr) 14100 Northgate Dr., Silver Spring, MD 20906 (F)

HOWARD, DARLENE V. (Dr) Georgetown Univ., Dept. of Psychology, Washington, DC 20057 (F)

HOWARD, JAMES H., JR. (Dr) 3701 Cumberland St., NW, Washington, DC 20016 (F)

‘HOWELL, BARBARA F. (Dr) 206 Baybourne Dr., Arnold, MD 21012 (F)

HOYT, JAMES, JR. (Mr) 8104 Tapscott Ct., Pikesville, MD 21208 (M)

HUANG, KUN-YEN (Dr) 1445 Laurel Hill Rd., Vienna, VA 22180 (F)

HUDSON, COLIN M. (Dr) 143 S. Wildflower Rd., Asheville, NC 28804 (EF)

HUGH, RUDOLPH (Dr) George Washington University Medical School, Microbiology Dept., 2300

Eye St., NW, Washington, DC 20037 (F)

HUHEEY, JAMES E. (Dr) 6909 Carleton Terr., College Park, MD 20742 (LF)

HUMMEL, JOHN N. (Mr) 200 Harry S. Truman Pkwy., Second Floor, Annapolis, MD 21401 (M)

HUMMEL, LANI S. (Ms) 9312 Fairhaven Ave., Upper Marlboro, MD 20772 (M)

HUNTER, WILLIAM R. (Mr) 6705 Caneel Ct., Springfield, VA 22152 (F)

HURDLE, BURTON G. (Mr) 6222 Berkley Rd., Alexandra, VA 22307 (F)

HURTT, WOODLAND (Dr) 7302 Parkview Dr., Frederick, MD 21702 (M)

HUTTON, GEORGE L. (Mr) 1086 Continental Ave., Melbourne, FL 32940 (EF)

IRVING, GEORGE W., JR. (Dr) 4601 North Park Ave., Apt. #613, Chevy Chase, MD 20815 (LF)

IRWIN, GEORGE R. (Dr) 7306 Edmonston Rd., College Park, MD 20740 (F)

ISBELL, HORACE S. (Dr) 3401 38th St., NW, Apt. #216, Washington, DC 20016 (F)

ISENSTEIN, ROBERT S. (Dr) 11710 Caverly Ave., Beltsville, MD 20705 (M)

JACKSON, DAVID J. (Dr) 1451 Siena Ave., Coral Gables, FL 33146 (NRF)

* JACKSON, JO-ANNE A. (Dr) 14711 Myer Terr., Rockville, MD 20853 (LF)

JACOX, MARILYN E. (Dr) 10203 Kindly Ct., Gaithersburg, MD 20879 (F)

JAMES, HENRY M. (Mr) 6707 Norview Ct., Springfield, VA 22152 (M)

JEN, CHIH K. (Dr) 10203 Lariston Ln., Silver Spring, MD 20903 (EF)

JENSEN, ARTHUR S. (Dr) 5602 Purlington Way, Baltimore, MD 21212 (LF)

JERNIGAN, ROBERT W. (Dr) American University, Dept. Mathematics and Statistics, 4400 Massa-

chusetts Ave., NW, Washington, DC 20016 (F)

JESSUP, STUART D. (Dr) 746 N. Emerson St., Arlington, VA 22203 (EF)

JOHNSON, DANIEL P. (Dr) P. O. Box 359, Folly Beach, SC 29439 (EF)

JOHNSON, EDGAR M. (Dr) 5315 Renaissance Court, Burke, VA 22015 (LF)

JOHNSON, PHYLLIS T. (Dr) 4721 East Harbor Dr., Friday Harbor, WA 98250 (EF)

JONES, HOWARD S., JR. (Dr) 3001 Veazey Terr., NW, Apt. #1310, Washington, DC 20008 (LF)

JONG, SHUNG-CHANG (Dr) American Type Culture Collection, 12301 Parklawn Dr., Rockville,

MD 20852 (LF)

210 WASHINGTON ACADEMY OF SCIENCES

JORDAN, GARY BLAKE (Dr) 13392 Fallen Leaf Rd., Poway, CA 92064 (LM)

JOYCE, PRISCILLA G. (Ms) 605 N. Emerson St., Arlington, VA 22203 (M)

KAISER, HANS E. (Dr) 433 Southwest Dr., Silver Spring, MD 20901 (M)

KANTOR, GIDEON (Mr) 10702 Kenilworth Ave., Garrett Park, MD 20896-0553 (M)

KAPER, JACOBUS M. (Dr) 115 Hedgewood Dr., Greenbelt, MD 20770 (F)

KAPETANAKOS, C. A. (Dr) 4601 North Park Ave., Apt. #921, Chevy Chase, MD 20815 (F)

KARP, SHERMAN (Dr) 10205 Counselman Rd., Potomac, MD 20854 (F)

KARR, PHILLIP R. (Dr) 5507 Calle de Arboles, Torrance, CA 90505 (EF)

KAUFMAN, H. PAUL (Lt. Col) P. O. Box 1135, Fedhaven, FL 33854-1135 (EF)

KAZYAK, KRISTIN R. (Ms) 2145 Hilltop Pl., Falls Church, VA 22043 (M)

KEARNEY, PHILIP C. (Dr) 8416 Shears Ct., Laurel, MD 20707 (F)

KEISER, BERNHARD E. (Dr) 2046 Carrhill Rd., Vienna, VA 22180 (F)

KESSLER, KARL G. (Dr) 5927 Anniston Rd., Bethesda, MD 20817 (F)

KIRK, KENNETH L. (Dr) National Institutes of Health, Bldg 8A, B1A02, Bethesda, MD 20892 (F)

KLEBANOFF, PHILIP S. (Mr) 6412 Tone Dr., Bethesda, MD 20817 (EF)

KLINGSBERG, CYRUS (Dr) 1318 Deerfield Dr., State College, PA 16803 (NRF)

KLINMAN, DENNIS MARC (Dr) 10401 Grosvenor PI., Suite #725, Rockville, MD 20852 (F)

KNOX, ARTHUR S. (Mr) 2008 Columbia Rd., NW, Washington, DC 20009 (M)

KNUTSON, LLOYD V. (Dr) Agricultural Research Center, Room 001, Bldg. 003, Beltsville, MD

20705 (F)

KRAMER, CAROLYN M. (Dr) M.R.A.D., The Gillette Company, Gillette Park, 5G-2, Boston, MA

02106 (NRF)

KROP, STEPHEN (Dr) 7908 Birnam Wood Dr., McLean, VA 22102 (F)

KRUGER, JEROME (Dr) 619 Warfield Dr., Rockville, MD 20850 (F)

KRUPSAW, MARYLIN (Mrs) 10208 Windsor View Dr., Potomac, MD 20854 (LF)

LANG, MARTHA E. C. (Mrs) Kennedy-Warren, 3133 Connecticut Ave., NW, Apt. #625, Washing-

ton, DC 20008 (EF)

LANG, SCOTT W. (Mr) 3640 Dorshire Ct., Pasadena, MD 21122-6469 (M)

LANG, TERESA C. H. (Mrs) 3640 Dorshire Ct., Pasadena, MD 21122-6469 (M)

LANGSTON, JOANN H. (Ms) 14514 Faraday Dr., Rockville, MD 20853 (F)

LAPHAM, EVAN G. (Mr) 2242 S.E. 28th St., Cape Coral, FL 33904 (EF)

LAWSON, ROGER H. (Dr) 10613 Steamboat Landing, Columbia, MD 21044 (F)

LEE, MARK A. (Mr) 5539 Columbia Pike, Apt. #407, Arlington, VA 22204 (M)

LEE, RICHARD H. (Dr) 5 Angola By The Bay, Lewes, DE 19958 (EF)

LEFTWICH, STANLEY G. (Dr) 3909 Belle Rive Terr., Alexandria, VA 22309 (LF)

LEIBOWITZ, LAWRENCE M. (Dr) 3903 Laro Ct., Fairfax, VA 22031 (F)

LEINER, ALAN L. (Mr) 850 Webster St., Apt. #635, Palo Alto, CA 94301 (EF)

LEJINS, PETER P. (Dr) 7114 Eversfield Dr., College Heights Estates, Hyattsville, MD 20782 (F)

LENTZ, PAUL LEWIS (Dr) 5 Orange Ct., Greenbelt, MD 20770 (EF)

LESSOFF, HOWARD (Mr) O.N.R. Europe, Box 39, FPO, New York, NY 09510-0700 (F)

LETTIERI, THOMAS R. (Mr) 10705 Hunters Chase Ln., Damascus, MD 20872 (M)

LEVIN, RONALD L. (Dr) 5012 Continental Dr., Olney, MD 20832 (F)

LEVINSON, NANETTE S. (Dr) American University, CT A-Hurst 206, Washington, DC 20016 (M)

LEVY, SAMUEL (Mr) 2279 Preisman Dr., Schenectady, NY 12309 (EF)

LEWIS, A. D. (Mr) 3476 Mount Burnside Way, Woodbridge, VA 22192 (M)

LEY, HERBERT L. (M.D.) 4816 Camelot St., Rockville, MD 20853 (EF)

LIBELO, LOUIS F. (Mr) 9413 Bulls Run Pkwy., Bethesda, MD 20817 (F)

LIEBLEIN, JULIUS (Dr) 1621 East Jefferson St., Rockville, MD 20852 (EF)

MEMBERSHIP DIRECTORY | 211

LIEBOWITZ, HAROLD (Dr) George Washington University, School of Engineering and Applied

Science, 2021 K St., NW, Suite #710, Washington, DC 20052 (F)

LINDSEY, IRVING (Mr) 202 E. Alexandria Ave., Alexandria, VA 22302 (EF)

LING, LEE (Mr) 1608 Belvoir Dr., Los Altos, CA 94022 (EF)

LINK, CONRAD B. (Dr) University of Maryland, Horticulture Dept., College Park, MD 20742 (F)

LIST, ROBERT J. (Mr) 1123 Francis Hammond Pkwy., Alexandria, VA 22302 (EF)

LOCKARD, J. DAVID (Dr) University of Maryland, Botany Depi., College Park, MD 20742 (F)

LOEBENSTEIN, W. V. (Dr) 8501 Sundale Dr., Silver Spring, MD 20910 (LF)

LONG, BETTY JANE (Mrs) 416 Riverbend Rd., Ft. Washington, MD 20744 (F)

LORING, BLAKE M. (Dr) 26889 Lancia St., Moreno Valley, CA 92388-4843 (EF)

LUSTIG, ERNEST (Dr) Rosittenweg 10, D-3340, Wolfenbuttel, Federal Republic of Germany (NRF)

LUTZ, ROBERT J. (Dr) 17620 Shamrock Dr., Olney, MD 20832 (M)

LYNN, JEFFREY W. (Prof) 13128 Jasmine Hill Terr., Rockville, MD 20850 (F)

LYONS, JOHN W. (Dr) 7430 Woodville Rd., Mt. Airy, MD 2177! (F)

MacDONELL, MICHAEL T. (Dr) 3939 Ruffin Rd., San Diego, CA 92123 (NRF)

‘MADDEN, ROBERT P. (Dr) National Institute of Standards and Technology, A-251 Physics Bldg.,

Gaithersburg, MD 20899 (F)

MAIENTHAL, MILLARD (Dr) 10116 Bevern Ln., Potomac, MD 20854 (F)

MALONE, THOMAS B. (Dr) 6633 Kennedy Ln., Falls Church, VA 22042 (F)

MANDERSCHEID, RONALD W. (Dr) 10837 Admirals Way, Potomac, MD 20854-1232 (LF)

MARCUS, MARVIN (Dr) 2937 Kenmore PI., Santa Barbara, CA 93105 (NRF)

MARTIN, EDWARD J. (Dr) 7721 Dew Wood Dr., Derwood, MD 20855 (F)

MARTIN, JOHN H. (Dr) 440 NW Elks Dr., Apt. #205, Corvallis, OR 97330-3749 (EF)

MARTIN, ROBERT H. (Mr) 2257 N. Nottingham St., Arlington, VA 22205 (EM)

MARTIN, ROY E. (Mr) National Fisheries Institute, 1525 Wilson Blvd., Suite #500, Arlington, VA

22209 (M)

MASON, HENRY LEA (Dr) 3440 S. Jefferson St., Apt. #823, Falls Church, VA 22041-3127 (EF)

MAYOR, JOHN R. (Dr) 3308 Solomons Ct., Silver Spring, MD 20906 (F)

McAVOY, THOMAS J. (Mr) 502 Burning Tree Dr., Arnold, MD 21012 (F)

McBRIDE, GORDON W. (Mr) 3323 Stuyvesant Pl., NW, Washington, DC 20015-2454 (EF)

McCRACKEN, ROBERT H. (Mr) 5120 Newport Ave., Bethesda, MD 20816 (LF)

McKENZIE, LAWSON M. (Mr) 1719 N. Troy St., Apt. #394, Arlington, VA 22201 (F)

McKINSTRY, PATRICIA A. (Ms) 11671 Gilman Ln., Herndon, VA 22070-2420 (M)

McNESBY, JAMES R. (Dr) 13308 Valley Dr., Rockville, MD 20850 (EF)

MEADE, BUFORD K. (Mr) 5903 Mt. Eagle Dr., Apt. #404, Alexandria, VA 22303-2523 (F)

_ MEARS, FLORENCE M. (Dr) 8004 Hampden Ln., Bethesda, MD 20814 (EF)

MEARS, THOMAS W. (Mr) 2809 Hathaway Terr., Wheaton, MD 20906 (F)

MEBS, RUSSELL W. (Dr) 6620 N. 32nd St., Arlington, VA 22213 (F)

MELMED, ALLEN J. (Dr) 732 Tiffany Ct., Gaithersburg, MD 20878 (F)

MENZER, ROBERT E. (Dr) 612 Silverthorn Rd., Gulf Breeze, FL 32561 (NRF)

MESSINA, CARLA G. (Mrs) 9800 Marquette Dr., Bethesda, MD 20817 (F)

MILLAR, DAVID B. (Dr) 1716 Mark Ln., Rockville, MD 20852 (F)

MILLER, CARL F. (Dr) P. O. Box 127, Gretna, VA 24557 (EF)

MILLER, LANCE A. (Dr) P. O. Box 58, Snickersville Pike, Middleburg, VA 22117 (F)

MITTLEMAN, DON (Dr) 80 Parkwood Ln., Oberlin, OH 44074 (NRF)

MIZELL, LOUIS R. (Mr) 8122 Misty Oaks Blvd., Sarasota, FL 34243 (EF)

MOORE, GEORGE A. (Dr) 1108 Agnew Dr., Rockville, MD 20851-1601 (EF)

MOORE, JAMES G. (Mr) CRS, Library of Congress, Washington, DC 20540 (M)

MORGAN, HARRY D. (Dr) 11001 Battlement Ln., Ft. Washington, MD 20744 (F)

MORRIS, ALAN (Dr) 5817 Plainview Rd., Bethesda, MD 20817 (F)

212 WASHINGTON ACADEMY OF SCIENCES

MORRIS, J. ANTHONY (Dr) 23-E Ridge Rd., Greenbelt, MD 20770 (M)

MORRIS, JOSEPH BURTON (Mr) 2010 Franklin St., NE, Washington, DC 20018 (EM)

MORSE, ROBERT A. (Mr) 5530 Nevada Ave., NW, Washington, DC 20015 (M)

MOSTOFI, F. K. (M.D.) 7001 Georgia St., Chevy Chase, MD 20815 (F)

MOUNTAIN, RAYMOND D. (Dr) 5 Monument Ct., Rockville, MD 20850 (F)

MUEHLHAUSE, C. O. (Dr) 112 Accomac St., Chincoteague, VA 23336-1401 (EF)

MUESEBECK, CARL F. W. (Mr) 18 N. Main St., Elba, NY 14058 (EF)

MULLIGAN, JAMES H., JR. (Dr) 12121 Sky Ln., Santa Ana, CA 92705 (NRF)

MUMMA, MICHAEL J., (Dr) 210 Glen Oban Dr., Arnold, MD 21012 (F)

MURDAY, JAMES S. (Dr) 6913 Raspberry Plain Pl., West Springfield, VA 22153 (F)

MURDOCH, WALLACE P. (Dr) 2264 Emmitsburg Rd., Gettysburg, PA 17325 (EF)

NAESER, CHARLES R. (Dr) 6654 Van Winkle Dr., Falls Church, VA 22044 (EF)

NAMIAS, JEROME (Mr) Scripps Institute of Oceanography, Univ. of California, Room A-024, La

Jolla, CA 92093 (NRF)

NEF, EVELYN S. (Mrs) 2726 N St., NW, Washington, DC 20007 (M)

NELSON, R. H. (Mr) Bethany Village, 512 Albright Dr., Mechanicsburg, PA 17055 (EF)

NEUBAUER, WERNER G. (Dr) 4603 Quarter Charge Dr., Annandale, VA 22003 (F)

NEUENDORFFER, J. A. (Dr) 911 Allison St., Alexandria, VA 22302 (EF)

NEUPERT, WERNER M. (Dr) Goddard Space Flight Center, Code 680, Greenbelt, MD 20771 (F)

NEWMAN, MORRIS (Dr) 1050 Las Alturas Rd., Santa Barbara, CA 93103 (NRF)

NICKUM, MARY J. (Mrs) 12174 Island View Circle, Germantown, MD 20874 (M)

NOFFSINGER, TERRELL L. (Dr) Route 1, Box 305, Auburn, KY 42206 (EF)

NORRIS, KARL H. (Mr) 11204 Montgomery Rd., Beltsville, MD 20705 (EF)

NYSTROM, ERIC O. (Mr) 10422 Cliff Mills Rd., Marshall, VA 22115 (M)

OBERLE, E. MARILYN (Ms) 58 Parklawn Rd., West Roxbury, MA 02132 (M)

OEHSER, PAUL H. (Mr) 9601 Southbrook Dr., Apt. #220 S, Jacksonville, FL 32256 (EF)

O’CONNOR, JAMES V. (Mr) 10108 Haywood Circle, Silver Spring, MD 20902 (M)

O’HARE, JOHN J. (Dr) 4601 O’Connor Ct., Irving, TX 75062 (NRF)

O’HERN, ELIZABETH M. (Dr) 633 G St., SW, Washington, DC 20024 (F)

OKABE, HIDEO (Dr) 6700 Old Stage Rd., Rockville, MD 20852 (F)

O’KEEFE, JOHN A. (Dr) Goddard Space Flight Center, Code 681, Greenbelt, MD 20771 (F)

OLIPHANT, MALCOLM W. (Dr) 1606 Ulupii St., Kailua, HI 96734 (EF)

OLIPHANT, V. SUSIE (Dr) 910 Luray Pl., Hyattsville, MD 20783 (M)

ORDWAY, FRED (Dr) 5205 Elsmere Ave., Bethesda, MD 20814 (F)

OSER, HANS J. (Dr) 8810 Quiet Stream Ct., Potomac, MD 20854 (F)

OSTAFF, WILLIAM ALLEN, (Mr) 10208 Drumm Ave., Kensington, MD 20895-3731 (M)

PANCELLA, JOHN R. (Dr) 1209 Viers Mill Rd., Rockville, MD 20851 (F)

PARASURAMAN, RAJA (Dr) Catholic University, Department of Psychology, Washington, DC

20064 (F)

PARMAN, GEORGE K. (Mr) 4255 Donald St., Eugene, OR 97405-3427 (EF)

PARRY-HILL, JEAN (Ms) 3803 Military Rd., NW, Washington, DC 20015 (M)

PARSONS, H. McILVANE (Dr) Human Resources Research Organization, 1100 S. Washington St.,

Alexandria, VA 22314 (F)

PAZ, ELVIRA L. (Dr) 172 Cook Hill Rd., Wallingford, CT 06492 (NRF)

PELCZAR, MICHAEL J. (Dr) Avalon Farm, P. O. Box 133, Chester, MD 21619 (EF)

PELLERIN, CHARLES J. (Dr) NASA, Code SZ, 600 Independence Ave., SW, Washington, DC 20546

(F)

MEMBERSHIP DIRECTORY | 213

PERKINS, LOUIS R. (Mr) 1234 Massachusetts Ave., NW, Apt. #709, Washington, DC 20005 (M)

PERROS, THEODORE P. (Dr) George Washington University, Chemistry Department, Washington,

DC 20052 (F)

PICKETT, WARREN E. (Dr) Naval Research Laboratory, Code 4692, Washington, DC 20375 (F)

PICKHOLZ, RAYMOND (Dr) 3613 Glenbrook Rd., Fairfax, VA 22031 (F)

PIEPER, GEORGE F. (Dr) 3155 Rolling Rd., Edgewater, MD 21037 (F)

PIKL, JOSEF M. (Dr) Meadowbrook Rd., Lincoln, MA 01773 (EF)

PITTMAN, MARGARET (Dr) 3133 Connecticut Ave., NW, Apt. #912, Washington, DC 20008 (EF)

PLAIT, ALAN O. (Mr) 5402 Yorkshire St., Springfield, VA 22151 (F)

PLANT, ANNE L. (Dr) 619 S. Woodstock St., Arlington, VA 22204 (M)

POLACHEK, HARRY (Dr) 11801 Rockville Pike, Apt. #1211, Rockville, MD 20852 (EF)

PONNAMPERUMA, CYRIL (Dr) 4452 Sedgwick St., NW, Washington, DC 20016 (F)

POST, MILDRED A. (Miss) 8928 Bradmoor Dr., Bethesda, MD 20817 (F)

POWELL, JAMES STANTON (Mr) 7873 Godolphin Dr., Springfield, VA 22153 (M)

PRESTON, MALCOLM S. (Dr) 10 Kilkea Ct., Baltimore, MD 21236 (M)

PRINCE, JULIUS S. (M.D.) 7103 Pinehurst Pkwy., Chevy Chase, MD 20815 (F)

PRINZ, DIANNE K. (Dr) Naval Research Laboratory, Code 4142, Washington, DC 20375-5000 (F)

PRO, MAYNARD J. (Mr) 7904 Falstaff Rd., McLean, VA 22102 (F)

PROCTOR, JOHN H. (Dr) 308 East St., NE, Vienna, VA 22180 (F)

PRYOR, C. NICHOLAS (Dr) 3715 Prosperity Ave., Fairfax, VA 22031 (F)

PURCELL, ROBERT H. (Dr) 17517 White Grounds Rd., Boyds, MD 20841 (F)

PYKE, THOMAS N., JR. (Mr) NOAA, FB #4, Room 2069, Washington, DC 20233 (F)

QUIROS, RODERICK S. (Mr) 4520 Yuma St., NW, Washington, DC 20016 (F)

RABINOW, JACOB (Mr) 6920 Selkirk Dr., Bethesda, MD 20817 (F)

RADER, CHARLES A. (Mr) Gillette Research Institute, 401 Professional Dr., Gaithersburg, MD

20879 (F)

RADO, GEORGE T. (Dr) 818 Carre Ct., McLean, VA 22101 (F)

RAINWATER, IVAN H. (Dr) 2805 Liberty Pl., Bowie, MD 20715 (EF)

RAMSAY, MAYNARD J. (Dr) 3806 Viser Ct., Bowie, MD 20715 (F)

RANSOM, JAMES R. (Mr) 107 E. Susquehanna Ave., Towson, MD 21204 (M)

RASKIN, ALLEN (Dr) 7658 Water Oak Point Rd., Pasadena, MD 21122 (F)

RATH, BHAKTA B. (Dr) 10908 Timbermill Ct., Oakton, VA 22124 (F)

RAUSCH, ROBERT L. (Dr) P. O. Box 85447, University Station, Seattle, WA 98145-1447 (NRF)

RAVITSKY, CHARLES (Mr) 1505 Drexel St., Takoma Park, MD 20912 (EF)

RAY, JOSEPH W. (Dr) 2740 Vassar Pl., Columbus, OH 43221 (NRF)

‘REDISH, EDWARD F. (Prof) 6820 Winterberry Ln., Bethesda, MD 20817 (F)

REED, WILLIAM DOYLE (Mr) 1330 Massachusetts Ave., NW, Thomas House, Apt. #624, Washing-

ton, DC 20005 (EF) j ;

REHDER, HARALD H. (Dr) 3900 Watson PI., Suite #2G-B, Washington, DC 20016 (F)

REINER, ALVIN (Mr) 11243 Bybee St., Silver Spring, MD 20902 (F)

REMMERS, GENE M. (Mr) 6928 Hector Rd., McLean, VA 22101 (M)

RESWICK, JAMES S. (Dr) 1003 Dead Run Dr., McLean, VA 22101 (F)

REYNOLDS, HORACE N., JR. (Dr) 14608 Pebblestone Dr., Silver Spring, MD 20910 (F)

RHYNE, JAMES J. (Dr) 2704 Westbrook Way, Columbia, MD 65203 (NRF)

RICE, ROBERT L. (Mr) 15504 Fellowship Way, Gaithersburg, MD 20878 (M)

RICE, SUE ANN (Dr) 6728 Fern Ln., Annandale, VA 22003 (M)

RICHMOND, ANNE T. (Mrs) 8833 Cold Spring Rd., Potomac, MD 20854 (F)

RIEL, GORDON K. (Dr) Naval Surface Weapons Center, White Oak Laboratory, Code R-41, Silver

Spring, MD 20903-5000 (LF)

214 WASHINGTON ACADEMY OF SCIENCES

RITT, PAUL E. (Dr) 36 Sylvan Ln., Weston, MA 02193 (F)

RIVERA, ALVIN D. (Dr) 4302 Star Ln., Rockville, MD 20852 (M)

ROBBINS, MARY LOUISE (Dr) Tatsuno House, A-23, 2-1-8 Ogikubo, Suginami-Ku, Tokyo 167,

Japan (EF)

ROBERTSON, A. F. (Dr) 4228 Butterworth Pl., NW, Washington, DC 20016 (F)

ROBERTSON, EUGENE C. (Mr) U. S. Geological Survey, 922 National Center, Reston, VA 22092

(M)

ROBERTSON, RANDALL M. (Dr) 1404 Highland Circle, SE, Blacksburg, VA 24060 (EF)

ROBSON, CLAYTON W. (Mr) 13307 Warburton Dr., Ft. Washington, MD 20744 (M)

RODNEY, WILLIAM S. (Dr) Georgetown University, Physics Dept., Washington, DC 20057 (F)

ROE, DONALD W. (Dr) 1072 Conestoga Estate, Harpers Ferry, WV 25425 (M)

ROLLER, PAUL S. (Dr) 4201 Butterworth Pl., NW, Washington, DC 20016 (EF)

ROSCHER, NINA M. (Dr) 10400 Hunter Ridge Dr., Oakton, VA 22124 (F)

ROSE, WILLIAM K. (Dr) 10916 Picasso Ln., Potomac, MD 20854 (F)

ROSENBLATT, DAVID (Prof) 2939 Van Ness St., NW, Washington, DC 20008 (F)

ROSENBLATT, JOAN R. (Dr) 2939 Van Ness St., NW, Washington, DC 20008 (F)

ROSENFELD, AZRIEL (Dr) 847 Loxford Terr., Silver Spring, MD 20910 (F)

ROSENTHAL, SANFORD M. (Dr) 12601 Greenbrier Rd., Potomac, MD 20854 (EF)

ROSS, FRANKLIN J. (Mr) 3830 N. Stafford St., Arlington, VA 22207-4513 (F)

ROSS, SHERMAN (Dr) 23 Glen Mary Rd., Bar Harbor, ME 04609 (EF)

ROSSI, PETER H. (Prof) 34 Stagecoach Rd., Amherst, MA 01002 (NRF)

ROTHMAN, RICHARD B. (Dr) 1510 Flora Ct., Silver Spring, MD 20910 (F)

ROTKIN, ISRAEL (Mr) 11504 Regnid Dr., Wheaton, MD 20902 (EF)

RUBLE, BRUCE L. (Mr) 4200 Davenport St., NW, Washington, DC 20016 (M)

RUTNER, EMILE (Dr) 34 Columbia Ave., Takoma Park, MD 20912 (M)

SAENZ, ALBERT W. (Dr) Naval Research Laboratory, Code 6603 S, Washington, DC 20375-5000 (F)

SALISBURY, LLOYD L. (Mr) 10138 Crestwood Rd., Kensington, MD 20895 (M)

SALLET, DIRSE W. (Dr) 4205 Tuckerman St., University Park, MD 20782 (M)

SAMUELSON, DOUGLAS A. (Mr) 1910 Wintergreen Ct., Reston, VA 22091 (F)

SANDERSON, JOHN A. (Dr) B-206 Clemson Downs, 150 Downs Blvd., Clemson, SC 29631 (EF)

SANK, VICTOR J. (Dr) 5 Bunker Ct., Rockville, MD 20854 (F)

SARMIENTO, RAFAEL A. (Dr) USDA, Federal Grain Inspection Service, P.O. Box 96454, Room

1631-S, Washington, DC 20090-1454 (F)

SASMOR, ROBERT M. (Dr) 4408 N. 20th Rd., Arlington, VA 22207 (F)

SASS, ARTHUR H. (Capt) RFD 6, Box 176, Warrenton, VA 22186 (M)

SAVILLE, THORNDIKE, JR. (Mr) 5601 Albia Rd., Bethesda, MD 20816 (LF)

SCHALK, JAMES M. (Dr) P. O. Box 441, Isle of Palms, SC 29451 (F)

SCHECHTER, MILTON S. (Mr) 10909 Hannes Ct., Silver Spring, MD 20901 (F)

SCHINDLER, ALBERT I. (Dr) 6615 Sulky Ln., Rockville, MD 20852 (F)

SCHLAIN, DAVID (Dr) 2-A Gardenway, Greenbelt, MD 20770 (EF)

SCHMIDT, CLAUDE H. (Dr) 1827 Third St., N., Fargo, ND 58102 (F)

SCHNEIDER, JEFFREY M. (Dr) 5238 Richardson Dr., Fairfax, VA 22032 (F)

SCHNEIDER, SIDNEY (Mr) 239 N. Granada St., Arlington, VA 22203 (EM)

SCHNEPFE, MARIAN M. (Dr) Potomac Towers, Apt. #640, 2001 N. Adams St., Arlington, VA 22201

(EF)

SCHOOLEY, JAMES F. (Dr) 13700 Darnestown Rd, Gaithersburg, MD 20878 (F)

SCHUBAUER, GALEN B. (Dr) 10450 Lottsford Rd., Unit #1211, Mitchellville, MD 20721 (F)

SCHULMAN, FRED (Dr) 11115 Markwood Dr., Silver Spring, MD 20902 (F)

SCHULMAN, JAMES H. (Dr) 4615 North Park Ave., Apt. #1519, Chevy Chase, MD 20815 (EF)

SCHULTZ, WARREN W. (Cdr) 4056 Cadle Creek Rd., Edgewater, MD 21037 (LF)

MEMBERSHIP DIRECTORY 215

SCOTT, DAVID B. (Dr) 10448 Wheatridge Dr., Sun City, AZ 85373 (EF)

SCRIBNER, BOURDON F. (Mr) 123 Peppercorn Pl., Edgewater, MD 21037 (EF)

SEABORG, GLENN T. (Dr) 1154 Glen Rd., Lafayette, CA 94549 (F)

SEEGER, RAYMOND J. (Dr) 4507 Wetherill Rd., Bethesda, MD 20816 (EF)

SEITZ, FREDERICK (Dr) Rockefeller University, 1230 York Ave., New York, NY 10021 (F)

SHAFRIN, ELAINE G. (Mrs) 800 Fourth St., SW, Apt. N-702, Washington, DC 20024 (F)

SHAPIRO, GUSTAVE (Mr) 3704 Munsey St., Silver Spring, MD 20906 (F)

SHEAR, RALPH E. (Mr) 1916 Bayberry Rd., Edgewood, MD 21040 (M)

SHEPARD, HAROLD H. (Dr) 2701 S. June St., Arlington, VA 22202 (EF)

SHERESHEFSKY, J. LEON (Dr) 4530 Connecticut Ave., NW, Apt. #400, Washington, DC 20008

(EF)

SHERLIN. GROVER C. (Mr) 4024 Hamilton St., Hyattsville, MD 20781 (LF)

SHIER, DOUGLAS R. (Dr) 416 Westminster Dr., Pendleton, SC 29670 (NRF)

SHOTLAND, EDWIN (Dr) 418 E. Indian Spring Dr., Silver Spring, MD 20901 (M)

SHRIER, STEFAN (Dr) 624A S. Pitt St., Alexandna, VA 22314-4138 (F)

SHROPSHIRE, W., JR. (Dr) Omega Laboratory, P. O. Box 189, Cabin John, MD 20818-0189 (LF)

SILVER, DAVID M. (Dr) Applied Physics Laboratory, 1110 Johns Hopkins Rd., Laurel, MD 20723

(M)

SILVERMAN, BARRY G. (Dr) 9653 Reach Rd., Potomac, MD 20854 (F)

SIMHA, ROBERT (Dr) Case-Western Reserve University, Department of Macromolecular Science,

Cleveland, OH 44106 (NRF)

SIMPSON, MICHAEL M. (Mr) Congressional Research Service/SPR/LM413, Washington, DC 20540

(LM)

SLACK, LEWIS (Dr) 27 Meadow Bank Rd., Old Greenwich, CT 06870 (F)

SLAWSKY, MILTON M. (Dr) 8803 Lanier Dr., Silver Spring, MD 20910 (EF)

SLAWSKY, ZAKA I. (Dr) 4701 Willard Ave., Apt. #318, Chevy Chase, MD 20815 (EF)

SMITH, BLANCHARD D., JR. (Mr) 2509 Ryegate Ln., Alexandria, VA 22308 (F)

SMITH, EDWARD L. (Mr) 11027 Earlgate Ln., Rockville, MD 20852 (F)

SMITH, MARCIA S. (Ms) 6015 N. Ninth St., Arlington, VA 22205 (LM)

SMITH, REGINALD C. (Mr) 7731 Tauxemont Rd., Alexandria, VA 22308 (M)

SNYDER, HERBERT N. (Dr) P. O. Box 1494, Tappahannock, VA 22560 (NRF)

SOLAND, RICHARD M. (Dr) George Washington University, SEAS, Washington, DC 20052 (LF)

SOLOMON, EDWIN M. (Mr) 3330 N. Leisure World Dr., Apt. #222, Silver Spring, MD 20906 (M)

SOMMER, HELMUT (Dr) 9502 Hollins Ct., Bethesda, MD 20817 (EF)

SORROWS, HOWARD EARLE (Dr) 8820 Maxwell Dr., Potomac, MD 20854 (F)

SOUSA, ROBERT J. (Dr) 56 Wendell Rd., Shutesbury, MA 01072 (NRF)

SPATES, JAMES E. (Mr) 8609 Irvington Ave., Bethesda, MD 20817 (LF)

SPECHT, HEINZ (Dr) Fairhaven, C-135, 7200 3rd Ave., Sykesville, MD 21784 (EF)

SPERLING, FREDERICK (Dr) 5902 Mt. Eagle Dr., Alexandria, VA 22303 (EF)

SPIES, JOSEPH R. (Dr) 507 N. Monroe St., Arlington, VA 22201 (EF)

SPILHAUS, A. F., JR. (Dr) 10900 Picasso Ln., Potomac, MD 20854 (F)

SPRAGUE, G. F. (Dr) 2212 S. Lynn St., Urbana, IL 61801 (EF)

SPROULL, JAMES D. (Mr) 416 Blair Rd., Vienna, VA 22180 (F)

STANLEY, WILLIAM A. (Mr) 10494 Graeloch Rd., Laurel, MD 20723 (M)

STAUSS, HENRY E. (Dr) 8005 Washington Ave., Alexandria, VA 22308 (F)

STEERE, RUSSELL L. (Dr) 6207 Carrollton Terr., Hyattsville, MD 20781 (EF)

STEGUN, IRENE A. (Miss) 62 Leighton Ave., Yonkers, NY 10705 (EF)

STEINBERG, ALFRED D. (M.D.) 8814 Bells Mill Rd., Potomac, MD 20854 (F)

STEINER, ROBERT F. (Dr) 2609 Turf Valley Rd., Ellicott City, MD 21043 (F)

STEPHENS, ROBERT E. (Dr) 4301 39th St., NW, Washington, DC 20016 (EF)

STERN, KURT H. (Dr) Naval Research Laboratory, Code 6170, Washington, DC 20375-5000 (F)

216 WASHINGTON ACADEMY OF SCIENCES

STEWART, T. DALE (Dr) 1191 Crest Ln., McLean, VA 22101 (EF)

STIEF, LOUIS J. (Dr) Goddard Space Flight Center, Code 691, Greenbelt, MD 20771 (F)

STIEHLER, ROBERT D. (Dr) 3234 Quesada St., NW, Washington, DC 20015 (EF)

STILL, JOSEPH W. (Dr) 1408 Edgecliff Ln., Pasadena, CA 91107 (EF)

STOETZEL, MANYA B. (Dr) Systematic Entomology Laboratory, Room 6, Bldg. 004, BARC-WEST,

Beltsville, MD 20705 (F)

STRAUSS, SIMON W. (Dr) 4506 Cedell Pl., Camp Springs, MD 20748 (LF)

STRIMPLE, HARRELL L. (Mr) 904 Bowery, Iowa City, IA 52240 (F)

SVOBODA, JAMES A. (Mr) 13301 Overbrook Ln., Bowie, MD 20715 (M)

SWEZEY, ROBERT W. (Dr) Clarks Ridge Rd., Route 3, Box 142, Leesburg, VA 22075 (F)

SYKES, ALAN O. (Dr) 304 Mashie Dr., Vienna, VA 22180 (M)

TALBERT, PRESTON T. (Dr) 400 Old Stone Rd., Silver Spring, MD 20904 (EF)

TASAKI, ICHIJI (Dr) 5604 Alta Vista Rd., Bethesda, MD 20817 (F)

TATE, DOUGLAS R. (Mr) Carolina Meadows Villa, Apt. #257, Chapel Hill, NC 27514-8526 (EF)

TAYLOR, BARRY N. (Dr) 11908 Tallwood Ct., Potomac, MD 20854 (F)

TAYLOR, JOHN KEENAN (Dr) 12816 Tern Dr., Gaithersburg, MD 20878 (F)

TAYLOR, LAURISTON S. (Dr) 10450 Lottsford Rd., Apt. #3011, Mitchellville, MD 20721-2734 (EF)

TAYLOR, WILLIAM B. (Mr) 4001 Belle Rive Terr., Alexandria, VA 22309 (M)

TEAL, GORDON K. (Dr) 5222 Park Ln., Dallas, TX 75220 (F)

TERMAN, MAURICE J. (Mr) 616 Poplar Dr., Falls Church, VA 22046 (EM)

THOMPSON, F. CHRISTIAN (Dr) 4255 S. 35th St., Arlington, VA 22206 (LF)

TOLL, JOHN S. (Dr) University Research Association, 1111 19th St., NW, Suite #400, Washington,

DC 20036 (F)

TOUSEY, RICHARD (Dr) 10450 Lottsford Rd., Apt. #231, Mitchellville, MD 20721-2742 (EF)

TOUSIMIS, A. J. (Dr) Tousimis Research Corp., 2211 Lewis Ave., Rockville, MD 20851 (M)

TOWNSEND, CHARLES E. (M.D.) 3529 Tilden St., NW, Washington, DC 20008-3194 (F)

TOWNSEND, LEWIS RHODES (M.D.) 8906 Liberty Ln., Potomac, MD 20854 (M)

TOWNSEND, MARJORIE R. (Mrs) 3529 Tilden St., NW, Washington, DC 20008-3194 (LF)

TRAUB, ROBERT (Col., Ret.) 5702 Bradley Blvd., Bethesda, MD 20814 (F)

TUNELL, GEORGE (Dr) 300 Hot Springs Rd., Apt. #124, Montecito, CA 93108 (EF)

TURNER, JAMES H. (Dr) 509 South Pinehurst Ave., Salisbury, MD 21801-6122 (EF)

TYLER, PAUL E. (M.D.) 1023 Rocky Point Ct., Albuquerque, NM 87123 (NRF)

UBERALL, HERBERT M. (Dr) Kenwood, Apt. #1417, 5101 River Rd., Bethesda, MD 20816 (F)

UHLANER, J. E. (Dr) 4258 Bonavita Dr., Encino, CA 91426 (NRF)

UTZ, JOHN P. (M.D.) Georgetown University Medical Center, 3800 Reservoir Road, NW, Washing-

ton, DC 20057 (F)

VAISHNAV, MARIANNE P. (Ms) P. O. Box 2129, Gaithersburg, MD 20879 (LF)

VAN ARSDEL, WILLIAM C., III (Dr) 1000 Sixth St., SW, Apt. #301, Washington, DC 20024 (M)

VAN COTT, HAROLD P. (Dr) 8300 Still Spring Ct., Bethesda, MD 20817 (F)

VAN TUYL, ANDREW H. (Dr) 1000 West Nolcrest Dr., Silver Spring, MD 20903 (F)

VAN VOORHEES, DAVID A. (Dr) 5526 Paxford Ct., Fairfax, VA 22032 (M)

VARADI, PETER F. (Dr) 4620 North Park Ave., Apt. #1605-W, Chevy Chase, MD 20815 (F)

VEITCH, FLETCHER P., JR. (Dr) P. O. Box 513, Lexington Park, MD 20653 (F)

VENKATESHAN, C. N. (Dr) P. O. Box 30219, Bethesda, MD 20824 (M)

VILA, GEORGE J. (Mr) 5517 Westbard Ave., Bethesda, MD 20816 (F)

VITAS, STEPHAN THOMAS (Dr) 2803 Cortland Pl., NW, Washington, DC 20008 (M)

VON HIPPEL, ARTHUR (Dr) 265 Glen Rd., Weston, MA 02193 (EF)

MEMBERSHIP DIRECTORY : 217

WAGNER, A. JAMES (Mr) 7568 Cloud Ct., Springfield, VA 22153 (F)

WALDMANN, THOMAS A. (M.D.) N.LH., Bldg. #10, Room 4N115, Bethesda, MD 20890 (F)

WALKER, CHRISTOPHER W. (Mr) Lake Rd., Box 2087, Middleburg, VA 22117 (M)

WALTON, WILLIAM W., SR. (Dr) 1705 Edgewater Parkway, Silver Spring, MD 20903 (F)

WARING, JOHN A. (Dr) 1320 S. George Mason Dr., Apt. #1, Arlington, VA 22204 (M)

WARRICK, EVELYNE J. (Ms) National Color Inc., 2700 Prosperity Ave., Fairfax, VA 22031-4703

(M)

WATERWORTH, HOWARD E. (Dr) 10001 Old Franklin Ave., Seabrook, MD 20706 (F)

WATSON, ROBERT B. (Dr) 1176 Wimbledon Dr., McLean, VA 22101 (EM)

WAYNANT, RONALD W. (Dr) 13101 Claxton Dr., Laurel, MD 20708 (F)

WEBB, RALPH E. (Dr) 21-P Ridge Rd., Greenbelt, MD 20770 (F)

WEBER, ROBERT S. (Dr) 4520 Marissa Dr., El Paso, TX 79924 (EM)

WEGMAN, EDWARD J. (Dr) George Mason University, 157 Science Tech. Bldg., Fairfax, VA 22030

(LF)

WEINBERG, HAROLD (Mr) 11410 Strand Dr., Bldg. 1-B, Apt. #314, Rockville, MD 20852 (F)

WEINER, JOHN (Dr) 8401 Rhode Island Ave., College Park, MD 20740 (F)

WEINTRAUB, ROBERT L. (Dr) 407 Brooks Ave., Raleigh, NC 27607 (EF)

WEISS, ARMAND B. (Dr) 6516 Truman Ln., Falls Church, VA 22043 (LF)

WEISSLER, ALFRED (Dr) 5510 Uppingham St., Chevy Chase, MD 20815 (F)

WEISSLER, PEARL G. (Mrs) 5510 Uppingham St., Chevy Chase, MD 20815 (EF)

WENSCH, GLEN W. (Dr) R.R. #1, Box 54, Champaign, IL 61821 (EF)

WERGIN, WILLIAM P. (Dr) 10108 Towhee Ave., Adelphi, MD 20783 (F)

WERTH, MICHAEL W. (Mr) 14 Grafton St., Chevy Chase, MD 20815 (EM)

WESTWOOD, JAMES T. (LCDR) 3156 Cantrell Ln., Fairfax, VA 22031 (M)

WHITE, HOWARD J., JR. (Dr) 8028 Park Overlook Dr., Bethesda, MD 20817 (F)

WHITELOCK, LELAND D. (Mr) 2320 Brisbane St., Apt. #4, Clearwater, FL 34623 (F)

WHITTEN, CHARLES A. (Mr) 9606 Sutherland Rd., Silver Spring, MD 20901 (EF)

WIENER, ALFRED A. (Mr) 550 West 25th Pl., Eugene, OR 97405 (F)

WIGGINS, PETER F. (Dr) 1016 Harbor Dr., Annapolis, MD 21403 (F)

WILMOTTE, RAYMOND M. (Dr) 2512 Que St., NW, Apt. #301, Washington, DC 20007 (LF)

WILSON, BRUCE L. (Mr) 1411 Highland Ave., Plainfield, NJ 07060-3143 (EF)

WILSON, CHARLES L. (Dr) P. O. Box 1194, Shepherdstown, WV 25443 (F)

WILSON, WILLIAM K. (Mr) 1401 Kurtz Rd., McLean, VA 22101 (LF)

WISTORT, ROBERT L. (Mr) 11630 35th PIl., Beltsville, MD 20705 (F)

WITTLER, RUTH G. (Dr) 2103 River Crescent Dr., Annapolis, MD 21403-7271 (EF)

WOLFF, EDWARD A. (Dr) 1021 Cresthaven Dr., Silver Spring, MD 20903 (F)

WOOD, LAWRENCE A. (Dr) 7014 Beechwood Dr., Chevy Chase, MD 20815 (EF)

‘WORKMAN, WILLIAM G. (Dr) Washington Street, P. O. Box 7, Beallsville, OH 43716 (EF)

WUERKER, ANNE K. (Dr) 887 Gold Spring Pl., Westlake Village, CA 91361-2024 (NRF)

WULF, OLIVER R. (Dr) 557 Berkeley Ave., San Marino, CA 91108 (EF)

WYNN, HARVEY (Mr) 6625 Lee Highway, Arlington, VA 22205 (F)

YAPLEE, BENJAMIN S. (Mr) 8 Crestview Ct., Rockville, MD 20854 (F)

YODER, HATTEN S., JR. (Dr) Geophysical Laboratory, 5251 Broad Branch Rd., NW, Washington,

DC 20015 (EF)

YOUMAN, CHARLES E. (Mr) 4419 N. 18th St., Arlington, VA 22207 (M)

ZELENY, LAWRENCE (Dr) 4312 Van Buren St., University Park, MD 20782 (EF)

ZIEN, TSE-FOU (Dr) Naval Surface Warfare Center, White Oak Laboratory, Code R44, Silver Spring,

MD 20903-5000 (F)

ZOCH, RICHMOND T. (Mr) Route 1, Box 930, Shelby, AL 35143 (F)

218 WASHINGTON ACADEMY OF SCIENCES

Necrology

Deceased Life Fellows/Members

Mr. Karl Hilding Beij Mr. Hajime Ota

Dr. F. G. Brickwedde Mr. John A. Rosado

Dr. Archibald T. McPherson Dr. Ramesh N. Vaishnav

The following fellows/members of the Academy deceased since the last publication of the WAS mem-

bership directory

Mr. Laverne S. Birks Mr. Richard S. Hunter

Dr. Harold R. Curran Dr. Marion B. Matlack

Dr. Roger W. Curtis Dr. Dudley G. McConnell

Dr. Roy C. Dawson Dr. Melvin R. Meyerson

Dr. Ashley B. Gurney Mr. Frank W. Reinhart

Dr. Milton Harris Dr. Frederick D. Rossini

Dr. Francis J. Heyden, S. J. Dr. William R. Van Dersal

Dr. Henry Hopp Dr. Werner K. Weihe

Mrs. Hope E. Hopps Dr. David A. Young, Jr.

Membership Distribution

Member Category N % Geographic Location N %

Fellow 300 44.8 Maryland 314 46.9

Emeritus Fellow 145 21.6 Virginia 133 19.8

Member 114 17.0 Other states 125 19.0

Life Fellow 46 6.9 District of Columbia 93 13.9

Non-resident Fellow 46 6.9 Outside U.S. 3 0.4

Emeritus Member 16 2.4

Life Member 3 0.4

Totals 670 100.0 O70. 100.0

———— eee

DELEGATES TO THE WASHINGTON ACADEMY OF SCIENCES,

REPRESENTING THE LOCAL AFFILIATED SOCIETIES

Pilmssopinea! Society of Washington. ..2. 2. oe. 2. ie wdc ae cnek occ s cet: .. Thomas R. Lettieri

Pmunopolosical Society Of- Washington ... 2.2.0.5... 0.0... 0% eee cdeeeee eas Belford Lawson III

Memerical SOCicty Ol WaSMINetOn: 2. Ps o 2 e ec ee cde eos ee ee he ee Kristian Fauchald

Sneaiesl society Of Washington (26.026. oo ec ec ek lade ce os we ok Sek Elise A. B. Brown

Paremomrical Society of Washington ... 20.01.0066... cease eee os F. Christian Thompson

EERMeE IE COGTADIING SOCIEIW! 2)..560.) 5-2 Ss ois Paes cee ed ee Melaka cusgre Stanley G. Leftwich

ie Pteal Society Of WaSMINGLON 22.00 255 bee ed chee te eee ee es James V. O’Connor

eee seciety Ol ine Strict Of Colmmbla 0.05 5 ek i ee ve cease abe John P. Utz

Praded Sectety.of Washington, DC 2. 30.02.08 6. .6c 0c eae ee uss Thomas G. Manning

Peer etcty Ol WaASMINOIOM 5. ores es oa) ee Bk PSs he. oad woes Muriel Poston

Seems American Foresters, Washington Section ........2.. 06.6.4. .0..5025. Eldon W. Ross

MN MPIOHESOCIEUY {Ol ENPINECEIS, 22: dos eace oc.oec 2.8 eo oi ods a vss wath Siew seb caine We Alvin Reiner

Institute of Electrical and Electronics Engineers, Washington Section ........ George Abraham

American Society of Mechanical Engineers, Washington Section ......... Clayton W. Robson

ioimnrwolopical Society of Washington ....-. 1.0.25. .2 0.60.20. seve lace Kendall G. Powers

American Society for Microbiology, Washington Branch .................. Herman Schneider

Society of American Military Engineers, Washington Post .................... James Donahue

American Society of Civil Engineers, National Capital Section .............. John N. Hummel

Society for Experimental Biology and Medicine, DC Section .............. Cyrus R. Creveling

ASMelmtemational, Washineton Chapter ...2.........26..0..0c2e ede seuss en Pamela S. Patrick

American Association of Dental Research, Washington Section ............. J. Terrell Hoffeld

American Institute of Aeronautics and Astronautics, National Capital

Se SAIUMMRMM Pe Oc ee cre Set NI cpt ine Ri Shells Sie hd sls Sissel ephyest Seek Reginald C. Smith

panieieanvercorolopical Society, OC Chapter . 2... 6. 00.2020. G bee e ee ees A. James Wagner

Poser ee SOIC Ol WaASHINStOM (5.2.25. sc ceo ote cose disci de oGe barrane Gena cos To be determined

Acoustical society of America, Washington Chapter ........6....0........-- Richard K. Cook

PincwGam Mulcicar Society, Washington Section ®... . 2... 6... sence ees eee ees eee Kamal Ara]

Institute of Food Technologists, Washington Section .................... George W. Irving, Jr.

American Ceramic Society, Baltimore-Washington Section .................. Curtis A. Martin

SERRE SOCIEL Yet sce ee Meee Sas whieh ube he Gade Sh eT sree Sls o Paul Natishan

Manin eranmaIstory Ol SCIENCE CIUD .. os od cos. coke e's Sage cane secon aes Albert G. Gluckman

American Association of Physics Teachers, Chesapeake Section ............. Robert A. Morse

Optical Society of America, National Capital Section ...................... William R. Graver

American Society of Plant Physiologists, Washington Area Section ............. Steven J. Britz

Washington Operations Research/Management Science Council .............. John G. Honig

Instrument Society of America, Washington Section ..................... .... Donald M. Paul

American Institute of Mining, Metallurgical and Petroleum Engineers,

Ream PAOMESCCIIOIM ee ey ek Gee il OLS oR oma a so Mido’ sto « David M. Sutphin

Cale ApieAl ASETONOMERS: (5.0. 53 6 o's ceele Piss eevee 6 oetee bees nee Robert H. McCracken

Mathematics Association of America, MD-DC-VA Section ..................... Alice Schafer

Pisinctoen Columbia Institute of Chemists .c2..... 0.6. cl eke kb he xe William E. Hanford

District of Columbia Psychological Association ..... EUR OER go (Tenia alot ms Tg Sue Bogner

Masoineion Faint Technology; Group), ..05.%.. £05 ose oe na Se en ew Lloyd M. Smith

American Phytopathological Society, Potomac Division .................... Kenneth L. Deahl

Society for General Systems Research, Metropolitan Washington

CORES PEEP ge SCN A Celt rs a aR 9 RR Uc John H. Proctor

Human Factors society; Potomac Chapter... 20.62. 2 oo ie See. Thomas B. Malone

Amencan Fisheries society, Potomac Chapter .. 2... <0... seers sce sk David A. Van Vorhees

Association for Science, Technology and Innovation .........:......2.......... Ralph I. Cole

SUE AStEEM SUOCIGIAPICAl SOUIGLY che coe ae oko eisoas te ook cee Seda Ronald W. Manderscheid

Institute of Electrical and Electronics Engineers, Northern Virginia

EC HOM ery eee ee Mee eer Ee Ore ulti d SOS. Sale viele Blanchard D. Smith

Association for Computing Machinery, Washington Chapter ............. Charles E. Youman

Washington! Stausiical S@cletys eerste sie os ees See oles Soe Sere Nancy Flournoy

Society of Manufacturing Engineers, Washington, DC Chapter .......... ney. James E. Spates

Institute of Industrial Engineers, National Capital Chapter ................... James S. Powell

Delegates continue to represent their societies until new appointments are made.

Mie

Washington Academy of Sciences 2nd Class Postage Paid

1101 N. Highland St.

Arlington, Va. 22201

Return Requested with Form 3579

at Arlington, Va.

and additional mailing offices.

b 2 Mee

Bie

Mes yi,

HECKMAN I

BINDERY INC. ,e

SEPT 99

= N. MANCHESTER,

Bound -To-Pleas® IhiniANA 46962

tng tanta d

fe0y eee

pets

mrs ee ey

eee

BAS ema

Mis Vauect

ery

Poros

whee 8 me el

rhs Fetawne +

yee

pee oases

Reeecen inert APHID ca MAAN om Oh Ate ce ~

Pie oad

PO BINED AUNTS

eres .

Byers eases

FS ek RaW ls

Spon abn

PoP AE AR: Mea Rap Os

ness Ob

we? BS.

PISO TEN

PRON AD Ny

ER Bete Se

Leta tnaher WNT DS oe

See et

sey #4. = al

+ NUS eR ¢

RPA TEEN ENUM Mee

Soe Tyy . Beware

a Petty ena ake maatedes spa nauey

Tara

wea

mae

Xe aS

HCPRERD UES PROD Up ana

4 opelh roe

WAY nee

“ere

wee

Pa vi

SN aye srt,

ree

Peete te

lee aR ner

Dubey

Mes Man lt? Meet

Sata od

Hae eam

AUTEM TR REN RS

EAD HON

NOUR A

sr PIR

AIR

Went’ s 4 arom bgiteate

12 Xeric

oie a Cit To a ee =

ves cae 2a

Se Bet Ma MNO et

Seca

hte

ena Re

UA ese Me aed

ADs

et Lae wah Speech ay

“90 V 88!

ead oe es Ce a

PROV Haws et PAN ORENATUN RA oe Rap

wey es

TA Pen 2

Lenore var

ay ee er

Pere rn

SPC ne es

x DO IY eh Te

FM ps er WIE «

Bat bit

(et x38 deg apes iy

Py een ts Pee

SPM Pee Ald Oe

ASM e ap Rit EN

Vt. ue al opt Oo

200718 SB oat

PNT pete dle ITE Shee USA PD pS MERCI TOMN Sah ost Mg LA yploe Knee

PRURLAR VI ine pont Raat

SEO g MLAS mane ereripks 2 Saaaee

Mupsusy

PLP B

MAEMO Se 92 ge OF

Sere eet Pee ent

Ecerernent

Pex Sire lene ADAP END Ne

MERU APRA #

HVE Se hee

Pee eee ree

FADS PAP AV y

BP PEELS Pagel

Der RORY

Wp ret y

SP UY GAVE La

PA ie a weta 44 ony

fe eee sega ay

Merck iptipy Ss. ag arotped we,

uss gy ch és saesanlti~ > spent naa >

Eat.

PRAY

Torte

EVEN Wer OF AES

qh betes

MPA prep Te

Zawihis Lp hy ded rd oe sep Maia ope sant

Menc ee

AOR ST neg a Metta,

2 TEE OSG Gta WE OSU E ce Le bag

Fosyinys

NEN, si

arte

MNS

saunas se nsagees pare Pt ere

So a eee re .

OOMGTMON eh gay A

Sy ame ae we be

Agere Oe Lane:

URE OF Ae

yk a ey

Ye eg has

PPV ON EAD

A AIMS yeaah Sune

sean NE Gz ee ony FAR ORE ape re Oras

seas Uae a ntedio FRYE NPS onze

tee be pk

Nanke Ces