The Bulletin of The Natural History Museum (formerly: Bulletin of the British Museum

(Natural History) ), instituted in 1949, is issued in four scientific series, Botany,

Entomology, Geology (incorporating Mineralogy) and Zoology.

The Zoology Series is edited in the Museum’s Department of Zoology

Keeper of Zoology Prof P.S. Rainbow

Editor of Bulletin: Dr B.T. Clarke

Papers in the Bulletin are primarily the results of research carried out on the unique and ever-

growing collections of the Museum, both by the scientific staff and by specialists from elsewhere

who make use of the Museum’s resources. Many of the papers are works of reference that will remain

indispensable for years to come. All papers submitted for publication are subjected to external peer review for

acceptance.

A volume contains about 160 pages, made up by two numbers, published in the Spring and Autumn.

Subscriptions may be placed for one or more of the series on an annual basis. Individual numbers and back

numbers can be purchased and a Bulletin catalogue, by series, is available. Orders and enquiries should be

sent to:

Intercept Ltd.

P.O. Box 716

Andover

Hampshire SP10 1YG

Telephone: (01264) 334748

Fax: (01264) 334058

Email: intercept @ andover.co.uk

Internet: http://www.intercept.co.uk

Claims for non-receipt of issues of the Bulletin will be met free of charge if received by the Publisher within 6

months for the UK, and 9 months for the rest of the world.

World List abbreviation: Bull. nat. Hist. Mus. Lond. (Zool.)

Zoology Series

ISSN 0968-0470 Vol. 65, No. 1, pp. 1-72

The Natural History Museum

Cromwell Road

London SW7 5BD Issued 24 June 1999

Typeset by Ann Buchan (Typesetters), Middlesex

Printed in Great Britain by Henry Ling Ltd., at the Dorset Press, Dorchester, Dorset

Bull. nat. Hist. Mus. Lond. (Zool.) 65(1): 1-13 Issued 24 June 1999

Phylogenetic relationships of Toad-headed

lizards . Agamidae) based on

morphology _

E.N. ARNOLD 9 JL 1999 — }

Department of Zoology, The Natural History Museum, Cromwell Road, London SW7 SBD, UK ae i

CONTENTS | GENERAL

MAE CVA TV GUT OOM ics runs owns dos Sakae canon ak geo esa ase eas ends Cane wa as totnns cause ounoncnw ana casbnassevunck'usessUchpiiiansnnaaUsdeyssdudsveasanedesennuns dv sacnsessiwe 1

ERE VAM CAS INA TIS OME zID SILO CET) LILLIES acacia Wes sg ae eae asta ns cha sak hc hand ch ann cp aca we nna aia sdemwddennnstceccnedannatens 1

MOM OLOcirall ChALACLELS USE (0) ESUIMNALE PINVIO SCD se cnc. ceatten noe anascucncccevnuacrapaneasvencceupuascnosnhsscWnascsrasaadeceicusvstvasae in cesadecngadeaxutastes 2

FBI O MEMEUIC AMALV SIS) sosec.cessasqcs-cevaeveconaendevsveviesscnacent<vestaqsssaunasaccacayodssvassveaneedeauesudvasaveterasssachaiscdvivsaevsseveszesudebsacasucvsdatasccscesteaastases 7

DDPTSCSTES SLO rpete eee ties cece Pace Ps eet nape Sn scents cain Mean sae aac say ebse ranean ee es ae Mean cc enc ofaectn assis eravvacstesSvatich saasuvoeass¥ebeves steevenee 9

PSCMMOWLE CS CMENES! 4-.nsstonuananottne ce rete wwe ateecaskicsssceworvencasesedsanesuceacive deecstaevsusvssivdacsaue th veveuacsocesuecsctvaiesbcedes sxackoseencvenessviesesendievaes 11

FELT HCE Secs ccaa aster raed eas ueOe sesacee ce fae cee eae exc er ate a ne was nes ps nome camera Sasthy conc ehvasancsdsctatenvate means Gciuswesesdasenavedneues 11

SyNopsIS. Phrynocephalus together with its sister-group, Bufoniceps is most closely related to other advanced Palaearctic and

African agamids. They have been regarded as the sister-group of all these species or derived from African Agama (Moody, 1980,

morphological data) or as the sister of Laudakia (Joger, 1991, albumin immunology) but reassessment of morphology suggests

a relationship to Trapelus. Parsimony analysis of 46 morphological characters, involving 54 derived states, of 25 species of

Phrynocephalus indicates that successive branches arising from the main lineage of the genus are as follows: P. mystaceus; P.

maculatus; P. arabicus; the P. interscapularis group — (((P. clarkorum, P. ornatus) ((P. euptilopus, P. luteoguttatus) (P.

interscapularis, P. sogdianus))); P. scutellatus; P. golubevi; P. reticulatus; P. raddei. There is then a group of 11 species in which

relationships are generally poorly resolved, although within this P. theobaldi, P. roborowskii and P. vlangalii are clearly closely

related to each other and perhaps to P. forsythii, and the tuberculated species, P. helioscopus, P. persicus, P. rossikowi and P.

strauchi may also form a clade. There is no clear morphological evidence that the northeastern species, P. axillaris, P. versicolor,

P. przewalskii. and P. guttatus (which also extends far to the west) form a holophyletic group. Phrynocephalus does not appear

to share its general phylogeographic pattern with other Asian reptiles and this may consequently result from dispersal rather than

vicariance events. The phylogeny suggests the ancestor of Phrynocephalus occurred inArabia-NW India area whence there were

three independent invasions of Central Asia: by the ancestors of PR. mystaceus, of P. interscapularis + P. sogdianus, and of P.

golubevi and its sister group, the latter later extending north and eastwards into Mongolia, China and Tibet. Phrynocephalus

appears to have primitively occupied aeolian sand habitats but to have spread to harder substrates from which sandy habitats were

sometimes reinvaded. Degeneration of the outer and middle ear occurred in the early history of Phrynocephalus but was partly

reversed in P. axillaris and the P. theobaldi group.

Arnold 1992 which was created for Phrynocephalus laungwalaensis

INTRODUCTION Sharma, 1978. Moody (1980) placed Phrynocephalus (including P.

laungwalaensis) with what at the time was usually called Agama

Daudin 1802, in his group 6 of the Agamidae. WithinA gama, as then

understood, this author recognised several separate genera: Agama

s. str., Xenagama Boulenger 1895, Pseudotrapelus Fitzinger 1843,

Trapelus Cuvier 1817 and Stellio Laurenti 1768. However, the name

Stellio is unavailable (Stejneger, 1933) and the assemblage it was

used to denote by Moody is paraphyletic, comprising distinct

Palaearctic and mainly African assemblages (Joger, 1991; Baig &

Bohme, 1997) of which the former is probably a clade and the

members of the latter more closely related to such taxa as Agama,

Pseudotrapelus and Trapelus (personal observations). Leviton,

Anderson, Adler & Minton (1992) argue for the use of Laudakia

Gray, 1845 for the Palaearctic forms, a course followed here. The

more recent suggestion (Henle, 1995), that Laudakia should be

ee confined to some members of this assemblage and the rest placed in

RELATIONSHIPS OF PHRYNOCEPHALUS Placoderma Blyth, 1854, requires more thorough assessment of the

relationships of these lizards before it is adopted. The name

Phrynocephalus is the sister group of the monotypic Bufoniceps Acanthocercus Fitzinger, 1843 is available for the remainder of the

Toad-headed agamids, Phrynocephalus Kaup 1825, are a found in

the mainly Palearctic desert regions of Asia, from Eastern Turkey

and Russia to Mongolia, and southwards to southern Arabia and

Pakistan. Species in the south and centre of the range of the genus

are, in the main, well defined but, in the northeast, boundaries

between them are often less clear and numerous nominal taxa have

been described (see e.g. Zhao & Adler, 1993). This makes the total

number of species in the genus uncertain but it is likely to be in

excess of 30. In this paper, an estimate of phylogeny is made for 25

of the better defined species using morphological characters, includ-

ing external features and some internal ones derived from the

skeleton, middle ear, shoulder muscles and abdominal arteries.

E.N. ARNOLD

Fig. 1 Anterior views of right nasal area of skulls showing differences in contribution of the maxilla (m) to the posterior wall of the narial opening.

a. Small, does not contact septomaxilla (s) (Bufoniceps laugwalaensis). b. More extensive contribution, especially dorsally, and broad contact with

septomaxilla (Phrynocephalus mystaceus). c. More extensive still, both dorsally and ventrally, broad contact with septomaxilla maintained (P.

euptilopus).

Fig. 2 Anterior views of right nasal area of skulls. a. Dorsal process of maxilla (m) tapering upwards, maxilla extending outwards below lateral process of

prefrontal bone (pf) which is large (P. euptilopus). b. Dorsal process of maxilla blunt above, maxilla not extending markedly outwards below lateral

process of premaxilla which is relatively small (P. persicus).

forms that Moody allocated to Stellio (Schatti & Gasperetti, 1994;

Henle, 1995; Baig & Bohme, 1997).

Unweighted Wagner tree analysis of the morphological data

presented by Moody (1980) indicated that Phrynocephalus was

derived from a paraphyletic Agama s. str., while compatability

analysis, and Wagner tree analysis where characters were weighted

according to their consistency index in an initial run, suggested that

Phrynocephalus was sister to all other members of Moody’s Group

6 (Moody, 1980).

Results of isozyme analysis have been interpreted as indicating

that Phrynocephalus is the sister group of Laudakia (Ananjeva &

Sokolova, 1990), a result in agreement with immunological studies

(Joger, 1991). In contrast, areassessment of morphology (pers. obs.)

suggests that the sister group of Phrynocephalus + Bufoniceps is

Trapelus. Shared features that appear derived within Moody’s Group

6 include the following: maxillae in contact beneath premaxilla,

lateral prefrontal processes very large, palatine roof of interorbital

canal narrow or absent, vomers fused, squamosal spatulate with no

hook-shaped projection on its lateral margin, presacral vertebrae

usually 22 or fewer; nostrils directed forwards rather than sideways,

no enlarged subocular scales (reversed in some Phrynocephalus),

external ear opening reduced in size, no spinous scales on dorsum of

neck (reversed in some Phryncocephalus), no caudal autotomy,

scales on tail not in regular whorls; nasal passage long and flexed,

depressor mandibulae muscle extends partly over tympanum.

MORPHOLOGICAL CHARACTERS USED TO

ESTIMATE PHYLOGENY

Skull

1. Contribution of the maxilla to the posterior wall of the narial

opening of the skull (Figure 1). Small, does not contact

septomaxilla (0); more extensive especially dorsally, broad con-

tact with septomaxilla (1); more extensive still both dorsally and

ventrally, broad contact with septomaxilla maintained (2).

2. Dorsal process of maxilla (Figure 2). Tapering upwards (0);

broad and ending bluntly above (1).

3. Maxilla extends clearly outwards below the anterior surface of the

lateral process of the prefrontal bone (Figure 2). No (0); yes (1).

4. Relationship of maxillary and nasal bones below the lateral

process of the nasal (Figure 3). Widely separated (0); more

narrowly separated (1); in contact (2).

PHRYNOCEPHALUS PHYLOGENY 3

Fig.3 Anterior views of right nasal area of skulls showing differences in arrangement of maxilla (m), septomaxilla (stippled)) and nasal bones (n). a.

maxilla and nasal widely separated below lateral process (Ip) of nasal (P. scutellatus). b. Maxilla and nasal more narrowly separated below lateral process

of nasal and the space filled by the septomaxilla (P. versicolor). c. Maxilla and nasal in contact below lateral process of nasal (P. persicus).

Fig. 4 View into anterior right orbit, showing width of posterior face of prefrontal bone (pf) relative to that of the posterior platine (pa) and variation in

lateral extension of the prefrontal relative to the infraorbital canal (io). a. Posterior face of prefrontal broad, extends laterally across infraorbital canal. b.

Posterior face of prefrontal narrow, does not extend across infrarorbital canal

CEES 2 epi ssoe

Fig. 5 Difference in size of the parietal foramen (black) in adults. a. Small, diameter less than distance from the lateral edge of the parietal bone (p) (P

scutellatus). b. Large, diameter more than distance from lateral edge of parietal bone (P. persicus).

5. Maxilla in contact with septomaxilla on surface of skull below (0); narrowed (1).

lateral process of nasal(Figure 3). No (0); yes but does not reach 9. Prefrontal bone extends laterally across infraorbital canal (Fig-

nasal (1); yes and reaches nasal (2). ure 4). No (0); yes (1).

6. Nasal bone projects laterally over maxilla anteriorly. No (0); 10. Size of parietal foramen in adults (Figure 5). Relatively small,

yes (1). its lateral diameter less than its distance from the lateral edge of

7. Size of lateral process of prefrontal (Figure 2). Relatively small the parietal bone (0); large, its lateral diameter more than its

(0); large and extended laterally (1). distance from the lateral edge of the parietal bone (1).

8. Width of posterior face of prefrontal bone in orbit relative to 11. Body of parietal bone relative to its supratemporal processes

width of posterior part of palatine (Figure 4). Relatively broad (Figure 6). Upper surface of body of parietal bone relatively flat

Fig. 6 Lateral profiles of supratemporal process (left) and body of

parietal bone (right), arrow indicates border between the two regions.

a. Upper surface of body of parietal bone running more or less smoothly

into upper margin of supratemporal process (P. mystaceus). b. Upper

suface of body of parietal abruptly raised relative to upper margin of the

supratemporal process (P. persicus). c. Similar, but upper surface of

body of parietal tuberculated (P. scutellatus).

and running more or less smoothly into upper margin of supra-

temporal processes which is relatively flat (0); upper surface of

body of parietal bone abruptly raised relative to upper margin of

supratemporal processes (1).

Other skeletal features

12. Number of scleral ossicles. Twelve (0); eleven (1); ten in some

individuals (2).

ERSTE Eo

KG FO EES S

HRT Vg

i) JM

Wek

Peers Lt a SI ee

SPE IO TETRIS |

mh S)

AT PED

og,

{}

C's

13%

14.

E.N. ARNOLD

Acrodonta including Agamidae usually have 12 scleral ossicles

in each eye instead of the usual lizard number of 14. While

Bufoniceps possesses 12 there is further reduction in

Phrynocephalus: most species and individuals have 11 ossicles

but some members of at least a proportion of species in the P

interscapularis group have 10. This is true of P. interscapularis,

P. luteoguttatus, P. sogdianus and P. ornatus. Occurrence of 10

ossicles may in fact be wider, but the other two members of the

P. interscapularis group (P. eutilopus and P. clarkorum) are

known from relatively few specimens, so checks on ossicle

number have been very limited in these.

Number of presacral vertebrae. Usually 22, occasionally 23 in

some species (0); usually 21, occasionally 20 (1).

Substantial data on presacral vertebral number are given by

Whiteman (1978) and my own observations confirm his. Excep-

tions to the usual numbers occur in some species but nearly

always constitute a small minority of not more than about 15%

of individuals.

Number of caudal vertebrae. Usually 40—50 or more (0); usually

less than 40 (1).

Again, my own observations confirm data given by Whiteman

(1978).

External features

16.

18.

Largest individuals exceed 60mm from snout to vent. Yes (0); no

(1).

Outline of body viewed from above. Robust and rounded (0);

more slender (1).

Position of nostrils relative to line joining anterior corners of

eyes when head viewed from in front (Figure 7). Nostrils clearly

below line (0); nostrils intersecting line or above it (2); interme-

diate (1).

Differences in position of the nostril are associated with differ-

ences in the conformation of the distal limb of the tubular nasal

vestibule. The proximal limb of the vestibule is more or less

vertical in all cases, running downwards from its connection

with the primary nasal chamber. Where the nostril is low, the

distal limb of the vestibule is relatively short and runs obliquely

upwards and outwards from the base of the proximal limb to the

nostril. In animals where the nostril is high the distal limb runs

more or less vertically upwards parallel to the proximal limb and

is about as long as this.

Number of internasal scales between the nasal scales (Figure 7).

Usually two or more — 0; usually one or nil — 1.

Fig.7 Anterior views of heads showing differences in nostril position and in number of scales between nasal scales. a. Nostrils lower and separated by

two or more internasal scales (P. theobaldi). b. Nostrils high and nasal scales in contact or separated by a single internasal scale (P. arabicus).

PHRYNOCEPHALUS PHYLOGENY

eee ey |

ey J ee

Fig. 8 Left side of head showing differences in number of horizontal rows of scales immediately above the supralabials counted below the anterior eye,

and in the size of the subocular and anterior temporal scales. a. 2 or 3 rows above supralabials, subocular and one or more anterior temporals enalarged

and elongate (P. clarkorum); b. 4-5 rows above supralabials, suboculars and anterior temporals not enlarged (P. golubevi).

Paann N

suena seeeeeesti

sud

cine

pea A

Fig.9 Ventral views of underside of head showing differences in scalation. a. Scaling more or less uniform (P. arabicus). b. Scaling heterogeneous, with

curved lateral row of enlarged scales, a large central patch of enlarged pointed scales, and scales at sides of throat, behind level of angle of mouth, very

small and granular (P. euptilopus).

19. Single internasal scale with a vertical keel. No (0); yes (1).

20. Number of horizontal rows of scales immediately above the

7AM

22.

24.

supralabials, counted below anterior part of eye (Figure 8).

Usually three rows, occasionally two (0); usually four or even

five rows (1).

Enlarged subocular scales (Figure 8). Not or only weakly

differentiated (0); one or more enlarged, keeled, antero-

posteriorly elongated scales (1).

One or more enlarged, diagonally keeled and elongated scales

on anterior temporal region (Figure 8). No (0); yes (0).

External ear opening. Present (0); absent (1).

A lateral row of enlarged throat scales beginning in mental area

and curving backwards and outwards usually to the vicinity of

the angle of mouth, separated from lower labial scales anteriorly

by one to three rows of scales (Figure 9). No (0); yes (1).

DE:

26.

Pile

28.

2).

Enlarged scales in curved lateral row on throat keeled. No (0);

yes (1).

A large central patch of enlarged pointed scales on throat, the

more postero-lateral ones directed outwards and backwards

(Figure 9). No (0); yes (1).

Scales at sides of throat, behind level of angle of mouth very

small and granular (Figure 9). No (0); yes (1).

Some scales on posterior temporal region and on sides of

anterior neck enlarged, elongate and pointed, and directed

outwards and upwards. No (0); yes (1).

Distinct enlarged, raised, often pointed tubercles on dorsum of

body. No (0); yes (1).

Tubercles are enlarged scales that project markedly above

the general level of the dorsal body skin. They may be sex-

ually dimorphic in Phrynocephalus, often being more strongly

we: A

aprons 4

'2e52

Paes

Fig. 10 Left side of tail base in P. roborowskii, showing enlarged spinose

scales.

30.

31.

32%

33).

34.

315),

developed in males than females. Tubercles are frequently

clumped, especially anteriorly, and there is considerable varia-

tion in the number associated in such groupings. Tubercle form

is also variable and is especially narrow, pointed and elongate in

P. forsythii, which also shows particularly strong sexual dimor-

phism. Although the presence of tubercles is usually a clear-cut

condition, their development is sometimes sporadic and weak.

For instance, many P. theobaldi lack them but a few animals

have somewhat enlarged scales that are raised and form weak

tubercles posteriorly.

Scales at sides of tail base distinctly enlarged and often spinose

(Figure 10). No (0); yes (1)

Horizontal fringe of pointed upturned scales on posterior sur-

face of proximal thigh. No (0); yes (1).

Subdigital lamellae on distal part of fourth toe of pes (Figure

11). With two or more keels or at least projections from the free

edges of the lamellae (0); with a single keel or none (1;

Narrow light longitudinal stripes often present on flanks. No (0);

yes (1).

Dark pigment frequent in mid-line area of belly in adults. No (0);

yes (1).

Distal tail often with substantial dark pigment at least ventrally,

where it may form transverse bars. No (0); yes (1).

Soft parts

36. Palatal flaps. Large (0); reduced or absent (0)

. Tympanum. Well developed and robust(0); reduced to a delicate

membrane (1); absent (2). This and other ear features of

acromiotrapezius occiput

episterno-

scapulodeltoideus

cleidomastoideus

E.N. ARNOLD

Fig. 11 Underside of fourth toes of pes (anterior edge above), showing

extent of lateral fringes of pointed scales and number of keels on

subdigital lamellae. a. Fringes small, especially anteriorly, two keels

distally (P. theobaldi). b. Large fringes, single keels (P. mystaceus).

Bufoniceps and Phrynocephalus are discussed further elsewhere

(Armold, submitted).

38. Pars inferior of extracolumella. Large (0); small or absent (1)

39. Pharyngeal opening of middle ear. Large, length 15-25% of

head length (0); distinctly reduced, length about 10-14% of

head length (1); minute or absent (2).

40. Episterno-cleidomastoideus muscle present (Figure 12). Yes (0);

very reduced (1); absent (2).

41. Episterno-cleidomastoideus muscle a single strap (Figure 12).

Yes (0); with two branches (1).

42. Episterno-cleidmasoideus muscle extends anteriorly to occiput

(Figure 12). No (0); yes (1).

43. Scapulodeltoideus muscle extends upwards immediately ante-

rior to insertion of acromiotrapezius muscle on scapula. No (0);

yes (1).

44. Origin of caecal artery on doral aorta (Figure 13). Anterior and

close to mesenterica cranialis artery and well posterior to coeliac

artery (0); close to and usually in front of coeliac artery, occasion-

ally behind (1).

The caecal artery, which arises from the dorsal aorta and sup-

plies the intestine, exhibits interspecific variation in the position

of its origin on the aorta, relative to the origins of the coeliac

artery, which runs to the stomach, and the mesenterica cranialis

artery, which like the caecal artery supplies the intestine (Henke,

1974). In at least some Sitana and Draco, and inAcanthocercus,

Fig. 12 Diagramatic representations of superficial muscles of the right shoulder and neck. a. Episterno-cleidomastoideus muscle a single strap not

extending to the occiput; no dorsal extension of scapulodeltoideus muscle anterior to insertion of acromotrapezius muscle. b. Episterno-cleidomastoideus

muscle divided, the anterior branch reaching the occiput; a dorsal extension of scapulodeltoideus muscle anterior to insertion of acromiotrapezius muscle

present.

PHRYNOCEPHALUS PHYLOGENY

a b c

caecal

coeliac coeliac coeliac

caecal

m. cranialis m. cranialis m. cranialis

Fig. 13 Variation in position of the origin of the caecal artery on the

dorsal aorta in Phrynocephalus. a. close to origin of mesenterica

cranialis artery. b, c. Close to origin of coeliac artery.

Xenagama, Agama s. str., Pseudotrapelus and Trapelus, the

caecal artery originates well posterior to the coeliac artery and

close to and anterior to the mesenterica cranialis artery (Figure

13a). Ina wide range of agamids, including Laudakia, the caecal

and coeliac arteries originate close together, with the former

usually, although not always, anterior (Figure 13b, c). (Informa-

tion from Henke, 1974 and personal observations).

Within Phrynocephalus some species exhibit an anterior origin

of the caecal artery, either a short distance in front of that of the

coeliac artery or, much less commonly, just posterior to it. In

contrast, the remaining members of the genus and Bufoniceps

Ro t=

oO = 2 o

o o

ny ag wm 2 ES gr 8. io 8 2

Deo as OFS

Hoe Oo oS US GU Ol! CUM US CUCU

3 o fF 5 as <~< «& ° 2 ‘3s 2 2

= = o oO =I a oO oO D Ss =)

= 5 > © © £ & =F = = ° 5 Oo

s s EP ce f=) (0 oO, t0° 2 = % “Se

[= (0) Efe No > et: Cn «nn > eo = ae « Sa» 2

88 98

P. reticulatus

show a posterior origin close to the mesenterica cranialis artery.

Other characters

45. Viviparous, giving birth to fully-formed young. No (0); yes (1).

46. Tail used frequently in intraspecific signalling. No (0); yes (1).

Hemipenial features

It has been suggested that features of the hemipenis delineate species

groups within Phrynocephalus (Semenov & Danayey, 1989). These

authors illustrate apparent differences in lobe length and in whether

calyces are present on the lobes. However, personal observations of

a wide range of species, including P. mystaceus, P. maculatus, P.

arabicus, P. euptilopus, P. interscapularis, P. helioscopus, P.

theobaldi, P. vlangalii, P. guttatus, P. versicolor and P. przewalskii,

suggest that the hemipenis in these forms is consistently deeply

lobed with a honeycomb structure on the outer lobe surfaces.

Possibly the differences described by Semyonovy and Danayev result

from examining hemipenes preserved in different stages of eversion.

PHYLOGENETIC ANALYSIS

The data set (Appendix 1) consists of 46 characters most of which

are binary but eight include three states. Trapelus and Laudakia were

used as alternative outgroups. Analysis was initially carried out

using the Hennig86 program (Farris, 1988) with the options ie- and

bb*, which apply branch swopping to a single tree certain to be of

minimum length. When characters were ordered and Trapelus used

as the outgroup, two trees of 110 steps were produced with a

consistency index of 0.49 and a retention index of 0.79. With

Laudakia as the outgroup two trees were again produced, with a

length of 112 steps, consistency index 0.48 and retention index 0.79.

P. raddei

P. przewalskii

P. guttatus

P. versicolor

P. persicus

P. rossikowi

P. strauchi

P. helioscopus

P. axillaris

P. forsythii

P. viangali

P. roborowskii

P. theobaldi

Fig. 14 Estimate of phylogeny of Phrynocephalus and Bufoniceps using Trapelus as an outgroup. Tree produced by parsimony analysis using branch and

bound on a tree guaranteed to be of shortest length. Figures indicate degree of bootstrap support, only that of 57% or above being shown.

ie)

Trapelus

Bufoniceps

P. mystaceus

P. maculatus

P. arabicus

P. ornatus

P. clarkorum

P. euptilopus

P. luteoguttatus

P. interscapulari

P. sogdianus

P. scutellatus

P. golubevi

P. reticulatus

P. raddei

P. przewalskii

P. guttatus

P. versicolor

P. persicus

P. rossikowi

P. strauchi

P. helioscopus

P. axillaris

P. forsythi

P. viangali

P. roborowskii

P. theobaldi

Fig. 15 Tree in Figure 14 after being subjected to successive approximations character weighting using Hennig86 program (Farris, 1988), resulting in P

scutellatus and P. golubevi being resolved as successive branches. Characters that define lettered nodes are as follows (brackets indicate some degree of

parallelism; R indicates reversal). A 17, 18, (32); B 1.1, 12.1, 23, 35, 37.1, 46; C 1.2, 44; D 15; E 37.2, 38, 39.2; F 3, 12.2?, (21), 24, (36), 42; G (14), 28, 43;

H 16, 22, 33; 125, 26, 27; J (8), 19, 31; K 13, 32R; L 10, 20; M (4), 17.2R; N 17.1R; O 18R, 44R; R (29); S 29; T 30; U 6, (8), (34), 38R, (39.2R), 45.

Trapelus

Bufoniceps

P. mystaceus

P. arabicus

P. maculatus

P. ornatus

P. clarkorum

P. euptilopus

P. luteoguttatus

P. interscapularis

P. sogdianus

P. scutellatus

P. golubevi

P. reticulatus

P. raddei

P. przewalskii

P. guttatus

P. versicolor

P. persicus

P. rossikowi

P. strauchi

P. helioscopus

P. axillaris

P. forsythii

P. viangali

P. roborowskii

P. theobaldi

Fig. 16 Conservative estimate of phylogeny for Phrynocephalus and Bufoniceps. Only nodes supported by two or more characters of low homoplasy are

shown.

PHRYNOCEPHALUS PHYLOGENY

In both cases the consensus has the same topology (Figure 14).

When all characters were unordered, trees of 102 steps were pro-

duced which are congruent with those where characters were ordered,

but with less resolution in the clade consisting of P. przewalskii and

its nearest relatives (the topology of this region of the tree is the same

as that shown in Figure 16.).

When all these analyses were repeated using the ‘heuristic search’

option of the PAUP 3.1.1 programme (Swafford, 1993), results were

identical. Bootstrapping (100 replicates), using this programme,

was also applied to the ordered tree rooted on Trapelus and nodes

with bootstrap support over 50% are indicated in Figure 14.

Use of the successive approximations character weighting option

in Hennig86 produced little change in the original tree based on

unordered characters and rooted on Trapelus, merely resolving the

trichotomy in the consensus tree involving P. scutellatus and P.

golubevi, by making them successive branches on the main lineage

of Phrynocephalus.

Principal states supporting nodes are shown in Figure 15. It will

be seen that some 13 nodes are supported by two or more conserva-

tive characters that show little or no homoplasy. The other nodes are

defined by single or noisy characters. A conservative tree recognis-

ing the nodes based on the former features, or with bootstrap support

above 50% (and in many cases both) is shown in Figure 16.

Several nodes on the main lineage of Phrynocephalus are quite

well supported and a number of other subclades can be recognised.

Thus six species constituting a holophyletic group with marked

internal structure form the Phrynocephalus interscapularis group

consisting of P. interscapularis, P. sogdianus, P. euptilopus, P.

luteoguttatus, P. clarkorum and P. ornatus. The clade has geographi-

cal coherence, occurring in western Pakistan, Afghanistan, eastern

Iran and adjoining central Asia. Another well defined clade, the P.

theobaldi group, includes P. theobaldi, P. roborowskii and the rather

more different P vlangalii. The similar tuberculated species, P

helioscopus, P. persicus, P. strauchi and P. rossikowi may form

another unit, although it lacks marked bootstrap support.

DISCUSSION

Biogeography

Ananjeva and Tuniyev (1992) speculate about the history and bioge-

ography of Phrynocephalus in the former USSR. Their complex

hypothesis is difficult to assess as it is not based on an estimate of

phylogeny for the species concerned and does not include other

members of the Phrynocephalus clade.

Phrynocephalus is a characteristic element of the deserts of

Palaearctic Asia, like the lacertid genus Eremias and the gecko

assemblage including Cyrtopedion, Agamura, Bunopus, and Crosso-

bamon etc. Its area cladogram is not shared with these other taxa and

there is substantial sympatry between species and species groups. It

therefore seems likely that parts of the genus dispersed into at least

some areas of its huge range. The estimate of phylogeny suggests that

the ancestor of the present species occurred in the south of the present

distribution of Phrynocephalus, possibly within the area running

from western Arabia to northwestern India. This region appears to

contain the primary range of Trapelus, which may be the sister of the

Bufoniceps + Phrynocephalus clade, and Bufoniceps itself occurs in

northwest India. Many of the basal branches of mainPhrynocephalus

lineage are found wholly or partly in this area, including P. maculatus

(Arabia to Pakistan), P- arabicus (Arabia), some members of the P.

interscapularis group (S.Afghanistan, SW. Pakistan) andP. scutellatus

(central and eastern Iran, S. Afghanistan and SW. Pakistan).

From this putative source area, there may have been at least a

triple invasion of the presently warm and arid lowland regions of

central Asia (Turkmenistan, Uzbekistan, Tadzhikistan, Kirgizstan,

southern Khazakstan): by the P. mystaceus and P. interscapularis-

sogdianus lineages and by the ancestor of P golubevi and the

members of its sister group (shown in Figure 15, 16). The latter

invasion has given rise to a series of taxa in the area ( including P

golubevi, P. reticulatus, P. raddei and the P. helioscopus group).

There was then apparently eastward spread: into the Tibetan

region, by the ancestor of the P. theobaldi group and perhaps P.

forsythii, and further north into Northwest China and southern

Mongolia. On the basis of morphology, it is not clear whether

extension into the latter region represents a single invasion and

radiation or independent invasion by several lineages.

A variety of additional movements by particular lineages has also

occurred. For instance, although within the P. helioscopus group P.

strauchi and P. rossikowi have relatively small allopatric ranges, P.

helioscopus is widespread in former Soviet Central Asia and the

very similar P. persicus on the soutwestern periphery of the range of

this species extends into eastern Turkey and Iran. P. guttatus now has

a broad distribution from northwest China westwards as far as the

north Caspian area.

Unfortunately, there is little or no fossil record of Phrynocephalus

and its immediate relatives. Material assigned to Phrynocephalus

has been reported from the Pliocene of eastern Turkey (Zerova &

Chkhikvadze, 1984), but the precise relationships of these fossils are

unknown and it is not even certain whether they represent a member

of the clade made up of all present species of Phrynocephalus or if

they fall outside this grouping.

This arrangement of branches on the main lineage of the

Phrynocephalus-Bufoniceps clade correlates with species distinct-

ness. As noted, the older southern side-branches comprise very well

differentiated taxa, whereas later ones in central Asia often involve

more similar species and this trend is especially marked among the

relatively recent, more terminal branches in the Northwest China-

Southern Mongolia region, where species are very variable, their

boundaries poorly defined and their taxonomy often confused.

Structural niche

Most members of the majority of genera in Moody’s Group 6

(Moody, 1980) climb to some extent. This is true of Laudakia, most

Acanthocercus andA gama s. stt., Pseudotrapelus and mostTrapelus.

Members of the latter genus, the likely sister-group of Bufoniceps +

Phrynocephalus, spend a lot of time on the ground but many of them

also climb in bushes. In contrast to these, Bufoniceps and

Phrynocephalus themselves are strictly ground-dwelling, a derived

condition.

There has been dispute as to whether the ancestral spatial niche of

Phrynocephalus is soft, wind-blown sand. This is suggested by

Chernov (1948), Whiteman (1978) and Semenov (1987), but Golubev

(1989) and Ananjeva & Tuniyev (1992) consider the group arose in

gravel and sandstone deserts. The estimate of phylogeny presented

here supports the former hypothesis, with Bufoniceps and three of

the four basal external branches of the main Phrynocephalus lineage

being found in loose-sand habitats. (References to use of soft-sand

habitats: P. mystaceus — Ananjeva & Tuniyev, 1992; P. arabicus —

Arnold, 1984, Gallagher & Arnold, 1988; P. clarkorum and P. orna-

tus — Clark, 1992; P. luteoguttatus — Minton, 1966; P. euptilopus —

Smith, 1935; P. interscapularis — Ananjeva & Tuniyev, 1992; P.

sogdianus — Bannikoy et al., 1979). Shifts to firmer ground occurred

in P. maculatus and independently in the ancestor of the clade

containing P. scutellatus and its sister group. There was some

subsequent shift back to looser substrates in P. guttatus (Ananjeva &

Tuniyev, 1992) and P. przewalskii.

Another indication that aeolian sand habitats are primitive is that

a number of features conferring performance advantage in such

environments first appear on the internal branch of the phylogeny on

which these habitats are entered, that is the ancestral lineage of the

Bufoniceps + Phrynocephalus clade. These are discussed below.

Changes in morphological features

Principal changes in morphology in the history of the Bufoniceps-

Phrynocephalus clade are listed in the caption of Figure 15. A high

proportion of the characters in the data set (Appendix 1) show a

single change on the phylogeny. Overt reversals occur in such

features as size (in P. euptilopus) and the pattern of arteries arising

from the aorta. Simple parallelisms are quite frequent in the remain-

ing characters, but few of these are really noisy.

Body size decreases early in the history of the main lineage of

Phrynocephalus. Many features that appear likely to confer per-

formance advantage in aeolian sand habitats develop at the base of

the Bufoniceps + Phrynocephalus clade and, as noted, are concur-

rent with entry into such habitats. These features include: lateral

fringes of elongate scales on the digits that prevent the feet sinking

into soft surfaces (Carothers, 1986); reduction of the keeling on the

digital lamellae, which may be less necessary to reduce heat intake

in soft-sand environments (Arnold, 1998); fringes of elongate scales

along the edges of the eyelids, countersunk jaws, valvular nostrils,

and a U-shaped nasal vestibule consisting of vertically parallel and

subequal proximal and distal limbs, all of which features appear to

exclude sand (Stebbins, 1943, 1944, 1948), although very long nasal

passages may also protect the main nasal cavity from desiccation;

skin covering the tympanum that may protect it from damage during

burial activity, and lateral prefrontal processes that possibly protect

the eyes during the same process.

Some of these features initially associated with aeolian sand

habitats persist in less basal forms that occur on firmer substrata.

Thus, toe and eyelid fringes and countersunk jaws occur in all

Phrynocephalus, although they are less marked in species that are

not found on loose sand. The outer limb of the nasal vestibule is

shortened in most firm-ground forms, a shift associated with the

changed position of the nostril (p. 5). This feature represents a

reversion towards the primitive condition found in other Group 6

agamids. It is also associated with increased contact between the

maxillary and nasal bones, either directly or via the septomaxilla.

These nasal features occur in more terminal Phrynocephalus species

on the main lineage of the genus and have developed in parallel in P.

maculatus.

Other changes loosely associated with shift to firmer substrates

include reduction in size of the lateral processes of the prefrontal

bones, reduction in number of presacral vertebrae, increase in

number of scale rows above the upper labial scales, increase in size

of the parietal foramen of the skull and reversal in the pattern of the

arteries arising from the aorta.

The high altitude P. theobaldi group is characterised by a number

of features, including viviparity, something that often develops in

cold conditions (Shine, 1985). Within this group, P. vlangalii devel-

ops a nostril structure that is even more reversed than in other

firm-ground forms.

The external and middle ear is heavily modified in the early

history of the main Phrynocephalus lineage, the tympanum disap-

pearing, the extracolumella decreasing in size and the pharyngeal

opening becoming very reduced or absent. These changes may be

associated with greater use of subterranean rather than aerial vibra-

E.N. ARNOLD

tion in hearing when lying under the sand. They partly reverse in the

P. theobaldi group and perhaps independently in P. axillaris. Cer-

tainly the former species do not usually bury directly in the substratum

and use permanent burrows instead (K. Autumn, pers. comm.)

Members of the P interscapularis group possess a range of

features that are rare or absent in other Phrynocephalus (see caption

of Figure 15); their functional significance is uncertain.

Behaviour

Phrynocephalus has a number of distinctive behaviour patterns. The

appearance of burial by fast lateral oscillation of the flattened body

(discussed by Arnold, 1995) is concurrent with entry into aeolian sand

habitats at the base of the Bufoniceps-Phrynocephalus clade and, like

some morphological features already discussed, is likely to be an

adaptation to this environment. In line with this, such shimmy burial

is best developed in more basal species (e.g. Bufoniceps — Sharma

(1978), P. mystaceus, P. interscapularis — Ananjeva & Tuniyev

(1992), P. arabicus, P. scutellatus, P. reticulatus (pers. obs.)). Lateral

oscillation often persists in species secondarily occurring on harder

substrata, for instance in P. maculatus (pers. obs) and P. helioscopus

(Ananjeva & Tuniyev, 1992). In such cases this behaviour may be

modified and not necessarily always used for burial.

When sprayed with water, P. helioscopus adopts a distinctive

posture in which the hindquarters are raised and the head lowered.

Any liquid on the back then moves forward by capillary action in the

channels between the scales (and probably by gravity when enough

water is present) towards the mouth where it is ingested (Schwenk &

Greene, 1987). Presumably, such behaviour permits advantage to be

taken of even minor precipitation and condensation, something

likely to be a significant benefit in the arid regions where P.

helioscopus lives. P. arabicus from the United Arab Emirates re-

sponds to spraying very similarly (pers. obs.). As these two species

are widely separated on the estimate of phylogeny for

Phrynocephalus, this stereotyped behaviour may well be more

widespread than presently known. It could not be demonstrated in

Trapelus flavimaculatus, also from the United Arab Emirates, so it

may be confined to Phrynocephalus and possibly Bufoniceps.

Phrynocephalus species are also distinctive in using the tail for

intraspecific signalling (e.g. Arnold, 1984; Ross, 1989, 1995). For

instance, it may be raised, curled upwards in the sagittal plane and

wagged laterally. Movements usually expose conspicuous markings

on the underside of the tail, such as a dark tip and transverse bars and

sometimes areas of bright pigment as well. Tail signalling has been

investigated for a number of Central Asian species by Dunayev

(1996), who recognises seven distinct ways in which the tail may be

used (Dunayev, Figure 3). Of the species considered in the present

paper, the following are listed as investigated: P mystaceus, P.

maculatus, P. interscapularis, P. sogdianus, P. reticulatus (as P.

ocellatus), P. raddei, P. strauchi, P. helioscopus, P. versicolor and P.

guttatus. When data for P. arabicus (Ross, 1995) is incorporated, it

is apparent that more basal forms on the main Phrynocephalus

lineage have less complex tail displays than the others. When the

seven display features are treated as two-state characters (absent or

present) and subjected to parsimony analysis on their own, they

produce the following consensus tree which is congruent with the

estimate of phylogeny based on morphology: (P mystaceus, P.

maculatus (P. arabicus (all other species))). However, the supposed

P. maculatus on which Dunayev’s observations were based are from

the small area of Tadjikistan where P. golubevi occurs, a species

which was previously not separated from P maculatus. If the

animals concerned are in fact P. golubevi, the tree based on tail

signalling is no longer congruent with that from morphology.

PHRYNOCEPHALUS PHYLOGENY

Ecological analogues of Phrynocephalus

Small diurnal lizards, that are sit-and-wait foragers, have high body

temperatures when active and in many cases signal with their tails,

are found in several desert systems. Apart from Phrynocephalus,

they include the agamids Ctenophorus and Tympanocryptis in Aus-

tralia, the phrynosomatid sand lizards in North America (Uma,

Callisaurus, Holbrookia and Cophosaurus), tropidurines in south

America (Leiolaemus), geckoes in southern Arabia and Somalia

(Pristurus) and lacertids in Southwest Africa (Meroles anchietae).

However, although they show significant parallels in morphology

and behaviour, these derived features are not necessarily assembled

in the same order (Arnold, 1994).

Nomenclature

As presently understood, Phrynocephalus is a well-defined clade

defined by six synapomorphies not found in closely related agamids

(numbers 1.1, 12.1, 23, 35, 37.1 and 46 in the present data set).

Besides lacking these, Bufoniceps, the sister taxon of

Phrynocephalus, possesses at least one apomorphy not found in the

latter genus, namely a very short tail. Golubev & Dunayev (1997)

suggested that Bufoniceps should be expanded to include P.

mystaceus, P. maculatus, P. arabicus, P. ornatus, P. clarkorum, P.

luteoguttatus, P. euptilopus, P. interscapularis and P. sogdianus.

These are all basal members of Phrynocephalus and their inclusion

in Bufoniceps would create a new grouping that is clearly paraphyletic

and reduce Phrynocephalus to a smaller and less well defined clade.

The suggestion should consequently be rejected.

ACKNOWLEDGEMENTS. I am grateful to Jens Vindum (California Acad-

emy of Sciences for the loan of material of Phynocephalus roborowskiii,

P. rossikowi and P. strauchi. Ed Wade drew Figures 7-11.

REFERENCES

Ananjeva, N.B. & Sokolova, T.M. 1990. The position of the genus Phrynocephalus

Kaup 1825 in agamids system. Trudy Zoologicheskogo Institituta, Akademiya Nauk

SSSR, Leningrad, 207: 12-21. (In Russian, English Summary)

Ananjeva, N.B. & Tuniyey, B.S. 1992. Historical biogeography of the Phrynocephalus

species of the USSR. Asiatic Herpetological Research, 4: 76-98.

Arnold, E.N. 1984. Ecology of lowland lizards in the eastern United Arab Emirates.

Journal of Zoology, London, 204: 329-354.

1992. The Rajasthan toad-headed lizard, Phrynocephalus laungwalaensis (Rep-

tilia: Agamidae), represents a new genus. Journal of Herpetology, 26: 467-472.

— 1994. Do ecological analogues assemble their common features in the same

order? An investigation of regularities in evolution, using sand-dwelling lizards as

examples. Philosophical Transactions of the Royal Society London B, 244: 277-290.

1995. Identifying the effects of history on adaptation: origins of different sand

diving techniques in lizards. Journal of Zoology, 235: 351-388.

1998. Structural niche, limb morphology and locomotion in lacertid lizards

(Squamata, Lacertidae); a preliminary survey. Bulletin of the Natural History Mu-

seum, London (Zoology), 64: 63-90

Submitted. Modes of ear reduction in iguanian lizards (Reptilis: Iguania),

different paths to similar ends.

Baig, K. J. & Bohme, W. 1997. Partition of the ‘Stellio’ group of Agama into two

distinct genera: Acanthocercus Fitzinger, 1843, and Laudakia Gray, 1845 (Sauria:

Agamidae). Herpetologia Bonnensis, 1997; 21-25.

Bannikoy, A.G., Darevskii, I.S., Ischenko, V.G., Rustamov, A.K. & Shcherbak, N.N.

1977. Opredelitel Zemnovodnykh i Presmykayushchikii Fauny SSR. Moscow:

Prosveshenie

Carothers, J.H. 1986. An experimental confirmation of morphological adaptation: toe

fringes in the sand-dwelling lizard Uma scoparia. Evolution, 40: 871-874.

Chernoy, S.A. 1948. Reptilia. In E. N. Pawlovsky and B. S. Vinogradov (eds) The

Animal World of the USSR, Vol.2 The Zone of Desert. Pp. 127-161. Moscow: USSR

Academy of Science Press. (In Russian)

Clark, R. 1992. Notes on the distribution and ecology of Phrynocephalus clarkorum

Anderson & Leviton 1967 and Phrynocephalus ornatus Boulenger 1887 in Afghani-

stan. Herpetological Journal, 2: 140-142.

Dunayey, E.A. 1996. On the possible use of the ethological features in the taxonomy

and phylogeny of toad agamas, Phrynocephalus (Reptilia, Agamidae). Russian

Journal of Herpetology, 3: 32-38.

Farris, J.S. 1988. Hennig86 program and documentation. Port Jefferson, New York:

published by the author.

Gallagher, M.D. & Arnold, E.N. 1988. Reptiles and amphibians of the Wahiba Sands,

Sultanate of Oman. Journal of Oman Studies Special Report No.3: 405-413.

Golubey, M.L. 1989. Phrynocephalus guttatus (Gmel) or Ph. versicolor Str. (Reptilia,

Agamidae): what Phrynocephalus species occurs in Kazakhstan? Vestnik Zoologii, 5:

38-46.

— & Dunayev, E.A. 1997. On the relationships between agamid genera

Phrynocephalus Kaup 1825 and Bufoniceps Arnold 1992. Abstracts of the Third

World Congress of Herpetology, 2-10 August 1997, Prague, Czech Republic. p. 86.

Henke, J. 1974. Vergleichend-morphologische Untersuchungen am Magen-Darm-

Trakt der Agamidae und Iguanidae (Reptilia: Lacertilia). Zoologische Jahrbiicher,

Anatomie, 94: 505-569.

Henle, K. 1995. A brief review of the origin and use of ‘Srellio’ in herpetology and a

comment on the nomenclature and taxonomy of agamids of the genus Agama (sensu

lato) (Squamata: Sauria: Agamidae). Herpetozoa, 8: 3-9.

Joger, U. 1991. A molecular phylogeny of the agamid lizards. Copeia, 1991: 616-622.

Leviton, A.E., Anderson, S.C., Adler, K. & Minton, S.A. 1992. Handbook to Middle

East Amphibians and Reptiles. Contributions to Herpetology, 8:\—252.

Minton, S.A. 1966. A contribution to the herpetology of West Pakistan. Bulletin of the

American Museum of Natural History, 134: 27-184.

Moody, S. 1980. Phylogenetic and Historical Biogeographical Relationships of the

Genera in the Family Agamidae (Reptilia: Lacertilia). Ph.D. Thesis, University of

Michigan.

Ross, W. 1989. Notes on ecology and behaviour with special reference to tail signalling

in Phrynocephalus maculatus (Reptilia: Agamidae). Fauna of Saudi Arabia, 10:

417-422.

— 1995. Tail signalling in populations of Phrynocephalus arabicus Anderson 1894

(Reptilia: Agamidae). Zoology in the Middle East, 11: 63-71.

Schatti, B. & Gasperetti, J. 1994. A contribution to the herpetofauna of southwest

Arabia. Fauna of Saudi Arabia, 14: 348-423.

Schwenk, K. & Greene, H.W. 1987. Water collection and drinking in Phrynocephalus

helioscopus: a possible condensation mechanism. Journal of Herpetology, 21: 134-139.

Semenoy, D.V. 1987. Systematics, ecology and behaviour of the Phrynocephalus of the

guttatus-group (Reptilia, Agamidae, Phrynocephalus). Thesis. Severtsov Insititute

of Evolutionary Morphology and Animal Ecology, Moscow.

& Danayey, Y.A. 1989. Morphology of the hemipenis and classification of lizards

of the genus Phrynocephalus (Reptilia, Agamidae). Zoologicheskii Zhurnal, 68: 59—

64.

Sharma, R.C. 1978. A new species of Phrynocephalus Kaup (Reptilia: Agamidae)

from the Rajasthan Desert, India with notes on its ecology. Bulletin of the Zoological

Survey of India, 1: 291-294.

Shine, R. 1985. The evolution of viviparity in reptiles: an ecological analysis. Biology

of the Reptilia, 15B: 606-694.

Smith, M.A. 1935. The Fauna of British India, Reptilia and Amphibia, 2 — Sauria.

London: Taylor & Francis.

Stebbins, R.C. 1943. Adaptations in the nasal passages for burrowing in the saurian

genus Uma. American Naturalist, 77: 38-S2.

Stebbins, R.C. 1944. Some aspects of the ecology of the iguanid genus Uma. Ecologi-

cal Monographs 14: 311-332.

Stebbins, R.C. 1948. Nasal structure in lizards with reference to olfaction and

conditioning of the inspired air. American Journal of Anatomy, 83: 183-221.

Stejneger, L. 1933. In Smith, M.A. Some notes on the monitors. Journal of the Bombay

Natural History Society, 35: 615-619.

Swofford, D.L. 1993. Phylogenetic Analysis Using Parsimony, version 3.1.1. Washing-

ton: Smithsonian Institution.

Whiteman, R.S. 1978. Evolutionary History of the Lizard Genus Phrynocephalus

(Lacertilia, Agamidae). M.A. Thesis: California State University, Fullerton.

Zerova, G.A. & Chkhikvadze, V.M. 1984. Review of Cenozoic lizards and snakes in

the USSR. News. Georgian SSRAcademy of Sciences, Biological Series, 10: 319-325

(In Russian).

Zhao, E-M & Adler, K. 1993. Herpetology of China. Contributions to Herpetology,10:

1-521.

Appendix 1 Data set for Phrynocephalus and its relatives. Figures above columns refer to characters listed on pp. 2—7. — indicates no data or character uncheckable or intermediate; v indicates

character variable.

11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Laudakia

Trapelus

Bufoniceps

P. mystaceus

P. maculatus

P. arabicus

P. ornatus

P. clarkorum

P. euptilopus

P. luteoguttatus

P. interscapularis

P. sogdianus

P. scutellatus

P. golubevi

P. reticulatus

P. raddei

P. rossikowi

P. strauchi

P. persicus

il?

P. helioscopus

P. forsythii

P. roborowskii

P. theobaldi

P. viangalii

P. axillaris

P. guttatus

P. versicolor

P. przewalskii

E.N. ARNOLD

PHRYNOCEPHALUS PHYLOGENY

Appendix 1 continued

27 28 29 300) sil 82 33 34 35 365 sy 38 39) 40) 41 42 43 44 45 46

Laudakia

Trapelus

Bufoniceps

P. mystaceus

P. maculatus

P. arabicus

P. ornatus

P. clarkorum

P. euptilopus

P. luteoguttatus

P. interscapularis

P. sogdianus

P. scutellatus

P. golubevi

P. reticulatus

P. raddei

P. rossikowi

P. strauchi

P. persicus

P. helioscopus

P. forsythii

P. roborowskii

P. theobaldi

P. vlangalii

P. axillaris

P. guttatus

P. versicolor

P. przewalskii

Bull. nat. Hist. Mus. Lond. (Zool.) 65(1): 15-21

Issued 24 June 1999

Rita sacerdotum, a valid species of catfish from

Myanmar (Pisces, Bagridae)

CARL J. FERRARIS, JR.

Department of Tehthyology, California Academy of Sciences, Golden Gate Park, San Francisco, CA 94118,

SYNOPSIS. Rita sacerdotum Anderson, 1879, is the valid name for the only species of the south Asian bagrid catfish genus Rita

that resides in Myanmar. This species is distinguished from other species of Rita by a comparatively short dorsal-fin spine that

never extends to the adipose fin base; palatal tooth patches, composed primarily of uniformly sized molariform teeth, that are in

broad contact across the midline anteriorly, but diverge posteriorly and terminate in an acute point; and a small eye that is only

about one-eighth to one-tenth the length of the head. Rita sacerdotum resides in the Sittoung and Ayeyarwaddy rivers, at least as

far north as Myitkyina, Kachin State. This species is redescribed and a new key to the species of Rita is provided.

INTRODUCTION

The south Asian catfish genus Rita is broadly distributed in India,

Bangladesh, Nepal, the Indus plain of Pakistan, and the Ayeyarwaddy

system of Myanmar. Jayaram (1966) provided a systematic account

of the species of the Rita, and subsequent reviews have been

provided by Misra (1976), Jayaram (1977, 1981), and Talwar &

Jhingran (1991). In each of these treatments, four species of Rita are

recognized, although the specific names used for some of the species

differ.

The identity of the single species of Rita that inhabits the

Ayeyarwaddy River, the focus of this paper, also varies among these

studies. This uncertainty has existed since the earliest accounts of

| the presence of Rita in Myanmar waters, its persistence due in part

| to the dearth of material available for study and an enigmatic

| proposal of a name for the Ayeyarwaddy species. I recently obtained

| additional specimens of this species, and this prompted me to

reexamine the question of their identity. The discovery of the

holotype of Rita sacerdotum Anderson, 1879, a specimen that has,

for all intents and purposes, been lost since the species was named,

| revealed that there is only one species of Rita in the Ayeyarwaddy

system, and that Rita sacerdotum is the valid name for that species.

METHODS AND MATERIALS

Measurements are all straight line distances. Specimen lengths are

all reported as standard length. Institutional abbreviations follow

| Levitonetal. (1985). Other abbreviations used herein are: HL — head

| length; SL — standard length.

When refering to previously published accounts of the region and

in the list of the specimens examined, I repeat the name Burma (or

Burmah), for the country now known as Union of Myanmar. In all

| other places, use Myanmar. Throughout the text, I use the officially

| accepted spellings: Yangon for Rangoon, Ayeyarwaddy for Irrawaddy,

Bago for Pegu, and Sittoung for Sittang or Sitang.

| MATERIAL EXAMINED

| Rita chrysea, 54 specimens, 61—205 mm.

INDIA: Orissa, Mahanadi River at Cuttack, K. Jayaram, CAS

54540 (6:107—130 mm, | cleared & stained). Mahanadi River basin,

Sonepur fish market, T. Roberts, 19-22 Feb 1985, CAS 61855

(43:61—205 mm). Mahanadi River at Amicut, Cuttack, K. Jayaram,

23 Oct 1954, SU 48799 (2:110—-118 mm). Bihar, Sheonath River at

Bisrampur, A. Herre, 13 Dec 1940, SU 41043 (1:106 mm).

Rita gogra, 6 specimens, 112—205 mm.

INDIA: Andhra Pradesh, Poona, Bombay Pres., A. Herre, 1940, SU

41044 (1:123 mm). Maharashtra, Godavari River, Nanden market,

K. Jayaram, 10 Feb 1955, USNM 11494 (1:112 mm). Karnataka,

Krishna River basin, Tungabahdra River or reservoir at Hospet,

Hampi, or Kampli, T. Roberts, 28 Jan—3 Feb 1985, CAS 62088

(5:138-205 mm).

Rita kuturnee, 8 specimens, 57-97 mm.

INDIA: Andhra Pradesh, Tungabhadra River, K. Jayaram, 10 Feb

1955, USNM 114950 (3:61—88 mm);Tungabhadra River, at Kurnool,

K. Jayaram, 10 Feb 1955, SU 48798 (2:57-59 mm). Karnataka,

Krishna River basin, Tungabahdra River or reservoir at Hospet,

Hampi, or Kampli, T. Roberts, 28 Jan—3 Feb 1985, CAS 62077 (1:59

mm). Maharashtra, Poona, A. Herre, 9 Apr, 1937 SU 34868 (2:83-

97 mm).

Rita rita, 16 specimens, 24-258 mm.

INDIA: Bihar, Ganges River at Patna, T. Roberts, Apr-May 1996,

CAS 92501 (1:61 mm). Uttar Pradesh, Allahabad, Ganga River, 8—

12-1974 [sic], USNM 317823 (1:214 mm). West Bengal, Hugli

River at Pulta, A. Herre, 23 Oct 1954, SU 34866 (10:70—94 mm, one

additional specimen in lot, not examined). Calcutta, A. Herre, 9 Apr

1937, SU 34867 (2:187-194 mm). Calcutta, A. Herre, 6 Apr 1937,

SU 14132 (1:258 mm). BANGLADESH: North Central Region,

Tangail District, Ganges River basin, 5 Nov 1992, CAS 92411 (1:24

mm).

Rita sacerdotum, 23 specimens, 22-690 mm.

MYANMAR: ‘3rd Defile of Irrawaddy River, Upper Burmah, Dr.

Anderson’, BMNH 1875.8.4.7 (1:690 mm, holotype of Rita

sacerdotum). Sittoung River, E. Oates, BMHN 1891.11.30:242

(1:285 mm), BMNH 1891.11.30.343 (1, disarticulated dry skeleton,

not measured). Ayeyarwaddy Division, Wa-ke-ma town market, 17

Sep 1996, Myint Pe, NRM 40631 (4:107—135 mm). Bago Division,

Bago market, 25 Oct 1997, C. J. Ferraris, Myint Pe, Mya Than Tun,

BMNH 1998.3.11.1 (1:150 mm), CAS 99210 (1:126 mm). Yangon

Division, Hlaing River, 31 Oct 1997, C. J. Ferraris, Mya Than Tun,

CAS 99309 (10:22—70 mm). Insein market (northern Yangon), July

1996, Myint Pe, AMNH 224490 (1:195 mm). Insein market, Nov

1997, Pe, C. J. Ferraris, Mya Than Tun, USNM 348211 (2:191—203

mm). Rangoon Market, A. Herre, 14 Nov 1940, SU 39869 (2:172—

184 mm).

HISTORY OF THE IDENTIFICATION OF THE

AYEYARWADDY RITA

Day (1873) provided the first mention of Rita from the Ayeyarwaddy

River in his account of the fishes of India and British Burma, under

the name Rita ritoides (Valenciennes, 1840), aname now considered

a junior synonym of Rita rita (Hamilton, 1822) (Jayaram, 1966).

Several years later Day (1877) included the Ayeyarwaddy within the

range of Rita buchanani Bleeker 1853, another junior synonym of

Rita rita. Day acknowledged that Rita ritoides might have been the

appropriate name for the species, but departed from his earlier use of

that name and, without explanation, used R. buchanani instead.

The name Rita sacerdotum was proposed in Anderson (1878

[1879]) for a species from the middle reaches of the Ayeyarwaddy

River. As noted by Jayaram (1966), several authors have attributed

the description of this species to Francis Day, presumably on the

basis of a statement in the book’s acknowledgements (Anderson,

1878 [1879]: xxiv) which states that Day ‘favored me with a list of

fishes collected on the First Expedition, and undertook the descrip-

tion of certain species’. However, the species described by Day are

those published elsewhere (Day 1870a, 1870b, 1871) and not the

ones that first appeared in Anderson (1878 [1879]). The style of

writing and the choice of anatomical characters are significantly

different from that of Day’s other published species descriptions. It

is important to note that the actual publication date for the species

description, and for the volume as a whole, differs from that on the

title page. A statement in the published corrigenda that follows the

title page clearly indicates that publication was unexpectedly de-

layed past 1878, the date on the title page, and was issued, instead,

in 1879.

Anderson’s (1878 [1879]) description of Rita sacerdotum was

based on his field observations of living examples of the species that

were treated as pets by the residents of a Buddist temple as well as

a single specimen that was secured and illustrated. The account was

published in a summary of an expedition to western Yunnan, along

with accounts of other species from Yunnan and ‘upper Burmah’.

Because of the title of the publication, some accounts have mistak-

enly cited the type locality of this species as Yunnan.

In neither Day’s (1888) Supplement to the fishes of India, nor his

modified and updated version of his earlier book (Day, 1889) is

Anderson’s Rita sacerdotum (or the other two fish species described

by Anderson) mentioned. The reason for this curious omission is

unknown. Itis possible, but highly unlikely, that Day was unaware of

Anderson’s book with its included species accounts. Day and

Anderson must have known each other, as evidenced by the above

mentioned acknowledgement of Day’s assistance by Anderson. Day

may have considered the species to lie outside the scope of his own

book, as it was described from Upper, rather than British, Burma.

For whatever reason, Day’s failure to include mention of Rita

sacerdotum in either of the two accounts he published on fishes of

southern Asia appears to have been a major factor in the subsequent

oversight of Anderson’s name.

Vinciguerra (1890) reported on a specimen of Rita from the

vicinity of Yangon, under the name Rita ritoides. He noted that his

specimen differed from the typical form of R. ritoides in the relative

length of the dorsal spine and the shape of the humeral process.

Vinciguerra compared his specimen with the description of Rita

sacerdotum, and decided that it too differed from his specimen on

C.J. FERRARIS, JR.

several features, but that the two specimens shared a comparatively

short dorsal-fin spine. On that basis, he concluded that two distinct

forms of Rita ritoides existed, one in Myanmar and one in India.

After a period of more than a half century without any mention of

Rita from the Ayeyarwaddy, Jayaram (1966) revised the genus Rita

and concluded that two species were found in Myanmar: Rita rita

and R. kuturnee (Sykes, 1839). Inclusion by Jayaram of Rita rita in

the fauna of Myanmar appears to be based solely on the literature

accounts of Day (1873) and Vinciguerra (1890). Jayaram tentatively

placed Rita sacerdotum into the synonymy of that species. All of the

specimens from the Ayeyarwaddy River, or elsewhere in Myanmar,

that were cited as having been examined by Jayaram were listed in

the account of Rita kuturnee. However, Rita kuturnee, and its widely

used junior synonym Rita hastata (Valenciennes, 1840), is a species

otherwise known only from the rivers of peninsular India. Talwar &

Jhingran (1991) doubted that R. kuturnee actually occurs in Myanmar,

even though Jayaram (1977, 1981) had continued to list it in

subsequent accounts of the distribution of that species.

Misra (1976) included Myanmar in the distribution of Rita rita,

but not that of R. kuturnee. In his abbreviated synonymy for R. rita,

there is no mention ofR. sacerdotum, and the publication ofAnderson

(1878 [1879]) is likewise missing from the literature cited. Talwar &

Jhingran (1991) similarly listed Rita rita as the only species of Rita

from Myanmar, but they tentatively included Rita sacerdotum in the

synonymy of that species.

IDENTITY OF THE AYEYARWADDY SPECIES

OF RITA

Although much of the recent literature suggests that the Rita species

inhabiting the Ayeyarwaddy River is Rita rita, the species in that

basin is, in fact, clearly distinct from R. rita. During this study,

specimens of R. rita, from various parts of the Ganges basin, the type

locality of the species, were found to exhibit characters lacking in

specimens from the Ayeyarwaddy. As first noted in Jayaram (1966),

the palatal teeth of theAyeyarwaddy Rita specimens are not arranged

in the broad, elliptical patches characteristic of R. rita but, instead, in

‘pear-shaped’ patches that tapered posteriorly nearly to a point

(Figure la). In addition, the dorsal-fin spine of R. rita is long and

stout with its length at least equal to the head length. The adpressed

spine usually extends well past the adipose-fin origin, at least in

large individuals. Day (1877) noted that the relative size of the spine

was apparently allometric, and that in small individuals it may only

equal the head length, but that in larger individuals it may exceed 1.3

times HL. In Ayeyarwaddy specimens, in contrast, the dorsal-fin

spine is never as long as the head and, more typically, it is shorter

than the head minus the snout, even in the largest specimens.

In contrast to the prevailing view, Jayaram (1966) identified the

Ayeyarwaddy specimens as R. kuturnee. It appears that his conclu-

sion is based primarily, but erroneously, on the similarity of the

palatal tooth patches in the two species. In placing the Ayeyarwaddy

Rita into R. kuturnee, he also looked beyond several striking differ-

ences between the two species. For example, the eye size of R.

kuturnee is significantly larger than that of the Ayeyarwaddy Rita. In

his diagnosis of R. kuturnee, Jayaram (1966) lists the eye size as

‘Eye 3.07 (2.70 to 4.70 of up to 8.80 in specimens from Burma) in

head length; 1.35 (1.00 to 1.50 or 3.90) in interorbital space width;

1.39 (1.00 to 1.50 or 3.00) in snout length.’ It is possible that Jay-

aram interpreted the consistant disparity in eye proportions between

the Indian and Ayeyarwaddy specimens as a result of allometric

growth in R. kuturnee. All of the specimens he examined from

ae Ra Set, ee a a

MYANMAR CATFISH

Fig. 1 Diagrammatic representation of tooth patches on jaws and palate of Rita species. A. Rita sacerdotum Anderson, 184 mm, SU 39869. Scale bar

= 1 mm.; B. Rita kuturnee (Sykes), 97 mm, SU 34868.

peninsular India are substantially smaller (36 to 103 mm) than any

of the specimens listed from Myanmar (184 to 318 mm). In the

Ayeyarwaddy specimens that I examined (22 to 285 mm), the eye

length is always 8 to 10 times in the head length. Curiously, despite

the observation that the Ayeyarwaddy specimens has small eyes,

Jayaram used the eye size of R. kuturnee to help distinguish it from

R. chrysea, a species for which he lists the eye diameter as ‘3.76

(2.83 to 5.22)’ in head length. Clearly he did not take into account

the Ayeyarwaddy specimens in this diagnosis of R. kuturnee.

The shape of the palatal tooth patches, a characteristic on which

Jayaram placed heavy emphasis, also differs between Rita kuturnee

and the Ayeyarwaddy form. The ‘pear-shaped’ tooth patches that

Jayaram (1966, Figure 1b) described and illustrated as characteristic

of R. kuturnee appear, in fact, to be those of the Ayeyarwaddy river

species, and not R. kuturnee. In all of the specimens of R. kuturnee

that I examined, the palatal tooth patches are slender, crescent-

shaped arches that are either separated at the midline (Figure 1b), or

meet only for the width of a single row of teeth. The palatal tooth-

patches in Rita kuturnee have stout conical teeth, larger in size than

those of the premaxilla, rather than the broadly rounded, or molari-

form ones that predominate in the palate of the Ayeyarwaddy species

(Figure la). As with the size of the eye, it is possible that Jayaram

assumed that his specimens of R. kuturnee from peninsular India

were juveniles, with incompletely developed palatal tooth patches,

and that the adult condition in the peninsular population is like that

in the Ayeyarwaddy specimens. Even in the smallest examined

specimens from the Ayeyarwaddy basin the palatal tooth patches are

broadly in contact across the midline and are composed primarily of

molariform teeth. Thus, I conclude that the Ayeyarwaddy form is not

conspecific with Rita kuturnee.

The Ayeyarwaddy River Rita population has never been consid-

ered conspecific with either of the two other Indian Rita species, R.

chrysea Day, 1877 and R. gogra (Sykes, 1839), and I have found no

reason to assign either name to theAyeyarwaddy fishes. Rita chrysea,

restricted to the Mahanadi River and nearby tributaries in Orissa and

considered to be the smallest species of Rita (Talwar & Jhingran,

1991), is characterized by a large eye (2.8 to 5.2 in HL) and by

having a broad, nearly rectangular, patch of molariform teeth that

extends across the midline of the palate (Jayaram, 1966). Rita gogra,

which is sometimes listed as Rita pavimentata (Valenciennes, 1840)

(e.g., Misra, 1976; Talwar & Jhingran, 1991), is known only from

rivers of the Deccan region of peninsular India, including the

Krishna, Harda, Godavari, Tungabhadra, Manjra, Bhima, and Mutha-

Mula (Jayaram, 1966). Although similar in overall appearance with

the Ayeyarwaddy Rita, R. gogra can be distinguished immediately

by the unusual shape of its head. The dorsal surface of the head,

posterior to the orbits, is dominated by a bilaterally symmetrical

swelling formed by massive extensions of the adductor mandibulae

muscle that cover the cranial roofing bones. All other species of Rita,

including the Ayeyarwaddy species, have the dorsal surface of the

cranium covered only with skin, through which the cranial roofing

bones can easily be palpated. In addition, the Ayeyarwaddy Rita can

be distinguished from R. gogra by the color of the mental barbel

(black in R. gogra, white in the Ayeyarwaddy species). The palatal

tooth-patch in R. gogra has finely conical teeth anteriorly and

increasingly large molariform teeth posteriorly (Jayaram, 1966).

Thus, it must be concluded that the Ayeyarwaddy Rita is not

conspecific with any of its Indian congeners. The only remaining

name that might apply is Rita sacerdotum, which was described

from the Ayeyarwaddy. The description and published illustration of

that species, however, only vaguely resembles a Rita, and character-

istics of the Ayeyarwaddy species are either absent from the

description, or in variance with the illustration. Although Jayaram

(1966) followed Vinciguerra (1890) in placing Rita sacerdotum into

the synonymy of Rita rita, he place the specimens he examined from

the Ayeyarwaddy into a second species of Rita. I have been unable,

so far, to find any specimens that represented a second species from

Myanmar. Clearly, an examination of the holotype of Rita sacerdotum

was necessary to determine whether it indeed represented a species

of Rita different from the one that I, and others, have observed.

THE HOLOTYPE OF RITA SACERDOTUM

ANDERSON

Anderson (1878 [1879]) did not indicate where the holotype of Rita

sacerdotum, or any of the other species described in the same paper,

were deposited. Although I expected to find the specimen in The

Natural History Museum, London, the holotype was not listed in its

type catalog, and there was no entry for R. sacerdotum in their

species catalog. In fact, no specimen of Rita collected by Anderson

was listed in the catalog. An exhaustive search through the registers

did uncover a Rita sacerdotum collected by Anderson, without any

indication that it was a holotype. With the assistance of the staff of

the Fish Section of the Zoology Department, the specimen was

C.J. FERRARIS, JR.

found among the collection of stuffed, dried fish specimens. Its

identity as the holotype was promptly made by comparsion of the

stated locality information and by direct comparision with the

published illustration.

It is puzzling that the specimen was never recognized as the

holotype of Rita sacerdotum. Although the specimen was registered

in 1875, prior to Anderson’s publication, the register entry (BMNH

1875.8.4.7) lists the name and is surrounded by entries for the other

species named by Anderson. It is even more suprising that although

the specimen was registered with the new name during Albert

Giinther’s tenure, he did not include the specimen in his personal

annotated copy of his catalog (Giinther, 1868) or annotate the

less, with the discovery of the holotype, it is now possible to clarify

some peculiar features in the illustration of Rita sacerdotum and,

with that information, finally resolve the identity of theAyeyarwaddy

Rita.

The holotype of Rita sacerdotum is a dried, stuffed specimen, 69

cm in standard length (Figure 2). The specimen appears to have been

placed on display at two different times, based on the fact that the

stuffed skin has two forms of wire attachments. One set of mounts,

extending from the ventral surface of the body, indicate that the

specimen was at one time mounted freestanding, probably on a

Fig. 2 Rita sacerdotum Anderson, holotype, 69 cm, BMNH 1875.8.4.7.

Fig. 3 Published illustration of holotype of Rita sacerdotum, reproduced from Anderson (1878 [1879], pl. 79, Fig. 3).

MYANMAR CATFISH

Fig.4 Rita sacerdotum Anderson, 150 mm, BMNH 1998.3.11.1.

Fig.5 Rita sacerdotum Anderson, 126 mm, CAS 99210.

wooden stand. A second pair of wires protrudes from the left side of

the body, suggesting that the specimen was mounted on a wall, with

the right side of the body on display.

The published illustration of the holotype (Figure 3) resembles

the mounted specimen quite closely, except for some damage to the

fins. Most importantly, the elongated caudal region of the body,

which is identical in proportion to that in the illustration, suggests

that the illustration was probably prepared from the dried mount

rather than the freshly collected specimen. The body of the specimen

is disproportionally long and the caudal region is far more slender

and cylindrical than other specimens of Rita from the Ayeyarwaddy

(Figures 4, 5). This unusual body form, and the illustration that

resulted from drawing the dried specimen, have made comparison

between the illustration and fresh specimens of the species problem-

atic. On close inspection, it appears that the body of the mounted

specimen must have been stretched well beyond the normal propor-

tions of the species when it was stuffed. As Anderson collected only

a single specimen, it is reasonable to assume that he or his taxider-

mist had no model to use to shape his specimen, once it was skinned

and the vertebral column removed. Although the general shape of

the body does not closely resemble the other specimens from the

Ayeyarwaddy, other features of the body are, in fact, quite similar

and clearly indicate that the holotype and the other available speci-

mens are conspecific. The shape of the palatal tooth patches, the

unusually short dorsal spine, and small eye combine to distinguish

this species from its congeners. All of the specimens that I have

examined exhibit this same suite of characters, albeit with some

ontogenetic variation. It appears therefore that there is only one

species of Rita in the Ayeyarwaddy system, and that the oldest

available name for that species is Rita sacerdotum.

DIAGNOSIS AND REDESCRIPTION OF RITA

SACERDOTUM

As noted above, the inaccurate taxidermic preparation of the holotype

of Rita sacerdotum made the specimen longer than it would have

been in life, and this precludes using the specimen for any propor-

tional measurements standardized against the body length. Therefore,

any statement in the description that relates a body measurement to

the standard length does not include the holotype.

Diagnosis

Rita sacerdotum is readily distinguished from all congeners by the

following combination of characteristics: eye small, its diameter 10—

13% head length; dorsal-fin spine length no greater than the length

of the head posterior to the snout; adpressed dorsal-fin spine does

not extend to adipose-fin origin; and palate with a single crescent-

shaped patch of primarily large, bluntly conical, teeth of

approximately uniform size.

Description

Body elongate, slightly compressed anteriorly, progressively more

compressed toward caudal fin. Body deepest at dorsal-fin origin, its

depth at that point approximately equal to distance from nasal barbel

to opercular margin; body depth decreases gradually to adipose-fin

origin, more rapidly thereafter. Least depth at caudal peduncle

approximately equals snout length. Skin of body and head covered

with thick coat of mucous, anchored by fine filamentous projections

from skin surface; filaments largest and most dense on chin and

opercular margin of head, and, especially, on lateral surface of body

ventral to dorsal fin.

Vent slightly anterior to anal-fin origin. Lateral line midlateral

and straight from past tympanum to hypural plate; anterior portion

of lateral line more dorsally situated; lateral line bent sharply in the

dorsal direction onto base of upper caudal-fin lobe posterior of

hypural plate margin. Lateral line pores extend laterally from canal,

through thick mucous coat. Anterior canal pores ramify and spread

in asymmetric pattern over pectoral-girdle elements and tympanum.

Cephalic canal pores similarly branch over dorsum of head and onto

opercle.

Head large, its length approximately 3% times in SL; head slightly

depressed, at pectoral-fin origin its depth approximately 80% its

width; head depth at orbit approximately 2/3 its width. Dorsal profile

of head straight from orbit to snout, slightly convex posteriorly;

ventral profile nearly straight. Mouth nearly terminal; upper jaw

slightly overhangs lower. Teeth in upper jaw conical and sharply

pointed, in 6 to 8 irregular rows. Tooth-bearing surface of premaxilla

long and nearly transverse, its long axis four to five times its short

axis. Tooth-bearing surface of mandible elongated, tapering

posteriorly. Teeth in lower jaw pointed and conical along anterior

margin of jaw, approximately equal in size to those of upper jaw; two

rows of bluntly rounded teeth, much larger in size than conical teeth,

present mesially; only blunt teeth present along posterior part of

mandible. Palate with coalesced tooth patch extending across mid-

line. Tooth patch convex anteriorly, concave posteriorly, with nearly

parallel lateral margins. Teeth on palate nearly all in form of bluntly

rounded pegs, slightly larger in diameter posteriorly, except for one

or two rows of somewhat smaller teeth along lateral and anteriolateral

margins of toothplate. Gill rakers 24 to 29; anterior 8 to 10 rakers on

lower arch rudimentary, shorter than intervening spaces; posterior

rakers moderately long and thick.

Eye small, ovoid, with long axis parallel to body length; long

diameter of orbit approximately 1/3 snout length, 1/5 interorbital

width, and equal to or slightly greater than 1/10 head length. Orbital

margin free.

Anterior naris situated along anterior margin of snout, its opening

a short tube, flared at margin, directed anteriorly. Posterior naris

remote from anterior naris, and slightly more laterally situated; its

anterior margin located midway between snout tip and anterior

margin of orbit. Naris surrounded by short rim, connected to nasal

barbel anteriorly.

Head with three pairs of barbels. Maxillary barbel extends from

fold between upper lip and skin of snout; barbel filamentous,

without fleshy attachment to snout. Maxillary barbel short, not

extending to margin of bony opercle. Nasal barbel short, its length

approximately equal to orbital diameter; adpressed barbel reaches

only to anterior margin of orbit. Ventral surface of head with single

pair of mandibular barbels; barbel originates at vertical through

anterior orbital margin; barbel filamentous, extending to, or nearly

to, vertical through pectoral spine origin.

Dorsal surface of supraoccipital, posttemporal and pterotic bones

granular, remainder of head covered with smooth skin. Adductor

mandibulae does not extend onto dorsal surface of cranium.

Upper lip with several rows of short papillae along margin;

papillae often multifurcated at tip. Lower lip broadly connected to

C.J. FERRARIS, JR.

skin of chin, separate laterally. Lip margin with papillae comparable

to those of upper lip, at least medially.

Opercular membrane free from isthmus at margin, but attached

more basally; membranes broadly connected across midline, but

separated posterior to isthmus connection. Branchiostegal rays 7 or

Dorsal-fin origin at approximately 40% of SL. Fin quadrangular,

first ray longest and approximately two times that of last ray; last ray

without membranous extension to body; fin margin straight. Fin

base approximately 1/2 of HL and shorter than interspace between

dorsal fin and adipose fin. Dorsal-fin spine stout, with sharply

pointed tip. Spine length equals head length minus snout, or approxi-

mately 15% SL. Anterior margin of spine produced into sharp keel,

without serrations; lateral and posterior surfaces smooth. Dorsal

spine preceded by fully formed spinelet. Dorsal fin preceded by

coarsely granular predorsal bone; lateral extent of predorsal bone

approximately equals that of supraoccipital spine. Dorsal fin rays

II,7; posterior two rays appear as one, split at base.

Adipose fin large; anterior fin margin straight, convex distally. Fin

extends posteriorly well past its posterior insertion.

Caudal fin deeply forked, lobes with acutely pointed tips; lobes

slightly asymmetrical, dorsal lobe longer and sometimes with

filamentous extension. Length of dorsal most primary ray approxi-

mately three times length of middle rays. Procurrant rays few, short,

not extending anteriorly onto caudal peduncle. Caudal fin rays

iL /eeshail

Anal fin quadrangular, anterior rays longest; posterior rays pro-

gressively shorter, fin margin straight. Last ray not connected to

caudal peduncle by membrane. Fin base short, approximately equal

to that of adipose fin. Anal-fin origin slightly posterior to vertical

through adipose fin origin. Anal-fin rays iv, 9-10.

Pelvic fin abdominal, its origin posterior to vertical through

posterior insertion of dorsal fin. First branched ray longest, follow-

ing rays only slightly shorter. Adpressed fin just reaches anal-fin

origin. Pelvic-fin rays 1,6.

Pectoral fin acutely pointed; first branched ray longest, its length

approximately three times posterior-most ray. Pectoral-fin spine

stout, sharply pointed at tip. Spine with short filament at tip, length

of filamentous extension approximately equals snout length. Outer

margin of spine produced into acute keel; keel very finely serrated

for basal quarter, smooth for remainder of its length; in small

specimens, most of spine margin covered with tiny transverse

serrations. Inner spine margin with densely packed, pointed, retrorse

serrations; serration height greater than length of space between

successive serrations. Humeral process acutely pointed posteriorly,

with a slightly rounded tip. In larger specimens, process becomes

more rounded posteriorly, as in holotype (Figures 2, 3). Surface of

humeral process granular, with granulations less coarse than those of

cranial surface. Pectoral-fin rays I,10 or I,11.

Coloration in preservative

Body gray, darker dorsally, gradually becoming lighter ventrally;

abdomen nearly white. Head dark gray dorsally, white ventrally;

transition between gray and white regions fairly abrupt, occuring

ventral to eye and approximately in line with maxillary barbel

origin. Operculum gray with white margin. Orbit surrounded by

distinct white ring. Maxillary barbel dark grey, mental barbel nearly

white.

Dorsal, anal, and pectoral fins pale, with broad black margin.

Pelvic fin uniformly pale or with some indication of dark margin.

Caudal fin with fine dark margin on middle rays; darkened margin

progressively larger toward lobe tips.

MYANMAR CATFISH

Distribution

Rita sacerdotum appears to be distributed widely through the

Ayeyarwaddy River basin of Myanmar. Specimens examined during

this study were all from the lower portions of the basin, but I

observed a few large individuals in markets as far upriver as Manda-

lay. In addition, Rita is seen occasionally in the Myitkyina market (U

Tun Shwe, pers. comm.). Outside of the Ayeyarwaddy basin and its

extensive delta, there is one record of specimens from the Sittoung

River (BMNH 1891.11.30.242—243), and two from the Bago River

(BMNH 1998.3.11.1 and CAS 99210).

Natural History

Little is known about Rita sacerdotum. While small specimens, up to

about 25 cm, are routinely found in markets of Yangon and smaller

delta villages, at least during the rainy season (April to September;

U Myint Pe, pers. comm.), large specimens are only rarely seen in

markets. All of the specimens examined during this study appear to

be juveniles and there is no published indication of the size of

maturity for this species.

Individuals as small as 22 mm were obtained from the tidal rivers

in the vicinity of Yangon in November, 1997. The presence of these

tiny individuals in the lower course of the river suggests that Rita

may reproduce in the estuarine part of the river and disperse more

widely throughout the river at a larger size. This idea is supported by

anecdotal reports that large numbers of large Rita appear at an island

pagoda, in the middle reaches of the Ayeyarwaddy River, for a short

period of time during monsoon season (U Nyi Nyi Lwin, pers.

comm.). This may be indicative of a spawning migration.

Examination of the gut contents of a few specimens revealed that

Rita sacerdotum feeds on a variety of aquatic and terrestrial inverte-

brates. Several specimens contained fragments of small glass prawns,

and others contained pieces of winged insects. A comprehensive

study of the food habits is not possible at this time due to the

relatively small number of specimens available and the fact that the

specimens in collections represent only juveniles.

KEY TO SPECIES OF RITA

1. Dorsal surface of head, between eyes and occipital spine, covered with

tckalayeriot muscle; pelvic fin BACK ........cerecncescnssecceseorseccessrsuceenses

Bees sate ses Rita gogra (rivers of the Deccan region of peninsular India)

Dorsal surface of head covered only with skin; pelvic fin pale ........ 2,

PS CAS HUTA Men NS GUE nsec avacnceesscseevaveceneseecceescoseeascsteesascisieunstsatorseces 3

een O40) Je EMM a auc, sn tevnseunannct teesuaies Mhousaanecisnstaenctanscessnedeientaess 4+

3. Dorsal-fin spine as long, or longer than length of head, adpressed spine

extends to or beyond adipose-fin origin (in specimens greater than 100

mm); palatal teeth in two elliptical patches, not meeting at midline;

teeth on posterior extent of lower jaw and palate molariform, much

ADOC Ra ATID Lett OIMUC EL Dest tee are oe cece ce fs sant car cmoSew sess de Slee ceaust sesso

Dorsal-fin spine no longer than head minus snout, adpressed spine not

reaching adipose-fin origin; palatal teeth in a single crescent patch that

extends across midline of palate; teeth on palate more or less uniform in

SUZ et OUICe IVAN) eeccceer ens Nearer tee ES os aca cas cbeca caveat sdenestintscendsscocess

Rita sacerdotum (Myanmar: Ayeyarwaddy and Sittoung River basins)

4. Palatal teeth in slender patches along lateral margin of palate, no larger

than teeth in upper jaw and not meeting at midline (Figure 1B), dorsal-

fin spine smooth anteriorly, except for few serrae basally ...............0..

ee tS Rita kuturnee (rivers of Deccan region of peninsular India)

Palatal teeth in large quadrangular patch that covers most of palate; teeth

large and molariform in middle of patch, smaller laterally; dorsal-fin

spine with single row of antrorse serrae, for at least basal 2/3 of spine

ACKNOWLEGEMENTS. Examination of specimens for this study was facili-

tated by David Catania and William Eschmeyer (CAS), Darrell Siebert and

Oliver Crimmon (BMNH), and Richard Vari (USNM). Tyson Roberts pro-

vided valuable insights during early stages of this study. The photograph of

the holotype of Rita sacerdotum was arranged by Darrell Siebert and taken by

Phil Hurst (BMNH). Al Leviton provided the photographic reproduction of

the published illustration of Rita sacerdotum, from his personal copy of

Anderson (1878 [1879]). Molly Brown drew figure | and Alison Schroeer

drew figure 5. Travel to Myanmar was undertaken as part of a series of

consultancies for the United Nations Food and Agriculture Organization.

Kent Carpenter and Dora Blessich of that organization were instrumental in

making these trips possible. Collection of specimens in Myanmar was made

possible by the Myanmar Department of Fisheries; several individuals,

including U Hla Win, U Nyi Nyi Lwin, U Myint Pe and U Mya Than Tun,

provided me with assistance and valuable information. Financial support to

travel to London and to examine specimens at The Natural History Museum

was provided by TWA and the Inhouse Research Fund of the California

Academy of Sciences. Without the help of all of these persons and organiza-

tions, this study could not have been undertaken. The manuscript was

improved by comments from James Atz, Nigel Merrett, Darrell Siebert and

Richard Vari.

REFERENCES

Anderson, J. 1878 [1879]. Anatomical and zoological researches: comprising an

account of the zoological results of the two expeditions to Western Yunnan in 1868

and 1875; and a monograph of the two cetacean genera, Platanista and Orcella. 984

pp., 84 pls. Bernard Quaritch, London.

Day, F. 1870a. Remarks on some of the Fishes in the Calcutta Museum. — Part I.

Proceedings of the Zoological Society of London 1869 (3): 511-527.

1870b. Remarks on some fishes in the Calcutta Museum — Part II. Proceedings of

the Zoological Society of London 1869 (3): 548-560.

1871. Monograph of Indian Cyprinidae. Parts 1-3. Journal of the Asiatic Society

of Bengal 40 (pt 2, no. 1-4): 95-142, 277-367, 337-367, Pls. 9, 21-23.

1873. Report on the fresh water fish and fisheries of India and Burma. 307 pp.

Office of the Superintendent of Government Printing, Calcutta.

1877. The fishes of India, being a natural history of the fishes known to inhabit the

seas and fresh waters of India, Burma, and Ceylon. Part 3:369-552, pls. 79-138.

Bernard Quaritch, London.

1888. Fishes of India, supplement, October 1888. pp. 779-816. Bernard Quaritch,

London.

1889. The fauna of British India, including Ceylon and Burma. Fishes, 1. xx +548

pp. Taylor and Francis, London.

Jayaram, K.C. 1966. Contributions to the study of bagrid fishes (Siluroidea: Bagridae).

1. A systematic account of the genera Rita Bleeker, Rama Bleeker, Mystus Scopoli,

and Horabagrus Jayaram. Internationale Revue der Gesamten Hydrobiologie 51 (3):

433-450.

— 1977. Aid to the identification of the siluroid fishes of India, Burma, Sri lanka,

Pakistan, and Bangladesh. I. Bagridae. Records of the Zoological Survey of India,

Miscellaneous Publications, Occasional Paper (8): 1-41.

1981. The freshwater fishes of India, Pakistan, Bangladesh, Burma and Sri Lanka

—a handbook. viii + 475 pp, 13 pls. Zoological Survey of India, Calcutta.

Leviton, A.E., Gibbs Jr, R.H., Heal, E. & Dawson, C.E. 1985. Standards in herpetology

and ichthyology: Part 1. Standard symbolic codes for institutional resource collec-

tions in herpetology and ichthyology. Copeia 1985 (3): 802-834.

Misra, K.S. 1976. The fauna of India and the adjacent countries. Pisces (second

edition), Vol. 3, Teleostomi: Cypriniformes: Siluri. xxi + 367 pp., 15 pls. Controller

of Publications, Delhi.

Talwar, P.K. & Jhingran, A.G. 1991. Inland fishes of India and adjacent countries. xx

+ 1158 pp. Oxford & IBH Publishing Co., Pvt. Ltd. New Delhi, Bombay and

Calcutta.

Vinciguerra, D. 1890. Viaggio di Leonardo Fea in Birmania e regione vicine. XXIV.

Pisci. Annali del Museo Civico de Storia Nationale de Genova, Milano, (serie 2a) 9:

129-362, pls. 7-11.

Bull. nat. Hist. Mus. Lond. (Zool.) 65(1): 23-29 Issued 24 June 1999

Indian Ocean echinoderms collected during

the Sindbad Voyage (1980-81): 4. Crinoidea

JANET I. MARSHALL CROSSLAND KX (S| Foe. |

Museum of Tropical Queensland, 70-84 Flinders St, Townsville, Queensland 4810 Australia

ANDREW R.G. PRICE

Ecology and Epidemiology Group, Department of Biological Sciences, University of Warwick, Coventry CV4

7AL, UK

CONTENTS

A EMCNCA ULE Ck area tany Meant ee cvcetsecas caso ace MU ae Pee EEO c casa int Sac taal cs dexdacisaavucysvedersusessausucescsseseseaeepeinseenaeecucsiciessaaqeseauasesours 23

Meare tiled hg cit AMIE REL OGS be cane Savi casheat scenes die pvawceseen-eu-t sa ssaxasceuauwannanvatttansdnsnedssdacacsnueedacanceictvaccsudssicecesaedvancssetbacsdeace coacstenscctangt vasesebe 23

Nea Screens tae ee Re ah A a oe ect maaan nec Bie cay cam te obo ve dexGael duesuaedomencs ates tenons cota yevaes vnsinbucn oye MRasesveaugdevvesanvadevestcuas 23

MEI CVI GSI OTM ec cos co sean cane cnia auc pean en Seo denned d Sencans znsaps GuaSaunctevadacse detudch saacabia ss soehtoesnes asec nana vaauesach cuiiaus dicoeSbauh dewtaa dosussesasdvepecdnes 29

PS TOC LE LSE IMC INS sen oye cdeaa nears aie el td ncn cae wanes encca nasigy eater cee cab sccccnnt ener cpncass ad uosisndede conse athersueesoxeseseWacecuunarasuassteeSevaaaseseseunses 29

INCE R CM COSI eens ce atu see we ncaa Saudaxosiusaiwanseamasionatanasde tedavesaadenesesvasduscacsarctabantsncdsaesanchdesseuttovsasscobreccsdelsacesadsaccndcceassocasccasadtevexateents 29

SYNOPSIS. Thirty species of shallow-water Crinoidea, representing eighteen genera in six families, are recorded from

collections made during the Sindbad Voyage (Oman to China) from the Lakshadweep (Laccadive) Islands, Sri Lanka and Pula

We Sumatra). Following the zoogeographic subdivisions of Clark and Rowe (1971), extensions of range are recorded for at

least six of the species: Clarkcomanthus albinotus (Indonesia/East Indies); Comanthus briareus (Sri Lanka area); Comanthus

gisleni (Sri Lanka area & Indonesia/East Indies); Comanthus suavia (Sri Lanka area & Indonesia/East Indies); Comanthus

wahlbergii (Maldive area, Sri Lanka area & Indonesia/East Indies), and Oxycomanthus bennetti (Indonesia/East Indies); and

possibly also Comaster parvicirrus (Sri Lanka area — doubt about earlier record) and Comaster multifidus (Maldive area —

specimens poorly preserved). In addition to the taxonomic treatment, ecological information for each crinoid species (habitat

types, depth range) is provided and broadly analysed.

INTRODUCTION

| Systematically, crinoid taxonomy has undergone relatively few

| changes since the monumental works of A.H. Clark (1915-1967),

the major exception to this being the recent revisions to the family

Comasteridae by Hoggett and Rowe (1986) and Rowe, Hoggett,

_ Birtles and Vail (1986).

| This paper is the fourth in a series reporting the collection of

| echinoderms made during a cruise by one of us (ARGP) across the

_ Indian Ocean from Oman to China. The expedition was undertaken in

a replica of an ancient Arab sailing vessel, “Sohar’. Systematic

accounts of the other echinoderm classes have already been published

(Price & Reid, 1985; Marsh & Price, 1991; Price & Rowe, 1996).

Thirty species of shallow-water crinoids from six families are

listed, including nine new distribution records. Generally, comments

| are made where the record extends or modifies a range of distribu-

| tion, or to clarify the identification. Where no comment is offered,

| the species was already known from the region and is widespread in

| the Indo-West Pacific.

| MATERIALS AND METHODS

Specimens were collected by one of us (ARGP) and other expedition

members atlocalities from Chetlat Island, Lakshadweep (Laccadives),

Sri Lanka and Pula Wé, Sumatra (Indonesia). Details of sampling

localities are shown in Figure 1. Sampling was undertaken principally

on coral reefs using scuba. At each locality, details of habitat type and

depth range were recorded, along with the number of individuals of

each species. The number of specimens collected is placed in

parenthesis after each sample number in the Material lists for each

species. Material was fixed and preserved using standard methods

(Lincoln & Sheals, 1979). Conditions on board and for specimen

storage on ‘Sohar’ were not as sophisticated as on modern research

vessels. Hence not all specimens returned were in good condition.

Specimens were identified by JIMC. Where the identification was

uncertain, due to the changes to crinoid taxonomy by Rowe, Hoggett,

Birtles and Vail (1986) and Hoggett and Rowe (1986), confirmation

was sought from one of the authors of those papers. In some cases,

subsequent re-examination of specimens has engendered doubt, and

this doubt is expressed in the text of this paper.

. Where three or more specimens of a species were collected,

representative specimens of that species were sent to the Singapore

Museum (SM), as the regional museum; otherwise material was

divided between the Natural History Museum (NHM), London, and

the Western Australian Museum (WAM), Perth.

Species are listed in families, and within each family, alphabeti-

cally by genus and species.

RESULTS

Throughout this account synonymy has been confined, where possi-

ble, to a single reference from which the original reference can be

traced.

J.I.M. CROSSLAND AND A.R.G. PRICE .

SO E 60 70 80 90 100

Muscat

OMAN

Chetlate ANDAMANS$

LACCADIVES «,° oe

SRI LANKA

NICOBARS#

Tangalla

MALDIVES ™

Ug Bau

Fig. 1 (a) Map of northern Indian Ocean showing sampling areas (@) during Sindbad Voyage, with insert (b) for Pula Wé Sumatra.

INDIAN OCEAN ECHINODERMS

Class Crinoidea

Family COMASTERIDAE

1. Alloeocomatella pectinifera (A.H. Clark, 1911)

SEE. Clark and Rowe, 1971:6—7; Hoggett and Rowe, 1986:122;

Messing, 1995: 644.

MATERIAL. NHM — 810501C/2 (2), 810505C/1 (1); WAM —

810425F/2 (1); SM — 810421 A/1 (1).

COLLECTION SITES.

Pula Wé, Sumatra.

Sabang Bay, Seukundo, and Ug Seukundo,

HABITATS AND DEPTH RANGE. Subtidal rock/coral, coral reef, on

gorgonian; 2—30m.

COMMENTS. The species was described by A.H. Clark (1911) and

placed (with reservation) in the genus Comissia, later to be included

in anew genus Alloeocomatella by Messing (1995). The species has

been found in the Maldives, Indonesia, the Great Barrier Reef

(GBR) of Australia, Papua New Guinea, New Caledonia and the

Marshall Islands.

2. Capillaster multiradiatus (Linnaeus, 1758)

SEE. Clark and Rowe, 1971:6—7.

MATERIAL. NHM — 810422E/3 (1), 810423C/4 (2), 810425F/2

(4), 810426B/1 (1), 810427D/3 (2), 810428D/2 (1), 810430A/1 (1),

810430A/7 (1), 810430A/23 (1), 810430A/30 (1), 810430A/31 (1),

810504A/2 (1); WAM — 810422C/3 (1), 810422D/9 (1), 810424D/5

(1), 810425B/1 (2), 810525E/1 (2), 810425E/2 (2), 810427D/1 (1),

810428D/1 (1), 810430A/11 (1), 810430A/22a (1), 810430A/25 (1

of 2), 810430A/24b (1), 810430A/29 (1); SM — 810422C/4 (1),

810422D/7 (1), 810424B/4 (1), 810425F/1 (4), 810425F/4 (2),

810425F/5 (1), 810426A/1 (1), 810430A/10 (1), 810430A/21b (1),

810430A/21d (2), 810430A/25 (1 of 2).

COLLECTION SITES. Klah, Seulakoe, Sabang Bay, Ug Murung, Ug

Tapa Gadja, Ug Seukundo and Rubiah, Pula Wé, Sumatra.

HABITATS AND DEPTH RANGE. Subtidal rock, coral reef, coral

rubble, fire coral, soft coral and gorgonian; 2—30m.

COMMENT. This species is well-known across the Indo-Pacific

region. However, its habitat varies; in some regions it inhabits

exposed situations, in others it is cryptic. Specimens have been

recorded from 0.5—1 m in Madagascar, to 77 m in the Bay of Bengal

(Clark, 1972). In this collection, habitat and depth also varied.

3. Capillaster sentosus (Carpenter, 1888)

SEE. Clark and Rowe, 1971: 6-7.

WAM = 810501C/4 (1).

Ug Seukundo, Pula Wé, Sumatra.

Subtidal rock; 20m.

MATERIAL.

COLLECTION SITE.

| HABITAT AND DEPTH.

_ 4. Clarkcomanthus albinotus Rowe, Hoggett, Birtles and

Vail, 1986.

SEE. Rowe, Hoggett, Birtles and Vail, 1986: 238.

MATERIAL. WAM -— 810428E/1(1).

COLLECTION SITES. Ug Tapa Gadja, Pula Wé, Sumatra.

HABITAT AND DEPTH. Soft coral, 2—10m.

COMMENT. This is a marked extension of range for this species,

previously only recorded from the Great Barrier Reef, Papua New

Guinea (Messing, 1994) and Japan.

5. Clarkcomanthus littoralis (Carpenter, 1888)

SEE. Rowe, Hoggett, Birtles and Vail, 1986: 236.

MATERIAL. NHM — 810423D/1 (broken), 810425C/3 (1),

810428C/4 (2); WAM — 810420A/2 (fragmented), 810430A/12 (1),

810501A/3 (1), 810504C/4 (1); SM —810421C/2 (1), 810428E/2

(1), 810430A/19 (1).

COLLECTION SITES. Klah, Seukundo, Ug Bau, Subang Bay, Ug

Tapa Gadja and Ug Seukundo, Pula Wé, Sumatra.

HABITATS AND DEPTH RANGE.

10m.

Subtidal rock/coral, coral reef; 2—

6. Clarkcomanthus luteofuscum (H.L. Clark, 1915)

SEE. Rowe, Hoggett, Birtles and Vail, 1986: 233.

MATERIAL. WAM - 810427D/1 (1).

COLLECTION SITES. Ug Murung, Pula Wé, Sumatra.

HABITAT AND DEPTH. Soft coral, 2—10m.

7. Comanthina nobilis (Carpenter, 1884)

SEE. Rowe, Hoggett, Birtles and Vail, 1986: 243.

MATERIAL. NHM — 810425A/18d (2), 810425D/1 (1), 810430A/

34 (1); WAM — 810425A/18e (2), 810430A/16 (1); SM — 810425A/

18 (1), 810430A/2 (1), 810501A/5 (2).

COLLECTION SITES. Ug Murung and Ug Seukundo, Pula Wé,

Sumatra.

HABITATS AND DEPTH RANGE.

30m.

Subtidal rock/coral, coral reef; 6—

8. Comanthina schlegelii (Carpenter, 1881)

SEE. Clark and Rowe, 1971:6—7; Rowe, Hoggett, Birtles and Vail,

1986: 244.

MATERIAL. NHM — 810422D/4 (1), 810423A/2 (1), 810430A/13

(1), 810504C/3 (1); WAM — 810423C/2a (1), 810430A/6 (1),

810504C/1 (1); SM — 810421C/1 (1), 810421C/4 (1), 810424D/3

COLLECTION SITES.

Pula Wé, Sumatra.

Seukundo, Klah, Ug Bau and Ug Seukundo,

HABITAT AND DEPTH RANGE. Subtidal rock and coral reef; S—20m.

9. Comanthus briareus (Bell, 1882)

SEE. Rowe, Hoggett, Birtles and Vail, 1986:218

MATERIAL. NHM — 810424B/5 (1); WAM — 810204A/8 (1).

COLLECTION SITES. Kalpitiya, Sri Lanka; Seulakoe, Pula Wé,

Sumatra.

HABITAT AND DEPTH RANGE. Coral reef; 3—5m, 20—30m.

COMMENT. The Sri Lankan specimen is a new locality record for

this species, and extends its range west into the Indian Ocean from

Indonesia.

10. Comanthus gisleni Rowe, Hoggett, Birtles and Vail,

1986.

SEE. Rowe, Hoggett, Birtles and Vail, 1986: 219.

MATERIAL. NHM -— 810124A/1 (1), 810204A/3 (1),810430A/14

(1), 810430A/22a (3), 810504B/1 (1); WAM — 810124A/1 (1),

810124A/6 (1), 810204A/2 (1), 810425F/1 (1), 810430A/15 (1);

SM — 810422D/5 (1), 810430A/4 (1), 810430A/14 (1), 810504A/1

(1).

COLLECTION SITES. Galle and Kalpitiya, Sri Lanka; Klah, Sabang

Bay, Ug Bau, Ug Seukundo and Rubiah, Pula Wé, Sumatra.

HABITATS AND DEPTH RANGE.

reef, soft coral; 2—20m.

Subtidal rock, coral rubble, coral

COMMENT. All these specimens represent extension of the range

of this species into the northern Indian Ocean. It has been recorded

from the coast of Western Australia, but otherwise only from the

Pacific Ocean coasts and islands, Thailand, Papua New Guinea and

Japan (Rowe et al., 1986: 221; Messing, 1994).

11. Comanthus mirabilis Rowe, Hoggett, Birtles and Vail,

1986.

SEE. Rowe, Hoggett, Birtles and Vail, 1986: 226.

MATERIAL. NHM —810501F/2 (1); WAM — 810427C/1 (1); SM —

810430A/5 (1).

COLLECTION SITES. Ug Bau and Ug Seukundo, Pula Wé, Sumatra.

HABITATS AND DEPTH RANGE.

30m.

Subtidal rock/coral, coral reef; 5—

COMMENT. The WAM specimen has 45 arms. The IIIBr series are

mostly 2; where an arm has broken off and regenerated there is

usually another devision series and extra arms, otherwise division

series beyond IIIBr are randomly distributed. Most pinnules beyond

P, are broken, so comb distribution further out cannot be ascer-

tained.

12. Comanthus parvicirrus (Miller, 1841)

SEE. Rowe, Hoggett, Birtles and Vail, 1986:211; Hoggett and

Rowe, 1986: 125.

MATERIAL. NHM-—810123A/1 (1),810123B/2 (1), 810203A/1 (1

of 3), 810206A/1 (1), 810212A/2 (1), 810420A/1 (1), 810425F/1

(1), 810427C/2 (1 of 3), 810428C/2 (1); WAM — 810126B/5

(1),810124A/6 (1);810203A/1 (1 of 3), 810204A/2 (1), 810204A/6

(1), 810425F/3 (1), 810427C/2 (1 of 3), 810428C/5 (1),810430A20a

(1); SM — 810125A/2 (1), 810126B/6 (1), 810203A/1 (1 of 3),

810204A/5 (1), 810204A/7 (1), 810425A/18d (1), 810427C/2 (1 of

3), 810430A/20c (1), 810430A/21a (1), 810501E/10 (1).

COLLECTION SITES. Galle, Hikkaduwa, Kandakkuliya, Kalpitiya,

Negombo and Unawatuna, Sri Lanka; Klah, Sabang Bay, Ug Murung,

Ug Bau, and Ug Seukundo, Pula Wé, Sumatra.

HABITATS AND DEPTH RANGE.

20m.

COMMENT. ‘The specimens from Sri Lanka may constitute a new

record for C. parvicirrus as it is now defined (Rowe et al., 1986),

Subtidal rock/coral, coral reef; 2—

J.1.M. CROSSLAND AND A.R.G. PRICE

depending on the correctness of H.L. Clark’s (1915) identification

of a specimen from the region.

13. Comanthus suavia Rowe, Hoggett, Birtles and Vail,

1986.

SEE. Rowe, Hoggett, Birtles and Vail 1986: 222.

MATERIAL. NHM -— 810501B/1 (1); WAM — 810123A/2(1),

810124A/6, 810501B/4 (1).

COLLECTION SITES. Galle, Sri Lanka; Rubiah, Pula Wé, Sumatra.

HABITATS AND DEPTH RANGE. Subtidal rock, coral reef; 5—20m.

COMMENTS. This is a major extension of range, as the species

was originally thought to be restricted to the northern Great

Barrier Reef and New Guinea. Two specimens, whose identity

was originally in doubt, have now been confirmed as this species.

One, 810124A/6, has 7—9 triangular comb teeth, recurved but

with bases not in contact; terminal tooth small, proximal tooth

usually saucer-shaped. Combs appear irregularly, e.g. on P,, P,, P,

ork), PPPs. The centrodorsal is stellate with cirrus buds and

cirrus scars, and subradial clefts are present. Specimen 810501B/4

has short combs with 4+2 teeth, triangular but not in lateral

contact, a saucer-shaped proximal tooth, and a smaller secondary

tooth on some pinnules.

14. Comanthus wahlbergii (Miller, 1843)

SEE. Rowe, Hoggett, Birtles and Vail, 1986: 228.

MATERIAL. NHM — 810123B/1 (1), 810421A/5(1); WAM —

810204A/2 (1); SM — 8101423B/3 (1), 810204A/4 (1).

COLLECTION SITES. Chetlat I., Laccadive Islands; Galle and

Kalpitiya, Sri Lanka; Seukundo, Pula Wé, Sumatra.

HABITATS AND DEPTH RANGE. Subtidal rock, coral reef; 3—30m.

COMMENT. The collection of C. wahlbergii from the Laccadives,

Sri Lanka and Sumatra, Indonesia fills the gaps in the distribution of

this species around the Indian Ocean.

15. Comaster multifidus (Miller, 1841)

SEE. Clark and Rowe 1971: 6.

MATERIAL. NHM-—810421A/4 (1); WAM — 810425E/1 (1); SM —

801210B/5 (1, fragmented), 801212A/2 (1, fragmented).

HABITATS AND DEPTH RANGE. Subtidal rock and coral; 10—30m.

COLLECTION SITES. Chetlat I., Laccadive Is; Sabang Bay and

Seukundo, Pula Wé, Sumatra.

COMMENT. This species is well known from Indonesia and north-

ern Australia, and from the South Pacific. The record from the

Laccadives is a marked extension of range, but identification is not

positive because of the condition of the specimens.

16. Oxycomanthus bennetti (Miiller, 1841)

SEE. Clark and Rowe, 1971:6-7; Rowe, Hoggett, Birtles and Vail,

1986:259.

MATERIAL. NHM — 810421B/3 (1), 810422D/9 (1),810423C/5

(1), 810423C/5 (1),810424D/2 (1),810424D/4 (1), 810425A/18b

(1), 810425A/18 (1), 810425E/1 (1), 810425F/2 (1), 810425F/3 (3),

810428B/2 (1); WAM — 810422D/6 (1), 810423C/2b (1), 810423C/

3 (2), 810423C/6 (1), 810424A/3 (1), 810424D/6 (1),810424D/7

INDIAN OCEAN ECHINODERMS

(1), 810425A/18c (1), 810427D/1 (2), 810428A/5 (1), 810428B/3

(1), 810504C/2 (1); SM—810421C/5 (1), 810422D/8 (1), 810423A/

1 (1), 810423C/1 (1), 810424D/1 (1), 810427D/1 (1), 810427D/2

(1), 810428A/6 (1), 810430A/8 (2),810430A/35 (1), 810430A/36

(1), 810501C/4 (2).

COLLECTION SITES. Seukundo, Klah, Ug Bau, Seulakoe, Sabang

Bay, Ug Murung and Ug Seukundo, Pula Wé, Sumatra.

HABITATS AND DEPTH RANGE. Subtidal rock/coral, coral reef, soft

coral; 2—30m.

COMMENT. Rowe ef al. (1986) do not record this species from

Indonesia, although it was recorded from Papua New Guinea by

Messing (1994); therefore this collection fills in the gap between the

Andaman Islands and the Philippines.

Family HIMEROMETRIDAE

17. Amphimetra tessellata (Miiller, 1841)

SEE. Clark and Rowe, 1971:6-7.

MATERIAL. WAM — 810425E/2 (1).

COLLECTION SITE. Sabang Bay, Pula Wé, Sumatra.

HABITAT AND DEPTH. On gorgonian; 10—20m.

18. Himerometra robustipinna (Carpenter, 1881)

SEE. Clark and Rowe, 1971:8—9.

MATERIAL. NHM — 810423E/1 (1).

COLLECTION SITES. Ug Bau, Pula Wé, Sumatra.

HABITAT AND DEPTH. Ona wreck, 5m.

Family MARIAMETRIDAE

19. Lamprometra palmata (Miiller, 1841)

SEE. Clark and Rowe, 1971: 8-9.

MATERIAL. NHM — 810425E/1 (1), 810504D/2 (1 of 2); WAM —

810124A/7 (1), 810212A/4 (1); SM —810425F/3 (1), 810504D/2 (1

of 2).

COLLECTION SITES. Galle and Unawatuna, Sri Lanka.

HABITATS AND DEPTH RANGE.

20m.

Subtidal rock, coral, coral reef; 2—

20. Oxymetra finschi (Hartlaub, 1890)

SEE. Clark and Rowe, 1971: 8-9.

MATERIAL. WAM — 810430A/33 (1).

Ug Seukundo, Pula Wé, Sumatra.

Subtidal rock; 12—13m.

COLLECTION SITE.

HABITAT AND DEPTH RANGE.

21. Stephanometra indica (Smith, 1876)

SEE. Clark and Rowe, 1971:8—9.

MATERIAL. NHM — 81020A/3 (1); WAM — 801212A/1 (frag-

mented); SM — 810421C/3 (fragmented).

COLLECTION SITES. Chetlat I., Laccadive Islands; Klah and

Seukundo, Pula Wé, Sumatra.

HABITAT AND DEPTH RANGE. Coral reef; 4-8m.

COMMENT. Even though two of the three specimens are frag-

mented, they are easily identifiable as this widely-distributed

Indo-Pacific species.

22. Stephanometra spinipinna (Hartlaub, 1890)

SEE. Clark and Rowe, 1971:8-9.

MATERIAL. NHM — 810423D/1 (1).

COLLECTION SITE. Ug Bau, Pula Wé, Sumatra.

HABITAT AND DEPTH. Coral reef; 2—8m.

Family COLOBOMETRIDAE

23. Cenometra bella (Hartlaub, 1880)

SEE. Clark and Rowe, 1971:10—11; Meyer and Macurda, 1980:88—

89.

MATERIAL. NHM — 810423C/7 (1).

COLLECTION SITE. Ug Bau, Pula Wé, Sumatra.

HABITAT AND DEPTH. On gorgonian; 10—20m.

24. Colobometra perspinosa (Carpenter, 1881)

SEE. Clark and Rowe, 1971:10-11.

MATERIAL. NHM — 810421B/5 (1), 810425F/2 (2); WAM —

810427C/1 (2), 810501D/3 (1); SM — 810424A/2 (1), 810501 A/1

(1), 810501B/2 (1).

COLLECTION SITES. Seukundo, Klah, Sabang Bay, Ug Murung

and Ug Seukundo, Pula Wé, Sumatra.

HABITATS AND DEPTH RANGE. Subtidal rock/coral, coral reef on

gorgonian; 2—30m.

25. Decametra brevicirra (A.H. Clark, 1912)

SEE. Clark and Rowe, 1971:10-11.

MATERIAL. WAM — 810425D/2 (1).

COLLECTION SITE. Sabang Bay, Pula Wé, Sumatra.

HABITAT AND DEPTH. Subtidal rock and sand; 25m.

COMMENT. Clark & Rowe (1971) implied that the main key char-

acteristic of D. brevicirra, the similarity in segment numbers in P,

and P,, distinguishing this species from its congeners D. mylitta

(A.H. Clark, 1912) and D. chadwicki (A.H. Clark, 1911), would not

‘hold good’ when more specimens from the type locality, the Bay of

Bengal, had been collected. This specimen, from Sumatra, clearly

has 10 segments on both proximal pinnules. It differs from the other

specimen of this genus collected in the same area, which clearly keys

out to D. parva (below) on the basis of having a higher cirrus

segment number. It may be time for a thorough re-examination of the

genus, as there are doubtless more records than there were in 1971.

26. Decametra parva (A.H. Clark, 1912)

SEE. Clark and Rowe, 1971:10-11.

MATERIAL. NHM — 810428A/11 (1).

COLLECTION SITE. Ug Bau, Pula Wé, Sumatra.

HABITAT AND DEPTH. Subtidal rock and coral, 20—30m.

27. Oligometra carpenteri (Bell, 1884)

SEE. Clark and Rowe, 1971:10-11.

WAM -— 810124A/7 (1).

Galle, Sri Lanka.

Subtidal rock; 10—15m.

MATERIAL.

COLLECTION SITE.

HABITAT AND DEPTH.

COMMENT. This is an extension of range for the species, which is

well known along much of the Great Barrier Reef and has been

recorded in Indonesia. This specimen has much less well-developed

keels on the proximal pinnules than in specimens from the GBR,

where the two species of the genus are quite distinct. However, the

pinnules are wider than long, and lack flaring of their distal ends of

segments. Only O. serripinna has been previously recorded from the

Sri Lanka area.

28. Oligometra serripinna (Carpenter, 1881)

SEE. Clark and Rowe, 1971:10-11.

MATERIAL. NHM -—810425D/2 (1); WAM — 810425E/2 (1); SM—

810422C/5 (1).

COLLECTION SITES. Klah and Sabang Bay, Pula Wé, Sumatra.

HABITATS AND DEPTH RANGE. Subtidal rock/sand, coral reef, on

gorgonian; 10—30m.

J.I.M. CROSSLAND AND A.R.G. PRICE

COMMENT. See above.

Family TROPIOMETRIDAE

29. Tropiometra carinata (Lamarck, 1816)

SEE. Clark and Rowe, 1971:10-11.

MATERIAL. NHM — 810126B/5 (4), 810428D/3 (1); WAM —

810123B/3 (1), 810213A/2 (4), 810428D/6; SM — 810123A/1 (1),

810212A/4 (3).

COLLECTION SITES. Galle, Hikkaduwa, Unawatuna and Tangalla,

Sri Lanka; Ug Tapa Gadja, Pula Wé, Sumatra.

HABITATS AND DEPTH RANGE. Subtidal rock/coral, coral reef; 3—

15m.

COMMENT. J. carinata is well known from Indian Ocean reefal

areas.

Family ANTEDONIDAE

30. Antedon parviflora (A.H. Clark, 1912)

SEE. Clark and Rowe, 1971: 10-11.

MATERIAL. NHM — 810425D/2 (1).

COLLECTION SITE. Sabang Bay, Pula Wé, Sumatra.

HABITAT AND DEPTH. Subtidal rock and sand, 25m.

Table 1 Regional distribution of crinoids from the Sindbad Voyage (names in parenthesis are equivalent zoogeographic subdivision of sampling area,

following Clark & Rowe, 1971)

Laccadives

(Maldive area)

Comanthus wahlbergii

Comaster multifidus

Stephanometra spinipinna

Sri Lanka

(Sri Lanka area)

Comanthus briareus

Comanthus gisleni

Comanthus parvicirrus

Comanthus suavia

Comanthus wahlbergii

Lamprometra palmata

Oligometra carpenteri

Tropiometra carinata

Pula Wé, Sumatra

(Indonesia/East Indies)

Alloeocomatella pectinifera

Capillaster multiradiatus

Capillaster sentosus

Clarkcomanthus albinotus

Clarkcomanthus littoralis

Clarkcomanthus luteofuscum

Comanthina nobilis

Comanthina schlegelii

Comanthus briareus

Comanthus gisleni

Comanthus mirabilis

Comanthus parvicirrus

Comanthus suavia

Comanthus wahlbergii

Comaster multifidus

Oxycomanthus bennetti

Amphimetra tessellata

Himerometra robustipinna

Oxymetra finschii

Stephanometra indica

Stephanometra spinipinna

Cenometra bella

Colobometra perspinosa

Decametra brevicirra

Decametra parva

Oligometra serripinna

Tropiometra carinata

Antedon parviflora

INDIAN OCEAN ECHINODERMS

DISCUSSION

The crinoids of the tropical Indo-west Pacific region (Africa,

Indonesia, Philippines, tropical Australia and the South Pacific)

are relatively well-documented (Clark & Rowe, 1971). The region

between the Red Sea and Indonesia has to date produced a rela-

tively depauperate crinoid record, but the reasons for this are not

clear. Unfortunately, the Sindbad collection does not resolve the

problem. The low number of crinoids in this collection from the

Laccadives and from Sri Lanka is probably due to a combination

of two factors: lower abundance and diversity of this group in the

localities collected, and limited sampling time available in those

regions. This situation is unfortunate, as the areas of the northern

Indian Ocean, except for the western fringe of Indonesia, are not

well-represented in any collections of echinoderms, so that spe-

cies and even generic distributions within the region are not

well-known. In fact, the majority of specimens collected during

the Sindbad Voyage are from around the small island of Pula Wé,

at the western tip of Sumatra, Indonesia. SE Asia is the region of

the Indo-West Pacific associated with greatest echinoderm species

richness (Clark & Rowe, 1971), and Indonesia in particular is

commonly regarded as the centre of distribution for coral reefs,

other invertebrate groups and marine tropical diversity in general

(Veron, 1995; Gray, 1997).

Crinoid records from this voyage’s collection are divided into the

different regions sampled in Table 1, which also shows the equiva-

lent zoogeographic subdivisions adopted by Clark & Rowe (1971).

The observed distribution is highly skewed, with all but two of the 30

species collected in Sumatra, eight in Sri Lanka and only three in the

Laccadives. Regional comparison based on more comprehensive

records, including Indian Ocean data of Clark and Rowe (1971) and

the results of James (1989) for the Laccadives (13 additional spe-

cies) and Sri Lanka (14 additional species), shows species

distributions to be less uneven. Nevertheless, the resulting pattern

reveals a progressive increase in species richness from the Maldive

area to Sri Lanka to East Indies/Indonesia, as suggested in the

Sindbad data (Table 1). However, the Laccadives, in particular,

probably remain undersampled. These islands are a prohibited area

under the control of India, and access will probably continue to be

restricted.

Range extensions, to the western fringe of Indonesia (Pula Wé)

and into the Indian Ocean, are recorded for at least six of the 30

crinoid species collected during the Sindbad Voyage, as follows:

Clarkcomanthus albinotus (Indonesia/East Indies); Comanthus

briareus (Sri Lanka area); Comanthus gisleni (Sri Lanka area &

Indonesia/East Indies); Comanthus suavia (Sri Lanka area & In-

donesia/East Indies); Comanthus wahlbergii (Maldive area, Sri

Lanka area & Indonesia/East Indies); Oxycomanthus bennetti (In-

donesia/East Indies); and possibly also Comaster parvicirrus (Sri

Lanka area — depending on validity of an earlier record) and

Comaster multifidus (Maldive area?— specimens poorly pre-

served).

Of the crinoids represented, Capillaster multiradiatus and

Oxycommanthus bennetti were the most common, each occurring in

19% of the samples, followed by Comanthus parvicirrus which

occurred in 9% of the samples. The first two species also occupied a

relatively wide range of depths (2-30 m) and habitats compared to

most other species collected. A more comprehensive ecological and

biogeographic assessment of echinoderms of Pula Wé, Sumatra is

currently in progress.

BY)

ACKNOWLEDGEMENTS. We are grateful to Dr R. Dalley, P. Hunnam, P.

Dobbs and D. Tattle for their considerable assistance during field work. One

of us (A.R.G.P.) would also like to thank T. Severin, leader of the Sindbad

Voyage, for the kind invitation to participate in the expedition which was

made possible by generous support from the Ministry of Natural Heritage and

Culture, Sultanate of Oman. Financial assistance to A.R.G.P. from the

Leverhulme Trust is gratefully acknowledged. Thanks are also due to Dr

Anne Hoggett (Lizard Island Research Station, GBR, Australia) for assist-

ance with identification of specimens in doubt; and to Dr Frank Rowe for

confirming several identifications.

REFERENCES

Clark, A.H. 1911.A new unstalked crinoid from Christmas Island. Annals & Magazine

of Natural History 7: 644-645.

Clark, A.H. 1915. A monograph of the existing crinoids. 1(1). Bulletin of the United

States National Museum, 82(1): 1-406.

Clark, A.H. 1921. A monograph of the existing crinoids. 1(2). Bulletin of the United

States National Museum, 82(2): xxv + 797, 57 pls.

Clark, A.H. 1931.A monograph of the existing crinoids. 1(3). Superfamily Comasterida.

Bulletin of the United States National Museum, 82(3): vii + 816, 82 pls.

Clark, A.H. 1941.A monograph of the existing crinoids. 1(4). Superfamily Mariametrida

(except the family Colobometridae). Bulletin of the United States National Museum,

82(4a):vii + 603, 61 pls.

Clark, A.H. 1947. A monograph of the existing crinoids. 1(4b). Superfamily

Mariametrida (concluded — the family Colobometridae) and Superfamily

Tropiometrida (except the families Thalassometridae and Charitometridae). Bulletin

of the United States National Museum, 82(4b):vii + 473, 43 pls.

Clark, A.H. 1950. A monograph of the existing crinoids. 1(3). Superfamily

Tropiometrida (the families Thalassometridae and Charitometridae). Bulletin of the

United States National Museum, 82(4c): vii + 383, 32 pls.

Clark, A.H. & Clark, A.M. 1967. A monograph of the existing crinoids. 1(5).

Suborders Oligophreata (concluded) and Macrophreata. Bulletin of the United States

National Museum, 82(5):xiv + 860, 53 figs.

Clark, A.M. & Rowe, F.W.E. 1971. Shallow-water Indo-West Pacific Echinoderms.

238p. British Museum (Natural History), London.

Clark, A.M. 1972. Some crinoids from the Indian Ocean. Bulletin of the British

Museum (Natural History) 24(2):73-156, 17 text figs.

Clark, H.L. 1915. The echinoderms of Ceylon (other than Holothurians). Spolia

zeylanica 10(37): 83-102.

Gray, J.S. 1997. Marine biodiversity: patterns, threats and conservation needs.

Biodiversity & Conservation 6: 153-175.

Hoggett, A.K. & Rowe, F.W.E. 1986. A reappraisal of the family Comasteridae A.H.

Clark, 1908 (Echinodermata: Crinoidea) with the description of a new subfamily and

a new genus. Zoological Journal of the Linnean Society 88:103—142, 3 figs.

James, D.B. 1989. Echinoderms of Lakshadweep and their zoogeography. Bulletin of

the Centre for Marine Fisheries Research Institute 43: 97-144.

Lincoln, R.J. & Sheals, J.G. 1979. Invertebrate Animals: Collection and Preservation.

150p. British Museum (Natural History) and Cambridge University Press.

Marsh, L.M. & Price A.R.G. 1991. Indian Ocean echinoderms collected during the

Sindbad Voyage (1980-81): 2. Asteroidea. Bulletin of the British Museum (Natural

History ), Zoology 57(1): 61-70.

Messing, C.G. 1994. Comatulid crinoids (Echinodermata) of Madang, Papua New

Guinea, and environs: Diversity and ecology, pp. 237—243. Jn David, B., Guille,

A., Feral, J-P & Roux, M. (eds). Echinoderms through time. Balkema. Rotterdam.

Messing, C.G. 1995. Alloeocomatella, a new genus of reef-dwelling feather star from

the tropical Indo-West Pacific (Echinodermata: Crinoidea: Comasteridae). Proceed-

ings of the Biological Society of Washington 108(3):436-450.

Meyer, D.L. & Macurda, D.B. 1980. Ecology and distribution of the shallow-water

crinoids (Echinodermata) of the Palau Islands and Guam (Western Pacific).

Micronesica 16: 59-99.

Price A.R.G. & Reid, C.E. 1985. Indian Ocean echinoderms collected during the

Sindbad Voyage (1980-81): 1. Holothurioidea. Bulletin of the British Museum

(Natural History ), Zoology 48(1): 1-9.

Price A.R.G. & Rowe, F.W.E. 1996. Indian Ocean echinoderms collected during the

Sindbad Voyage (1980-81): 3. Ophiuroidea and Echinoidea. Bulletin of the British

Museum (Natural History ), Zoology 62(2): 71-82.

Rowe, F.W.E., Hoggett, A.K., Birtles, R.A. & Vail, L.L. 1986. Revision of some

comasterid genera from Australia (Echinodermata: Crinoidea) with descriptions of

two new genera and nine new species. Zoological Journal of the Linnean Society

86:197-277, 10 figs.

Veron, J.E.N. 1995. Corals in Time and Space. University of New South Wales Press,

Sydney.

Bull. nat. Hist. Mus. Lond. (Zool.) 65(1): 31—50

On the hybrid status of Rothschild’s Parakeet

Psittacula intermedia (Aves, Psittacidae)

PAMELA C. RASMUSSEN KX (Si CTobE.1)

NHB 336 MRC 114, Smithsonian Institution, Washington, D.C. 20560-0131, USA

NIGEL J. COLLAR

BirdLife International, Wellbrook Court, Girton Road, Cambridge CB3 ONA, UK

SYNOPSIS. The name Psittacula intermedia was attached to seven dataless specimens sent from India to England between 1895

and 1907, six of which are now at the American Museum of Natural History, the other being at The Natural History Museum,

Tring, U.K.Their origins and taxonomic standing have long puzzled authorities, since they look intermediate between male Plum-

headed Parakeet P. cyanocephala and Slaty-headed Parakeet P. himalayana, and no definite field records exist. Although a hybrid

origin has been suggested, intermedia has recently been considered a valid species on the bases of: (a) uniformity of characters;

(b) a single origin; (c) a non-captive origin; (d) an old description of hybrid cyanocephala x himalayana which does not match

intermedia; (e) reports of captive intermedia in the 1990s; and (f) biochemical analysis of captive birds.

For this study, we examined all published intermedia specimens. For hybrid diagnoses we compared morphology of adult

males qualitatively and mensurally with the putative parental species, including also Grey-headed Parakeet P. finschii and

Blossom-headed Parakeet P roseata. We examined six live adult hybrid cyanocephala x himalayana bred by two different

aviculturists, as well as one live bird in India claimed to be intermedia, and we considered published avicultural evidence.

Our analyses showed all the defences of the specific status of intermedia to be wanting, as follows: (a) considerable variation

exists in the original material; (b) the specimens could not all have had a single origin; (c) six of the seven specimens showed signs

of captivity; (d) the 65-year-old account of cyanocephala x himalayana only furnishes passing descriptions of juveniles, and is

therefore not comparable with the adult intermedia specimens; (e) all the specimens examined in Bombay are hybrid

cyanocephala x krameri, while other captive intermedia in Austria and India are of uncertain provenance (but the former appear

to becyanocephala x finschii); and (f) the biochemical analysis was seriously flawed, most importantly in that the specimens used

were not intermedia but hybrid cyanocephala x krameri.

Neither cyanocephala nor himalayana shows any morphological characters incompatible with being parent to intermedia, and

all features of the latter are explained by a combination of the two former species. Moreover, mensurally the AMNH intermedia

fall midway between cyanocephala and himalayana. All known male cyanocephala x himalayana possess plumage features and

measurements matching AMNH’s five adult male intermedia, while the previously undescribed female hybrid has the head paler

than himalayana and drabber than female cyanocephala. This evidence leaves no doubt that intermedia is a hybrid of

Issued 24 June 1999

cyanocephala and himalayana.

INTRODUCTION

Rothschild’s or the Intermediate Parakeet Psittacula intermedia was

described from a single dataless specimen (Rothschild 1895) backed

up by a later series which also lacked data but had been exported

from Bombay (Rothschild 1907, Hartert 1924). However, apart from

being listed in Peters (1937), this parakeet was overlooked until

mentioned by Ripley (1953) as a species of Indian origin. Of

subsequent authors who have considered the status of intermedia

and treated it as taxonomically valid, only Biswas (1959) had

examined Rothschild’s entire series. Walters (1985) had access only

to a single specimen in the bird collections of The Natural History

Museum, Tring, U.K. (BMNH); in Bombay, Sane (1975, 1977, Sane

et al. 1987) had only his own captive birds; while Inskipp and

Inskipp (1995) simply reviewed the literature on intermedia; and

Bhargava (1998) had only his own specimens in India. Psittacula

intermedia has been accepted uncritically as a species by several

authors (e.g. Howard and Moore 1991, Monroe and Sibley 1993).

However, Salvadori (1907), in reference to the type, had stated that

intermedia was ‘. . . not improbably established on a hybrid!’.

Immediately after Salvadori’s comment, Rothschild (1907) men-

tioned having obtained six more specimens, which he maintained

*.. Should certainly dispose of any doubt regarding the distinctness

of intermedia’ . Conversely, Husain (1959), on the basis of the single

skin at the BMNH, considered that intermedia was a hybrid between

Plum-headed Parakeet Psittacula cyanocephala and Slaty-headed

Parakeet P. himalayana, while Forshaw (1973) concluded the same

after examination of the series at the American Museum of Natural

History, New York (AMNH); Wolters (1975) also treated it as a

probable hybrid. Still other authors have remained undecided as to

its status (Peters 1937, Ali and Ripley 1969, 1981, Wirth 1990);

Juniper and Parr (1998) and Collar (1997) tentatively gave it a

species account pending publication of the present study. Psittacula

intermedia is currently listed as a globally threatened species with

IUCN status Vulnerable, but for which no specific threats have been

identified (Collaretal. 1994), although its possible extermination by

collectors has been suggested (Walters 1985).

The fact that the phenotype of Psittacula intermedia places it

midway between cyanocephala and himalayana was acknowledged

in both the original description and in the specific epithet chosen

(Rothschild 1895). Since then, no characters of intermedia have

been identified that either differ from those of cyanocephala or

himalayana or are not manifestly intermediate between these two

(contra Inskipp et al. 1996, whose cited references nowhere

demonstrate non-intermediacy). Moreover, the area of origin of

intermedia has never been accurately pinpointed, despite fairly

extensive subsequent ornithological work in many presumably likely

areas; even Rothschild (1907) admitted that “speculations as to its

exact locality were useless, as these collections contained forms

exclusively found in the Eastern Himalaya as well as others occur-

ring only in the north-western portions of India.’ A statement by Ali

and Ripley (1969) concerning intermedia — ‘never consciously seen

alive in the wild state by any ornithologist’ — remains true today.

Recent reports within India (R. Bhargava pers. comm. 1996, in litt.

1997, 1998; Ahmed et al. 1996, Anon. 1997, 1998a, b; Mookerjee

1997; Bhargava 1998; Them 1998), and three birds identified as

intermedia (Sane 1975, 1977, Sane et al. 1987, S. R. Sane pers.

comm. 1997), do not clear up the mystery, as all refer to captive birds

of uncertain provenance, and some of these are of problematic

identification as well (Rasmussen and Collar 1998, Bhargava 1998).

Appeals for information (Rothschild 1907, Sane 1977, Wirth 1990,

Inskipp and Inskipp 1995) have not led to the discovery of a wild

population.

If intermedia were a typical diurnal, noisy Psittacula, it would be

a most unusual bird, not only for having escaped the attentions of

field ornithologists for over a century in one of the best-known parts

of tropical Asia, but also for showing complete intermediacy in

numerous characters between two clearly differentiated congeners.

There are two possible explanations for this double circumstance:

the first is that it is an extremely rare species and therefore requires

concerted conservation attention; the second is that it is not a species

at all, but a hybrid. Only the second explanation accounts for both of

its unusual traits. In this paper we reexamine the evidence for

specific status vs. hybrid origin of intermedia, based on plumage and

mensural analyses both of museum specimens and of newly located

captive birds of known parentage.

METHODS

SPECIMENS EXAMINED

Eight museum specimens have been published in the primary litera-

ture as Psittacula intermedia, although one (AMNH 621545) has

been considered an immature himalayana (Biswas 1959, 1990,

Forshaw 1973). Each was thoroughly examined, photographed, and

measured for this study: AMNH 621539 (holotype), 621540-621542,

621544-621545; BMNH 1980.3.1; and BNHS (Bombay Natural

History Society) 26758. In addition, we examined another

uncatalogued specimen belonging to Mr. Sane, as well as a photo-

graph of three unaccessioned specimens in the possession of R.

Bhargava.

In the early 1930s, Rothschild’s entire series of intermedia went

to AMNH along with most of the rest of his collection. Subse-

quently, one (BMNH 1980.3.1, formerly AMNH 621543) was

exchanged to the then British Museum (Natural History) (M.P.

Walters, pers. comm. 1997), where it had already resided since 1959

on long-term loan (Knox and Walters 1994). The BMNH specimen

was lent to AMNH so that we could compare it there with the

remainder of the series.

A colour transparency of AMNH 621540 (placed with the speci-

mens; date and photographer unknown) taken sometime after 1973

— based on an accompanying note: “Psittacula ‘intermedia’ believed

to be a hybrid himalayana x cyanocephala see Forshaw (1973: 336)’

— shows that it had long central rectrices when the photo was taken,

but these were lacking when the specimen was first photographed by

PCR in 1993, and it cannot now be determined if the rectrices were

fully grown. Estimates of lengths of the tail and of the yellow tip of

the central rectrix of AMNH 621540 were made from the transpar-

ency, in which the subject is 1/3 natural size and photographed from

the side.

BNHS 26758 is essentially dataless (label data: male, aviary bird,

P.C. RASMUSSEN AND N.J. COLLAR

S. R. Sane, Bombay, 12/90), as is Sane’s second uncatalogued

specimen; these are two of the three birds examined from his

collection. The first may or may not be the specimen described in

Sane (1975, 1977), referred to as having died in 1978, and as being

in the BNHS collection (Sane et al. 1987), but if ‘12/90’ refers either

to date of death or to date of accession it can hardly be the same

individual.

The only other specimens reputed to be intermedia of which we

are aware are those preserved by R. Bhargava, and we have seen

photos of three of those. However, adult female and especially

immature intermedia (see below) would readily escape notice among

series of similar congeners, and may well exist undetected in mu-

seum collections.

We assessed variability among the eight published putative

intermedia specimens in the plumage and mensural characters listed

in Tables 1-4 and the Appendix.

PHOTOGRAPHIC EVIDENCE

Most of the published information on intermedia was recently

summarized by Inskipp and Inskipp (1995). Through perusal of the

avicultural literature we located an additional, previously unrecog-

nized, published photograph of an intermedia-like bird, and

correspondence with aviculturists and researchers (after the main

Statistical analyses for this paper were complete) resulted in addi-

tional unpublished information, including the location of several

more captive birds, some of documented parentage.

HYBRID DIAGNOSES

For hybrid diagnoses (Graves 1990), plumage, other external char-

acteristics, and measurements of adult males were compared among

the species of Psittacula that either had been suggested previously as

possible parental taxa (Husain 1959, Forshaw 1973, Wolters 1975)

or for which the phenotype of the presumptive hybrids indicated the

likelihood of those species being involved. Sane’s birds were com-

pared indirectly with the other intermedia specimens (Table 1) and

with series of adult male cyanocephalaand Rose-ringed Parakeets P.

krameri (Table 2), while other intermedia were compared with

series assembled at the National Museum of Natural History (USNM)

of each of the potential parental species: cyanocephala (n = 21),

himalayana (n = 26), Grey-headed Parakeet P. finschii (n = 17),

krameri (n= 10), and Blossom-headed Parakeet P. roseata (n= 11).

Additionally, all adult males of these species in the collections of

AMNH, the Academy of Natural Sciences of Philadelphia (ANSP),

the Museum of Comparative Zoology (MCZ), and the University of

Michigan Museum of Zoology (UMMZ), as well as several smaller

collections, were examined and measured; their plumage characters

(which did not differ materially from those of the series assembled

at USNM) were not included in the analyses, but their measurements

are included in the Appendix and in the statistical analyses. A few

unsexed specimens with plumage characteristics diagnostic of males

were included. All other species of Psittacula were ruled out as

potential parental species as they have plumage characters strongly

incompatible with the phenotype of specimens reputed to be

intermedia.

Mensural characters of adult males (listed in Appendix) were

used to evaluate which (if any) of the species listed above could

potentially be parental species of the intermedia specimens. Measure-

ments taken as far as possible for each specimen were: culmen

length (from distal edge of cere); height and width of maxilla (upper

mandible, at distal edge of cere); minimum distance between nares;

width of (lower) mandible; wing length (straightened and flattened);

shortfalls of each primary (P1—P10, with P1 outermost) from

wingpoint; for P1, distance from notch on inner web to feather tip,

maximum width, and width at notch; widths of P2—5, each taken at

TAXONOMIC STATUS OF PSITTACULA INTERMEDIA

ajeymeds Apysiys

diy 0} ysinyq wuoyrun

anjq-ystusess oped

feondiyyja ‘ayerapour

SUISSTU

SUISSTUL

[eondiyja ‘ayesopour

MoleU

ay1YM-YsIMmoy[od

anyjq-ystusel3

[eondiyya ‘ayerapour

MOLIeU

MOTJ9A ound

uda13 [OOS

pajurod ‘ayeopour

MOLIeU JOyeI

MO][aX asosuLd

anyq sep

[eondiyja ‘aresapour

a1e1apour

MO]J9A ayed

anyjq [eAor

arenbs ‘proiq

peoiqg

MO]JA oped

aniq yep

aienbs ‘ayerapoul

oye19pour

MOT[OA

anyjq [eo

orenbs ‘a\e1apour

SUISSTU

SUISSTUL

SUISSTUI

orenbs ‘prog

dn [x Jo pearg

Ta Jo dn yo inojop

Ty JO uonsesprut Jo mojoD

TIVE

€d jo adeyg

YSIS ISIS auou ouou quasaid Suryor| quasoid juasoid quasaid == sy19A09 BuIMIoddn 0} a8uy ont g

osu Ajsni 1YsI]s ouou auou ouou quasaid juasaid quasaid quasoid quasaid yoyed Jopynoys Jo souasa1g

ONIM

YSIS YSIS Suryory] suons suons ayeJapoul suons aye1opou a}e1Opoul duini uo a8un anyq Jo yua}xq

SAT[O-MOTI[AA IYS1IG SAT[O-MOT[OA JYSUG AIO-[9A JOyIEp “|S UdAIB [OOD = IATO-"]9A TYP "|S 9ATJO-"J9A TYP ‘IS SAT[O-MOT]OA SAT[O-MOTJAA —- DAT[O-MOT[OA yorqg Jo 10[oD

aye1apoul ayerapoul aye1apoul auou Moe AIDA MoLIeU AIDA peoig A100 aye1apoul peroiq Je]JOO uo anjq Jo JuaX|

UPIPLITA UPIPLITA uaeis-anjq ayed =usess-anjq Aysnp = uder8-anjq ajed = uaaas-anjqayed §=—s uaauS-anyq ayed ~=—uaa3-anyq aed uaais-anjq aed Ie][oo adeu jo 10joD

AGO

anyjq uRe[NIa9 YystIq «= aN{q URayNIIO 1YsLUqQ Aois-any]q Aois Aysnp anjq-Aois anjq-Aais anjq-Aois anjq-Aois anjq-Aois adevu/uMmoio IeaI JO INojOD

ystXois oeyt]-ystAois yurd-ystmoy[aA ude13-MO][ah yuid-ysrjdind yuid doap yuid-ystjdind yuid-ysyjdind = yurd-ystjdind SaI0] JO INO[OD

ajayduio9 aa[duios auou auou A ouou auou ouou auou a[qipueul uo yep Jo Juax_

por 1ysuq por 1y311q aeJopou ouou suons ayed A190 aed suons suons PI[IXew UO pos/asuvIO Jo yJdaq

pol [je por [je aSurIO ouou asurIO 7% jewruru asur10 Y aSUPIO % asuvl0 % —«R| [Xe UO pol/asuPIO Jo JUaIX_

avaH

ynpe ynpe qnpe onyeUUUT y[Npe {inpeqns ynpe ynpe ynpe ssejo os

pansojeyeoun 8SL9T Ve 086l SPSIc9 brsic9 CrsSIc9 IvS1c9 OvSIc9 6€S 179 Jojourey)

oues SHN4@ HNW4 HNWYV HNNV HNNAV HNWV HNWV HNWNV

“DIPAWAAJUI DINIVDIJIS_ O} paynginye susuttoads [enprlArpur Jo ssojoeseyo sayneyyenb [euslayxq | I[Qey,

P.C. RASMUSSEN AND N.J. COLLAR

Table 2 Qualitative characters used in hybrid diagnosis between Psittacula cyanocephala, P. krameri, and S. R. Sane’s specimens and living bird.

Character P. cyanocephala Sane’s birds

Maxilla all yellowish all red

Mandible black black

Cere shape moderately wide, rounded narrow, nearly straight

Cere colour medium grey pale greyish-horn

Orbital skin dark grey fleshy whitish

Lores pattern

rather wide, not prominent

no line

Forehead reddish-purple

Auriculars and central face reddish-purple

Lower face purplish

Rear crown shining mauve

Lower border neck collar as nape

Nape bright bluish-green

Upper wing coverts variably bluish-green

Shoulder patch maroon

Underwing coverts pale turquoise-blue

Rump variably bluish-green

Uppertail rich dark blue

Tail tip white, moderate width, spatulate

Foot dark pinkish-grey

rather narrow, very prominent

slight line

greenish tinge

dull purplish-blue

cerulean blue

broken orange-chestnut

Viridian

slightly bluish

slight tinge on one

slightly bluish

pale blue-green

concolorous with rest or

narrow whitish; slightly spatulate

pale pinkish-grey

P. krameri

all red

black

narrow, straight

whitish

orange

rather narrow, very prominent

strong line

green

bluish-green

lime green

powder blue

nearly complete rose-orange

slightly bluish-green

lacking bluish

absent

yellowish-green

lime-green

concolorous with rest, not spatulate

whitish

tip of next shortest feather; lengths of longest (R1, central) and next

more lateral (R2) rectrices (both taken from insertion of central

rectrices); maximum width of yellow or white tip and approximate

maximum width at distal end of blue or green area of R1 (with

feathers flattened out); distances between tips of each rectrix (except

R1) of one side and the next shortest (next more lateral) rectrix;

widths of each rectrix of one side at the tip of the next shortest one;

approximate distance from tips of R1 and R2 to definite blue or

green part of feather (= length of pale tip); tarsus length; minimum

width of tarsus; length of claw of middle toe (from distal edge of

scute); length of hindclaw (from distal edge of scute). Feathers in

sheath or in a damaged or heavily worn state were not measured, and

if there was a difference in length between rectrices of a pair, the

longer one was measured. Maximum skull width was measured over

skin and compressed feathers for specimens in which palpation and/

or x-rays indicated that the widest portion of the skull was intact and

not padded with stuffing.

Two specimens from the Rothschild series (BMNH 1980.3.1 and

AMNH 621545) showed very different plumage and mensural

characters from each other and from the remainder of the specimens

in this series, and so were treated as unknowns in the analyses.

AMNH 621545, although thought a female intermedia by Rothschild

(1907), was considered by Biswas (1959, 1990) and Forshaw (1973)

to be an immature himalayana, the latter opinion being shared by us

after examination. We therefore compared its plumage characters

with known immature himalayana and finschii, and its measure-

ments with nine juveniles (sexes combined) of the former.

Univariate statistics and principal components analyses (PCAs)

using correlation matrices were run on external and skeletal meas-

urements using SYSTAT for Windows (Version 5.0) on an

IBM-compatible PC. Variables for PCAs were chosen to allow the

inclusion of selected individual intermedia specimens without esti-

mation of missing data, which would be inadvisable owing to the

small sample size of intermedia.

EVALUATION OF ORIGIN OF SPECIMENS

To test the idea that the Rothschild Collection series of intermedia

had the same origin — an argument first put forward long ago by

Hartert (1924) — we compared preparation styles and materials used

among these specimens by external examination and study of x-rays.

We also compared them with native-prepared (e.g., ‘Bombay prepa-

ration’, ‘India’, and ‘Madras’) skins of other Psittacula species

(cyanocephala, himalayana, roseata, finschii) at AMNH, MCZ, and

USNM. To permit analysis of certain aspects of preparation styles

and materials used, radiographs (x-rays) were taken of the Rothschild

intermedia specimens, and of native skin specimens of himalayana,

finschii, and cyanocephala for comparison. X-rays (ventral and

lateral views) were made of the AMNH and BMNH intermedia

(including the putative immature himalayana) by M. N. Feinberg,

Department of Ichthyology, AMNH (30 kV and 3 mA for 2 min,

using Kodak Industrex-M Ready-pack film), and for the other

specimens by PCR in the Fish Division, National Museum of

Natural History (USNM; 25 kV and 5 mA for 30 sec, using Kodak

Industrex SR film).

To evaluate whether the Rothschild Collection series originated

from wild, not captive birds — an argument used by Biswas (1959) to

support species status — we examined the Rothschild specimens for

presence of: overgrown bill and claws; broken remiges and rectrices;

overly worn feathers due to delayed moult; abrasion damage to

feathers of the type resulting from repeated contact with cage bars;

and dirt on plumage, bill, and feet consistent with a confined

environment.

EXAMINATION OF CAPTIVE BIRDS

We examined five living adult hybrids belonging to Mr M. Sedgemore

that are the progeny of an experimental pairing of a male

cyanocephala and a female himalayana. The female parent, which

died in the nest shortly after producing the second of two hybrid

broods in successive years, was considered unsalvageable as a

specimen; the male parent died more recently and the skin is

preserved as BMNH 1998.33.2.We took hand-held photographs and

aviary videotape of all five hybrids, as well as several measurements

(taken by PCR while the birds were held by Sedgemore) of bill,

wing, and tail. All the hybrids were in some stage of moult, so certain

measurements could not be taken. The recently moulted central

rectrices of the single female hybrid are now at the BMNH, and

Sedgemore also gave us several photographs of the hybrids, both as

juveniles with their parents and as adults. PCR also examined and

videotaped the single live bird claimed to be intermedia remaining in

Sane’s collection in December 1997. We also studied photos sent by

TAXONOMIC STATUS OF PSITTACULA INTERMEDIA

MrL. Critchley of yet another adult male hybrid and its parents. This

bird was one of five hybrids in a single brood that Critchley

incidentally produced by housing a male cyanocephala in a mixed

aviary with a female himalayana, but he sold the other four hybrids

to a pet shop in the U.K. before they attained adult plumage. Photos

of all captive birds discussed in this paper are on file both with

BirdLife International HQ and the senior author, and selected photos

showing each specimen will appear on a colour plate accompanying

a short article on Psittacula intermedia (Rasmussen and Collar in

press).

RESULTS

SANE’S CAPTIVES

BNHS 26758 and Sane’s other two birds, one stuffed and one alive

as of December 1997, are all captive, dataless males identified as

intermedia. However, all proved to be considerably different from

the Rothschild specimens, possessing characters of both cyano-

cephala and krameri, but none inconsistent with their being hybrids

between the latter two species (Tables 1 and 2; Appendix). Sane’s

birds have much bluer, paier sides to the head and a greener mid- and

hindcrown than do any of the Rothschild Collection intermedia; like

the latter they have mainly lilac cheeks, but many feathers of the

head are multicoloured, at least on the two skin specimens. On the

sides of the crown and edge of the black moustache, most of the

individual feathers have peach-coloured bases and blue tips; on the

cheek the bases tend to be peach and the tips lilac; while the feathers

of the centre of crown and nape have green centres and blue tips. The

specimens have entirely black lower mandibles and all-red upper

mandibles except for the paler tips. Their soft-part colours and the

Table 3 Component loadings for PCAs of (A) bill width, wing, and tail

measures for a model including juvenile Psittacula himalayana; (B) bill,

wing length, and rectrix 5 measures, including Sane’s mounted

specimen and three of Sedgemore’s living birds; and (C) head and wing

measures, including BNHS 26758 (Sane’s study skin).

Factor

A B (G

Measurement 1 2 1 2 1 2 3

HEAD

Culmen length - - 0.93 0.03 0.82 -0.49 0.06

Maxilla height - - 0.96 0.06 0.87 -—0.36 0.10

Bill width 0.67 0.06 0.95 0.05 0.85 —0.38 —0.20

Skull width - - - = 0.84 -—0.24 —0.02

WING

Wing length 0.95 0.11 0.93 -0.12 0.92 -0.20 0.12

P3 shortfall - - - — -0.09 0.81 0.06

P4 shortfall ~ - - - 0:67 . 0:57 0:24

PS shortfall - - = - 0.86 0.41 0.09

P6 shortfall 0.93 0.16 — - 0.92 0.29 0.10

P7 shortfall 0.93 0.24 —- - O95 0n16" 10115

P8 shortfall 0.97 0.14 —- - O95 O21) 10M

P9 shortfall 0.96 0.19 —- = OMOGTy HOMs One

P10 shortfall 0.96 0.19 — - OO5— OS OkkG

P1 notch length - - - - 0:68) —0'55, 10512

Pl maximum width 0.68 —0.43 -

P1 notch width - - - - 0.36 —0.57 —0.45

P2 width 0.62 -0.59 — - 0.58 0.56 —0.42

P3 width - - — _ 0.66 0.31 —0.49

P4 width = - - - 0.81 0.11 —0.34

PS width - - - - 0.86 -—0.16 —0.27

TAIL

R1 width 0.83 0.02 —- - - - -

R2 width 0.75 -0.50 0.56 0.76- —- - -

feathering at the bill base are all unlike AMNH intermedia. The

maxillae of Sane’s birds are smoothly rounded on lateral view and

not particularly robust proximally, being very similar in shape to

krameri, not himalayana. Two of the three individuals bear no

indication of the red shoulder patches (and they are very vague in the

third) that are shown by both male cyanocephala and himalayana,

and that are present in five of Rothschild’s six adult intermedia, but

that are always lacking in male krameri. However, of all the features

in which Sane’s three birds differ from AMNH intermedia, none is

more telling than the broken orange-chestnut neck ring of the former

(Table 2), which (assuming that the birds are hybrids) can hardly

have come from any source other than krameri or the much larger

Alexandrine Parakeet P. eupatria. Also, in both of the individuals

with the central rectrices present, the feathers have very small pale

tips, consistent only with the latter two species.

In 1990, one of two ‘intermedia’ then alive in Sane’s possession

was photographed in Bombay by R. Wirth. The published photo

(Wirth 1990) shows a bird very similar to BNHS 26758 and Sane’s

uncatalogued specimen, and from the date it may be either the bird

still living as of 1997 or the uncatalogued specimen; it possesses the

same suite of features consistent with its being a hybrid krameri x

cyanocephala (or possibly krameri x roseata). A description of the

second live bird was not provided, but Sane considered both to be

intermedia, and Wirth (in litt. 1997) noticed no differences between

11.5%, 1.0; R4 vs. rectrix widths

Factor 2

Factor 1 70.8%, 6.4; size (R4 uncorrelated)

Fig. 1 Identity of AMNH 621545 with immature Psittacula himalayana:

graphs of individual scores (circles), group means (triangles), and 95%

confidence intervals (open ovals) of factor scores from principal

components analysis (PCA) on measurements of adult male P.

cyanocephala (C, grey-filled circles), P. himalayana (H, white), AMNH

P. intermedia (black), immature P. himalayana of both sexes (diagonal

hatching); P. roseata (R, diagonal cross-hatching), both populations of P.

finschii (F, horizontal cross-hatching), one of Sane’s specimens

(checkered), and P. krameri (K, horizontal bars). A polygon outlines the

scores for AMNH P. intermedia specimens due to small sample size.

Summary statistics presented in the axis labels are percent variance

explained and eigenvalues, respectively, followed by important measures

for each axis. Component loadings for PCA are given in Table 3A.

the two. PCR examined all three of Sane’s birds in the space of two

days and concluded that they lacked salient differences, all being

apparent krameri x cyanocephala hybrids.

IMMATURE SPECIMEN

AMNH 621545, considered to be a female intermedia by Rothschild

(1907), but thought by others to be an immature himalayana (Biswas

1959, 1990, Forshaw 1973), shows no relevant plumage differences

from the series of immature himalayana at AMNH with which we

directly compared it, nor from others at USNM and other museums

with which photos of it were compared. In a PCA of several

measurements, AMNH 621545 falls within the 95% confidence

limits of immature himalayana of both sexes (Figure 1, component

loadings in Table 3, summary statistics in Appendix). In body

plumage 621545 resembles himalayana in being generally cooler

green and less yellowish than immature finschii, although some

juveniles of the two species overlap in this. AMNH 621545 is unlike

juvenile hybrid cyanocephala x himalayana (see below) in its

bigger, duskier maxilla, brighter green nape, lack of yellowish

collar, bluer-green overall body colour, and especially in its bright

green upper tail surface with a bright yellow tip. Thus, on the basis

of both plumage and measurements, all evidence supports the

hypothesis that AMNH 621545 is an immature himalayana, and we

therefore exclude this specimen from further analyses.

BMNH SPECIMEN

Comparison of photos and measurements of BMNH 1980.3.1 with

those of the adult male AMNH intermedia showed that the former

has several differences from all other intermedia, although it is part

of the Rothschild series, and despite the seemingly inexplicable fact

that the BMNH specimen was the one upon which Husain (1959)

based his conclusion that intermedia was a hybrid himalayana x

cyanocephala. This bird was therefore lent to AMNH for our

comparisons, where we confirmed (Table 1, Appendix) that it has a

smaller maxilla with only a slight reddish tinge basally (less than in

allAMNH birds except 621542, the specimen said by Biswas [1959]

to be completing post-juvenile moult); it has a slightly duller head

with paler reddish-purple on the face (washed yellowish in front of

the eye) and paler greyish-blue on the crown and nape; and it totally

lacks maroon shoulder patches. Its P3 is narrower and less squared

at the tip than in all adult AMNH specimens except 621544. The tail

is greener at the base, more turquoise for most of its length, and has

the pale tip whiter and shorter. This specimen is the only one of the

Rothschild series that has a measurable, fully grown tail, so its

rectrix length cannot be directly compared with the otherintermedia.

The salient differences between BMNH 1980.3.1 and typicalroseata

are: the former lacks reddish shoulder patches; it has an entirely pale

lower mandible (though this is nearly all-pale in a few roseata;

Table 4); its hindneck has a turquoise tinge; it has a slightly broader

tail tip; the front of its face is slightly redder; and its P3 tip is broader

(Table 1). In most statistical analyses, BMNH 1980.3.1 falls within

the roseata and cyanocephala groups (Figures 2-5).

REMAINING JNTERMEDIA SPECIMENS

The other five AMNH specimens attributed to intermedia (including

the holotype), and also Bhargava’s three specimens, are quite similar

to one another. However, although most previous authors (Rothschild

1907, Hartert 1924, Husain 1959) have treated the first five under

one description as if they were identical, they are in fact variable in

most of the characters that separate them from any of the putative

parental species (Table 1). Only Biswas (1959) mentioned variation

among these five, but even he called them ‘exceedingly similar’. All

have fairly large bills with varying amounts of orange at the base of

the maxilla. All have nearly or entirely pale lower mandibles,

P.C. RASMUSSEN AND N.J. COLLAR

although AMNH 621544 has a broad black stripe down one side of

the lower mandible. Each has the front of the face bright purplish- to

deep pink, grading into the duller grey-blue crown, nape, and lower

portions of the face. All have a pale blue-green collar, but this is

highly variable in breadth and prominence, even allowing for differ-

ences in preparation. In addition, all have a bluish wash of variable

strength on the wing coverts and/or rump. Only three of the speci-

mens now have the central rectrices present, and in none of these

(contra Biswas 1959) are they fully grown (this cannot now be

determined in AMNH 621540, the fourth intermedia that once had

central rectrices, but the fact that they are now missing from this

specimen suggests they were loosely attached and thus moulting).

Thus, original tail lengths presented in previous treatments — Biswas

(1959): 185, 202, 221mm; Husain (1959): about 220 mm; Forshaw

(1973): 167-195 mm (mean = 180.7, n = 3), 206 (n = 1) — would be

expected to be too short. However, we measured the central rectrices

of the three specimens in which they are now mostly grown as 157,

200 and 170 mm, and the now-missing rectrices of AMNH 621540

were estimated at ca. 200 mm. The central rectrices of all four of

these birds show (or showed) long, at least moderately broad, pale to

pure yellow tips and dark or royal blue upper tail surfaces. The

breadth and length of the yellow tail tip are quite variable, however,

and the length of the yellow R1 tip of AMNH 621540 is estimated to

have been 48 mm, compared with a mean tip length of 41.6 mm for

the others (Appendix). From the photograph of Bhargava’s two

specimens for which the central rectrices are present (both photo-

graphed next to a cm rule), these rectrices appear to be ca. 234 and

229 mm (although it cannot be determined from the photos whether

these rectrices are full-grown), while the yellowish tips are ca. 51

and 54 mm, respectively.

We found no external qualitative characters in AMNH intermedia

or Bhargava’s specimens that differ from those exhibited by at least

one member of one of the two species groups (roseata/cyanocephala

and finschii/himalayana), or that are not intermediate between them

(Table 4). Among the potential parental species, roseata exhibits the

most plumage features incompatible with the AMNH intermedia

phenotype, while finschii also has a few characters inconsistent with

intermedia, mostly in tail shape and colour. Neither cyanocephala

norhimalayana shows any plumage features incompatible with their

being parental species of AMNH intermedia, and a combination of

the former two readily explains all plumage features of the latter.

STATISTICAL RESULTS

Summary statistics for measurements of the putative intermedia

specimens (with the BMNH specimen treated separately),

Sedgemore’s hybrids, Sane’s specimens, and comparative samples

of the five putative parental species are given in the Appendix. For

almost all measures, theAMNHintermedia are intermediate between

cyanocephala and adulthimalayana, and in many cases also between

others of the putative parental species. Bivariate scatter plots of

selected measurements overwhelmingly demonstrate this pattern,

e.g. Figure 2A showing culmen length from cere vs. culmen width,

in which allAMNH intermedia and Sedgemore’s hybrids fall between

the cyanocephala/roseata pair and the himalayana/finschii pair.

Furthermore, krameri is larger than, and Sane’s two specimens are as

large as, the himalayana/finschii pair. A slightly different pattern is

shown in Figure 2B (wing length vs. culmen length): here again,

AMNH intermedia and Sedgemore’s hybrids fall between

cyanocephala/roseata and himalayana but, because of the shorter

wing of finschii compared with himalayana, there is slight overlap

between finschii and intermedia. Psittacula krameri is similar in

wing length to himalayana but is bigger-billed, and Sane’s birds fall

between cyanocephala/roseata and krameri, being considerably

TAXONOMIC STATUS OF PSITTACULA INTERMEDIA 37

Culmen width

Culmen from cere

Wing length

Pale tip length of R2

R2 distal width

Fig. 2 Bivariate scatter plots of measurements of putative parental species and hybrids: (A) culmen length from cere vs. culmen width; (B) wing length

vs. culmen length from cere; (C) R2 pale tip length vs. R2 tip width. Symbols are as for Figure 1, with the addition of the BMNH intermedia specimen

(white square) and Sedgemore’s hybrids (black squares).

[a4

—]

(e)

jaa)

[24

oO.

“QUO UI posuN-MoOT[AA ApYSI]S,, suduToeds WOM ¢€ UT YSIN]| gs “JoyJOuR UI YstuseIs poysem ‘suouroads uJOM ¢€ UI [[NC,

Seyur] Wg wos uswIoads ou ur [[Np A19A ‘adeu Jo a8pa JOMOT Iv JsosuONs sXemye ‘a]quueA, ‘Aoid aed sa10] sey YZ/Tp ‘eyed s[qrpueul Jo a10UW JO ¢ aAvy | [/E, ‘a[ed afqrpurU Jo a10UI JO Ff DALY ()Z/Sq :9UO UT aseq Jv Yyoyed YsTyor|_,

‘sotoadssadns 10yj0 ay} JO JAqUIOUI JOYIIE JO sIa}OeILYO YIM UOHeUIquIOS ul adMousYd syT]-pIpaustajUI Ue 0} 9INGLNUOS 0} ATOYI] JOU S1O}ORIeYS sa\oueP (,,) YSU}se UY

$90} Ue) JaAars ‘Iayep

Aais ayed

UI}.

MOT[EA posuny

-a5uvIO ‘snooneys,.

ysorym 0} uMoIg aed,

ajdind-ysnyq [np

ugeis-ysinyq ojed

[eordiyja ‘ayesopour

uoorew yep

Ud9I8-YsIMo]]OA

UddIS-YsIMOTIAA JYSLIG

(apuewi<<) usaI3

-MOTJOA uouTUTOId,.

poulyep-][om ‘Moreu

on] q-ystuseis WYysLIq

$90} Ue) J9AOIS ‘oy.ep

Agis oyed

peoiq

MO][OA oind 7Yy311qG

UMOIG yep

onyq-ystjdind [np

ugeis-ysinj{q ajed

arenbs ‘aye1apow

uooreul yep A19A

gU99013 [009

ua913 [009

(a]JUeUUI=) dAT[O [OOS

prog 0) ae1opoul

ugais-ysinjq ojed 04

anjq-ystuseis ayed

$90] se dues

Aois-ASOJ WINIpoUul

ayejmeds Apysiys ‘ayeropour

MOTO ayed

anyjq [eAo1

ugeis-ysinj{q oyed

ysuenbs ‘aye1apour

uooreul

posun-onjq

u9d13-YsIn]q

(epueUt=) aAtfo-moj[a4 [INP

oye. 1Opoul

udo1s-anyq ayed

$20} se OURSZ,

Aois-ASO1 WnTpoul

peoig 0} MOeU JOyIeI

MoOT]9A dnd 0} ayed

yor lq

anyq yep 0} [eAor

uad13-ysin{q ajed

(%) orenbs ‘(%) aye19poul

UOOJPUU-pru

(%) pesun-onyq

u99.15-YsIn]q

(apuru=) sArfo-Moy[ad [Np

peoig AIA 0} MOLIeU

uso13-an{q ojed

Woy Yepe

Adis yrepy

ur}.

MO]IOA oped AroA

yor[q

anjq-astonbin} daap..

udd13 MOTIOA,.

jeondiyja ‘morreu,,

woo yep

ugois oind,,

udeis yYysLq sind,

(afURUI=) dATTO

Quou,.,

uddI3 SUIT].

$20} sv ures

Aogis-ysryuid yep

aye1opoul

yoy

yor lq

anyq 4ep You

anjq-astonb.in) ayed

arenbs ‘prog A19A

uooJeW 1Y31I1q

udo13-ysintq A[qeiiea

udeis-ysinjq A[qeiiea

(epueUIZ) aArTo 1YSTIq

Modeu AT[ensn

ude18-YsIn]q IYs11q

INo[Oo MET

INO[O9 90],

Ou Jo dn jo peg

Ta Jo dn yo mojo5

O[PPIUU Iv yeYs TY JO nojoD

Ty JO wonsesprur Jo mojoD

TIV.L

SJIOAOO SUIMIJOpUN JO INO[OD

€d Jo adeys

yoyed 3UIM pal Jo INO[OD

$}19A09 IOSsa] JO INO[OD

ONIM.

duini jo mojo

Japsog apueUl/Ie[[oD Jo INojoD

Ie]JOO UO anyq Jo Waxy

Ieyjoo adeu jo mojop

xAdOg

Aais Ayeys Wem Aoais Aye]s ploo onjq-Aois anjq-Aois PII x ,SANPU SUTUTYS adeu/uMolO Ied1 JO MO;OD

Kors Ayeys Agis posun-useis yurd-ysijdind yurd-ysijdind ajed ¢ Yysiq ¢ yuld-uowyes,. per IYys11q 0} (g) yuId Yys1Iq S910] JO INO[OD

MOUeU MOLIeU a)8.19pour a1eIapouls, OpIM ATOANLIAI apm AToAyeIo1 ULYS [PIIGIO o1eq JO PIM

poprArp Aped ‘opi poprAip Apued ‘opin UOISIAIP JYSI[S “OprM Ayey UOISIAIP JYSI[S ‘apr Aparey UOISIAIP JYSI]s 10 OU ‘apiM AyIIey UOISTAIP JYSI[S 10 OU ‘apm ATATey 2199 Jo adeys

ujoy-asue10 aed asurio ajed uwioy Wem Woy WAeM 0} [eNOU yspyor[ qs. pAois-UunI paul 9199 JO INOJOD

UIOY-YsTMoT[oA WIOY-YsTMoT[oh WIOY-YsIMo][oA UIOY-YsTMo][oA > ed (¢] orig 2[qIpuew Jo mMojoD

Aavay ayes Aavoy ayeJOpoul a}e.1opoul [jews [yews PI[IXBU JO 2ZIS

pol-osue10 o8URIO-pol asurl0 osuvio UIOY-YsIMoy[oh (.0Z/6L) UIOY-YstMoy[oA PI[IXeBU JO aseq JO INO[OD

avaH

nyosuyf q puvkvjouny J spuigdy s,o10ulaspas pIpawiaqul HNINV DIDASOL Y pypydaz0uv4o gq

soroadsiadns ,popeoy-Ajr]s,

Dipawuiajul J

sotoadsiadns ,poperoy-wossojg,

“T Q[QRL, 998 YOIyM JO} ‘Vipaiwsajui ynq [ek 10} payeoipurl UONeLIeA,

‘(L] =u) nyosurf gq pure ‘(97 = Uv) DuDkyjouny gq (+ =U) Splaghy s,a10Wespag ‘vipauiaqul q HNIWY ‘(11 = 4) vivasos J “(17 =U) vjpydasoundo vjnovyisg JO sisouseIp pliqdy UI pasn siajoweY) fp FUL.

TAXONOMIC STATUS OF PSITTACULA INTERMEDIA

K—S— 10

H-=_ =— 42

juv. H—=—j—— 7

Eastern F—==- 30

Western F == 20

Sane 2

Sedgemore = 5

In@5

pet 40

R—=Z=— 35

el At aro

Factor 1 92.1%, 15.8; bill size

Fig. 3 Graph of means (squares), standard deviations (heavy bars), and

ranges (narrow bars) of individual scores for putative parental species

and hybrids on Factor 1, the only significant axis in a PCA of culmen

length (component loading 0.95), maxilla height (0.96), and maxilla

length (0.97). The number to the right of the range bar is n. Symbols are

as for Figure 1 and 2; eastern and western finschii are included as

separate groups.

larger than intermedia. In the plot of the distal width of rectrix 2 (R2)

vs. the pale tip length of that feather (Figure 2C), intermedia fall

betweencyanocephala/roseata andhimalayana, but not mostfinschii.

Psittacula krameriis unique in its combination of a broad, very short

pale tail tip, and in this Sane’s specimen is intermediate between

krameri and cyanocephala/roseata, and certainly not himalayana.

On a PCA of three bill measures (Figure 3, Table 3) selected to

allow inclusion of as many specimens and live hybrids as possible,

the only significant axis was Factor 1, a very strong size axis. On this

axis, cyanocephala and roseata had the smallest mean factor scores,

with the BMNH intermedia slightly larger. The other putative

intermedia and known hybrids fell between these and the succes-

sively larger finschii and himalayana groups. Psittacula krameri

was much the largest, and Sane’s specimens were the largest of the

putative hybrid groups, again showing the influence of krameri.

In a PCA for which variables were selected to allow inclusion of

one of Sane’s specimens (Figure 4A, Table 3), the AMNH intermedia

and Sedgemore’s birds group near each other, and between the

widely spaced roseata/cyanocephala and himalayana groups, but

overlap considerably with finschii. Sane’s bird, however, falls

between the roseata/cyanocephala and krameri groups.

Another PCA for which the variables selected allowed inclusion

of Sane’s other specimen (Figure 4B, Table 3) showed AMNH

intermedia grouping out halfway between cyanocephala and

himalayana, with the mean of roseata falling out more distantly. In

this case, the second Sane specimen falls out much closer to

cyanocephala than to krameri.

WING AND TAIL FORMULAE

In mean shortfalls of each primary tip from the wingpoint (Figure 5A),

the AMNH intermedia are completely intermediate between

himalayana and cyanocephala. However, BMNH 1980.3.1 is very

like the mean of cyanocephala in pattern of primary shortfalls from

the wingpoint. AMNH intermedia are closer in mean primary short-

falls to roseata than to cyanocephala (and thus less intermediate

between roseata and himalayana; Figure 5B), but neither roseata

norfinschii could be ruled out as parental species on this basis alone.

Sane’s single specimen on which these characters are measurable is

nearly intermediate in primary shortfall pattern between

cyanocephala andkrameri (Figure SC), although again these data do

not rule out some other parental combinations.

On mean widths of primaries, AMNH intermedia were again

intermediate between himalayana andcyanocephala except in width

of P2, a measurement that is highly dependent on shortfall of P3

(Figure 6A). In spacing between tips of rectrix pairs 3-6 (Figure 6B),

17.7%, 1.1; R2 tip length vs. width

Factor 2

a. S 0 l 2 3

Factor 1 71.5%, 4.3; size (R2 w, tip | uncorrelated)

// Sane2 9 Sra

/ s ¢

@ F

—————

3.0%, 15.9; bill and Plw, notch vs. P3s, P2w

Factor 2

2 -1 “0

Factor 1 12.1%, 63.7; size (P3s, Plw uncorrelated)

Sse | Q)

Fig. 4 Identity of AMNH intermedia with Sedgemore’s hybrids,

distinctness from Sane’s specimens, and intermediacy of all the above

between putative parental species: graphs of individual scores on

Factors | and 2 from PCAs on measurements of adult males of putative

parental species and hybrids. Symbols are as for preceding figures.

Summary statistics of PCAs are given in Table 3B and C. (A) Variables

chosen to allow inclusion of Sedgemore’s hybrids and Sane’s first

specimen; (B) variables chosen for inclusion of Sane’s second specimen.

40 P.C. RASMUSSEN AND N.J. COLLAR

80; == himalayana

j ; 15 —

70| ™ = AMNH intermedia

Aj oe cyanocephala

— BMNH intermedia =

50 |

: m= 10 —

30 =

A %,

: | | | | | |

1 (max) 1 (notch) 2 3 4 5

Primary No.

os mum roseata

= "70

= = = AMNH mtermedia

& 60) = finschii

Be =~ 80

= 50 =

: E

= 40 Z

: 2 60—

e380 =

: 3

S 20 2

2 ae

a 10 E

B ry

8 20—

A B TITTLE

| | | l

* == 2 3 4 5

krameri Rectrix No.

== Sane specimen e

60| sini cyanocephala cS

” a en ;

E I

= 7.5 |— “ton, 3 --" aes eosin rl

4} “Ny, , OS hd 2 yo!

C 5.0 ) ) | ) | :

WAL ead jeep eub Ness Nee) 12A/ | Ney Nee) eal) 1 (tip) 1 (mid) : 3 4 5

Primary No. Rectrix No.

Fig. 5 Graphs of mean shortfalls of primaries from wingpoint for: (A) Fig. 6 Remex and rectrix width and spacing for Psittacula cyanocephala,

Psittacula cyanocephala, P. himalayana, and P. intermedia (BMNH P. himalayana, and P. intermedia: (A) mean widths of primaries,

specimen separate); (B) P. cyanocephala, P. roseata, and P. intermedia maximum and at notch for P1 (outermost), and width of P2—PS at next

(BMNH specimen separate); (C) P. krameri, P. cyanocephala, and one innermost remex; (B) mean spacing between rectrices; (C) mean rectrix

of Sane’s specimens. Full data are presented in Appendix. widths at next more lateral rectrix. Full data are presented in Appendix.

TAXONOMIC STATUS OF PSITTACULA INTERMEDIA

the AMNH intermedia are again totally intermediate, but BMNH

1980.3.1 differs from intermedia and is very like roseata (Appen-

dix). In rectrix widths (Figure 6C), AMNH intermedia are more or

less intermediate, although the central rectrices are somewhat closer

in width to those of cyanocephala, while the outer rectrices are

closer to those of himalayana; BMNH 1980.3.1 differs here as well

(Appendix).

STYLES OF SPECIMEN PREPARATION

Examination of preparation styles of the Rothschild intermedia

(Figure 7) other than the type (which predated the others) but

including the juvenile himalayana (AMNH 621545), showed that

one (AMNH 621541) has a longish neck and understuffed throat (vs.

short necks and breast nearly touching the bill on the others); another

(AMNH 621542) has the bill slightly open, while in three (AMNH

621544—-5, BMNH 1980.3.1) the maxilla is extended far beyond the

mandible (vs. naturally positioned in AMNH 621540-1). The wings

are positioned far forward and low on the body on two (AMNH

621544—5), and low but to the sides on two others (AMNH 621540—

1) vs. well-positioned to the sides on the remaining two (BMNH

1980.3.1, AMNH 621542). The body is compressed dorsoventrally

in all but one (BMNH 1980.3.1). The tail is twisted in relation to the

body in all but two (BMNH 1980.3.1, AMNH 621541); one with a

twisted tail (AMNH 621542) has one of its central rectrices rotated

180° and its rectrices are spread, while they are folded tightly in the

others. One specimen (AMNH 621542) is filled with dirty cotton,

while another (AMNH 621544) has the body made of a tightly but

roughly wound ball of coarse brown fibres each about 0.5 mm in

width, and the rest of the Rothschild specimens are stuffed with

rough bundles of straw. Support sticks in two (AMNH 621542,

621544) are thin (ca. 3 mm diameter), whittled, and orange-brown;

thicker (3.4 mm), rougher, and grey-brown in another (AMNH

621540); and very thick (ca. 7 mm), coarse, crudely broken, and

dark brown in yet another (AMNH 621545), while sticks are not

visible externally in the other specimens.

All the Rothschild intermedia share the following external prepa-

ration features: the eyes are not stuffed and are dried shut; the breast

is crudely stuffed so that the feathers of the upper breast are pushed

outwards and upwards; the abdominal incision is rough; the tibiotarsi

are broken medially and the feet were not secured, now being

entirely missing in three (presumably having been lost after prepara-

tion). Strangely, the only published photograph of any of the

Rothschild intermedia specimens (AMNH 621540, in Arndt 1996)

was digitally enhanced to add in a lifelike eye and periorbital skin,

even though it lacks a wing on the side photographed.

Radiographs (Figure 8) elucidate additional pertinent preparation

features of the Rothschild intermedia specimens: AMNH 621541

and 621545 both had similar loose-woven cloth wound around the

top of the support sticks and pushed into the open back of the skulls,

while the others have little or no stuffing in the skulls. The X-rays

confirm the similarity between the straw used in stuffing of five

specimens (AMNH 621539-621541, 621545, BMNH 1980.3.1),

and show that straw is lacking in two others (AMNH 621542,

621544). The body of AMNH 621540 is fusiform, while in AMNH

621544 the rear body is nearly empty. BMNH 1980.3.1 lacks a

support stick altogether. In some of the specimens (AMNH 621540-

1, 621545) the support stick is jammed into the braincase, while in

AMNH 621544 the tip lies between the orbits, and in AMNH

621542 it projects into the base of the maxilla. In all specimens, the

wings are positioned carelessly and variably, and those of the

holotype are positioned differently to the rest. Similarities among

the specimens visible in the x-rays include: most or all of the radii

and the entire humerus have been removed; much of the back of the

skull was removed but in an inconstant manner; the bones were often

haphazardly broken and bone chips are embedded inside five speci-

mens; and sacral vertebrae were left in five specimens.

Other dataless specimens examined by us that had been prepared

in this characteristic native skin style (the ‘Bombay preparation’:

Rothschild 1895) include the following abnormally plumaged

specimens: a partial lutinocyanocephala(AMNH 454031); a yellow-

tinged (flavistic?) cyanocephala(AMNH 621491); and a near-lutino

krameri (AMNH 454030). Bombay preparation Psittacula skins

with normal plumage at AMNH include: three cyanocephala

(621490, 621492, 621537); five himalayana (621551-621554,

621556); two finschii (621550, 621557); and one krameri (621337).

Further Bombay preparation skins are now in other collections (e.g.

MCZ 383245). No skins of the Bombay preparation were found

among the AMNH series of other Indian Psittacula species, al-

though other presumed native skin styles are represented among

them.

EVIDENCE FOR CAPTIVE ORIGIN

We found the following features among the Rothschild intermedia

that are consistent with their having been held in captivity: (1)

breakage or damage of primaries in three (AMNH 621542, 621544—

5); (2) loss of central rectrices on AMNH 621539 and ongoing

replacement of central rectrices on at least three others (AMNH

621541—2, 621544 and presumably 621540, showing that a high

proportion of the sample is in moult); (3) irregular dark worn areas

on feathers of the carpal area and heavily frayed wing coverts in

AMNH 621544; (4) a featherless patch on the left side of the upper

breast, and damaged feathers on the forehead of AMNH 621544 and

shoulder of AMNH 621542; and (5) dirt on the feathers of the breast

and/or belly of four (AMNH 621539, 621544—-5, BMNH 1980.3.1),

dirt on the right wing of BMNH 1980.3.1, and apparent whitewash

on the upper tail surface of BMNH 1980.3.1.

PHOTOGRAPHS AND OTHER REPORTS OF CAPTIVE INTERMEDIA

A photograph of a parakeet published in Herrmann (1994) as a male

cyanocephala is instead much like the five AMNH adult male

intermedia specimens. This individual (which has now been sold)

and its mate (which has died and for which no details are available)

were said to be wild-caught from an unknown locality and were held

in captivity in Austria, where the photograph was taken by F. Pfeffer

in 1985 (T. Arndt, in litt. 1997). Based on a copy of Pfeffer’s

photograph, the male differs from cyanocephala (and agrees with

AMNH intermedia) in having two-thirds of the upper mandible red-

orange and fairly large; a pale yellowish lower mandible; a large

head with more extensive, slatier-blue areas on the rear face and

head; a weaker blue wash on nape, wing coverts, and rump; and the

tail tip longer and yellowish. From the photograph of the male, it

appears to differ from the AMNH intermedia in having a paler, less

pure yellow tail tip; narrower central rectrices; a yellow-green area

between the mantle and the hindcollar; and a dark maroon shoulder

patch.

A second male (also of unknown provenance but from around

1985) was recently located in captivity at Turnersee Bird Park in

Austria by R. Low (in litt. 1997) and F. Pfeffer (T. Arndt in litt.

1997), and was almost immediately published (with colour photos)

as a true intermedia (Fuchs 1997). Photos of this bird (which is

missing parts of its toes) from all three above sources show it to be

very similar to the previously mentioned captive bird in Austria,

except that its bill looks smaller and less orange-red. Some of the

photos clearly show very narrow central rectrices with very long

pale yellow tips and whitish shafts encroaching into the blue portion

at least as far proximally as the level of the R2 tips. R. Low (in litt.

1997) stated it was the same size as the female cyanocephala with

42 P.C. RASMUSSEN AND N.J. COLLAR

Fig. 7 Photo of all Rothschild Psittacula intermedia skins to show preparation styles: (A) ventral view of (left to right) AMNH 621539 (holotype),

621540-2, 621544-5, BMNH 1980.3.1; (B) lateral view.

TAXONOMIC STATUS OF PSITTACULA INTERMEDIA

Fig.8 X-rays of all Rothschild Psittacula intermedia skins to show preparation styles and materials. Ventral view of (left to right, top row) AMNH

621544, BMNH 1980.3.1, AMNH 621539 (holotype), 621540, (bottom row) 621545, 621542, 621541.

which it was kept, and that the owners informed her that it had been

bred from cyanocephala at Vogelpark Turnersee. However, the

article about this individual (Fuchs 1997), which claimed that it was

a great rarity, does not mention its parentage nor even the possibility

that intermedia is a hybrid.

Sane (1977) recalled that in 1972 or 1973 he had seen a bird

similar to his first ‘intermedia’ , but that it had been purchased by H.

H. Jamsaheb, Nawanagar, India. He had also been told in about 1976

that someone in Britain had two or three intermedia, and that 8-10

had been imported into Holland in 1976. An individual was offered

for sale in 1976 as ‘probably the only known specimen in captivity

in the world’ (Inskipp and Inskipp 1995). However, we know of no

documentation for any of these claimed intermedia.

In a poster session at the 1996 BirdLife Asia Conference in

Coimbatore, India, R. Bhargava exhibited a photo of a captive bird

that closely matched the AMNH specimens already examined by

PCR. Bhargava (pers. comm. 1996) informed us that such birds were

not rare in the pet trade within India, but that the illegality of this

trade makes documentation difficult. He had obtained as many as

five individuals from traders at one time, although one of these has

since been ringed and released at a news conference (Anon. 1998)

and three have died and been preserved as specimens. Bhargava’s

three specimens appear from photographs to be indistinguishable

from AMNH intermedia, but definite, verifiable data on their prov-

enance appear to be lacking.

KNOWN HYBRID CYANOCEPHALA X HIMALAYANA

After the publication of Walters (1985), Sedgemore informed M. P.

Walters (pers. comm., 1995) that he had crossed himalayana and

cyanocephala in captivity and obtained intermedia-like hybrids.

Beginning in the late 1970s Sedgemore tried pairing captive

himalayana andcyanocephala to determine whether Husain’s (1959)

hypothesis was correct (M. Sedgemore, in litt. 1997), but his adult

male cyanocephala (age unknown) and five-year-old female

himalayana refused to bond. In the mid-1980s he tried pairing

different individuals of the same species, but again without result

(M. Sedgemore, in litt. 1997). Then in 1991 he housed an immature

female himalayana and immature male cyanocephala together, and

these showed pair behaviour that year but did not breed. In 1992, of

three eggs laid, two were clear and one was fertile but did not hatch,

and in 1993 three more eggs were produced, only one of which

hatched but the chick died at two weeks of age. Finally, in 1994, all

three eggs laid hatched and the chicks fledged, as did chicks from

two of three eggs laid in 1995 (M. Sedgemore, in litt. 1997). All the

hybrids — four males and one female — were still alive and in adult

plumage when we saw them in September 1997.

Photos of Sedgemore’s hybrids with their parents show that the

mother was a typical himalayana and the male parent a typical

cyanocephala, the latter being additionally confirmed by the speci-

men. All these photos and our direct examination, photographs, and

measurements confirm the identity of the hybrids with AMNH

intermedia. The heads of all four adult male hybrids were coloured

as in the five AMNH adult specimens, with only slight variability

among them. Their ceres were pale fleshy horn; their maxillae

orange-red on the basal two-thirds and with yellowish tips; their

lower mandibles were pale; their eyering skin was pale greyish; and

their feet and claws were pale greyish-pink. All had a broad pale

greenish-blue hindcollar, lesser wing coverts, ramp, and underwing

coverts, the undersurfaces of the rectrices yellow with the outer

webs bluish proximally, and the tail tips slightly broadened and pale

yellowish from above.

The single female hybrid was less distinctive but still possessed

characteristics which should enable recognition of specimens or

P.C. RASMUSSEN AND N.J. COLLAR

captives. It would be immediately distinguishable from adult

himalayana or finschii by its much paler, duller grey head and lack

of a narrow black collar, and from males of the above two species by

its lack of a maroon wing patch. Its head was drabber grey than adult

female cyanocephala, with a pale area in front of the eye much as in

juvenile cyanocephala; its upper mandible was heavier and was

strongly tinged orange at the base; its lower mandible was pale; it

had a slight yellowish collar on the sides of the neck (paler and duller

than in adult female cyanocephala and less ochraceous than in

female roseata); and it had long, slightly broadened, pale yellow tail

tips, which extend much farther proximally on the feather than in

cyanocephala.

Sedgemore (1995) briefly described the juveniles of the first

brood. His photos of four of these same hybrids as fresh-plumaged

juveniles show that they would be difficult but not impossible to

distinguish from those of either parental species. All four hybrids

had a pale area on the front of the face (on the forehead, lores, and

area around the bill base); greenish-grey auriculars; a dull green

crown and nape; and a pale yellowish-green collar contrasting with

the head and mantle. From below, the tails of the hybrids were

narrowly yellow-tipped. The upper tail surfaces are visible only on

two individuals, in which they were greenish-blue with a pale yellow

tip. Note that the latter character disagrees with Tavistock’s (1933)

statement (see below) that the tails of his hybrids were brighter blue

than in young cyanocephala and were white-tipped. A photo of one

of Sedgemore’s 1995 hybrids in direct comparison with its 1994-

hatched brother (an adult by then) shows that the juvenile has the

upper tail surface more turquoise than the adult male. The young

hybrids differed from juvenile himalayana by their smaller bills,

their lack of blackish blotches at the bases of the maxillae, their

duskier ceres, yellower collar, bluer upper tail surface, and narrower

rectrices, and from juvenile cyanocephala by their duller, less

yellow collar, and yellowish tail tips. It is uncertain whether juvenile

hybrids between cyanocephala and himalayana could reliably be

discriminated from young finschii, which would be more similar in

size and proportions.

Successful hybridization of male cyanocephala x female

himalayana was also achieved earlier by Critchley, who placed an

adult male cyanocephala whose mate had just died in a mixed aviary

with a placid adult female himalayana that had lived in a pet shop for

several years (L. Critchley, in litt. 1998). The birds paired up during

the first spring (1989) that they were kept together; from all five eggs

laid, chicks hatched and were reared successfully. The fates of four

of the young which were sold to a pet shop are unknown, but the

remaining male moulted into adult plumage in August 1990 and was

in Critchley’s aviaries as of May 1998. Several photos, including one

of the adult hybrid with its parents, show that Critchley’s hybrid

matches in every plumage detail the AMNH intermedia series and

Sedgemore’s male hybrids, and confirm the specific identity of the

parents. Finally, a male cyanocephala and female himalayana also

hybridized successfully in the aviaries of Mr E. Beale (now de-

ceased, M. Sedgemore, pers. comm. 1997), and the pair reared two

chicks in 1980. These had yellow tail tips as juveniles (Beale 1981),

but this brief description does not enable evaluation of whether they

match Rothschild’s intermedia in other respects.

DISCUSSION

SANE’S CAPTIVES

Recognizing that the bird in Sane’s collection photographed by

Wirth (1990) was different from the AMNH material, Arndt (1996)

stated that ‘specimens caught on the plain of the Indian state of

TAXONOMIC STATUS OF PSITTACULA INTERMEDIA

Uttar Pradesh and described in literature as Psittacula intermedia

do not probably belong to this species’. It has been noted else-

where (Inskipp and Inskipp 1995) that the insufficient descriptions

and the lack of photos in Sane et al. (1987) prevent independent

evaluation both of their putative intermedia and their statement

that ‘each year between 1979 and 1984, one or two live specimens

of this species were available in the Indian bird market most of

which could not however be acquired by us. . . . The following

facts, however, show their true identity: (a) all Sane’s birds share

characters of krameri and cyanocephala, but not himalayana; (b)

Sane (1975, 1977) himself described his immature bird as looking

like a hybrid krameri x cyanocephala, with ‘some reddish colour

resembling somewhat that on the nape of the Rose-ring Parra-

keet’; (c) trappers informed Sane that his birds were caught with,

and were natural hybrids between, krameri and cyanocephala

(Sane 1975, 1977); and (d) Sane et al. (1987) were confused

about soft part coloration, post-mortem changes, and sexual di-

morphism in their birds (see below, also Inskipp and Inskipp

1995). Sane (1977) also stated that ‘the call of my bird is more

like a Ringneck than a Blossom-head’ [here meaning

cyanocephala). This evidence, coupled with our new data on

Sane’s three birds, demonstrates that all the intermedia reported

by Sane (1975, 1977) and Sane et al. (1987) are krameri x

cyanocephala rather than intermedia. Sane (1977) identified his

birds as intermedia from the description in Biswas (1959), and

related that when Biswas saw Sane’s first bird, he indicated that it

appeared to be intermedia but that measurements were needed for

confirmation. Sane later took measurements which he said ‘con-

firmed [sic] with the original’ (Sane ef al. 1987).

The lack of red shoulder patches in two of Sane’s birds virtually

rules out the possibility of eupatria (in which both sexes have

shoulder patches) as a parental species, but it is consistent with

krameri. In addition, the great size difference between eupatria and

cyanocephala makes their hybridization unlikely, and the bill shapes

of the Bombay specimens and live bird do not resemble that of

eupatria. The only other possible combination that could result in

such a phenotype would be krameri x roseata. However, all Sane’s

birds have a slight bluish tinge to the nape and wing coverts, so

roseata could hardly be a parent, since this colour is lacking on both

krameri and roseata.

The assumption by Sane et al. (1987) of reversed sexual dimor-

phism in intermedia based on the presumed sex of two of their living

captives has already been called into question by Biswas (1990), and

Inskipp and Inskipp (1995) also queried this as well as their state-

ments on post-mortem changes in bill coloration. Now that we know

that these birds are krameri x cyanocephala (explaining why two of

them lack shoulder patches), the only remaining published corrobo-

ration for Sane ef al.’s (1987) hypothesis of reversed sexual

dimorphism is, by their own account, ‘another male in the collec-

tion, which . . . had mated with a female Roseringed parakeet.

However, the eggs laid were infertile. Their ‘subadult female’

intermedia must presumably have been sexed by inference on the

basis of its shoulder patches after Sane er al. (1987) had concluded

that those lacking this feature were males. Its measurements were

then compared with those published for AMNH intermedia, which

Sane et al. (1987) termed ‘female’, but which surely must be males.

However, even though male krameri lack shoulder patches, male

cyanocephala possess them, so some male krameri x cyanocephala

could have patches as well, and thus Sane et al.’s (1987) sexing of

their subadult bird as a female on this basis is not upheld. In addition,

the unaccessioned stuffed specimen (presumably the ‘female’)

showed only a hint of a shoulder patch. Thus, even if Sane’s birds

really were intermedia and even if intermedia was a valid species, it

is untenable to presume that it would exhibit reversed sexual dimor-

phism with respect to other Psittacula species.

Clearly, the specimens (number not stated) whose blood was used

in the electrophoretic analyses reported in Sane et al. (1987) must

have been krameri x cyanocephala, rendering the results inapplica-

ble to the question of the taxonomic status ofintermedia. In addition,

the four loci examined do not form an acceptable sample, the

methods of analysis and interpretation of results are problematic (R.

Fleischer, pers. comm. 1997), and the major differences claimed

between intermedia and other Psittacula species are improbable,

especially given the apparent parentage of the individuals sampled.

BMNH SPECIMEN

One of the original Rothschild Collection intermedia specimens,

BMNH 1980.3.1, was the one on which Husain’s (1959) analysis of

the hybrid origin of intermedia was primarily drawn. It is, moreover,

more like the illustration of male intermedia in Inskipp and Inskipp

(1995) than are any of the AMNH intermedia. However — and

despite Hartert’s (1924) assertion that specimens in the Rothschild

Collection are alike - BMNH 1980.3.1 differs in several respects

from the remaining five adults (Table 1). It is closer in overall

appearance to adult maleroseata than are the others, and is mensurally

similar to both roseata and cyanocephala; except for the pale lower

mandible and lack of shoulder patches it could be a hybrid

cyanocephala x roseata. \t could also be an F2 hybrid, or if bred in

captivity, a trigen. Its tail tips are broader than in either roseata or

finschii, while the turquoise upper tail surface and tail tip coloration,

shape, and length are as in roseata. The slightly brighter red on the

front of its face than in roseata cannot be explained as roseata x

himalayana ot finschii. However, its complete lack of reddish shoul-

der patches is unique among the intermedia series, is not due to

feather loss, moult, or immaturity, and defies ready explanation. On

present evidence we cannot resolve its parentage, but it does not

appear to be an Floffspring of a cyanocephala x himalayana cross.

EVIDENCE FROM THE AMNH SPECIMENS

As far as we can determine, all characters of adultAMNH intermedia

are either (a) shared with the himalayana/finschii species pair or the

roseata/cyanocephala pair, or (b) intermediate between one or both

members of these two species groups. If for the moment we accept

intermedia as a hybrid (to be further substantiated below), then we

must assume (as did Husain 1959) that one member of each of the

above species pairs was the parental species. Psittacula roseata

cannot have been involved, as its facial coloration is not bright or

deep enough to result in an intermedia phenotype, and its P3 is much

too narrow (Table 4). In additionroseata lacks bluish on its hindneck,

wing coverts, and rump, while most AMNH intermedia have the

blue tint in these areas stronger than on the himalayana/finschii pair,

and the upper tail surface of roseata is a paler, greener blue than in

the other species and in AMNH intermedia.

However, hybridization of cyanocephala with either himalayana

or finschii would involve none of the problematic characters of

roseata. Nevertheless, finschii has narrow central rectrices that make

it unlikely to be a parental species, whether mated withcyanocephala

or roseata, since the mean distal width of the central rectrices is

greater in intermedia than for any of those three species (Appendix;

see also Husain 1959). In addition, a finschii x cyanocephala or

roseata cross could hardly result in the bright yellow tail tips of

AMNH intermedia (as already noted by Husain 1959). Finally,

unlike intermedia, finschii has a bright yellow-green band above the

mantle and pale shafts on the upper tail surface. Incidentally, the tail

tips of the individual illustrated as himalayana in Inskipp and

Inskipp (1995) are actually those of finschii.

Mensural and statistical analyses show the intermediacy of AMNH

intermedia between one or both members of the himalayana/finschii

and roseata/cyanocephala species pairs in every character set exam-

ined. Because roseata and cyanocephala are similar in size and

proportions, roseata is not mensurally ruled out as a parental spe-

cies, but it is ruled out on plumage (see above). However, finschii is

smaller than himalayana, particularly in wing characters, and its tail

proportions are different from any other species and AMNH

intermedia, so it is unlikely to have been parent to the latter. We

know of no cases in which a good species, which is intermediate

between two patently different congeners in numerous phenotypic

characters, totally lacks distinctive features of its own. Also, it is

scarcely conceivable that a wild species would duplicate exactly the

character states found in known hybrids between two quite distinct

taxa. Thus both plumage and mensural analyses very strongly

support the hypothesis that AMNH intermedia are of hybrid origin,

and this is further validated by their identity with known hybrids

between himalayana and cyanocephala.

ARGUMENTS PREVIOUSLY USED IN FAVOUR OF SPECIFIC STATUS

Hartert (1924) stated that if intermedia were a hybrid, “so many

specimens would not very likely have come at the same time,* and

one would expect them to vary, but they are all alike’, with foot-

note 2 disclosing that ‘Our six males were selected by Mr.

Dunstall, a dealer in feathers, from a greater number of these

birds, he told us’. These statements have often been repeated as

evidence of specific status (Biswas 1959, Walters 1985, Inskipp

and Inskipp 1995), but both are flawed. First, the holotype did not

originate with the other intermedia. Second, one of the six re-

maining specimens is a typical immature himalayana (see above),

leaving only five intermedia supposedly of similar origins and

identity. Hartert’s assertion that these came from a greater number

(though not “a much greater number’, contra Walters 1985) in the

possession of and selected by Dunstall implies that Hartert him-

self did not see additional intermedia but had taken the London

plumassier’s word for it. There is nothing to indicate that Dunstall

would have recognized the difference between intermedia and

cyanocephala, and (our third objection) there seems no compel-

ling evidence that he actually had more specimens of intermedia:

the ‘greater number of these birds’ may have referred to the rest of

a shipment of other Psittacula parakeets, a possibility supported

by one of the six ‘intermedia’ being a juvenile himalayana. There

was a considerable millinery trade in Psittacula skins around this

time (Hartley 1907), and only a very small percentage would have

ended up in reference collections.

Fourth, preparation styles and materials used in the five Dunstall

intermedia plus the immature himalayana differ strikingly among

the skins. These differences strongly suggest that, although all are

native skins, they were not all prepared at the same time and place.

They may have come to Rothschild’s museum at the same time, but

not necessarily so to Dunstall or his supplier. Additional support for

staggered acquisition of the material lies in the fact that native skins

of cyanocephala and himalayana strongly resemble, in style and

materials, not only those of intermedia but also those of finschii,

which (given their distribution) must have originated farther east.

Conversely, one adult intermedia (AMNH 621541) and the imma-

ture himalayana are so similar in preparation materials as to make it

highly probable that they were prepared together. Many of the

‘Bombay preparation’ parakeet skins very likely came from bird

markets to which captive birds had been brought from afar. Indeed,

the incidence of at least three partial lutino specimens of this same

preparation suggests selective breeding for this trait, which has long

been highly desired by Indian aviculturists (Greene 1884).

After restating Hartert’s (1924) contentions that intermedia is a

P.C. RASMUSSEN AND N.J. COLLAR

valid species, Biswas (1959) indicated he had concluded the same

independently. However, his only further evidence was as follows:

‘Besides, if they were man-made hybrids, they would necessarily

have been cage birds. But the character of their toes does not indicate

this. Psittacula intermedia may, therefore, be regarded as a genuine

wild species’. Besides the obvious fact that bird hybrids are not

necessarily ‘man-made’, Biswas (1959) gave no indication of which

features of the toes were found inconsistent with captive origin.

Our examination showed that the intermedia specimens exhibit to

varying degrees several conditions consistent with their having been

in captivity under suboptimal conditions (Harrison and Harrison

1986). It thus seems likely that most or all of the intermedia

specimens in existence were captives for some period immediately

prior to their death. In India, Psittacula parakeets have long been

extremely popular cagebirds (Finn 1906, Ali 1927, Dharma-

kumarsinjhi 1954, Sinha 1959), and although most are taken from

nests (Hume 1890), others are captive-bred commercially (H. S. A.

Yahya, pers. comm. 1997), and they have long been bred for the

Indian aristocracy (Greene 1884). Mutations in particular are bred in

captivity in India (S. R. Sane, pers. comm. 1997), garnering up to Rs.

20,000 (Ahmed 1997). From the prices of cagebirds considered to be

intermedia (Rs. 2,000 vs. less than Rs. 25 for ordinary cyanocephala:

A. Rahmani in Inskipp and Inskipp 1995), it is self-evident that

captive-breeding of such hybrids would be well worth the trouble.

Walters (1985) drew attention to previously overlooked descrip-

tions of captive-reared cyanocephala x himalayana (Tavistock

1932-1938) which do not match Forshaw’s (1973) description of

intermedia. Based on this, as well as Hartert’s (1924) arguments,

Walters (1985) concluded that intermedia could not be a hybrid

between those species and must therefore be a valid species, and in

this he has been followed by most recent authors. Presumably also

on the basis of this captive-breeding event, Arndt (1996) stated ‘it

has been discovered that hybrids between [himalayana and

cyanocephala| differ considerably from Psittacula intermedia. . . .

However, a reevaluation of the aviculturist Tavistock’s writings

shows some relevant discrepancies.

First, although Tavistock (1932—1938) did repeatedly pair a male

himalayana with a female cyanocephala, rearing at least seven

young over a period of five years, he published only a very brief

description of just two of those young, which were nestmates

(Tavistock 1933), and did not describe their adult plumage. Thus

there is no assurance that the other hybrids resembled these two, nor

is there information enabling comparison between his hybrids and

AMNH intermedia.

Second, while it is true that Tavistock’s (1933) description of the

two young cyanocephala x himalayana as having white tail tips

does not match the specimens of intermedia (the discrepancy noted

by Walters), other statements Tavistock made throw doubt upon his

entire account. His remark that ‘they resemble young Plumheads,

but their central tail feathers are brighter blue with white tips and

their heads have a dusky tinge’ is nonsensical, as young cyanocephala

do have white tips to their tails, and their heads may have a dusky

tinge much like juvenile himalayana. A\so, it is counterintuitive that

hybrids would have brighter blue central rectrices than those of

young cyanocephala, since juvenile cyanocephala have these feath-

ers considerably bluer than do juvenile himalayana, which are

green-tailed. Sedgemore stated that his juvenile cyanocephala x

himalayana had ‘blue green’ upper tail surfaces (Sedgemore 1995),

and this is confirmed by photos of them as juveniles. The central

rectrices of the adult plumage of both parental species are bluer than

in the juvenile plumage, while those of adult cyanocephala are bluer

and less purple than for adult himalayana. Since the two young

hybrids described by Tavistock (1933) had only hatched that year,

TAXONOMIC STATUS OF PSITTACULA INTERMEDIA

presumably in the spring or summer of 1933, they could scarcely

have moulted into diagnostic adult rectrices before Tavistock’s

article went to press. These inconsistencies indicate that little weight

should be given to Tavistock’s rather off-hand description.

Although not in connection with intermedia, Low (1992: 118)

mentioned hybrids bred by the Duke of Bedford (then the Marquess

of Tavistock), stating that the Duke had paired a male finschii with a

female Blossom-headed Parakeet. However, Tavistock (1932) spe-

cifically stated he used a male ‘Hodgson’s Slaty-headed Parakeet’

and female Plumheads (Tavistock 1932—1938).Whilecyanocephala

has often gone by the common name of Blossom-headed Parakeet,

to our knowledge roseata (earlier known as rosa) has not been called

Plum-headed Parakeet, so he probably used cyanocephala.

“‘Hodgson’s Slaty-headed’ can refer only to nominate himalayana,

not finschii. However, whether or not Tavistock had used true

himalayana, some progeny of a cross between yellow-tipped and

white-tipped parents might well show white tail tips, and in any case

the tail tips of finschii are glaucous yellow, not white.

GEOGRAPHIC PROVENANCE

Biswas (1959) indicated that the Rothschild Museum label of the

type specimen of intermedia states ‘India Nat. Skim’ and considered

it uncertain whether ‘native skin’ or ‘Native Sikkim’ (= then-autono-

mous Sikkim) was meant. This was then taken by Ripley (1961) and

Ali and Ripley (1969) as indicating that the type probably originated

in Sikkim. However, the type’s original label (the only one borne by

the specimen, in Hartert’s handwriting) clearly reads ‘Nat. Skin’, by

which was meant the Bombay preparation of these trade skins, and

thus there is no evidence pointing to Sikkim as the region of origin.

On the basis of Rothschild (1895) and Hartert (1924), it has

also been assumed that intermedia is from the Western Himalayas

(Forshaw 1973, Sibley and Monroe 1990), an idea reinforced by

Sane et al.’s (1987) birds that we now know are krameri x

cyanocephala, although the latter were reputedly from the plains

just to the south (Sane 1977, Sane et al. 1987, Knox and Walters

1994). However, there is no basis for this assumption regarding

either lot of Rothschild Collection skins that contained inter-

media specimens. In the description, Rothschild (1895) stated that

the type came to him with two skins of Palaeornis schisticeps

and, because it was shipped from Bombay, it most likely came

from the “Western Provinces’. However, later Rothschild (1907)

stated that speculation was useless, as the same shipment con-

tained birds from various parts of the Himalayas; still later, Hartert

(1924) stated that the birds evidently came from some part of the

Himalayas. Subsequently it has been assumed without comment

(Biswas 1959, Ali and Ripley 1969, Walters 1985) that by

schisticeps Rothschild meant Psittacula himalayana of the west-

ern and central Himalayas. However, Rothschild did not

differentiate between himalayana and the eastern form, finschii,

both of which were then known as schisticeps, in his description

of intermedia. Since at least two native skins of finschii of the

‘Bombay preparation’ (AMNH 621550, 621557) are present in

the Rothschild Collection, but were not identified as such until

later in a different hand (the former specimen is listed in the

AMNH register as himalayana ssp., the latter erroneously as

Ph. himalayana), it is by no means certain which form was meant

by Rothschild, and the lack of a register for his collection prior

to its accession at AMNH makes it impossible to determine this

now.

CAPTIVE INTERMEDIA OF UNKNOWN PROVENANCE

The male ‘intermedia’ located in Austria are probably both hybrids

between cyanocephala and finschii, as indicated by the uniformly

narrow central rectrices with very long pale yellow tips and pale

shafts midway up the feathers, the yellow-olive band between the

bluish nape collar and olive-green mantle, and the bright yellowish-

green underparts. None of these features is consistent with

himalayana as a parental species. Also, R. Low (in litt. 1997)

thought the Turnersee bird was the same size as the female

cyanocephala with which it was kept, which further supports finschii

rather than himalayana as a parental species, as does this individu-

al’s small bill. Both of the Austrian ‘intermedia’ have the front of the

face bright rose-red, a feature incompatible with roseata being one

of the parental species.

KNOWN CYANOCEPHALA X HIMALAYANA

Sedgmore’s captive hybrids of known parentage are virtually iden-

tical in both plumage and measurements to AMNH intermedia. The

slight mensural differences shown in Figure 3 are almost certainly

due to measurement error, as the live birds had to be measured with

great care to avoid injuring them, and thus they are probably slightly

too large. Also, slight shrinkage of museum specimens is well-

known. The identity of these hybrids with the type and only known

series of intermedia cannot be ascribed to coincidence.

CONCLUSIONS

There is no evidence that intermedia is a valid species, and there is

abundant circumstantial and unambiguous direct evidence that the

AMNH series is comprised of hybrid himalayana x cyanocephala

specimens. The discovery in the 1990s of more birds matching the

phenotype of AMNH intermedia does not negate the above, particu-

larly as they may well originate in captivity. In addition, Rothschild’s

original series not only contained a juvenile himalayana, but also

another hybrid of uncertain parentage; Sane’s ‘intermedia’ are from

a third hybrid combination (krameri x cyanocephala), and the two

cage birds in Austria are probably from a fourth (finschii x

cyanocephala). Thus the literature refers entirely to birds putatively

of four different hybrid combinations, and the supposed species

Psittacula intermedia has no taxonomic standing.

ACKNOWLEDGEMENTS. Special thanks go to Mr M. Sedgemore of

Codsall, near Wolverhampton, U.K., for sending photos of his captive

hybrids, allowing us personally to examine the birds, and sharing his notes

with us. In addition we owe thanks to T. Arndt, R. Wirth, F. Pfeffer, R. Low;

to R. P. Pr¥ys-Jones, M. P. Walters, and M. Adams, The Natural History

Museum, Tring, U.K. (BMNH); C. Blake, M. LeCroy, P. Sweet, and M. N.

Feinberg, American Museum of Natural History (AMNH); L. Bevier, Acad-

emy of Natural Sciences of Philadelphia (ANSP); D. E. Willard, Field

Museum of Natural History (FMNH); R. A. Paynter, Jr., Museum of Com-

parative Zoology, Harvard University (MCZ); R. B. Payne and J. Hinshaw,

University of Michigan Museum of Zoology (UMMZ); G. R. Graves and S.

L. Olson, National Museum of Natural History (USNM); A. Rahmani, S.

Unnithan, S. R. Sane, and A Aktar, Bombay Natural History Society

(BNHS); R. Bhargava and H. S. A. Yahya, Aligarh Muslim University; T. P.

and C. Inskipp, K. Kazmierczak, L. Critchley, Cumbria, U.K.; and P. J. K.

McGowan. Funding for PCR’s travel to Bombay on two occasions was

provided by the Research Opportunities Fund, National Museum of Natural

History, and by British Airways, through M. Sitnik and T. Shille, Office of

Biodiversity Programs/NOAHS, Smithsonian Institution. The manuscript

was substantially improved by comments from G. R. Graves, R. P. Prys-

Jones and M.D. Gottfried.

REFERENCES

Ahmed, A. 1997. Live bird trade in northern India. A TRAFFIC India Trade Study.

New Delhi: World Wide Fund for Nature — India.

Ali, S. 1927. The Moghul emperors of India as naturalists and sportsmen. Part 2.

Journal of the Bombay Natural History Society 32: 34-47.

Ali, S. & Ripley, S.D. 1969. Handbook of the birds of India and Pakistan. Vol. 3. 1st ed.

Bombay: Oxford University.

Anon. 1997. Three Indian birds go extinct, says [sic] experts. Free Press J., 31 May

1997.

Anon. 1998a. Birdwatcher finds world’s rarest parrot in UP. The Asian Age, 25 Feb.

1998.

Anon. 1998b. Lost and found. New Scientist No. 2153 (26 September 1998): 21.

Arndt, T. 1996. Lexicon of parrots. Bretten, Germany: Arndt Verlag.

Beale, E. 1981. Rothschilds [sic] Parrakeet (hybrid) Psittacula intermedia. Parrot

Society 15: 33.

Bhargava, R. 1998. The rediscovery of the Intermediate Parakeet or the Rothschild’s

Parakeet Psittacula intermedia from northern India. Abstract 106. Ostrich 69 (3 &

4): 396-397.

Biswas, B. 1959. On the parakeet Psittacula intermedia (Rothschild) [| Aves: Psittacidae].

Journal of the Bombay Natural History Society 56: 558-562.

Biswas, B. 1990. Taxonomic status of Psittacula intermedia (Rothschild). Journal of

the Bombay Natural History Society 86: 448.

Collar, N.J. 1997. Family Psittacidae (parrots). Jn J. del Hoyo, A. Elliott, and J.

Sargatal, eds., Handbook of the birds of the world. Vol. 4. Sandgrouse to Cuckoos, pp.

280-477. Barcelona: Lynx Edicions.

Collar, N.J., Crosby, M.J. & Stattersfield, A.J. 1994. Birds to watch 2. The world list

of threatened birds. BirdLife Conservation Series No. 4. Cambridge, U.K.: BirdLife

International.

Dharmakumarsinjhi, R.S. 1955. Birds of Saurashtra, India. Dil Bahar, Bhavnagar,

Saurashtra, India: Privately published.

Finn, F. 1906. Garden and aviary birds of India. Calcutta: Thacker, Spink and Co.

Forshaw, J.M. 1973. Parrots of the world. \st ed. New York: Doubleday and Co.,

Garden City.

Fuchs, W. 1997. Der Rothschilds Edelsittich. Aktuelles aus der Vogelwelt 7/97: 224—

225.

Graves, G.R. 1990. Systematics of the ‘“Green-throated Sunangels’ (Aves: Trochilidae):

valid taxa or hybrids? Proceedings of the Biological Society of Washington 103(1):

6-25.

Greene, W.T. 1884. Parrots in captivity. London: George Bell and Sons.

Harrison, G.J. & Harrison, L.R. 1986. Clinical avian medicine and surgery, includ-

ing aviculture. Philadelphia: W.B. Saunders Co.

Hartley, E.A.H. 1907. Ruthless importation of foreign birds. Avicultural Magazine

(2)5: 221-222.

Hartert, E. 1924. Types of birds in the Tring Museum. B: Types in the General

Collection, IV. Novitates Zoologicae 31: 112-134.

Herrmann, K. 1994. Der Pflaumenkopfsittich Psittacula cyanocephala. Gefiederte

Welt 118: 170-171.

Howard, R. & Moore, A. 1991. A complete checklist of the birds of the world. 2nd ed.

London: Academic Press.

Hume, A.O. 1890. The nests and eggs of Indian birds. 2nd ed. London: R. H. Porter.

Husain, K.Z. 1959. Is Psittacula intermedia (Rothschild) a valid species? Bulletin of

the British Ornithologists’ Club 79: 89-92.

Inskipp, C. &. Inskipp, T.P. 1995. Little-known Oriental bird: Intermediate Parakeet

Psittacula intermedia. Oriental Bird Club Bulletin No. 22: 28-31.

Inskipp, T.P., Lindsey, N. & Duckworth, W. 1996. An annotated checklist of the birds

P.C. RASMUSSEN AND N.J. COLLAR

of the Oriental region. Sandy, U.K.: Oriental Bird Club.

Juniper, T. & Parr, M. 1998. Parrots: a guide to parrots of the world. New Haven: Yale

University.

Knox, A.G. & Walters, M.P. 1994. Extinct and endangered birds in the collection of

The Natural History Museum. Tring, U.K: British Ornithologists’ Club Occasional

Publications No. 1.

Low, R. 1992. Parrots: their care and breeding. 3rd rev. ed. London: Blandford.

Monroe Jr., B.L. & Sibley, C.G. 1993. A world checklist of birds. New Haven and

London: Yale University.

Mookerjee, A. 1997. Second life for parakeet. Indian Express, 30.11.97.

Peters, J.L. 1937. Check-list of birds of the world, vol. 3. Cambridge, Massachusetts:

Harvard University.

Rasmussen, P.C. & Collar, N.J. 1998. Reanalysis of specimens and captive birds

proves hybrid origin of Psittacula intermedia. Abstract 145. Ostrich 69 (3 & 4): 412.

Rasmussen, P.C. & Collar, N.J. in press. Little-known Oriental non-bird. Oriental

Bird Club Bulletin.

Ripley, S.D. 1953. Consideration of the origin of the Indian avifauna. National Institute

of Sciences of India 7: 269-275.

Ripley, S.D. 1961. A synopsis of the birds of India and Pakistan together with those of

Nepal, Sikkim, Bhutan and Ceylon. \st ed. Bombay: Bombay Natural History

Society.

Rothschild, W. 1895. On a new parrot. Novitates Zoologicae 2: 492.

Rothschild, W. 1907. [Exhibition of specimens of Palaeornis intermedia]. Bulletin of

the British Ornithologists’ Club 19: 49-50.

Salvadori, T. 1907. Notes on the parrots. VII. /bis (9)1: 122-151.

Sane, S.R. 1975. Rothschild’s Parrakeet. Avicultural Magazine 81: 238-239.

Sane, S.R. 1977. The Rothschilds [sic] Parrakeet (Psittacula intermedia): an appeal.

Parrot Society 11: 136-137.

Sane, S.R., Kannan, P., Rajendran, C.G., Ingle, S.T. & Bhagwat, A.M. 1987. On the

taxonomic status of Psittacula intermedia (Rothschild). Journal of the Bombay

Natural History Society 83 (Centenary Suppl.): 127-154.

Sedgemore, M. 1995. Slatey-headed Parakeet (Psittacula himalayana himalayana)

(Lesson). Parrot Society 29: 187-188.

Sibley, C.G. & Monroe, Jr, B.L. 1990. Distribution and taxonomy of birds of the

world. New Haven: Yale University.

Sinha, R.P.N. 1959. Our birds. Delhi: Ministry of Information and Broadcasting,

Publications Division.

Tavistock, the Marquess of 1932. 1932; the things that didn’t come off and the new

arrivals. Avicultural Magazine (4)10: 248-256.

Tavistock, the Marquess of 1933. Breeding notes for 1933. Avicultural Magazine

(4)11: 379-387.

Tavistock, the Marquess of 1934. 1934: the things that did come off. Avicultural

Magazine (4)12: 255-261.

Tavistock, the Marquess of 1935. Notes for 1935. Avicultural Magazine (4)13: 227—

229.

Tavistock, the Marquis of 1936. Breeding notes for 1936. Avicultural Magazine (5)1:

258-262.

Tavistock, Lord 1937. Breeding results for 1937. Avicultural Magazine (5)2: 334-350.

Tavistock, the Marquess of 1938. Facts and figures. Avicultural Magazine (5)3: 83—

88.

Them, P. 1998. Rothschilds Edelsittich erstmals im Freiland entdeckt? Papageien 1/

98: 6.

Walters, M.P. 1985. On the status of Psittacula intermedia (Rothschild). Journal of the

Bombay Natural History Society 82: 197-199.

Wirth, R. 1990. Systematik: Der Rothschild-Edelsittich Psittacula intermedia.

Papageien 4/90: 126-127.

Wolters, H.E. 1975. Die Vogelarten der Erde. Hamburg: Paul Parey.

iss}

Ss)

)

fe)

8 611-78

Cc 1*9'6

8 °89-6'F

L'0*6'S

8 081-96

Oierecel

LY 8I-S0I

CEFE VL

L €9C-0°CC

CIF VC

L ‘6 18-L'69

QEtCOL

6 O'EL-E'EL

6 CFL'89

6 ‘1S9-6'€S

9°09

6 F'9S-0'LP

6 CFE CS

6 9 SP-S'LE

SCHL

6 6 VE-S 87

CCF le

6 812-991

S1+S 61

6 L'e-0'1

6 0FEC

6 00-00

0'0+0°0

(sui

9'0FIT

OL “€8I-S9l

LISELI

Ol 697-6 €7

O'1F9'ST

OL €TI-OTI

VOFL TI

Ol VSI-VtI

vos v1

6 ‘O'0I-S'9

Gast

Ol VrI-V7Zl

LOFV El

Ol VvC-6 77

V'OF8°EC

LOUDLY

6 CS TI-S'6/9 8 11-01

9'0FS'01/S 06 01

8 OL-L t/t °6:9-9'S

O'1+€°9/6'0FS'°9

OL 86-7 L/ZI “L'0I-8'L

L'0¥1'8/8'0¥9'°8

SI VSI-SEI/II C9I-S TL

90FS TI/0' IFC FI

€€ OIZ-SHI/LI LCC-L OL

DIF 8/S [+061

LI 9'6L-1'L9/11 “6'8L-0'0L

VEF8 CLIO EFC PL

81 VIL-6 19/11 “8'7L4'€9

0'€*8°99/8'C+P'L9

81 LY9-S PS/ZI “6 €9-S'SS

9°C+9'6S/8'T+0'09

61 O'09-L'EP/El “P9S-L’ Lr

€ €*6 19/9 CFT CS

61 9'0S-H'8E/EI “8 LP-S'6E

8 (C¥6 Eb/P C+0 PP

61 LIt-lOe/el TVLE-L le

CCFO VE/D IFT VE

61 “S'87-0'0C/EI “€ 9C-C'I7

OCF VC/9' IFL'E7

VE SVS tC Sa)

CIF E/0 TFL €

SI 0'0-00/EI “17-00

0°0*0°0/9'0*7'0

Cl Tr-O'0/EI ‘9° L-0'0

OIFL T/T CFS

61 ‘SSI-IPl/0e “6SI-LrI

CCTV 8PL/E €+9'CS I

81 S2-€°C7/1T ‘O'L7-L 7

L'0¥0'PC/0' 1#C' VT

rl “S TI-V01

1S'0*6 01

02 ‘VP I-CEL/IE 9 FI-O'EI

€OF8'€1/7 0F8'El

07 “L'S-6'F/DE “8°S-h'b

0¢ 6CI-WII/IE VEI-L II

VOFE TI/E OFS TI

Ot VEC-E 6I/CE “PH HC-C 07

6 0FP 17/0 TFL 17

A/M Myosuif

(L “8°01-0'6) 91 “S'ZI-LO1

(9:01 01) 9'°0F¥' TI

(8 ‘0'6-7'9) LI *L'8-9'9

(6:0F7' L) L'0F8'L

(L ‘9°6-¥'L) SZ ‘TOI-1'8

(L'0F¥'8) 9'0¥1'6

(6 ‘€9I-L7I) SC ‘LLI-8'71

(V1F9'P1) T1FE'ST

(L ‘8°6I-1'ST) IE “€'¥2-WLI

(VIFP LD STFLIZ

(6 9°SL-8'L9) ST “E'88-6'EL

(1h IL) S€F0'18

(6 °6'99-L’79) LZ “8° 18-0°89

(9: TFE'P9) O'EFS'EL

(6 ‘S'6S-I'PS) 87 ‘TIL-€°6S

(L'1#'9S) 9°T#L'S9

(6 “6'6F-8 Eh) LZ “T'79-1'TS

(6: 1#8 9p) €°T#O'9S

(6 ‘60-7 SE) LZ ‘1S’ Er

(8° 1#0'8€) €TF89F

(L “p 1€-0'97) 97 “8 EF-E'EE

(O'CFL'87) Vth LE

(8 °8°CZ-6 LI) LZ ‘O'0E-8' EZ

(8'1*6'61) SIFI97

(6 ‘C 8-4’) €% TL-E'E

(ph 1#8'S) OTFE'S

(6 ‘C'I-0'0) St ‘L'I-0'0

(F070) S070

(8 ‘I I-0'0) #7 ‘0°S-0'0

(F'0FT'0) S1FTT

(6 ‘O91-OS1) 8E “L8I-VS1

(CEFT'SST) O'9FT LOI

(¢ ‘0'87-7'ET) OE ‘8°8Z-6'EZ

(6° TF1°S7) O'1#h'STZ

CVG GOD 86

(LOFTTD S'0*601

(Gy SIEVE)! Gy G@SiSsicl

(9 0F7'FL) SOFT FTI

(8 ‘€'9-0'S) 8E ‘T9-8'P

(F'OFL'S) p'OFS'S

(6 ‘S'EI-8' 11) Gr ‘8'EI-E 11

(L:0F8°T1) 9°0F8°T1

(L ‘TV €7-8°61) Ih ‘O'P7-L'61

(TIFT 17) 6 0FL 17

pupavjpuny q

(ainjeurut) I[Npe

[CC

EA ai!

c0*9'T

LSI-8¢1

POSTS

O'ET

hetss 0.

VOF8' II

Ligt=G sl

L0*9'¢e1

OCI-8 TIT

L0*6 TI

LCC-8 IT

90FT IC

suoutsads

ours

€ 8vI-vrl

[cre 91

ST bl-€7I

LOFC EL

¢ TEI-9TI

€ 0*8 Cl

CS 0C-L'61

rm Oa OY

splgsy

$,d10WISpPag

CS TI-6'6

O1+0TI

¢ € 01-89

V1+08

SW RSi

(6'8) 7 1+8°8

y €9I-E v1

(6+1) 6 0F6'F1

C 66I-1'81

(p81) LOFT 6!

CS 6 LLVTL

(1'L9) © CFL GL

¢ “L'69-S'99

(Z'19) T 1#8°L9

¢ “1'09-0°9¢

(L'pS) L1F8°8S

¢ TCS-S' Lh

(8°Sp) LIFEOS

CC Ere Ib

(€°LE) O LF rtp

¢ 9rE-S TE

(6°97) TIFIEE

C ‘lr7-L 91

(E81) O'€FS'IZ

c V'9-0'r

(O'€) O1FH'S

b 00

($0) 0'0*0'0

61-00

(0'0) 0'1+6'0

¢ ‘9SI-CSI

(O'CH1) 9 IFS'EST

¢ 0'97-0'E7

(LT) VIFE GT

v TTI-8'6

(O'0L) 01601

¢ LEI-O'El

(VTL) TOFE EL

Cc 6rer

(8°€) TOF9'D

C SCIFI

(E11) OFS TT

¢ ‘90C-S 61

(0°81) F'0*0'07

(1'€'0861 HNN)

vipauatul J HNWV

Ol T67L

JOFES

Il P9-€¢

vOF8'S

Ol 867L

LOFI'8

CL SHI-6'I1

LOFC El

€€ OCS HI

IFT SI

8€ “LTL-S'09

9°€+0'99

BE “8S9-CHS

CeF9'6S

LE €09-C'8r

CeEFO'ES

8E O1S-C IP

LCFS Sp

LE O'PP-L EE

OCFOLE

LENGE SZOIEG

O'C¥8°8C

9E €S7-0'ST

crv 6l

LE 98-0'€

S179

LeeLiGa0i0)

6 0FL0

1€ L700

6 0F6°0

Or “S StI-67I

SrFe ET

SE “LC V6

6 0FS IT

€7 ‘9OII-S'8

8'0*9'6

8€ “LT7I-6 01

C0F8 TT

6€ SS-L'E

VOFS TY

8€ T1I-S6

yOre OL

ce “TOC-E SI

Wie /E\8

DIDASOA qd

GE WIS O

0' 196 YIPIM Ed

97 ‘SL-EP

8':0F7'9 WIPIM 7d

GZ ‘86-7 L yojou qe

9'0FP'8 YPM Td

pt ‘€SI-6'8 YIpIM

pIFlel wmurnxeur [q

I€ PIZ-S's1 aoursip dy

C1081 -yojou [dq

97 ‘S'OL-6'LS [reyoys

9'TF9'P9 Old

87 ‘7 S9-S PS [epoys

b TFC 6S 6d

6c ‘ELS-L 8b [repoys

UTFS 7S 8d

87 ‘6 6b-1 th [reptoys

6 IFESP Ld

O€ ‘S'Ih-9'rE [reyioys

8 1F9'LE 9d

O€ “STE-6'ST [repoys

C 1¥6 87 Sd

I€ ‘77191 Treproys

CIFL 61 td

6¢ ‘T9-1'7Z [reyztoys

6 OFE'F €d

87 ‘0'I-0'0 [repsoys

7000 Zd

6c ‘I'+-0'0 [reproys

€ 19] Id

br OSI-971 yySug]

O'S¥S' SET BUIA\

SHOINAY GNV ONIN

Or ‘L'b7-S 07 YIpIa

L1¥9'% 11S

€Z ‘L'11-9'8 pia

6 0F8'6 a[qipuey

OF ‘8 ZI-TII yIpIM

€'0F0'T1 PI[IXeI

Eb ‘8'S-9'E YpIM

vOFL'P [eLeN

Ob ‘b11-8'6 qysIoy

vOFL OI BIpEXe

Or ‘661-91 yysue]

8'0F8'LI uauynD

davdH

pipydaz0uvio gq dINSedI[

“UOTIOAS SpOUIAY] Ul USATS s]UIUTOINSPOU

jo suondussaq ‘sususads nipawuiaiul HNWY JO sonsneis Areuruins SUIMOT[OJ Sasayjuared UI UDAIS [°€'O861 HNING 10} sjuowiainseayy] “(uv ‘odue1) UONRIASp plepur}s F USUI Sv UDAIS Ore [TY “awnsy

d pure ‘Ajaansadsau ‘suonejndod usajsva pur wiajsom wo suoumtoads myosuyf q ‘sasayjuesed ut sapruaant way) ‘pupdpjpuiy g ynpe ‘(7 = UV) UONIET[OD S,auRg Woy SudUTIDads ‘(a[eUIay 9UO SuIpNyoUt)

UON}DA][O9 S,a1OWASpas url spugdy sAq] ‘VIpauiajul J HNINY ‘Vivasos q ‘vjpydasouvd9 vjnovyIsg aeu ynpe JO (wu ‘YSIIY = Y ‘IPPIM = M “U)Sug] = |) sjUoWoINsvoul Jo sonsNeis AeuUINg xipueddy

P.C. RASMUSSEN AND N.J. COLLAR

QE B-SL 6 E8-8'S/L ‘*S'8-1'S

€0F6L 8: 0FP'9/0 1+9°9

8 ‘O'11-0'8 SI ‘Pp 6-9'L/ZI “b'6-6'L

0'1*6'6 S OFF’ 8/P 0FS'8

8 0'S=6'E OL ‘Tt-1'€/S ‘0'b-6'7

SOFT € 0F9 €/P OF PE

OL ‘L'81-8'91 91 6 SI-E TST ‘O'LI-CTI

9°0F9'LI 6 0FL'b/6 0Ft'SI

8 ‘P'b-0'7 bl ‘OS-PI/IZ ‘O'EL-E'LI

L'0F8'7 6 6F8'87/L EFT Th

8 ‘1'b-0'0 EI ‘O€I-06/ET “0'S9T-0'S8

(6 4 OCT FLLOL/9'6IFS STI

L‘6O€I-V Il 8 ‘Y'8-T'L/S ‘L'6-9'9

6 0FETI v0F8'L/T 1F78

6 ‘871-66 ROU aEOe BSa kL

O1F8'TT SOFT L/L'OFL'L

6 ‘€TI-L'8 6 ‘6 LV 9b “P8-TL

TIFT TI SOFT L/S 0F9'L

6 ‘T'€1-9'6 SI ‘O'6-€'S/8I ‘8'8-L'b

VIFL01 6 0FS'9/T 1#8°S

e SI ‘19-8 P/6I ‘9°9-€'€

= p OFS S/LOFI'S

9 ‘P8-9°9 El ‘p'S-6E/ST “€'9-O'P

SOFL'L v0F9 p/LOFTS

6 ‘7C-LI OL ‘SEI-€ 8/61 ‘O'SI-8L

8 1761 SIFSOVETFUTI

6 ‘OC-SI OL ‘€OI-1'9/6l ‘€'€I-V'L

6 1F9°L1 O'1Fb'8L/9' 196

6nge-L1 LI ‘T91-9'6/91 ‘181-901

EVFE TT 8 1F9°TI/p TF6 El

6‘Le-I€ —_ LT ‘0'69-L PI /PE ‘O'SL-L'FI

O'PF6 LE P S18 11/8 €1Fh'0S

8 ‘OL-€T SI ‘OpI-€6/p ‘IZI-O1I

€ OCF16S 8 PIF8 91 V/L PFT PII

8 “SS7-S9I 91 ‘O9LI-ELT/61 “P81-86

OOF POI O'VIFL CEL/8°7#OLET

8 ‘pST-C7Z OL ‘ILZ-1IZ/TI ‘TOE-TET

vOlFr LEZ L'61F9 0S7/6'8IFL P97

Serial Ol ‘O'II-Z 01/8 ‘8'ZI-V01

SOFTEI 9'0FT 01/8 OFE II

8S TZI-€01 6 ‘8 7I-V' 01/9 ‘TZI-T TI

80FS II 8 OFF IT/€0F9 TI

MauUD1y A/M Hyosuyf

(8 ‘7 L-€'9) E% “V'L-€'9

(€0FL'9) €0FO'L

(8 ‘S'6-1'L) 67 ‘S'‘0I-7'8

(L'0¥9'8) S 076

(v ‘T'¥-0'€) IZ “€'F-1'€

(v'0FS'€) €0FV'E

(9 ‘L'SI-8'r1) TE “P'8I-0'E1

(V'OFP'SL) VIFE'ST

6£ O'SS-0'ST

GOFO'8E

8€ O'SOI-0'8S

LOI+¥ 18

(8 ‘0'0I-0'L) OI ‘8'0I-8'8

(8'0*S'8) L 0FP'6

(L‘9'8-S°9) FI ‘9°6-8°L

(L'0¥9'L) 9'0¥9'8

(p T6-L'L) OTS O1-O'L

(9'0¥S°8) 6'0¥6'8

(C ‘L'6-1°8) Lt “6 6-T'L

(8'0FL'8) L0FV'8

be O'OI-O'L

L'OFS'8

O€ ‘811-08

60*L6

(8 ‘v'€I-7'6) TE “9'LI-0'01

(VIFEL1) O'CFS'EL

(9 ‘1'OI-1'9) €€ ‘T'9I-O'L

(S 1¥9°8) 6 1FEOI

(9 ‘T'ZI-S‘OD) LE “T'It-0°6

(QOFT II) SSFEST

(9 ‘707-6 V1) 8E O'E9-O'TE

(6: 1*6'91) 6 9F1 br

OT ‘TII-IL

C 6F9' 8

Or O8I-SII

CCIFI 9FI

Lt 19C-LOT

O'SIFL VEC

(S *‘L7I-O'1D) 91 LvI-L' 11

(9'0*6'I1) 8°0*8 CI

(p ‘TTI-L'8) 91 “TrI-C OL

(C1+¥ 01) O'1FS 71

DUDADIDULIY

(QInyeUUT) J[Npe

(OWE

Lc¥0'S1

6CI-6 Cl

0'0*6 Cl

OLI-S'SI

Pease)

[tet]

9LI

OCI

VIT

suawtoads

ours

b O0StOLT

Tel+l Ce

b 0'79-00€

S CIF 8P

b 0L-09

S0FCT9

€ 01-08

LT 1+t6

7 O'0S-0'cE

OCIFS IV

b 99I-StI

8°8+0'SST

[ry ‘911-96]

v €L7-9VT

9 T1*0°C9C

spuighy

$,d10WISpoIg

6 SESH

(0'9) L:0*0'8

€ 88-€8

(SL) 70F9'8

€ 1-67

L0F0'€

€ 6vI-L v1

(9°S1) 1:08 'r1

SW GaaGC

(S11) PTFE STZ

t ‘Tpb-16€

(6'€£) SEFO' Ih

¢‘Oll-8

O IFES

CP 6-9'L

L0F78

C‘L6-VL

1 1#78

¢ ‘0'6-0'L

(9°9) L'0F8'L

€ ¢lL79

(S'S) S0FO'L

eo FREL

(19) TI1FE8

¢ TVEI-C ll

(701) 8'0F6 TI

C‘SOI-LL

(€°6) O1*6'8

¢ ‘7 EI-T 01

(68) TIFL II

¢ ‘O'rh-8'81

(1°07) O';01F7'SE

¢ ‘8eI-901

(ZOI) V€1F0O°L7Z1

(+781) +007 “+OLT +ZS1

C*LTI-€ ll

C OFT CI

¢ 6 7I-001

(5 av, |1)!

(1°€'0861 HNN)

Dipausajul q@ HNNV

Cc ‘OL-19

v0+9°9

€l C6-OL

9'0*6'L

Il 1t-L7

VOFCE

cE 6 SI-O'TI

6 0F9'E1

9€ ‘O'EI-0'S

OCFLL

le “LS7-€L

OVFS OT

Ol ‘76-99

SO0FEL

OL €8-8'S

L:0*0'L

OL 18-7

8079

87S L-E'S

COFCO

TE 89-1 P

9'0FTS

0¢ L'9-6'€

DOFES

LE ‘Vel-c9

eC IFe Ol

9€ ‘1'SI-0'9

S1FS8

9€ VEI-L9

CIFL6

CES Sh-8 SI

SSFE VC

LI VILL €v

BLFB CS

LE “I€1-68

18+9'€OT

€€ 6CC-I¥1

VSI+I 991

OL ‘O'11-06

9°0*C 01

Ol v0I-S'8

C OFE 6

DIDISOL I

€1 OL-G'S

COFI9

€C O6-1TL

C 0F6'L

£7 SP VT

VOFEE

€€ 19l-V7l

60FC V1

Iv ‘CET-E'S

SvFl El

9€ “S$ 8E-0 TI

LOFE6C

vo S6-EV

C1+9'L

St £6-9'S

60F0'L

Sc LOI-wS

O'1+8°9

€€ L8-1S

L'0+0'9

0€ “6 L-6'F

L'0*0°9

Lt TOI-6' 4

CIFLL

ce PrI-69

L1+9'6

CE CIE

Tcre Ll

0€ LOI-C9

c 1+9°8

tr L'6r-0'01

V6FS VE

€t 60I-LY

6 TI+9L'88

LE 61-16

eLIFl SII

Ce “TET99I

8 SIFT C07

91 071-66

9°00 TT

v7 811-88

LO0*V OL

pjpydas0uvio q

y3ug]

Me[SPuTH

y3ue]

MPTSPTIAL

UIPIM sns.iey

WINUITUTy

wsue]

snsiey,

aI TGNIH

yysue] dn 7y

wsugy dn 1y

IpIM Ca

WIpIM ra

WPIM ca

WPIM Ta

(pI

jewrxoid ,.y

Teasip Ta

aoudIAyFIp

9a-Sa

oUdIOTFIP

cara

QOUudIOTIIP

pa-ca

2oudIOTFIpP

eu-ca

JoudIATJIpP

ca-la

ysug]

ysugy

SSOMLOTA

(PIM Cd

WIPIM td

amnseayy

panuyuo) xipusddy

Bull. nat. Hist. Mus. Lond. (Zool.) 65(1): 51-72

Issued 24 June 1999

A review of the genus Bargmannia Totton, 1954

(Siphonophorae, Physonecta, Pyrostephidae)

PR.PUGH xX (SIEC)\.

SouthaMpton Oceanography Centre, Empress Dock, Southampton, Hants, SO14 3ZH, UK

CONTENTS

ITEICYCLC TOM reeks se cesorsoses=cexaceccnrstdectscsscaueacrscesageesscsancvercesuscsssanasoe>

Family Pyrostephidae Moser, 1925 ..............:.::.scesccssecserseeseee

Genus Bargmannia Totton, 1954 .....eeceeesessetseeeeseeneeeeeeee

Bargmannia elongata totton, 1954 .......:cccccsseesesseseereees

BOREMANMIG GMOENG SP. TWOV: :...00:s-tn-+vasensedoxcegavedostvss-n>s

Bargmannia lata Mapstone, 1998 ...........ccsesescenseneeees

(BAF 2IIANIIGIEISAS | SP MONS «ose cae ctaca-Sedececaedscnnrsda-eadecersnode

TAS GIS SIO faced ee execs ceed o waaawns damnsen te dase oavadenns itches snensuevesastet eonee

PN GRAIO WE SELCIUS ceceasd. catenins saat ernesntaveecsnevevns qaracsatnasebonsacadepc

PR GECHOMGE Sa sase cre arn zest ssa cc suendnaae ta evavanserih cise sone sesatquses veaetssseecvar>

Sor oS nt oes ate hee estan ea a aac euc Recalls ua caca ths dus lyeeten 00

NoPE rowan E este nas eee vance dated ec cet str stresccesstssaostee 00

Be ae ee cea reece nn acta octa cn ds cadaver nisi atone -veaicen dns 00

Waeannanild Wane es ex we uaa uae Tuas savas eens Ws pccnsdaceasceccatncstevdsceset suesaanasacver 00

Lecce Cece REREAD OA ae CACC CR ETE DP REE ee ae 00

So esacanapie vances Coren iatan ees sartusccvansustvagsvesiipecundunesiuncvesrsearaestecreereonrrcae 00

des geetavtnastevepoetacone rover strcrorcen-trevttensadecnsiccrsaereseg nanny sasteeenntnasttterets 00

SYNOPSIS. Two new species of the physonect siphonophore genus Bargmannia are described, and B. elongata Totton (1954)

and B. lata (Mapstone, 1998) are redescribed. The status of the genus and its retention within the family Pyrostephidae are

discussed.

INTRODUCTION

Totton (1954) established the genus Bargmannia, named after his

colleague Dr Helene Bargmann, to include the single species, B.

elongata; nectophores of which he had found in thirteen Discovery

samples, plus one from the Michael Sars Expedition (Leloup, 1955).

Because the structure of the nectophores differed so markedly from

those of all other known physonect siphonophores, Totton did not

give a detailed description of them; remarking only that the lateral

radial canals on the nectosac had straight courses. Totton (1965)

later noted that, although B. elongata was one of the most easily

recognised siphonophore species, nothing more had been published

on it since its original description. In fact, by the time of publication

of Totton’s monograph, only Alvarifio (1963, 1964) had mentioned

it; and then only in lists of siphonophore species collected in the

western Pacific. Totton included a brief description of a further

specimen collected at Discovery St 4246 (37°50'N, 13°22'W), re-

marking on the orange coloration of the stem.

Since that time several authors have reportedly identified this

species from various collections. However, examination of both

Totton’s material and that from more recent Discovery collections

(Mackie, Pugh & Purcell, 1987) appeared to indicate that Totton’s

(1954, 1965) illustrations of Bargmannia elongata could be

referred to two species, and that his material also included a third

species. However, it was not until submersibles collected speci-

mens of this genus that this contention could be proved beyond

doubt. Study of this submersible material, together with that from

the Discovery collections, shows that there are at least four spe-

cies that may be referred to the genus Bargmannia. The second

species that Totton illustrated under the name B. elongata has

recently been described under the name B. lata. More detailed

descriptions of both these species, together with descriptions of

two previously undescribed species, are given herein.

Totton (1954) did not refer the genus Bargmannia to any of the

physonect families, although his description appears at the end of a

section dealing with various species of the family Agalmatidae.

Later, Totton (1965) placed the genus in the family Pyrostephidae,

which previously had been monotypic for the species Pyrostephos

vanhoeffeni Moser, 1925. However, his diagnosis of that family

applied only to the genus Pyrostephos, and included such features as

marked bends in the dorsal and lateral radial canals on the nectosac

of the nectophores. This character alone would exclude the genus

Bargmannia. Since then, Stepanjants (1967) placed the genus in the

catch-all family Agalmatidae, whereas Daniel (1974) retained it

within the family Pyrostephidae. Now that intact specimens have

been collected by submersibles it is possible to review the systematic

position of the genus Bargmannia. It is concluded that, for the

presentat least, it should be retained within the family Pyrostephidae,

the diagnosis of which is adjusted accordingly.

Family PYROSTEPHIDAE Moser, 1925

DIAGNOSIS. Long-stemmed physonect siphonophores. Necto-

phores with large triangular thrust block; with lateral wedge-shaped

processes reduced or absent. With apico-, infra- and vertical (meso-)

lateral ridges; apico-laterals divide above ostial level. Adaxial wall

of nectosac lacking musculature; deeply hollowed. Long pallial

canal; short pedicular canal, giving rise, on nectosac, to only dorsal

and ventral radial canals; lateral radial canals arise separately from

dorsal. Dorsal and lateral radial canals either looped or straight.

Tentillum with straight (or twisted, but not tightly coiled) cnidoband;

lacking an involucrum; with terminal filament. Dactylozooids either

absent or modified to form peculiar palpacle-less oleocysts. Indi-

vidual specimens of single sex (dioecious), with gonophores budded

one from another to form a small gonodendron; female gonophores

contain two or more eggs.

REMARKS. In Pyrostephos vanhoeffeni, the triangular thrust block

is best seen on smaller nectophores. On larger, preserved ones it is

bent up dorsally (see also Discussion section).

Genus BARGMANNIA Totton, 1954

DIAGNOSIS. Pyrostephids with distinctive elongate nectophores.

Mature nectophores with large, triangular thrust block; without

apical wedge-shaped processes; with extensive ventro-lateral wings.

Basic ridge pattern may be augmented by additional ridges branch-

ing from apico-laterals. Nectosac basically cylindrical; dorsal and

ventral radial canals straight; lateral radial canals arise separately,

but in close proximity, from the dorsal canal. Pheumatophore with-

out apical pore.

Siphosome diffuse; devoid of fully formed dactylozooids. Bracts

specifically variable in shape. Each cormidium; with simple tenta-

cle-like structure attached to stem midway between successive

gastrozooids; with single gonodendron; with four bud-like struc-

tures (?vestigial dactylozooids) with sexually dimorphic arrangement.

Second tentacle and fifth bud occasionally present proximal to a

gastrozooid.

REMARKS. The meso-lateral ridges on the nectophores, as referred

to in the above diagnosis, are homologous with the vertical lateral

P.R. PUGH

ridges, as defined by Pugh and Youngbluth (1988), found on the

nectophores of certain agalmatid species. In these latter species

these ridges run vertically, or slightly obliquely, between the apico-

and infra-lateral ridges, although they may not reach the latter.

However, in Bargmannia spp. their arrangement is strikingly differ-

ent in that they have a very oblique course; and it is the infra-lateral

ridges that may or may not join them basally. For these reasons the

term meso-lateral ridges will be used herein.

In contrast, the outer of the two branches of the apico-lateral

ridges should not be compared with the lateral ridges of agalmatid

species, as defined by Pugh and Youngbluth (1988). They more

closely resemble the near-ostial branching of the apico-laterals in

agalmatid species such as Lychnagalma utricularia (Claus, 1879)

(see Pugh & Harbison, 1986) and Halistemma transliratum Pugh &

Youngbluth, 1988, which also possess normal lateral ridges.

The long, median canal that runs up the thrust block (see Figure

2), just below its ventral surface, has been variously referred to as a

pallial (e.g. Daniel, 1974) or a pedicular canal (e.g. Daniel, 1985). In

accord with the definitions given by Totton (1965) here the canal will

be referred to as the pallial canal; and the short canal, passing

through the mesogloea from the stem to the nectosac, the pedicular

canal.

Recently, it has been brought to my attention (Dr S. Haddock,

personal communication) that the generic name Bargmannia was

used by Herre (1955) in a description of a genus of an extinct sala-

mander. Bargmannia Totton, 1954 clearly has priority of publication.

Fig. 1 Bargmannia elongata. A. Photograph (reproduced by kind permission of Larry Madin, WHOI) of live specimen collected during Alvin Dive 961.

B. Photograph (reproduced by kind permission of Steve Haddock, UCSB) of live specimen collected during JSL I Dive 2673. Nectosomal length c. 9 cm.

BARGMANNIA REVISION

Bargmannia elongata Totton, 1954

(Figures 1—S)

Bargmannia elongata Totton 1954 (Text-Figure 28 A-—D only);

Totton 1965 (Figure 45, A—D only); Kirkpatrick & Pugh, 1984:

Figure 11.

HOLOTYPE. BMNH 1952.11.19.7, designated by Totton (1954):

one nectophore from Discovery II St. 699; 14° 27.3'N, 30° 02.3'W;

14-v-1931; 0-370m. The specimen was figured by Totton (1954,

text-Figure 28 C, D; 1965, Figure 45 C, D).

PARATYPES. As designated by Totton (1954): eighteen nectophores

from the same sample as the holotype. BMNH 1952.11.19.8—25.

MATERIAL EXAMINED. The holotype and paratype material have

been re-examined in order to establish to which of the presently

recognised Bargmannia spp. the name elongata should be applied.

Totton’s (1954) other material also has been re-examined and,

although all the material is in poor condition, it appears that only the

nectophores from two other Discovery stations belong to this spe-

cies. These are St. 681 (21°13'S, 29°55.25'W; 1—v—1931) where a

TYFYV net was fished over a depth range 1500—1000m; and St. 107

(43°03'S, 17°03'E; 4—xi-1926) where the net used was a N450 and

the depth range was 850—950m. The nectophore from the former of

these stations was figured by Totton (1954, text-Figure 28 A, B;

1965, Figure 45 A, B). The other nectophore, from Discovery St.

1769, also illustrated in the same figures (E, F) does not belong to B.

elongata, but to B. lata.

Several nectophores of this species have been found in more

recent Discovery collections, as is discussed below. However, the

major part of the redescription will be based on two specimens

collected by DSRV Alvin off San Diego, California, U.S.A. in 1979,

during Dives 961 (32°14'N 117°22'W; 5-ix-1979; water depth

833m) and 966 (33°04'N 118°16'W; 8-ix-1979; water depth 747m).

The Alvin Dive 961 specimen, preserved in Steedman’s solution, has

been deposited in The Natural History Museum London (BMNH

1998.2163). The exact depths of collection for both Alvin specimens

were not recorded.

DIAGNOsIS. Nectophores with central thrust block broadly rounded

or obliquely truncate apically. Pair of short ridges, directed toward

mid-line, branch from apico-laterals where latter bend out sharply at

a right angle. Outer branches of apico-laterals end, basally, on, or

just apical to, enlarged processes lateral to ostium. In preserved

specimens ostium opens dorso-basally and nectosac, with appar-

ently dense musculature, has distinct dorso-ventral undulations. The

ratio of the overall length of the nectophore to the length of the

nectosac averaged 1.31. Delicate, foliaceous bracts; typically with

patches of ectodermal cells on distal half of dorsal surface.

DESCRIPTION. A photograph of the living specimen collected

during Alvin dive 961 is shown in Figure 1A. By the time it was

taken, in a tank on board the mother ship, several nectophores had

become detached and the siphosomal stem had contracted. A second

living specimen, collected during Johnson-Sea-Link (JSL) I Dive

2673 (27°02.7'N, 85°01.5'W, depth 780m), is shown in Figure 1B.

PNEUMATOPHORE. The pneumatophore measured c. 2.2 mm in

length and 1 mm in width, but was distorted and ruptured. No

pigmentation was apparent. In theAl/vin dive 961 specimen, the main

gas cavity, the pneumatosaccus (height 1.8 mm), was separated from

the small gas secreting region, the pneumadenia, by a narrow collar.

Below the pneumatophore was a long stalk, up to 7.6 mm in length.

Immediately above the nectosome, this stalk narrowed and was

flattened to form a hinge-like structure, which could facilitate the

513)

use of the pneumatophore as a means of orientating the animal.

NECTOPHORE (Figures 2—3). The nectophores had a biserial, stag-

gered arrangement down the nectosome (Figure 1). Forty two

nectophores were found with the Alvin dive 961 specimen, though

many were small or immature; and 26, mostly mature ones, were

found with the A/vin dive 966 specimen. The mean dimensions, for

the fully developed nectophores of each specimen, were:- length:

21.29 + 0.93 mm and 16.49 + 0.75 mm; width: 9.58 + 0.56 mm and

7.40 + 0.32 mm; and the ratios of total length of the nectophore to the

length of the nectosac were 1.29 + 0.02 and 1.34 + 0.04, respectively.

For net collected nectophores, damage and distortion by preserva-

tion, particularly to their basal halves, made it difficult to assess this

ratio accurately.

The nectophores of the dive 966 specimen were noticeably smaller

than those from dive 961 but, as will be seen in the description of the

following species, the size range of the nectophores can vary greatly

between individual specimens. In general, the thrust block was

roundly, and often slightly asymmetrically, truncate (Figure 2A, tb;

2B), although for a few of the nectophores of the smaller specimen

it was distinctly tapered. The latter was also apparent on several net

collected nectophores where the apex of the thrust block was

drawn out to form a small digitiform process that could be folded

over ventrally.

Fig. 2 Bargmannia elongata. A. Upper, B. lower, and C. lateral views of

mature nectophore. Scale bar = 5 mm. bi, bo: inner and outer branches

of apico-lateral ridge; mp: mouth-plate; n: nectosac; o: ostium; pc:

pallial canal; pedc: pedicular canal; ral, ril, rml: apico-, infra- and meso-

lateral ridges; sb: side branch; tb: thrust block.

The basic Bargmannia ridge pattern is supplemented by a pair of

short ridges (Figures 2A, sb; 3 A) that branch from the apico-laterals

(Figure 2A, ral) at the point where the latter bend sharply, through

90°, away from the mid-line. This sharp bend typically can be seen

in less well preserved specimens and is characteristic for this

species. The side branches are directed, for a short distance, toward

the deep median furrow. In many specimens, particularly net col-

lected material, they were difficult to discern but often can be seen

after staining. Basally, the inner branch of each apico-lateral ridge

curves inwards and then down to reach the ostium (Figure 2 A, bi),

except for immature nectophores (Figure 3A) where it ends slightly

above that level. Each outer branch (Figures 2A, bo; 3C) typically

terminates on or just above one of the small, but prominent, lateral

processes on either side of the ostium.

Basal extensions of the meso-lateral ridges form the baso-lateral

margins of the bilobed mouth-plate (Figure 2B, mp; 2C), each lobe

being thickened ventrally, particularly toward the mid-line. Basally,

the two lobes typically overlap and unite, in the mid-line, at about

half the height of the mouth-plate (Figure 3C). The lower nerve tract

(see Mackie, 1964), which can be traced down the nectophore,

beneath its ventral surface in the mid-line, recurves at this point and

P.R. PUGH

continues obliquely to the baso-ventral margin of the ostium (Figure

3C). In immature nectophores the mouth-plate is not thickened and

has a U-shaped emargination in the mid-line (Figure 3A, B) whichis

deepest in the youngest nectophores.

Above the mouth-plate, the basal extensions of the meso-lateral

ridges curve round toward the mid-line, on the ventral surface of the

nectophore (Figure 2 B), before looping back outwards as the meso-

laterals proper (Figure 2C, rmil). The infra-laterals are weakly

defined in the region where they divide from the meso-laterals, and

in younger nectophores clearly terminate before reaching the latter

(Figure 3 B). The meso-laterals curve up, obliquely, across the

lateral surface to reach the junction with the other main ridges at a

level slightly below the apex of the nectosac (Figure 2C). The

connection with the other ridges is weak, and often the meso-laterals

appear to end slightly below the junction, as was found for younger

nectophores (Figure 3 A).

The infra-laterals (Figure 2B, ril) demarcate the ventral margins

of the thickened walls of the more basal part of the ventro-lateral

wings. In lateral view these wings are slightly emarginate in outline.

Apical to where the infra-laterals curve up to join the other ridges,

the wings remain well developed and are thickened with mesogloea.

Fig. 3. Bargmannia elongata. A. Upper and B. lower views of young nectophore; C. detail of ostial region of mature nectophore. Scale bar = 1 mm.

BARGMANNIA REVISION

Fig. 4 Bracts of Bargmannia elongata. Scale bar = 1 mm.

This thickening diminishes in the region of the thrust block, but there

is still a shallow median gutter that enfolds the nectosomal stem in

the region of attachment of the nectophore (Figure 2 B).

In the preserved nectophores, the nectosac is a dorso-ventrally

undulating tube (Figure 2B, n; 2C), with prominent dorso-lateral

extensions in the mid region, and ventro-lateral ones both apically

and basally. However, this arrangement is not apparent in the

nectophores of the living animal (Figure 1). The nectosac is broadest

at about two-thirds its length, narrowing slightly towards its apex. It

has a distinct apical emargination; U-shaped in the younger

nectophores (Figure 3 A). Typically, the ventral, adaxial region

towards the apex of the nectosac is distinctly undercut and, from a

level just basal to the point of insertion of the pedicular canal, its wall

is devoid of musculature (Figure 2 B). The musculature of the

remainder of the nectosac appears well developed and gives it a

distinctly opaque appearance. The ostium, in the preserved material,

opens onto the dorso-basal (abaxial) surface (Figure 2C, 0) and is

roughly rhomboidal in shape. However, this probably is distortion

due to preservation (see Figure 1). In the Alvin specimens it has a

large velum, with a relatively small central opening, but in net

collected material often the velum is destroyed. The lateral walls of

the ostium extend out to form lateral processes (Figure 3C) that,

typically, are covered by patches of ectodermal cells of varied size.

Further such patches are present on the ventral margin of the velum,

but not on the dorsal margin, except for the youngest nectophores.

Some, if not all, of these cells probably produce bioluminescent

material since this has been found to be the case in another

Bargmannia spp. (Dr S. Haddock, personal communication).

The long pallial canal (Figure 2B, pc) extends up into the median

thrust block, where it ends with a short dorsal inflection into the

P.R. PUGH

Fig.5 Bargmannia elongata. A. Young tentilla with stenoteles (st) at proximal end of cnidoband (magn. 50x); B. Part of siphosome showing three

siphosomal tentacles (¢) and several buds (b) (magn. 16x); C. Male gonophores (magn. 30x).

mesogloea. At its base the lower nerve tract can be seen to leave its

proximity and to continue down beneath the ventral surface of the

nectophore to reach the ostium. The short pedicular canal (Figure

2B, pedc) extends through the mesogloea, from the base of the

pallial canal, to the nectosac. There it gives rise to only the dorsal and

ventral radial canals. The lateral canals arise separately, but in close

proximity to each other, from the dorsal canal, and initially are

directed toward the apico-lateral margins of the nectosac. They then

continue down the lateral margins of the nectosac and, although their

courses show undulations (Figures 1C, 2C), they are merely follow-

ing the dorso-ventral undulations in the nectosac itself; the latter

being a preservation artefact.

The youngest nectophores (Figure 3A, B) typically show the

absence of a median thrust block, and the apico-lateral margins are

demarcated by the apico- and infra-lateral ridges. The basal portions

of the apico-lateral ridges are particularly well marked, and the inner

branches are distinctly broadened, often appearing almost bifurcate

at their basal ends, which lie just above the ostium (Figure 3A).

There are two short tracts of cells extending out from the lateral

processes of the ostium just ventral to the outer branch of the apico-

lateral ridges. These could not be discerned in the mature nectophores.

BRACT (Figure 4). The bracts are extremely delicate, foliaceous

structures, the largest of which measures 9 mm in length. The dorsal

surface is slightly convex, the ventral one slightly concave. For many

the proximal region is bent up dorsally, or one side is folded over the

other resulting in a distinct asymmetry. The bracteal canal extends,

approximately in the mid-line and in close proximity to the ventral

wall, to about four-fifths the length of the bract. The distal end of the

bract is slightly truncate and bears two lateral processes, which vary

in shape from merely rounded corners to distinct teeth. The region

between them usually is roundly pointed. Additional processes may

be present on the lateral margins of the bract. Again these can form

distinct teeth, but quite often are indiscernible. The maximum

number of lateral processes found was two on one side, and one on

the other. The distal half of the dorsal side of the bract is dotted with

distinctive patches of small round ectodermal cells. These patches

are densely packed on the smallest bracts; but more spread out on the

larger ones, where some patches have been lost by abrasion. These

cells probably are sites of bioluminescence.

GASTROZOOID AND TENTACLE. The larger gastrozooids in the

Alvin material measured up to 10 mm in length. They are brown in

colour, in their preserved state, and are comprised of a short, narrow

basigaster, to the base of which the tentacle is attached; a large,

expanded stomach, the inside of which is covered with thickened

patches of endodermal cells; and a long proboscis, with longitudinal

endodermal hepatic stripes. Several younger, smaller gastrozooids

BARGMANNIA REVISION

also are present, which are largely colourless and transparent, with

only small patches of endodermal cells in the stomach region.

No mature tentilla remained with the specimens. The immature

tentilla (Figure SA) conformed to the basic Bargmanniadesign, with

the cnidoband ranging from being straight to curved or slightly

twisted. There was a maximum of only six large nematocysts,

probably stenoteles, thatmeasurec. 120 by 80um, irregularly arranged

onthe proximal region of the cnidoband. Abouthalf the circumference

ofthe cnidoband is covered withrows of two othertypes of nematocysts;

one is ovoid, measuring c. 16 x 11 um; and the other is spherical,

measuring c. 8.5um in diameter. Similar nematocysts are also present

on the terminal filament. It has not been confirmed that these

nematocysts are the acrophores and desmonemes that are typically

found on the terminal filaments of many physonect species. It is,

however, unusual to find such small nematocysts on the cnidoband.

The terminal filament obviously can extend to a considerable extent,

but in the preserved specimens it is generally tightly coiled.

SIPHOSOMAL TENTACLES AND BUDS (Figure 5B). Midway between

each gastrozooid a peculiar tentacular structure is attached directly

to the siphosomal stem. In the preserved specimens they are usually

tightly coiled, but some relaxed ones can reach lengths of 8 mm.

Along one side there is a biserial arrangement of spherical

nematocysts, c. 13 um in diameter, similar to those on the cnidoband

of the tentacle. The gastrovascular canal is lined by an irregular

honeycomb of large endodermal cells.

In addition to the siphosomal tentacle, small bud-like structures

were noted protruding from the stem. Because the siphosome in both

specimens was tightly coiled it was not possible to assess the precise

disposition of these buds. However, their arrangement may be

similar to that which will be described for the following species.

GONOPHORE (Figure SC). Both the Alvin specimens are male and

bear numerous gonodendra at various stages of development. The

gonophores measure up to 4 mm in length, including the pedicel.

They appear to bud one from another to form a small gonodendron.

If the gonophores becomes detached, their thin-walled pedicels

remain, giving the false impression of the presence of gonopalpons.

Again, since the stem is highly contracted, it is difficult to ascertain

their exact disposition.

REMARKS CONCERNING THE IDENTIFICATION OF BARGMANNIA

ELONGATA. Complete and well-preserved specimens of B. elongata

easily can be distinguished from the other Bargmannia spp., particu-

larly as the form of the bracts is very distinctive. For the nectophores,

the arrangement of the apico-lateral ridges, with their distinct right-

angled bend and the presence of the short extra ridges branching

from them, also are characteristic features. However, in the case of

net collected material, which is usually damaged or distorted, the

nectophores of this species may be difficult to distinguish from those

of the following species, as is discussed further after its description.

Bargmannia amoena sp. nov.

HOLOTYPE. BMNH 1998.2164, preserved in Steedman’s solution,

collected during JSL Il Dive 1458 off Dry Tortugas, Florida;

24°00.6'N 82°17.4'W; 3.ix.1987; 841m.

ie S&S pe YY gut

}

Fig.6 Bargmannia amoena sp. nov.. Photographs (taken by Ron Gilmer) of live specimen collected during JSL II Dives 968 (A) and 1687 (B).

Nectosomal lengths: A. c. 5 cm, B. c. 7 cm.

P.R. PUGH

Fig. 7 Bargmannia amoena sp. nov. A. Upper, B. lower, and C. lateral views of mature nectophore from type specimen collected during JSL II Dive 1458.

Scale bar = 5 mm.

PARATYPE. BMNH 1998.2165, preserved in Steedman’s solution,

collected during JSL I Dive 2636 off The Bahamas; 25°53.2'N

77°48.3'W; 5—xi-1989; 890m.

MATERIAL EXAMINED. 67 specimens have been collected during

40 dives by the submersibles JSL I and II. Of these, 52 have been re-

examined for this description. Some of the material originally

ascribed to B. elongata by Totton (1954) probably belongs to this

species. In addition, some poorly preserved nectophores have been

found in recent Discovery collections.

DIAGNOsIS. Apico-lateral ridges of nectophores smoothly curved,

without pronounced bends; their outer branches terminating well

above the ostium, before reaching the relatively small lateral ostial

processes. No additional ridges. in smaller specimens central thrust

block pointed with small digitiform process apically; in larger ones,

latter folded over ventrally so that, in upper view, thrust block

appears roundly truncate. In preserved specimens, ostium opens

basally. Nectosac more translucent than that of B. elongata. Ratio of

overall length of nectophore to that of nectosac averages 1.42,

varying slightly according to the size of specimen. Bracts of two

types; both delicate and foliaceous, with two pairs of lateral teeth;

without patches of large ectodermal cells.

DESCRIPTION. Photographs of living specimens collected during

JSL Il dives 968 and 1687 are shown in Figure 6. The specimens

from the JSL collections fall within three size classes, based on the

length of the mature nectophores, but also reflected by the degree of

sexual maturity. All the smaller specimens were colourless, while

the largest ones had bright orange-red basigasters; the basal part of

the gastrozooid.

PNEUMATOPHORE. The pneumatophore measured approximately

3 mm in height and 1.5 mm in width, but was highly distorted and

BARGMANNIA REVISION

Fig. 8 Bargmannia amoena sp. nov. A. Upper, B. lower, and C. lateral views of mature nectophore from small specimen collected during JSL II Dive 976.

Scale bar = 1 mm.

ruptured by the expansion of the gas within it. No pigmentation is

apparent. The pneumatophore is inserted onto the apical end of a

long stalk that, depending on the degree of contraction, can be 5—6

mm in length. As in B. elongata, this stalk is flattened at its base,

where it joins the nectosome, to form a hinge-like structure.

NECTOPHORE (Figures 7-9). The nectophores had a biserial, stag-

gered arrangement down the nectosome (Figure 6). The number of

nectophores found with each specimen varied from 5 to 32. Depend-

ing on the mean length of their nectophores, these specimens can be

divided into three size categories. Seven specimens, all collected

during the same cruise in 1984, bore c. 10 relatively small nectophores

whose lengths were less than 8 mm. The mean length, for the mature

nectophores, was 7.41 + 0.43 mm; the mean width 3.10 + 0.22 mm;

and the ratio of the overall length to that of the nectosac averaged

1.41 + 0.06. None of these specimens was sexually mature. The bulk

of the specimens was included in second size category, where the

length of the mature nectophores ranged from 9 to 19 mm. These

specimens bore distinct, but immature, gonophores. Each specimen

averaged about 20 nectophores, whose mean length was 13.70 +

1.86 mm; mean width 6.71 + 0.88 mm; and the ratio of the overall

length to that of the nectosac averaged 1.42 + 0.06. Finally six

specimens had even larger nectophores and were sexually mature.

They averaged 13.5 nectophores, whose mean length was 20.94 +

2.36 mm: mean width 10.58 + 2.18 mm; and the ratio of the overall

length to that of the nectosac averaged 1.44 + 0.05.

As was the case for B. elongata, the apex of the thrust block of the

smaller specimens was drawn out to form a small digitiform process

(Figure 8A). In the larger specimens, this process usually became

folded over onto the ventral side of the nectophore (Figure 7C), so

that, in upper view, the thrust block appeared roundly truncate

(Figure 7A).

The apico-lateral ridges are, in their preserved state, smoothly

curved and have no pronounced bend or side branches (Figures 7

& 8), as was found for B. elongata. After these ridges divide, the

inner branches extend obliquely down to reach the ostium; while

the outer branches curve down the sides of the nectophore, but

peter out well above ostial level. The latter is particularly marked

on the smaller nectophores (Figure 8C).

Basal extensions of the meso-lateral ridges form the baso-lateral

margins of the mouth-plate (Figures 7 & 8). The structure of the

mouth-plate varies with the size of the mature nectophore. In the

smallest specimens, the mouth-plate is only slightly truncate basally

(Figure 8A). In the middle size range of specimens, the mouth-plate

becomes more and more emarginate and, in the largest ones, it has a

narrow U-shaped median indentation stretching up to the ostium

(Figure 7A). The mouth-plates of the immature nectophores of all

sizes of specimens show the same features as the corresponding

mature ones (Figure 9).

Above the mouth-plate, in the small and medium sized speci-

mens, the basal extensions of the meso-lateral ridges curve slightly

in toward the mid-line (Figure 8B), before curving out again to form

the meso-laterals proper. In addition the infra-laterals do not unite

with the latter. On the largest specimens, there is no inward curve of

the meso-laterals (Figures 7B, 9B), but the infra-laterals have a very

weak connection with them (Figure 7B); However, the apical junc-

tion of the meso-laterals with the other ridges is always clearly

defined. The arrangement of the infra-lateral ridges, in the small

(Figure 8C) and medium sized specimens, is very similar to that

described for B. elongata. However, in the largest specimens, the

ventro-lateral wings are more extensive in the region where the

infra-laterals curve up to join the other ridges. The ventral margins of

these wings are distinctly emarginate.

The nectosac, in its preserved state, appears as a dorso-ventrally

undulating tube; but this is probably a preservation artefact. The

dorso-lateral extensions, in the mid region of the nectosac, are

slightly more extensive than in B. elongata. At its apex the nectosac

has a shallow U-shaped indentation, and the adaxial wall is distinctly

undercut and devoid of musculature. On the remainder of the

nectosac the musculature appears much less dense that of B. elongata,

and the nectosac is considerably more translucent. The arrangement

of the pallial and pedicular canals, and the radial canals on the

nectosac is similar to that of B. elongata.

In the preserved specimens, the ostium opens almost basally and

has a large velum. Its lateral walls are only slightly extended to form

small lateral processes. The pattern of the patches of ectodermal

cells is similar to that of B. elongata, but the cells are more uniform

in size, and the patches more diffuse laterally. In addition, there are

two ventro-lateral patches of deeply granulated cells that are rela-

tively large and almost spherical.

The youngest nectophores (Figure 9) typically show the absence

of a median thrust block. The inner branches of the apico-lateral

ridges reach the ostium. The degree of emargination of the apex of

P.R. PUGH

Fig. 9 Bargmannia amoena sp. nov. A. Upper, B. lower, and C. lateral

views of young nectophore from specimen collected during JSL II Dive

1449. Scale bar = 2 mm.

the nectosac is variable, according to the developmental stage. It

ranges from a narrow, median U-shaped indentation to a marked

emargination across most of the width of the nectosac. As noted

above the shape of the mouth-plate varies according to the size of the

specimen. On either side of the ostium there is a tract of small

ectodermal cells extending up toward the end of the outer branch of

the apico-lateral ridges. These tracts are longer than those seen on

the young nectophores of B. elongata and, again, are difficult to

discern on the adult nectophores.

BRACT (Figure 10). There are three pairs of bracts per cormidium.

Each is thin and leaf-like, with a slight thickening in the central

region of the proximal half. The dorsal surface is slightly convex,

and the ventral one slightly concave. In general their size is in

proportion with that of the nectophores, with those of the largest

specimens measuring up to 18 mm in length. No patches of ectoder-

mal cells were observed. However, in each cormidium, each

successive pair of bracts tends to be slightly larger than the pair

proximal to it. The proximal part of each bract is slightly asymmetri-

cal to allow for insertion onto the stem. The bracteal canal extends to

about two-thirds to four-fifths the length of the bract. It remains in

close contact with the ventral wall of the bract at all times.

There is much variation in the shape and form of the bracts, but

two basic types can be distinguished; both having two pairs of lateral

teeth. In one type, which make up the first two pairs of bracts in each

cormidium, the bracts are relatively symmetrical. The more distal

pair of lateral teeth are very variable in shape, ranging from being

virtually absent to being quite marked (Figure 10A, B, D). In the

second type (Figure 10C, E), which are the distal pair, the bracts are

asymmetrical, and the bracteal canal can have a distinct proximal

curve. The distal pair of lateral teeth are well developed and closer

BARGMANNIA REVISION

Fig. 10 Bracts of Bargmannia amoena sp. nov. Scale bars: A, B, C = |

mm, D, E=2 mm.

together than on the first type, so that the distal end of the bract is

relatively narrow. One of the proximal pair of teeth is usually more

developed than the other, and on that side the lateral wall of the

proximal part of the bract often extends out as a rounded notch.

GASTROZOOID AND TENTACLE (Figure 11A, C). The largest

gastrozooids measure up to 10 mm in length. In the preserved state

they are suffused with a brown coloration with, in the largest

specimens, the basigaster having bright orange-red pigmentation.

The latter (Figure 11C, bg), typically, is cup-shaped, enclosing the

base of the stomach region, and is covered in large rounded ecto-

dermal cells. The stomach region (Figure 11C, s) appears relatively

thin and the endodermal hepatic villi can be seen within. The

proboscis region can be extended to some distance.

The tentacle can be several centimetres in length. It is a simple,

narrow, unsegmented tube, bearing a haphazard and irregular arrange-

ment of the two sorts of small nematocysts that are also found on the

tentilla. In the present specimens, only a few tentilla, up to 10,

remain attached close to its base. In their preserved state, the tentilla

(Figure 11A) typically are highly contracted and are comprised of a

short pedicel; an irregularly twisted cnidoband; and, for the most

part, a regularly coiled terminal filament. The cnidoband is a simple

tube that, in life, is generally straight or slightly curved, and can

extend to a length greater than 0.5 cm. One side of the cnidoband

appears to consist of a primitive elastic strand. It is not tightly folded,

as is the case in some other physonect species, but a few pleats are

present. The other side of the tentillum is comprised of numer-

ous rows of small nematocysts of two types, as was the case in

B. elongata. These are ovoid, measuring 20 x 14 um and 12 x 11 um,

and occur in roughly equal proportions and possibly in alternat-

ing rows, although this could not be determined with certainty.

Similar nematocysts are found along the length of the terminal

filament. Again, it has not been determined whether these nemato-

cysts are the acrophores and desmonenes found in other physo-

nect siphonophores. At the proximal end of the cnidoband there is a

paired series of up to 26 stenoteles that measure 135 x 105 um.

SIPHOSOMAL TENTACLES AND BUDS (Figure 11B,C). As several of

the siphosomal stems of the specimens examined remained relaxed,

it was possible to study the disposition of the siphosomal tentacles

and buds in detail. The primary siphosomal tentacle (Figure 11B, #)

is inserted midway between successive gastrozooids and can be

tightly coiled or extend to several millimetres in length. As in

B. elongata, its surface is covered in large ectodermal cells and there

is a paired series of nematocysts along one side.

On each cormidium there are, at least, four solid bud-like struc-

tures, whose arrangement is sexually dimorphic. In the female

specimens (Figure 11B), the first bud (b/) lies a short distance distal

to the gastrozooid (gz/), while the second (b2) is inserted about one

quarter the length of the cormidium. The gonodendron is then

inserted between the latter and the central siphosomal tentacle (7).

The third bud (b3) lies a short distance distal to this tentacle, and the

last (b4) is inserted immediately proximal to the next gastrozooid

(gz2). In the male specimens (Figure 11C), the gonodendron is

situated immedietly distal to the gastrozooid. The first bud (b/) then

lies distal to the gonodendron at about one quarter the length of the

cormidium; that is approximately in the same position as the second

bud on the female specimens. The second bud lies immediately

proximal to the central siphosomal tentacle; and the third midway

between that tentacle and the next gastrozooid. The fourth, as in the

female specimens, is inserted immediately proximal to the next

gastrozooid. These arrangements pertain in the great majority of the

specimens examined, but in the largest ones another tentacle, and

possibly another bud, are found in close proximity to the fourth bud.

Usually, this tentacle is much smaller than the central tentacle, but

otherwise appears to be identical; including the double row of

nematocysts.

GONOPHORE. (Figure 11B, C). As noted above, the degree of

sexual maturity of the specimens appears to be directly related to

their size, as assessed by the length of the nectophores. Thus in the

smallest specimens, at most, only gonophore buds can be seen. The

major group of medium sized specimens have better developed

gonophores, while the largest are obviously sexually mature. All

seven of the largest specimens are male.

There is only a single gonodendron per cormidium. In male

specimens the gonodendron lies immediately distal to a gastrozooid

and proximal to the first siphosomal bud. The mature male

gonophores (Figure 11C, mg) measure up to 5.5 mm in length and

1.1 mm in diameter The female gonophores (Figure 11B, fg) are

attached to the stem by a short stalk that is inserted approximately

midway between the second siphosomal bud and the central

siphosomal tentacle. Between one and six gonophores, in various

stages of development, are attached to it by short pedicels. Each

62 P.R. PUGH

Fig. 11 Bargmannia amoena sp. noy. A. Mature tentillum, with stenoteles (st) at base of cnidoband (magn. 25x); B. Cormidium of siphosome, with

gastrozooids detached (gz' and gz’ mark their attachment points) showing the siphosomal tentacle (t), four buds (b'*) and female gonophores (fg) (magn.

25x); C. Male gonophores (mg) attached just distal to gastrozooid, with its basigaster (bg) and stomach (s), and proximal to the first bud (b’) (magn.

12.5x).

BARGMANNIA REVISION

gonophore contains two eggs. This is a highly unusual situation as,

according to Carré and Carré (1995) all other physonect siphono-

phores have only one egg in each gonophore. However, Totton

(1965) states that the gonophores of Pyrostephos vanhoeffeni con-

tain from three to five eggs.

DISTRIBUTION. Much of Totton’s (1954, 1965) Bargmannia

material, from early Discovery collections, is so poorly preserved

that it is difficult to be certain to which species it belongs. However,

as noted earlier, the nectophores from Discovery Sts. 699 and 681,

from the South and North Atlantic Ocean respectively, belong to B.

elongata; as does that from Discovery St. 107 from south of South

Africa. Several other of his nectophores probably belong to B.

amoena, but this has not been established with certainty.

The great majority of the 8500+ nectophores of Bargmannia spp.

that have been identified from over 300 recent Discovery samples,

mainly from the North east Atlantic Ocean, belong to either B.

elongata or B. amoena. However, these identifications were made

before it was realised that two similar species were present. A re-

examination of some of the material, however, typically showed that

the material was too poorly preserved for specific identification.

However, it was clear that B. amoena, not B. elongata, was the

predominant species of the genus at c. 44°N, 13°W, where an

extensive series of collections was made (Pugh, 1984). Collectively,

both species have a widespread distribution in the North-east Atlan-

tic Ocean; from the equator to 60°N, with possibly B. elongata being

more common at lower latitudes and B. amoena at higher ones.

Nectophores have been collected at all depths from the surface to

4520 m, but the vast majority were found in samples from between

200 and 600 m.

Most of the 67 specimens of B. amoena collected by the sub-

mersibles JSL I and Il came from a relatively small area in the region

of The Bahamas, from 25°03' to 26°36'N and 77°23’ to 78°44'W.

Five others were collected near the Dry Tortugas, between Florida

and Cuba, at c. 24°30'N, 83°45'W. All were collected over a wide

depth range, from 435 to 910 m, with a mean depth of 625 + 130 m.

This mean depth is slightly deeper than the depth range for both B.

elongata and B. amoena found in Discovery net collections. How-

ever, both figures probably are biased because, in the case of the

submersible, most observations and collections were made within

the 600-900 m depth range, while at 44°N, 13°W for instance, most

of the net sampling was concentrated in the top 600 m of the water

column.

REMARKS CONCERNING THE IDENTIFICATION OF BARGMANNIA

AMOENA. Complete and well-preserved specimens of B. amoena

easily can be distinguished from the other Bargmannia spp. as they

have very distinctive bracts. However, if only poorly preserved

nectophores are present, then there may be some difficulty in

separating this species from B. elongata. They cannot be confused

with B. gigas, because of the relatively enormous size of the latter’s

nectophores; and should not be confused with B. lata. The much

narrower nectosac of the latter species, together with the greater

depth of the furrow between its deep lateral wings, should easily

distinguish it.

As noted above, the best feature distinguishing B. elongata and B.

amoena is the arrangement of the apico-lateral ridges. InB. elongata

these have a pair of side branches, running down toward the mid-

line, while in B. amoena such side branches are absent. In addition,

in B. elongata, at the point where these side branches arise, the

apico-laterals bend sharply outwards, through 90°, while in B.

amoena the apico-laterals only curve gently away from the mid-line.

Unfortunately, it is this region of the nectophore that most often

becomes distorted in poorly preserved specimens and these distin-

guishing features can be masked. This can result in the distinct,

right-angle bend in the apico-lateral ridges of B. elongata coming to

resemble the much smoother curve in B. amoena or, conversely,

those of the latter species becoming distorted to form distinct bends

resembling those of the former. The side branches to the apico-

laterals in B. elongata often are difficult to discern, but usually show

up after staining.

Another obvious difference, despite its subjectiveness, is the

density of the musculature on the nectosac. Nectophores with almost

opaque nectosacs appear to belong to B. elongata, while those with

translucent nectosacs belong to B. amoena. In addition, the ratio of

the total length of the nectophore to that of the nectosac may be of

use. In B. elongata this ratio, in well-preserved specimens, averages

1.31, while in B. amoena it averages 1.42. However, with poorly

preserved material, particularly when the basal half of nectophore is

damaged, both measurements could be underestimated, which would

lead to a corresponding increase in the ratio.

Other features of the nectophore that might be considered include

the fact that in B. amoena the outer branch of the apico-lateral ridges

peters out higher above the ostium than in B. elongata. Also, the

lateral processes to the ostium are much larger in B. elongata. In

addition, the angle at which the ostium opens is very characteristic

in well-preserved material. In B. elongata the ostium is directed

dorso-basally while in B. amoena it opens basally. However, all

these features might be difficult to discern in poorly preserved

nectophores The structure of the mouth-plate may be important but,

as has been shown for B. amoena, this may vary according to the size

of the specimens. More well preserved specimens of B. elongata are

needed in order to assess this. Similarly, this applies to the arrange-

ment of the meso-lateral ridges and their basal extensions; and to the

profile of the ventral margins of the ventro-lateral wings.

ETYMOLOGY. Amoena is Latin for ‘pleasing, delightful’.

Bargmannia lata Mapstone 1998

Bargmannia elongata Totton 1954 (partim) (text-Figure 28, E-F

only), 1965 (partim) (Figure 45 E-F only).

Bargmannia lata Mapstone 1998: 141-146, figs 1-3.

HOLoTyYPE. In the collections of the British Columbia Provincial

Museum, BCPM 996-203-1; one nectophore and one bract collected

at St. LC10 (48°22.4'N 126°20.2'W; 24-iv-1987; 700-Om) off

British Columbia, Canada;

PARATYPES. As designated by Mapstone; 1. 7 nectophores and 7

bracts (BCPM 996-204-1#1), and 2. 6 nectophores and 6 bracts

(BCPM 996-205-1#2) from the same sample as the holotype; 3. 11

nectophores (BCPM 996-206-1#3), and 6. 1 bract (BMNH

1996.1239-1240#6) from St.A4 (48°15'N 126°40'W; 21.iii.87; 500

m); 4. 8 nectophores (0O-700m; BCPM 996-207-1#4), and 5. 14

nectophores and 2 bracts (BMNH 1996.1234-1238#5) from St.

LB17 (47°56.5'N 126°26.1'W; 21.11.87; 700m).

MATERIAL EXAMINED. One specimen collected during Alvin Dive

966 off San Diego, California, U.S.A.; 33°04'N 118°16'W; 8-ix—

1979; water depth 747m. The depth of collection of the specimen

was not recorded.

Two nectophores collected at Discovery St. 1769, and figured by

Totton (1954, Text-Figure 28, E, F; 1965, Figure 45, E, F) as

B. elongata; 33°43.3'S 8°38.5'E; 20—v—1936; 1000-750 m; NHML

1957.5.15.110.

In addition, the specimens that Totton included under the name

B. elongata have been re-examined. Although not all are in good

condition, the following appear to belong to B. lata:-

Table 1 Geographical distribution of Bargmannia lata from recent Discovery collections.

Station Date Position

8565# 1 1—viti-74 3°03.1'N 23°14.3'W

6662#37 21-11-68 10°34.9'N 19°43.7'W

6662#32 20-11-68 10°45.3'N 19°51.7'W

6662#30 19-11-68 10°47.4'N 19°52.7'W

6662#15 16-11-68 10°57.0'N 19°56.6'W

6662#20 17-11-68 10°57.5'N 19°49.0'W

6662#22 17-11-68 10°57.6'N 19°57.3'W

6662#16 16-11-68 10°59.4'N 19°52.1'W

7824#39 10-11-72 11°01.1'N 19°55.8'W

6662#10 15-11-68 11°03.1'N 19°59.2'W

6662# 7 14-11-68 11°04.6'N 19°48.2'W

6662# 8 15-11-68 11°08.2'N 19°47.8'W

7831# 1 16-11-72 13°18.4'N 25°33.2'W

7803# 2 19-11-72 18°06.4'N 25° 8.1'W

11261#16 28—vi-85 31°13.1'N 25°18.3"W

8281#29 17-i1-73 31°42.5'N 63°43.6'W

11794#36 26-vi-88 47°14.2'N 19°31.2"'W

11794#83 2-vii-88 47°17.9'N 19°21.4"°W

11794#31 26—vi-88 47°27.1'N 19°18.0'W

12096# 2 3-vi-90 47°57.8'N 16°49.6'W

4 nectophores from John Murray Expedition St. 34; 13°05.6'N,

46°24.7'E; 16—x—33; 0-1022 m. BMNH 1949.11.10.378; and

4 nectophores from Discovery St. 206; 16°36'S, 6°25.1'W; 1—v—

37; 1900-1500 m. BMNH 1957.5.15.111.

Several nectophores and bracts also have been identified from

more recent Discovery material from the N.E. Atlantic (Table 1).

DIAGNOSIS. Nectophores with relatively long median thrust block;

with extensive ventro-lateral wings, emarginate on ventral margins.

Nectosac a relatively short, narrow tube without any pronounced

dorso-ventral undulations; squarely truncate apically. Ratio of over-

all length of nectophore to that of nectosac, on average, exceeds

1.59. Bracts large, robust, distally truncate; with semicircular dorsal

ridge connecting tips of baso-lateral processes and delimiting a

dorso-basal facet; with prominent tooth on outer lateral margin.

DESCRIPTION. A photograph of the live specimen collected by

Alvin and taken on board the mother ship is shown in Figure 12.

Unfortunately, prior to photography, several nectophores had be-

come detached and the siphosomal stem had contracted. No

pigmentation is apparent in the preserved specimen, but the original

colour photograph indicates that the whole of the endodermal lining

of the stem was suffused with an orange-red colour.

PNEUMATOPHORE. The highly distorted pneumatophore of the

Alvin specimen measured 2 mm in height. It is borne on a very short,

but probably highly contracted, stalk.

NECTOPHORE (Figures 13-15). A total of 16 nectophores was

found with the Alvin specimen. The mean dimensions for the fully

developed nectophores of this specimen were:- length 19.44 + 2.73

mm (range 14.79—23.86 mm); width 9.72 + 1.07 mm, and the ratio

of total length to that of the nectosac was 1.59 + 0.09. Nectophores

of the earlier Discovery material, in the NHM, are somewhat larger

with an average length of 24.94 + 6.02 mm (range 17.74-31.94

mm). The ratio of total length to that of the nectosac also was slightly

greater ratio (1.67 + 0.10; n = 10); the same as that for Mapstone’s

(1998) material. Similarly, the nectophores of more recent Discov-

ery material also are larger: length 22.89 + 4.04 mm (range 15.8—29.6

mm; n = 48). These nectophores have the greatest length:nectosac-

length ratio of 1.73 + 0.11. However, this increase in ratio probably

P.R. PUGH

Depth (m) Nects Bracts

700— 800 2)

1060-1300 24 16

1210-1450 3 4

730-795 17

600— 695 DH

810-900 8

610— 680 25 3

810— 890 18 DB

895-1000 3)

910-985 4

715— 800 33

910-985 Ti

10-1000 12

1015-1250 13

1000-1100 16

1259-1500 26

1200-1300 10

1300-1395 7

2500-2750 17 2

1100-1200 16 20

Fig. 12 Bargmannia lata. Photograph (reproduced by kind permission of

Larry Madin, WHOJ) of live specimen collected during Alvin Dive 966.

Nectosomal length c. 5 cm.

BARGMANNIA REVISION

Fig. 13 Bargmannia lata. A. Upper, B. lower, and C. lateral views of mature nectophore collected during DSRV Alvin Dive 966. D. Ventral view of the

apex of another nectophore. Scale bar = 2 mm.

is due to the fact that the base of the nectophore frequently is

damaged, resulting in an underestimate of both measurements, and

a consequent increase in their ratio.

The central thrust block forms an extensive triangular process

whose apex is often roundly pointed Figure 13A). However, on

several nectophores one side is drawn out to form a small digitiform

process that may be folded over laterally or ventrally (Figure 13D).

The ridge pattern conforms to the basic Bargmannia design, with no

extra ridges being present.

From their junction with the meso- and infra-lateral ridges on the

‘shoulder’ of the nectophore, the apico-laterals are directed ob-

liquely toward the mid-line. They closely approach each other, and

continue for some distance in a basal direction; leaving a narrow

median furrow between them. At about one quarter the length of the

nectophore, in the Alvin material, they rapidly curve out laterally,

before giving rise to the typical inner and outer branches (Figures

13-15). The inner branch curves obliquely toward the mid-line and

joins the ostium on its dorsal surface. The outer branch curves down

and then round and ends on the lateral margin of ostium, although it

can be difficult to discern basally. The angle between the apico-

lateral ridge and its inner branch is acute (Figure 13A). However, in

less well preserved nectophores, this pronounced angle is not always

apparent (Figure 14B) and the inner branch can appear as a simple

continuation of the main ridge.

The mouth-plate is small and made up of two rounded lobes that

unite in the mid-line, slightly basal to the ostium. The ventro-lateral

margins of these lobes are, as usual, formed by basal extensions of

the meso-lateral ridges. Above the ostium, on the ventral surface of

the nectophore, these basal extensions curve round toward the mid-

line, before looping back out as the meso-laterals proper and

continuing apically. After a relatively long distance, in comparison

with B. elongata and B. amoena, the infra-laterals branch from them

(Figures 13C, 15C). The meso-laterals then continue obliquely up

and across the lateral margins of the nectophore to join the apico-

and infra-laterals on the ‘shoulder’ of the nectophore. The junctions

with the other ridges, both apically and basally, are obvious, unlike

in B. elongata.

In the basal two-thirds of the nectophore, the infra-lateral ridges

form the ventral margins to the ventro-lateral wings (Figure 13C,

15C). These wings are relatively large in comparison with those of

B. elongata and B. amoena, occupying more than half the depth of

the nectophore. They are distinctly emarginate in the mid region of

66 P.R. PUGH

‘ -

= oo ---

-- =<

‘

-—

L---

Noes

-2

ere

Fig. 14 Bargmannia lata. A. Lower and B. upper views of nectophore from Discovery St. 1769. Scale bar = 5 mm.

Fig. 15 Bargmannia lata. A. Upper, B. lower and C. lateral views of young nectophore collected during DSRV Alvin Dive 966. Scale bar = 1 mm.

BARGMANNIA REVISION

the nectophore. In the region where the infra-laterals leave the

ventral margins of the ventro-lateral wings, the latter begins to

thicken toward the mid-line. These thickened, rounded, unridged

lateral walls continue up to the apex of the nectophore and, on the

thrust block, delimit a narrow gutter that enfolds the nectosomal

stem (Figures 13B, 14A).

The nectosac is a relatively short and narrow tube without any

marked dorso-ventral undulations (Figures 13, 14) in the preserved

specimens. Its apex lies approximately on a level with the ‘shoulder’

of the nectophore. The nectosac occupies only c. 38% of the width

of the nectophore, as measured across its ‘shoulder’. This results

from the fact that the extensive ventro-lateral wings not only increase

the depth of the nectophore, but also increase its relative width. The

apex of the nectosac is squarely truncate, without any marked

indentation. Its adaxial surface is distinctly undercut and, typically,

is devoid of musculature. The remaining musculature on the nectosac

appears much less dense in comparison with that of B. elongata.

The canal system follows the basic Bargmannia plan. The long

pallial canal ends, apically, with a short dorsad inflection into the

mesogloea. On the nectosac the pedicular canal gives rise to only the

dorsal and ventral radial canals. In contrast to B. elongata and B.

amoena, in the preserved material the lateral radial canals have

straight courses down the lateral margins of the nectosac. However,

afi

wire er

in life, their courses appeared to be slightly undulating (Figure 12).

In the original colour photograph there are indications that the

pallial, dorsal, ventral, ostial ring, and proximal parts of the lateral

canals were suffused with a light orange-red colour.

The ostial opening, in the preserved nectophores, typically is

displaced slightly dorsad and has a well-developed velum, but no

pronounced lateral processes. There are no marked patches of

ectodermal cells, although some nectophores show a single row

around the basal half of the ostium and/or a short, narrow band of

small cells that lies just dorsal to the outer branch of the apico-lateral

ridge. These cells, again, are presumed to be sites of biolumines-

cence.

BRACT (Figure 16). Only seven bracts were retained with theAlvin

material. However, because of their very characteristic shape, sev-

eral more have been identified from recent Discovery material. The

bracts measured from 13.5 to 27 mm in length and were remarkably

robust. They had a convex dorsal and a concave ventral surface. In

the Alvin specimen, there were two types of bract, with one type

being represented by only a single small bract. The key feature that

distinguishes them is the presence of only a single lateral tooth on

the outer margin (Figure 16A, B) of the larger ones; while the

smaller one has lateral processes on both sides (Figure 16C). The

Fig. 16 Bracts of Bargmannia lata collected during DSRV Alvin Dive 966. Scale bar = 2 mm.

larger ones also were distinctly asymmetric proximally; and the

bracteal canal made a right-angled bend. These differences may be

just the result of growth, or may related to their point of attachment

on the cormidium, as was noted for B. amoena. Both types of bract

are distally truncate, and possess a semicircular dorsal ridge that

delimits a rounded distal facet. The ridge connects the tips of two

distal processes.

The shape of the distal margin of the larger bracts varied consid-

erably. On some, the inner margin of one of the distal process

extended up the ventral side of the bract forming a flap-like struc-

ture; while on others this flap was absent. Another small ventral flap

can be present, approximately in the mid line, in the proximal half of

the bract. The bracteal canal lies just above the ventral surface of the

bract and ends close to its distal margin. The original colour photo-

graph of the Alvin dive 966 specimen indicated that, in life, some of

the bracteal canals had an orange-red pigmentation like that of the

remainder of the stem.

GASTROZOOID AND TENTACLE. Only a small portion of the proxi-

mal end of the siphosome was preserved from the Alvin specimen, so

that only a few young gastrozooids were present. These measured up

to 7 mm in length, and showed no distinguishing features. No

pigmentation is apparent in the preserved material, but in life they

had a deep red pigmentation.

The tentacles attached to the gastrozooids mainly bore young

tentilla; with a c. 1 mm pedicel; a 2.5 mm cnidoband, apparently

devoid of large nematocysts; and a c. 4 mm uncoiled terminal

filament, apparently without a terminal process. However, a few

more mature tentilla had cnidobands extending to more than 8 mm,

with 4-6, possibly more, large nematocysts (stenoteles), arranged

biserially and arranged alternately, at their proximal ends. Small

nematocysts, possibly of two types, are present throughout the

cnidoband and terminal filament.

SIPHOSOMAL TENTACLES AND BUDS. The peculiar tentacular proc-

esses, previously noted in specimens of B. elongata and B. amoena,

are present on the small part of the stphosomal stem remaining but,

because of the contracted state of the latter, it was not possible to

ascertain their exact disposition. They are narrow, delicate struc-

tures, measuring up to c. 4 mm in length, and are covered in large,

rounded ectodermal cells. Small nematocysts are present but, with-

out destroying the tentacle, it was no possible to assess whether they

had a biserial arrangement, as noted for the previously described

species. Siphosomal buds also appear to be present, but their arrange-

ment could not be discerned.

GONOPHORE. A few loose male gonophores are present with the

Alvin material. They are identical to those previously described for

B. elongata and B. amoena.

DISTRIBUTION. A total of 288 nectophores and 27 bracts of B. lata

have been found in recent Discovery collections (Table 1). The data

indicated that, in the North-east Atlantic Ocean, B. lata was more

commonly collected at lower latitudes and at deeper mesopelagic

depths; with a mean depth of c. 1000 m. Totton’s material came from

two sites in the SouthAtlantic Ocean and one in the Gulf of Aden; the

Alvin material came from off San Diego, California, USA; and

Mapstone’s (1998) from off British Columbia, Canada; thus indicat-

ing a widespread geographical distribution for this species.

REMARKS CONCERNING THE IDENTIFICATION OF BARGMANNIA LATA.

B. lata can now be easily distinguished from other Bargmannia spp.

The nectophores are relatively broad, with extensive ventro-lateral

wings, that are markedly emarginate along their ventral margins.

The median thrust block is relatively long so that the ratio of the total

P.R. PUGH

length of the nectophore to that of the nectosac is high, c. 1.6, or

more for net collected material, as compared with c.1.31 in B.

elongata and c. 1.41—1.44 in B. amoena. The nectosac appears as a

relatively short, narrow, straight-sided tube, without any pronounced

dorso-ventral undulations, and squarely truncate apically. It occu-

pies only c. 38% of the width of the nectophore, as compared with

45% in B. elongata. The large, robust bract, with a semicircular

dorsal ridge connecting the tips of the baso-lateral processes, also is

distinctive.

Despite these differences it is clear that Totton (1954, 1965), had

little reason to suspect that he was dealing with at least two

Bargmannia spp., particularly as he had so few, poorly preserved

nectophores. However, with the collection of complete specimens of

Bargmannia spp. by submersibles the specific differences between

the various types of nectophore that Totton illustrated can now be

established.

Bargmannia gigas sp. nov.

HOLOTYPE. BMNH 1998.2166 one nectophore, preserved in

Steedman’s solution, collected at Discovery St. 8560#2 (0°03.1'N,

22°44.2'W; 27-vii-1974; 1510-2000 m; RMT8 net).

PARATYPES. Three nectophores, preserved in Steedman’s solu-

tion, from the same Discovery sample. BMNH 1998.2167-69.

MATERIAL EXAMINED. The type and paratypes, and a further ten

nectophores from the same station, which are retained in the Discov-

ery collections at the Southampton Oceanography Centre. All the

nectophores are presumed to have originated from a single speci-

men.

DIAGNOSIS. Nectophores relatively enormous, up to 52 mm in

length; with large ventro-lateral wings; with small mouth-plate

deeply divided. Basic ridge pattern supplemented by three pairs of

ridges, all dividing from apico-laterals; two pairs short, directed

toward mid-line; third pair directed laterally. Nectosac without

dorso-ventral undulations; apex only slightly emarginate; ostial

opening large. Ratio of overall length of nectophore to that of

nectosac averages 1.63.

DESCRIPTION. NECTOPHORES (Figures 17 and 18). The relatively

enormous nectophores varied in size from 14.5 x 8 mm (length x

width) for the smallest, immature one, to 52 x 20 mm, respectively,

for the largest. The mean dimensions for the fully developed

nectophores were length: 41.0 + 6.96 mm and width: 19.1 + 2.52

mm, and the ratio of total length to that of the nectosac was 1.63 +

0.10. The nectophores, in their present state of preservation, are

devoid of pigmentation and, in most cases, the muscular lining of the

nectosac has become detached and/or lost. The large, thickened,

central thrust block is roundly truncate.

The basic pattern of the prominent ridges conforms with that of

the genus, and both the inner and outer branches of the apico-laterals

appear to reach the dorso-lateral margins of the ostium. In addition

there are three pairs of ridges that branch from the apico-laterals

(Figures 17A, 18). Two of these pairs of ridges are very short and run

down into the shallow median gutter, towards the mid-line. One pair

arises at a level of about two-fifths the length of the nectosac, while

the other originates from the inner branches of the apico-laterals,

close to the ostium. The other pair arises from the outer branches of

the apico-laterals and extends up the sides of the nectophore between

the apico- and meso-laterals. These ridges peter out approximately

at the mid-length of the mature nectophore. Below them the lateral

walls of the nectophore often show prominent thickenings (Figure

iG):

BARGMANNIA REVISION

4 * ee,

ee ee eee

-<--

Fig. 18 Bargmannia gigas sp. nov. Upper views of A. smallest, and B. slightly larger nectophores. Scale bar = 0.5 cm.

The broad, but relatively short, mouth-plate consists of two

rounded processes whose basal margins are marked by basal exten-

sions of the meso-lateral ridges. These ridges peter out, without

apparently uniting, on the lower surface of the nectophore a short

distance above ostial level. The infra-lateral ridges branch from the

meso-laterals approximately on a level with the mid-length of the

nectosac. The meso-laterals then run obliquely up the sides of the

nectophore to join the other ridges, approximately on a level with the

top of the nectosac. The infra-laterals form the ventral margins of the

ventro-lateral wings up to a level just above the top of the nectosac.

They then bend through 90° and run up to join the apico- and meso-

laterals (Figure 17C). The thickened ventro-lateral wings are well

developed and enclose a deep groove, which at its deepest occupies

half the height of the nectophore (Figure 17C). They are roundly

truncate apically at about four-fifths the length of the nectophore.

The nectosac is a long tube, with only a slight apical emargination,

that occupies most of the main body of the nectophore, and has no

obvious dorso-ventral undulations. It is distinctly undercut adaxially

and is presumed to have a muscle-free zone in that region, although

this could not be established with certainty. The ostial opening is very

large and is only slightly directed towards the upper surface. The

pedicular canal (Figure 17B) typically only gives rise to the dorsal and

ventral radial canals. The course of all the canals is straight.

Typically, the youngest nectophores show the absence of a central

thrust block (Figure 18A), but with a clearly defined ridge pattern. A

slightly larger nectophore shows the gradual development of the

thrust block and the ventro-lateral wings (Figure 18B).

REMARKS CONCERNING THE IDENTIFICATION OF BARGMANNIA GIGAS.

B. gigas is known only from the nectophores of what is presumed to

be a single specimen, collected in the equatorial Atlantic at a depth

of 1510—2000m. The nectophores easily can be distinguished from

other Bargmannia sp. by their incredible size and the distinctive

pattern of ridges.

ETYMOLOGY. The specific name gigas refers to the giant size of

the nectophores.

DISCUSSION

As was noted in the Introduction, the content of the genus Barg-

mannia is debated. Although Totton (1965) included it in the family

Pyrostephidae, some of the characters that he listed in his diagnosis

of that family apply exclusively to the scope accorded to the genus

Pyrostephos, which is monotypic for P. vanhoeffeni. In particular,

these are the looping of the lateral radial canal on the nectosac of the

nectophore, and the three to four marked bends of the dorsal canal.

In Bargmannia all the canals are held to be straight, or only slightly

sinuous. Other characters, such as the number of tentilla on the

tentacle, and the structure of the bracts and gastrozooids probably

are more specific than familial. However, in both genera the

nectosome is long but again this is not a good familial character.

At first glance, the nectophores of Pyrostephos vanhoeffeni (see

Totton, 1965, Figure 41) and Bargmannia species look strikingly

different. However, there are several similarities. Specimens of P

vanhoeffeni have been collected recently by SCUBA divers (G.R.

Harbison, personal donation) and by net (Pages, Pugh & Gili, 1994).

It is apparent from these that the mature nectophores can vary greatly

in size; ranging from 8 x 5 mm (length x width) in the SCUBA

collected material (Figure 19B) to 15 x 18mm, respectively, in the net

collected specimens (Figure 19C). Such large size ranges of the

mature nectophores of physonect species havenotoften been observed,

P.R. PUGH

although such is so in Nanomia bijuga (delle Chiaje, 1841) (Pugh,

pers. obs.). Itis also known to be so in at least two Bargmannia spp. In

B. amoena (Figure 19A), the size variation of mature nectophores is

even greater than that of P. vanhoeffeni, ranging from c. 6 x 3 mm

(length x width) to 25.5 x 12.5 mmrespectively. Although the general

shape of mature Bargmannia nectophores does not change with size,

it appears that that of P. vanhoeffeni does. In the smaller specimens of

the latter (Figure 19B) there is a large triangular thrust block,

reminiscent of that on mature Bargmannia nectophores. However, in

the larger, preserved specimens (Figure 19C) the thrust block is folded

upwards producing a deep transverse furrow on the dorsal surface, just

basal to it. Neither P. vanhoeffeni nor Bargmannia spp. have large

apico-lateral processes.

The nectophores of both genera have the same basic ridge pattern;

comprising apico-, infra, and vertical (meso-) laterals, but no lateral

ridges. In addition, in both, the apico-laterals divide into two branches

close to the ostium. The inner branch (‘frontal ridge’) of the larger

nectophores of Pyrostephos vanhoeffeni (see Totton, 1965, Figure

AQ) is relatively short, in comparison with Bargmannia spp., and

directed only toward the mid-line. However, the present material,

particularly that of the smaller specimens, shows that these ridges

can curve round basally and continue for a short distance towards the

ostium before petering out. Nonetheless, the species of these two

genera are not the only physonects to show this basic pattern of

ridges. It is also found on the nectophores of two others namely,

Frillagalma vityazi Daniel, 1966 (see Pugh, 1998) and Erenna

richardi Bedot, 1904 (P.R.Pugh, personal observation). In addition,

an even simpler arrangement, in which the vertical lateral ridges are

absent, is found in two Marrus species, namely M. antarcticus

Totton, 1954 and M. orthocanna Kramp (1942). For these, the

branching of the apico-lateral ridges is weak and difficult to discern.

A third species, namely M. orthocannoides, that Totton (1954)

include in the latter genus probably does not belong there as its

nectophores do not have an adaxial muscle-free zone on the nectosac.

Species referred to both Bargmannia and Pyrostephos have an

adaxial zone on the nectosac of the nectophore that is muscle-free and

deeply embayed. In addition, the lateral radial canals arise separately

from the dorsal canal. These appear to be important characteristics. Of

the other species previously mentioned Marrus antarcticus and M.

orthocanna show all of these characters. However, in Frillagalma

vityazi, there is no deeply embayed, muscle-free adaxial zone; al-

though the lateral radial canals do arise separately from the dorsal one,

albeit very close to the point of insertion of the pedicular canal (Pugh,

1998). Further, this species has many marked differences from the

others under consideration and need not be considered further in this

discussion. Erenna richardi does have a muscle-free zone, but it lies at

the apex of the nectosac, whichis not deeply embayed adaxially. Thus,

from the basic arrangement of the ridges and nectosac, the nectophores

of Bargmannia, Pyrostephos and Marrus species are very similar.

Another common feature is that they all have relatively short pedicular

and relatively long, ascending pallial canals. But how do their

siphosomal elements compare?

Most siphonophores are believed to be hermaphrodite (mono-

ecious), bearing both male and female gonophores. However,

specimens of Physalia, the Portuguese Man O'War, and probably all

other cystonect siphonophores, are single sexed (dioecious). It

should be noted that Mackie, Pugh & Purcell (1987, p.100) used the

terms monoecious and dioecius erroneously. In physonect

siphonophores, species of the benthic family Rhodaliidae appear to

be dioecious (Pugh, 1983), as are Marrus antarcticus, Pyrostephos

vanhoeffeni (Totton, 1965), and from the present study Bargmannia

spp. According to Andersen (1981) M. orthocanna is monoecious,

but the male gonophores he illustrated were only minute, bud-like

BARGMANNIA REVISION

Fig. 19 Nectophores of A. Bargmannia amoena sp. nov. (magn. 10); B, C. Pyrostephos vanhoeffeni collected by SCUBA (B, magn. 11x) and by net (C,

magn. 7.5x).

structures. Since only female gonophores could be identified on

submersible collected specimens, this point could not be checked

(P.R.Pugh, personal observation). Whether Erenna richardi is

monoecious or dioecious remains unknown. Nonetheless, it is of

interest to note that, of the physonect species whose female

gonophores are known, only those of P. vanhoeffeni and B. amoena

contain more than one egg; 3—5 in the former species (Totton,

1965) and two in the latter.

The structure of the tentillum is another feature in which close

similarities between Bargmannia species and Pyrostephos vanhoeff-

eni appear. In both the cnidoband is straight, or slightly twisted, but

not tightly coiled, and is without a basal involucrum. In addition,

they both have long terminal filaments. Even more striking is the

presence of large nematocysts, probably stenoteles, only in the

proximal region of the cnidoband of both species. However, those of

Bargmannia spp. are considerably larger than those of P. vanhoeffeni,

which measure c. 40 x 28 mm. Further, the other nematocysts

present on the cnidoband and terminal filament are very similar. Two

types of small nematocysts were found in Bargmannia spp. and

similar ones, measuring 13-17 x 9.5—10.5 ym and 6.5 x 6.5 um,

were found in P. vanhoeffeni. Although the tentillum of Erenna has

a straight cnidoband, and that of Marrus is loosely coiled or straight,

the types and distribution of the nematocysts are quite different. The

cnidoband of Evenna is massively armed with two types of elongate

nematocysts, measuring c. 160 x 37 um and c. 35 x 18 um, while the

terminal process appears to be devoid of any nematocyst. The

cnidoband of Marrus contains heteronemes and haplonemes, meas-

uring c. 55 x 20 um and c. 35 x 7 um of the type often seen in other

agalmatid species. The terminal filament of the latter species con-

tains only small nematocysts, probably desmonemes, acrophores or

anacrophores, measuring c. 16 x 9.5 um and c. 10 x 10 um. These

differences in the nematocyst types alone seem sufficient to indicate

that Bargmannia and Pyrostephos are more closely related to each

other than either is to Marrus or Erenna.

Despite all these similarities between Bargmannia spp. and Pyro-

stephos vanhoeffeni, there are at least two major differences:

Bargmannia spp. are the only physonect siphonophores known to

have siphosomal tentacles, though apolemiid species have nectosomal

ones; they also lack dactylozooids, although the bud-like structures

may be vestigial ones. In addition, P vanhoeffeni is the only species

known to have highly modified dactylozooids, the oleocysts, without

palpacles. The only other species in which dactylozooids are thought

to be absent is Marrus orthocanna (Andersen, 1981). However,

Totton (1965) reported that palpons are present on the gonodendra of

M. antarcticus. Further work needs to be carried out on well-preserved

specimens of theseMarrus species in order to investigate this apparent

difference, and whether each is monoecious or dioecious.

Although there are major differences between Bargmannia spp.

and Pyrostephos vanhoeffeni, there appear to be sufficient similari-

ties to warrant the retention of the genus Bargmannia in the family

Pyrostephidae. The alternative would be to propose a new family for

it, since the genus certainly does not fit neatly into the family

Agalmatidae. This might also apply to the genera Marrus and

Erenna but their species are too little known.

ACKNOWLEDGEMENTS. [am extremely grateful to Drs Richard Harbison

and Edie Widder for inviting me to participate in several cruises involving the

use of submersibles, and for donating the siphonophore material collected to

me. I thank Dr Paul Cornelius for his helpful comments on the manuscript.

P.R. PUGH

REFERENCES

Alvarino, A. 1963. Chaetognatha, Siphonophora, and Medusae in the Gulf of Siam and

the South China Sea. Report on the results of the NAGA Expedition — South East Asia

Research Project. Scripps Institution of Oceanography: 104-108.

1964. Report on the Chaetognatha, Siphonophorae, and Medusae of the

MONSOON Expedition to the Indian Ocean. Report. Scripps Institute of Oceanog-

raphy. SIO Ref. Ser. 64/19: 103—108,209-212.

Andersen, O.G.N. 1981. Redescription ofMarrus orthocanna(Kramp, 1942) (Cnidaria,

Siphonophora). Steenstrupia 7: 293-307.

Carré, C. & Carré, D. 1995. Ordre des Siphonophores. pp. 523-596 in Traité de

Zoologie. Anatomie, Systématique, Biologie. Tome III. Fascicule 2. Cnidaires.

Cténaires. ed. D. Doumenc. Masson, Paris.

Daniel, R. 1974. Siphonophora from the Indian Ocean. Memoirs of the Zoological

Survey of India 15(4): 1-242.

—— 1985. Coelenterata: Hydrozoa Siphonophora. The fauna of India and adjacent

countries, Zoological Survey of India, 440 pp.

Herre, W. 1955. Die Fauna der miozanen Spaltenfiillung von Neudorf a.d.Match

(CSR.), Amphibia (Urodela). Sitzungsberichte der Osterreichische Akademie der

Wissenschaften, Mathematisch-naturwissenschaftliche Klasse. Abteilung I, 164:

783-803.

Kirkpatrick, P.A. & Pugh, P.R. 1984. Siphonophores and Velellids. Synopses of the

British Fauna (New Series) 29: 1-154.

Leloup, E. 1955. Siphonophores. Report on the Scientific Results of the ‘Michael Sars’

North Atlantic Deep-Sea Expedition 1910 5(11): 1-24.

Mackie, G.O. 1964. Analysis of locomotion of a siphonophore colony. Proceedings of

the Royal Society of London, B 159: 366-391.

—— Pugh, P.R. & Purcell, J.E. 1987. Siphonophore biology. Advances in Marine

Biology 24: 97-262.

Mapstone, G.M. 1998. Bargmannia lata, an undescribed species of physonect

siphonophore (Cnidaria, Hydrozoa) from Canadian Pacific waters Zoologische

Verhandelingen 323: 141-147.

Pages, F, Pugh, P.R. & Gili J-M. 1994. Macro- and megaplanktonic cnidarians

collected in the eastern part of the Weddell Gyre during summer 1979. Journal of the

Marine Biological Association of the United Kingdom 74: 873-894.

Pugh, P.R. 1983 Benthic Siphonophores. A review of the Family Rhodaliidae

(Siphonophore, Physonectae). Philosophical Transactions of the Royal Society of

London B 301: 165-300.

—— 1984. The diel migrations and distributions within a mesopelagic community in

the North east Atlantic. 7. Siphonophores. Progress in Oceanography 13: 461-489.

— 1998. A re-description of Frillagalma vityazi Daniel 1966 (Siphonophorae,

Agalmatidae). Scientia Marina 62: 233-245.

& Harbison, G.R. 1986. New observations on a rare physonect siphonophore,

Lychnagalma utricularia (Claus, 1879). Journal of the Marine Biological Associa-

tion of the United Kingdom 66: 695-710.

& Youngbluth, M.J. 1988. A new species of Halistemma (Siphonophora,

Physonectae, Agalmidae) collected by submersible. Journal of the Marine Biologi-

cal Association of the United Kingdom 68: \1—14.

Stepanjants, S.D. 1967. Siphonophores of the seas of the USSR and the north western

part of the Pacific Ocean. Opredeliteli po Faune SSSR 96: 1-216.

Totton, A.K. 1954. Siphonophora of the Indian Ocean together with systematic and

biological notes on related specimens from other oceans. Discovery Reports 27: 1—

162.

— 1965. A Synopsis of the Siphonophora. London: British Museum (Natural

History).

GULDE FOR AUTHORS

Aims and scope. The Bulletin of the British Museum (Natural

History) Zoology, was established specifically to accommodate

manuscripts relevant to the Collections in the Department of Zool-

ogy. It provides an outlet for the publication of taxonomic papers

which, because of their length, prove difficult to publish elsewhere.

Preference is given to original contributions in English whose

contents are based on the Collections, or the description of speci-

mens which are being donated to enhance them. Acceptance of

manuscripts is at the discretion of the Editor, on the understanding

that they have not been submitted or published elsewhere and

become the copyright of the Trustees of the Natural History Mu-

seum. All submissions will be reviewed by at least two referees.

Submission of manuscripts. Initially three clear, complete cop-

ies should be submitted in the style and format of the Bulletin. The

text must be typed double-spaced throughout, including references,

tables and legends to figures, on one side of A4 paper with 2.5 cm

margins. All pages should be numbered consecutively, beginning

with the title page as p. 1. SI units should be used where appropriate.

Whenever possible a copy of the text, once the paper has been

accepted, should also be provided on floppy disc (see below). Discs

should only be sent after final acceptance, as papers generally need

revision after refereeing. If it is impossible to provide an appropriate

disc please ensure that the final typescript is clearly printed.

Authors are requested to ensure that their manuscripts are in final

format, because corrections at proof stage may be charged to the

author. Additions at proof stage will not normally be allowed. Page

proofs only will be sent.

Word-processor discs. _ Please follow these instructions.

1. Ensure that the disc you send contains only the final version of

the paper and is identical to the typescript.

2. Label the disc with the author’s name, title of the paper and the

word-processor programme used. Indicate whether IBM or Apple

Mac (IBM preferred).

3. Supply the file in the word-processor format; if there is a

facility to save in ASCII please submit the file in ASCII as well.

4. Specify any unusual non-keyboard characters on the front

page of the hard copy.

5. Do not right-justify the text.

6. Do not set a left-hand margin.

7. Make sure you distinguish numerals from letters, e.g. zero (0)

from O; one (1) from | (el) and I.

8. Distinguish hyphen, en rule (longer than a hyphen, used

without a space at each end to signify ‘and’ or ‘to’, e.g. the Harrison—

Nelson technique, 91-95%, and increasingly used with a space at

each end parenthetically), and em rule (longer than an en rule, used

with a space at each end parenthetically) by: hyphen, two hyphens

and three hyphens, respectively. Be consistent with rule used paren-

thetically.

9. Use two carriage returns to indicate beginnings of paragraphs.

10. Be consistent with the presentation of each grade of heading

(see Text below).

Title. The title page should be arranged with the full title; name(s)

of author(s) without academic titles; institutional address(es); sug-

gested running title; address for correspondence.

Synopsis. Each paper should have an abstract not exceeding 200

words. This should summarise the main results and conclusions of

the study, together with such other information to make it suitable

for publication in abstracting journals without change. References

must not be included in the abstract.

Text. All papers should have an Introduction, Acknowledge-

ments (where applicable) and References; Materials and Methods

should be included unless inappropriate. Other major headings are

left to the author’s discretion and the requirements of the paper,

subject to the Editors’ approval. Three levels of text headings and

sub-headings should be followed. All should be ranged left and be in

upper and lower case. Supra-generic systematic headings only

should be in capitals; generic and specific names are to be in italics,

underlined. Authorities for species names should be cited only in the

first instance. Footnotes should be avoided if at all possible.

References. References should be listed alphabetically. Authori-

ties for species names should not be included under References,

unless clarification is relevant. The author’s name, in bold and lower

case except for the initial letter, should immediately be followed by

the date after a single space. Where an author is listed more than

once, the second and subsequent entries should be denoted by a long

dash. These entries should be in date order. Joint authorship papers

follow the entries for the first author and an *&’ should be used

instead of ‘and’ to connect joint authors. Journal titles should be

entered in full. Examples: (i) Journals: England, K.W. 1987. Certain

Actinaria (Cnidaria, Anthozoa) from the Red Sea and tropical Indo-

Pacific Ocean. Bulletin of the British Museum (Natural History),

Zoology 53: 206-292. (ii) Books: Jeon, K.W. 1973. The Biology of

Amoeba. 628 p. Academic Press, New York & London. (iii) Articles

from books: Hartman, W.D. 1981. Form and distribution of silica in

sponges. pp. 453-493. In: Simpson, T.L. & Volcani, B.E. (eds)

Silicon and Siliceous Structures in Biological Systems. Springer-

Verlag, New York.

Tables. Each table should be typed on a separate sheet designed to

extend across a single or double column width of a Journal page. It

should have a brief specific title, be self-explanatory and be supple-

mentary to the text. Limited space in the Journal means that only

modest listing of primary data may be accepted. Lengthy material,

such as non-essential locality lists, tables of measurements or details

of mathematical derivations should be deposited in the Biological

Data Collection of the Department of Library Services, The Natural

History Museum, and reference should be made to them in the text.

Illustrations

DRAWINGS — Figures should be designed to go across single (84 mm

wide) or double (174 mm wide) column width of the Journal page,

type area 235 x 174 mm. Drawings should be in black on white stiff

card with a line weight and lettering suitable for the same reduction

throughout, ideally not more than 40%. After reduction the smallest

lettering should be not less than 10 pt (3 mm). Tracing paper should

ideally be avoided because of the possibility of shadows when

scanned. All artwork must have bulletin, author and figure number

included, outside of the image area, and must be free of pencil, glue

or tape marks.

PHOTOGRAPHS — All photographs should be prepared to the final size

of reproduction, mounted upon stiff card and labelled with press-on

lettering (eg Letraset). They can be mounted on white or black

background; a black background must be evenly black all over; any

background must be free of all pencil and glue marks within the

image area. All figures should be numbered consecutively as a single

series. Legends, brief and precise, must indicate scale and explain

symbols and letters. Photos, when components of figure-plates

should be abutted, trimmed as regular rectangles or close trimmed

up to edge of specimen. Joins etc. can be removed at the scanning

stage but at extra cost. Cropping instructions, if any, should be

indicated on an overlay or marked on a photocopy of the figure.

SIZE — Maximum size of artwork for use of flatbed scanners is A3.

Larger artwork has to be reduced photographically prior to scanning,

therefore adding to expense.

Symbols in text. Male and female symbols within the text should

be flagged somehow within curly brackets to enable setter to do a

swift global search.

Reprints. 25 reprints will be provided free of charge per paper.

Orders for additional reprints can be submitted to the publisher on

the form provided with the proofs. Later orders cannot be accepted.