UNIVERSITY

JUL 2 4 1984

• 36

fc/*

FIELDIANA

Botany

Published by Field Museum of Natural History

Volume 36, No. 12 January 29, 1976

A Partial Revision of Paullinia (Sapindaceae)

for

Ecuador, Peru, and Bolivia. Part I.

DONALD R. SIMPSON

SUMMARY

The species of Paullinia known from the Central Andean countries and adjacent

areas are treated, and discussions presented on the taxonomy of some of the more

problematical taxa. Only the taxa of section Paullinia (**Neurotoechus Radlk.) are

treated here, the remaining taxa to be covered in a subsequent publication. The

section is divided into several new infrasectional taxa, two new species and two new

varieties are described, and the status of several taxa are changed. The identification

key is supplemented by two plates illustrating features of capsule, inflorescence, and

stem morphology used in the key to species. Three of the new taxa are illustrated by

a full plate each.

I. INTRODUCTION

Paullinia is a genus of the Sapindaceae having between 150 and

200 species, all restricted to the Western Hemisphere with the

exception of P. pinnata L. which occurs in both the American and

African Tropics. Within the Sapindaceae, Paullinia is most closely

related to the large genus Serjania, the two being separated solely

by characters of the fruit; in Paullinia the maturing pistil becomes

a dehiscent capsule, in Serjania it develops into three samaras.

The present treatment is a revision of the species found in

Ecuador, Peru, and Bolivia, with occasional comments on taxa

found outside that area. In the "Flora of Peru" (Fieldiana: Bot.,

vol. 13, part III-A, no. 2. 1956) Macbride included a total of 54

species of Paullinia that had been collected in Peru or that could

be expected there.

Only the species of Section I, Paullinia (=Sect. Neurotoechus

Radlk.) are treated here; the remainder of the genus will be treated

Library of Congress Catalog Card Number: 75-22521

US ISSN 0015-0746

Publication 1225 125

126 FIELDIANA: BOTANY, VOLUME 36

in part II, to be published later. Of Macbride's 54 species, 11 belong

definitely to section Paullinia. Five species of this section that were

not included by Macbride are now known from Peru, but since

several of the species that Macbride listed are reduced or modified

in this revision, the results are, for Peru, a total of 16 taxa

belonging to only 12 species. For Ecuador, this treatment gives four

taxa belonging to three species; and for Bolivia, six taxa belonging

to five species, with the probable occurrence there of two more

species.

II. NOTES ON THE BIOLOGY OF Paullinia

With the exception of Radlkofer's detailed studies in the

anatomy of the Sapindaceae, little has been published on either the

structure or functional biology of Paullinia. Some of the

morphological features found in this genus suggest adaptation for

specialized biological functions, especially those related to pollina-

tion and dispersal mechanisms.

The following notes include some of the hypotheses I have

formed in the course of this study, to which are added references to

pertinent literature. Any hypothesis about the probable function of

morphological features must, of course, remain conjectural pending

study of living plants in the field. I am aware, too, that their

consideration is not central to this taxonomic study, but it is hoped

that discussions of these matters may encourage others to initiate

studies in the functional biology of these tropical taxa.

There are aspects of the biology of Paullinia other than those

mentioned below that deserve some discussion but their inclusion

must be deferred to part II of this study. Among these are the very

interesting forms of dispersal mechanisms that seem to have

evolved in this genus. Discussion of a possible phylogeny of the

species of Paullinia is also deferred to part II.

1. ANATOMY AND MORPHOLOGY

Because of Radlkofer's special interests in plant anatomy, the

family Sapindaceae is, from the standpoint of anatomy and

morphology, probably one of the more thoroughly studied of the

large tropical families of dicotyledons. In one of his larger

publications he presented a new classification of the Sapindaceae

(Radlkofer, 1890) based, at least in part, on the results of his

anatomical studies. His system of classification in the genus

Paullinia (Radlkofer, 1895) made extensive use of the anatomical

SIMPSON: REVISION OF PAULLINIA 127

features of stem, leaf, and capsule, in addition to the standard use

of floral morphology.

His classification divided the genus into 13 sections that were

distinguished primarily on the basis of anatomical structure, with

special emphasis on the anatomy of the capsule. This system of

classification is a relatively natural one, but the morphological and

anatomical characters on which it is based are somewhat difficult to

recognize. This is partly because some of these characteristics occur

only rarely outside the Sapindaceae, consequently, are unfamiliar to

most taxonomists; and partly because of a lack of illustrations of

critical features. Finally, difficulty may arise because Radlkofer's

descriptions are replete with terminology unfamiliar to taxonomists.

It is apparent that most taxonomists who have published new taxa

of Paullinia since Radlkofer's monograph have generally not

understood his sections nor the anatomical and morphological

features by which he distinguished them. Certain of these features,

especially those that occur in sect. Paullinia, deserve some

discussion.

A distinctive appearance of the dried capsule is the primary

feature on which section Paullinia is recognized. The capsule's

outer surface is covered with a system of ridges or nerves that are

closely packed, parallel to each other, longitudinally oriented but

curving away from the lines of dehiscence and converging on the

midline of each valve (pi. I). All species of the section exhibit this

feature and only in occasional immature fruits is it indistinct.

Another feature of taxonomic significance found in section I is

the presence of composite wood. This is an unusual development of

the vascular system of the stem and branches that apparently

occurs only in a few lianaceous genera. Within this section it is

most markedly developed in certain species of subseries Alatae (pi.

II). The presence and degree of development of this feature is

roughly correlated with the occurrence of other distinguishing

features but most striking is its correlation within subseries Alatae

with ecological conditions. It is most strongly developed in such

species of the lowland tropical rain forest as P. exalata, P.

largifolia, and P. alata ssp. loretana in the western part of the

Amazon Basin. In the Pacific and Caribbean drainages, P. alata ssp.

loretana again shows highly developed composite wood, especially

in the areas of almost continuous high humidity such as the pre-

montane forest of Western Ecuador, and from the Choco of

128 FIELDIANA: BOTANY, VOLUME 36

Colombia through Panama to the Caribbean lowlands of Costa

Rica.

By contrast, P. elegans ssp. elegans of the seasonably dry

forests and savannahs of eastern and southern Brazil, Paraguay,

southeastern Bolivia, and northeastern Argentina has a form of

composite wood in which the peripheral bundles have little

secondary growth, remaining small while the central bundle grows

normally.

In the remaining subseries and series of section Paullinia, the

peripheral bundles are either weakly developed or entirely lacking

(wood simple). All these have species that occur in the wet lowland

rain forest of Amazonia, but, on the basis of other characteristics, I

am inclined to think that these species probably evolved outside

that humid forest area, having entered during either the Pleistocene

or Holocene, while several other taxa seem still to occupy only

areas peripheral to the rain forest.

2. POLLINATION BIOLOGY

The literature of pollination biology contains little mention of

Paullinia, but in the course of this study I have encountered

structural features very suggestive of adaptation for specialized

pollination mechanisms. Possibly the earliest mention of Paullinia

in pollination literature was by Muller (1873). The statement, as

translated by Thompson (1883), was to the effect that the flower of

Serjania cuspidata St. Hil. "resembles the labiate type and like

Paullinia L. is protandrous." In fact, the flowers of S. cuspidata, as

in many species of Paullinia, are functionally unisexual.

The flower is zygomorphic; the filaments are of unequal length

and as a group curve downward against the two adaxial sepals then

recurving upward to bring the anthers in contact with the hooded

scales on the upper (abaxial) petals. In this respect they resemble

the position of the stamens as found in many species of Cassia. The

four petals are asymmetric, grouped toward the top (abaxial side) of

the flower, and each bears a large scale or flap arising on the inner

surface near the base. The scales are often nearly as large as the

petals and are generally of two types; the two upper scales are

markedly crested or hooded and usually more or less bilaterally

symmetrical, the two lateral scales are usually asymmetrical, often

are furnished with long hairs on the inner surface and a fringed

margin, and the apical hood or crest is less well developed or in

some species entirely lacking. The apex of the crests are of thicker,

SIMPSON: REVISION OF PAULLINIA 129

perhaps glandular tissue. I have reasoned that the hood of the scale

may contain accumulated nectar within, derived either from glands

within the crest or from glands in the fleshy lobes of the disc that

project up into the hooded scales. If this assumption is correct then

it is likely that the pollinator enters the flower from above (i.e.,

pollination is sternotribic, the pollen being deposited on the dorsal

surface of the pollinator) forcing a way between the anthers and the

upper scales and perhaps gaining leverage by grasping the hairs and

fringe of the lateral scales. Pollination mechanisms such as this are

well known for many papilionoid legumes. While in staminate

flowers the reduced, non-functional pistil is small and symmetrical,

in functionally pistillate flowers the pistil is larger, curves toward

the two lower sepals, the style recurving upward to bring the

stigmas in close proximity to the upper scales; again the

resemblence to the flowers of many of the Leguminosae is apparent.

In most species the flowers are borne in spicate, racemose, or

sometimes paniculate thyrses which are produced singly in the leaf

axils or at the apex of a shoot. In contrast to this are a few species

in which the inflorescence is always ramiflorous or cauliflorous, i.e.,

produced on older, leafless branches or mainstems. Some examples

are P. alata, P. exalata, and P. hemiptera, all of section Paullinia,

and in other sections of the genus such species as P. trilatera, P.

metensis, and P. tenera.

In these species the inflorescences are produced at the nodes on

usually leafless, older stems. These inflorescences are generally

borne in clusters of two to several at a node, each being a very

condensed, shortened, paniculate thyrse. As a result of the

condensation and clustering, the mature flowers seem to be borne

in a hemispheric arrangement. The combination of cauliflory and

the hemispheric inflorescences are very suggestive of certain bat

pollinated plants. Perch requirements of the bats probably select for

cauliflory and ramiflory on sturdy older branches. It also seems

reasonable to assume that the much condensed inflorescences are

shortened to keep the flowers within reach of the perch area. These

ideas are only assumptions, and it should be remembered, as Faegri

and Van der Pijl (1966, p. 166) have pointed out, that some cases of

cauliflory are related to seed dispersal by bats rather than to

pollinator requirements. Discussion of dispersal biology is deferred

to the next publication on the genus.

In other species of Paullinia, for example, P. scaberula R. E.

Schultes, P. turbacensis, and some undescribed taxa such as

130 FIELDIANA: BOTANY, VOLUME 36

Woodson & Schery 163, from Panama, the cauliflorous

inflorescences are somewhat elongated, not forming a densely

flowered hemispheric cluster. I am still unable to formulate any

ideas about the adaptive significance of this latter form of

cauliflory.

III. GUIDE TO THE TAXONOMIC TREATMENT

1. CITATION OF TAXONOMIC LITERATURE

An exhaustive review of the literature on Paullinia has been

presented by Radlkofer (1896, 1931-1934) and it is not necessary to

repeat it here. In the following text, literature citations will be

limited to: 1) the protolog, 2) the two Radlkofer publications, and

3) for those taxa not treated by Macbride (1956) it is so indicated.

Throughout the Sapindaceae Macbride listed a number

immediately following the date of the protolog, and separated from

it by a semicolon. That number corresponds to the page on which

that species is described by Radlkofer (1931-1934) in the

' Tflanzenreich."

2. CITATION OF GEOGRAPHICAL DATA

In citation of collection localities it has sometimes seemed

useful to add to the terse information from the specimen label,

placing my additions or comments in brackets. This is done, in most

cases, either to pinpoint a collection site or to clarify some element

of confusion. An example of the latter is the citation under P. alata

ssp. alata of Llewellyn Williams 7262 from San Roque, in Dept.

San Martin. One might suppose this to be San Roque de Cumbasa,

located about 12 km. from Tarapoto and accessible by road, but this

is not the case. Rather, it refers to San Roque (on present day maps

usually only "Roque") near the upper Rio Sisa about 50 km. from

Tarapoto. At the time that Mr. Williams collected there in 1930,

the principal route between Tarapoto and Moyobamba, the

departmental capital, passed through the latter-named San Roque.

At present the highway passes to the east of Tabalosas leaving all

the upper Rio Sisa drainage (including Roque) relatively in-

accessible.

Many of the collection sites for tropical South American

specimens are very difficult to locate. This is especially true of

many older collections, such as those of Pearce, Haenke, and some

of Martius' collections, but is true as well for some twentieth

century collectors. This difficulty is due to a number of things,

SIMPSON: REVISION OF PAULLINIA 131

including lack of comprehensive gazetteers; the impermanence of

small settlements and villages (especially in the areas of tropical

rain forest); difficulties in transcribing place names derived from

local dialects; and confusion in orthography (e.g., Sangaban, in

Dept. Puno, Peru has been variously cited as San Gaban, San

Gavan, and Sangavan). To treat these matters properly, they

should be dealt with in a gazetteer-like compilation of collection

localities, but that also is beyond the scope of this study.

3. CITATION OF HERBARIUM SPECIMENS AND PHOTOGRAPHS OF

SPECIMENS

Among the citations of specimens are citations of photographs

from Field Museum's botanical "phototype" collection1. Most of

these are photographs of New World collections found in European

herbaria and were photographed by Macbride or his assistants from

1929 through 1939. Not all are photos of type specimens, but those

not types are often referred to as "authentic" specimens, i.e.,

specimens cited in critical or important works. These are cited by

negative number and by reference to the herbarium where the

specimen was housed when photographed. Thus, "(Photo ex M: F

neg. 6002)" refers to negative no. 6002, a photograph of a Martius

collection (the type of P. trilatera Radlk.) in the herbarium at M

(Botanische Staatssammlung, Munich, Germany). Of special impor-

tance are the photographs of specimens in the Berlin herbarium

(over 5,900 negatives), many accompanied by fragments taken from

the sheet photographed. Positive prints from all the negatives are

mounted on herbarium paper and interfiled in the herbarium at

Field Museum.

Herbarium specimens will be cited as follows: Woytkowski 6630

(F, MO-2, US) to indicate I have examined the one herbarium sheet

of Woytkowski 6630 that exists in the herbarium at Field Museum,

the two sheets at the Missouri Botanical Garden, and the single

sheet in the United States National Herbarium.

The abbreviations for herbaria are those of the Index

Herbariorum (Lanjouw and Stafleu, 1967). Certain collections are

cited by abbreviations as follows:

U.C.B.P. — A series of collections made by Mildred Mathias,

Dermot Taylor, and Jose Schunke Vigo, in a project sponsored by

the Departments of Botany and Pharmacology of the University of

1 Field Museum has several photograph collections, each with a separate series of

negative numbers. Therefore, when requesting prints, please indicate "Botany

Phototype" followed by the negative number.

132 FIELDIANA: BOTANY, VOLUME 36

California at Los Angeles. The collection numbers of this series are

distinct from those of Jose Schunke Vigo's personal collection

number series. Thus, U.C.B.P. (Jose Schunke V.) 5739 is a PaulUnia

collected in 1962, while Jose Schunke V. 5739 is a species of

Marantaceae collected in 1972.

Goodspeed Exped. - Collections made by several collectors

working under the direction of T. H. Goodspeed of the University of

California, Berkeley. Between 1935 and 1947 there were four

expeditions to the Andean countries of South America, involving

several collectors or teams of collectors. The collection numbers

were those of the expedition series and were continuous through the

four expeditions. The collections from these expeditions are often

cited by collector and number with no indication that the number

belongs to the Goodspeed Expeditions series rather than to the

collectors own number series.

Subsequently, there were three more expeditions under the

direction of Paul Hutchinson. These did not follow the number

sequence of the earlier expeditions, using instead, Hutchinson's own

collection numbers beginning with P. C. Hutchinson No. 2 on the

fifth expedition.

The first five Goodspeed Expeditions were reviewed, with the

collectors and itineraries summarized, by Goodspeed and Stork

(1955).

IV. TAXONOMIC TREATMENT

Paullinia L. emend. Radlk. in Engler & Prantl, Naturl.

Pflanzenfam. Ill, 5: 305. 1895. [For synonymy see Radlk.

Monographia Paullinia pp. 71-76, in Abh. Math.-Phys. Cl. Konigl.

Bayer. Akad. Wiss. 19: 67-381. 1896.]

Lianas or scandent shrubs; stems with one to several distinct vascular bundles

and referred to as being simple (when only one) or composite (when more than one).

Leaves compound (rarely a unifoliolate leaf may be produced as an anomaly on an

otherwise 3-foliolate or 5-foliolate plant); mostly pinnate or bipinnate, sometimes

ternate or bitemate, or rarely palmate (in sect. Castanella)', stipules usually present,

persistent or deciduous; petiole and rachis often wing-margined; leaflets membranous

to coriaceous, often pellucid punctate or lineate, often having a network of semi-

pellucid lactiferous canals that are orange-colored by transmitted light, tertiary

venation reticulate or often clathrate, margin usually serrate, serrate-dentate, lobate,

or rarely subentire, the teeth often gland tipped. Inflorescence generally single and

apical or in the leaf axil, sometimes several in a fascicle then usually produced at the

nodes of older, leafless stems (ramiflorous or cauliflorous); basically a thyrse but

usually the branches modified into cincinni which may be borne on a racemose,

5cm

PLATE I. Capsules of sect. Paullinia. a, P. alata ssp. loretana; from Wurdack

1977. b, P. elegans ssp. neglecta; from U.C.B.P. (Schunke) 6200 and Li Williams

6626. c, P. elegans ssp. elegans; from AOf. Job 960 (coll. in Argentina), d, P.

tumbesensis; from Dodson & Thien 1285. e, P. obovata var. obovata; from

Wcryf/jows&t 5375. f, P. obovata var. pofymorpha; from J. Steinbach 7102. g, P.

bracteosa; from U.C.B.P. (J. Schunke V.) 6314. h, P. eriocarpa; from

5634.

133

134 FIELDIANA: BOTANY, VOLUME 36

paniculate, or even a corymbose inflorescence; the cincinnus usually subtended by a

bract, the flowers by bracteoles. Flowers polygamous or often unisexual, zygomorphic

(usually bilaterally symmetric), sepals 5 or sometimes 4 by fusion of the two

lowermost, imbricate, free or fused near the base; petals 4, imbricate, free, each

bearing on the abaxial surface a petaloid, usually hooded and fringed, glandular

scale; scales of the two lateral petals asymmetric, the hood lacking or less developed

than on the upper scales; scales of the two upper petals usually markedly hooded,

bilaterally symmetrical or less often asymmetrical, sometimes provided with one or

two conical or knobby or variously shaped apical projections; disc present, glandular,

inequilateral; stamens 8, length slightly unequal, somewhat connivent and slightly

upturned toward the apex thus forming a keel-like cluster; pistil often short stipitate,

ovary 3-locular, one ovule in each cell, style 1, apically 3-lobed. Fruit a dehiscent

capsule, stipitate or sessile; valves septicidal, dorsally winged or wingless. Seeds with

a usually lustrous black seedcoat and partly enclosed in a fleshy, usually white aril.

Section I. Paullinia. Section Neurotechus Radlk. Monogr. Paull.

108. 1895. Pflanzenreich IV, 165: 223. 1931.

The principal distinguishing features of this section are based on the morphology

of the mature capsule as follows; capsule wall markedly striated (especially in the

dried state) with numerous, parallel, fine ridges and grooves; the ridges formed by

elongate, fibrous bundles in the mesocarp; striations longitudinally oriented but in at

least the upper half of the capsule diverging from the lines of dehiscence (the margins

of the valves) to converge along the mid-line of each valve; capsules stipitate except

in series Eriocarpae.

Although the following characters may not serve to distinguish this section from

all others of the genus, they are often helpful in identification of herbarium

specimens. The leaves ternate or 5-foliolate pinnate; no known species of this section

have more than 5 leaflets nor are decompound. Stipules usually longer than broad,

acute to acuminate (except P. eriocarpa q.v.). Wood simple or composite.

Fasciculate, cauliflorous inflorescences in several species. Sepals all free (in some

species of the other sections the two lower sepals may be fused along their adjoining

margins in varying degrees).

Infrasectional subdivisions: On the following pages the species

are grouped in three series, one of which is divided into three

subseries. The subseries are to some extent intergradient but do, I

believe, represent natural clusters of closely related species. I have

recognized the species clusters at the hierarchical levels of subseries

and series rather than that of subsection and section, as more

correctly reflecting the close natural relationships between the

species clusters of section Paullinia. It may be that series

Eriocarpae should be elevated to subsectional status, but I believe

that the information now available does not yet support such a

change.

SIMPSON: REVISION OF PA ULLINIA 135

KEY TO THE SUBDIVISIONS OF SECTION Paullinia

A. Capsules sessile or rarely subsessile, fusiform to ellipsoidal to ovoid, sparsely

pubescent (in P. leiocarpa) to densely lanate (in P. eriocarpa), valves moderately

thick; wood composite; branchlets sparsely to densely pubescent; leaves 5-foliolate

pinnate. Trinidad and Panama to Brazil and Peru series C. Eriocarpae

AA. Capsules usually stipitate, fusiform to obovate to pyriform, glabrous or

sometimes sparsely pubescent; wood simple or composite; branchlets glabrous to

moderately pubescent (but never densely lanate); leaves unifoliolate to 5-foliolate

pinnate.

B. Capsule valves mostly 5-8 mm. thick; placenta persistent; capsule large, mostly

3-6 cm. long; wood mostly composite; leaves 5-foliolate pinnate. Mexico and the

West Indies to Bolivia and Brazil series B. Obovatae

BB. Capsule valves thin, mostly 2 mm. thick or less; placenta dehiscing with the

valves or sometimes persistent; capsules small or large; wood simple or

composite; leaves 1- to 5-foliolate. Distributed throughout the range of the

genus series A. Paullinae

Consisting of the following three subseries:

C. Capsules narrowly cylindrical or narrowly claviform, often falcately curved;

wood simple to weakly composite; leaves 3-foliolate, rachis not winged.

Mexico and the West Indies south to the Caribbean Coast of South America

and southeast along the coast to Pard State, Brazil subseries 1. Paulliniae

CC. Capsules broadly claviform to obovate to pyriform, usually straight; wood

simple or composite; leaves 1- to 5-foliolate, rachis winged or wingless.

D. Capsules short (mostly 15-20 (30) mm. long); capsule wall thin, somewhat

pliable (often compressed out of shape in dried material) and leathery;

wood mostly composite. Costa Rica south to Brazil, Uruguay and

northern Argentina subseries 2. A la tae

DD. Capsule long (mostly 2.5-3.5 cm. long); capsule wall thin but rigid;

wood simple or composite. Distributed throughout the range of the genus.

subseries 3. Pinnatae

ARTIFICIAL KEY TO SPECIES OF SECTION Paullinia

Illustrations of characters used in the key indicated by asterisk (*).

1. Inflorescences fascicled, condensed, and cauliflorous (pi. I, fig. a).

2. Inflorescences single and axial (pi. I, figs, b, c).

3. Simple wood (pi. II, figs, d, e).

4. Composite wood (pi. II, figs, a, b, c).

5. Clathrate tertiary venation (pi. V).

A. Leaves with winged petiole and rachis (wing sometimes very narrow, e.g., P.

spicata).

B. Inflorescences condensed and cauliflorous* (clustered on the older, mostly

leafless stems), thyrsoid-paniculate usually about as broad as long.

136

FIELDIANA: BOTANY, VOLUME 36

5mm

PLATE II. Stem cross-sections illustrating principle variations found in section

Paullinia. Wood composite in figures a, b, and c; wood simple in figures d and e. a, P.

alata ssp. loretana; from Wurdack 1977 (F). b, P. elegans ssp. neglecta; from Vargas

16076 (US), c, P. eriocarpa; from Woytkowski 5069 (US), d, P. clavigera var. bullata;

from U.C.B.P. (Mathias & Taylor) 5530 (F). e, P. hemiptera; from LI. Williams 6856

(F).

C. Wood simple*; capsules clavate (endemic to the area around Tarapoto, Peru).

7. P. hemiptera

CC. Wood composite*; capsules usually pyriform.

D. Leaves trifoliolate (central and western part of Amazon Basin).

1. P. largifolia

DD. Leaves 5-foliolate (widely distributed; along the Pacific Coast from

Panama to southern Ecuador and in the Amazon Basin from about 1,000 m.

alt. along the east side of the Andes, eastward to Para State, Brazil, and

southward to Sta. Cruz, Bolivia) 2. P. alata

BB. Inflorescences borne singly* in the axils of the leaves or terminal on the

branch, racemose or spicate.

E. Inflorescence bracts broad and rounded or broadly acute at the tip, never

acicular, at least 3 mm. wide, usually wider; wood simple*, stems usually stout.

F. Branchlets, inflorescences, and leaves beneath pubescent; capsules

estipitate, ellipsoidal or fusiform, densely lanate with yellow-golden

pubescence; lower bracts of inflorescence often as broad as long (central and

eastern Colombia south to southeast Peru) 13. P. eriocarpa

FF. Branchlets and leaves glabrous or branchlets sometimes tomentose;

capsules usually stipitate; inflorescence bracts always with length several

SIMPSON: REVISION OF PA ULLINIA 137

times the width (Costa Rica and Panama to Amazonian Bolivia and Brazil).

11. P. bracteosa

EE. Inflorescence bracts acuminate, usually acicular; wood composite* or

simple*.

G. Valves of the capsules (3) 5-8 mm. thick; wood simple* (Amazonia and the

Guianas) 12. P. imberbis

GG. Valves of the capsules ca. 2 mm. thick; wood simple* or composite*.

H. Petiole and rachis narrowly winged, wing 0.5-1.0 mm. wide on each side;

wood inconspicuously composite* (western Ecuador and Dept. Tumbes,

Peru) 8. P. tumbesensis

HH. Petiole and rachis broadly winged, at least 3 mm. on each side; wood

composite* or simple*.

I. Leaves large, mostly (15) 20-40 cm. long; leaflets mostly 12-18 cm. long

by (4) 5-7 cm. wide; petiole wing mostly 5-8 mm. on each side; wood

inconspicuously composite* to simple* (Mexico to Brazil).

6. P. clavigera

II. Leaves small, mostly 11-18 cm. long; leaflets mostly 5-12 cm. long by 2-

3.5 cm. wide; petiole wing mostly 3-5 mm. on each side; wood

inconspicuously composite* (northeastern Argentina, Paraguay and

probably adjacent parts of Brazil and Bolivia).

5. The narrow-leaved form of P. pinnata

AA. Leaves with petiole and rachis wingless.

J. Inflorescences clustered on older, mostly leafless stems (cauliflorous)*, thyrsoid-

paniculate, condensed and usually about as broad as long; wood complex*

(central and western parts of the Amazon Basin) 3. P. exalata

JJ. Inflorescences single*, axillary or terminal on mostly leafy branches, spicate or

racemose; wood composite* or simple*.

K. Leaflet with tertiary venation clathrate*; capsule valves mostly 5-8 mm.

thick; wood simple* (Peru and Bolivia) 10. P. obovata

KK. Leaflet with tertiary venation clathrate* or reticulate; capsule valves

mostly less than 2.5 mm. thick; wood simple* or composite*.

L. Leaflet with tertiary venation reticulate; wood composite* 4. P. elegans

LL. Leaflet with tertiary venation clathrate*; wood simple* 9. P. spicata

Series A. Paulliniae.

Capsules stipitate, mostly pyriform (seed-bearing part sometimes subglobose) or

in a few species narrowly clavate or cylindrical and sometimes falcate-curved (e.g., P.

cururu)', capsule wall often thin, usually 2 mm. or less in thickness, in some species

showing the effects of compression in dried, pressed specimens (e.g., P. alata and P.

elegans). Leaves trifoliolate or 5-foliolate, pinnate; tertiary venation of the leaflets

138 FIELDIANA: BOTANY, VOLUME 36

generally reticulate, never markedly clathrate (?); petiole and rachis winged or

wingless.

Subseries 1. Paulliniae.

Capsules narrowly cylindrical or narrowly claviform, usually falcate-curved,

capsule wall somewhat rigid.

This subseries is erected to contain P. cururu L., the type

species of the genus (judging by a photograph of the type of P.

nitida H.B.K. it may also belong here). It is distributed from

Mexico and the West Indies, to the Caribbean and Atlantic Coasts

of Columbia, Venezuela, the Guianas and Brazil (Territory Amapa

and Para State), but does not occur in Ecuador, Peru, or Bolivia

nor in areas adjacent to their borders.

Subseries 2. Alatae D. Simp, subser. nov.

Capsulae pyriformes; partibus semeniferentibus plerumque subglobosis et

aliquant um radialiter trilobatis; valvis illis subseriebus 1 et 3 menus rigidis.

Capsules pyriform; the seed-bearing part often subglobose and radially 3-lobed;

the wall less rigid than in subseries 1 and 3, such that in dried material the wall often

seems to have been squashed in the drying press. Wood usually composite.

This subseries should probably include, in addition to the

species given below, P. densiflora J. E. Smith, P. fasciculata Radlk.,

and P. macrophylla H.B.K.

Type species: P. alata (R. & P.) G. Don.

1. Paullinia largifolia Radlk. Bot. Jahrb. 37: 149. 1905.

Pflanzenreich IV, 165: 241, no. 2. 1931. Not included in "Flora of

Peru."

Liana; the stem with composite wood, 3-angled in cross-section. Leaves

trifoliolate; petiole winged, 8-11 cm. long, maximum width (including wings) 6-12

mm.; terminal leaflet 15-24 cm. long by 10-15 cm. wide, broadly ovate, obtuse to

truncate at base; lateral leaflets elliptic, broadly acute at base, in the type 18 cm.

long by 10.5 cm. wide (fide Radlk.), in the Peruvian collection cited below, as small as

8 cm. long by 5.2 cm. wide; all leaflets subentire, obsolete callose dentate,

chartaceous to subcoriaceous, glabrous above except pubescent on veins, uniformly

short pilose beneath; petiolules in type ca. 8 mm. long, in Peruvian collection ca. 2

mm. long (i.e., leaflets subsessile). Inflorescences condensed, clustered on older,

leafless stems. Fruit unknown.

Type: Ule 5816 (B, destroyed?).

PERU: [Dept. Loreto: Prov. Alto Amazonas;] edge of forest,

Fortaleza, [near] Yurimaguas, LI. Williams 4317 (F, US).

BRAZIL: Amazonas: [village of] Belem [on the Rio] Jurua

Miry, Ule 5816 (photo ex B: F neg. 5606).

SIMPSON: REVISION OF PA ULLINIA 139

Easily distinguished by trifoliolate leaves with winged petiole,

cauliflorous, condensed inflorescences and striate, wingless capsules.

The Peruvian collection closely resembles the type, differing

only in the leaves being smaller in all dimensions except width of

winged petiole. Radlkofer describes the leaflet texture as charta-

ceous but from the photograph of the Berlin type it is certainly as

thick as that of the Peruvian specimen which I prefer to describe as

subcoriaceous.

2. Paullinia alata (R. & P.) G. Don, Gen. Syst. 1: 660. 1831. Radlk.,

Monogr. Paull. 125, no. 3. 1895. Pflanzenreich IV, 165: 242, no. 4.

1931. Simarillaria alata Ruiz et Pavon, Flora Peruv. et Chil. IV: t.

340. 1802 [the unpublished text of vol. IV is reproduced in Anal.

Inst. Hot. Cavanilles vol. 12, 13, 14, 15; for S. alata see vol. 12: 158-

159. 1953]. P. rhizantha Poepp. & Endl. Nov. Gen. & Sp. 3: 36, pi.

243. 1844. Radlk. Monogr. Paull. 127, no. 4. 1895. Pflanzenreich IV,

165: 243, no. 5. 1931.

Stem wood usually markedly composite and 3-angled. Leaves pinnate with 5

leaflets; petiole and rachis winged. Inflorescences short, condensed, fascicled on the

older stems. Fruit wall markedly striate, usually only 1-2 mm. thick in dried

specimens.

In the Pflanzenreich, Radlkofer notes that P. rhizantha may

not be distinct from P. alata. An isotype specimen of P. rhizantha

(Poeppig 2239) at F resembles the isotype of P. alata (Ruiz &

Pavon s.n.) at F more than any other specimen in the herbarium!

The separation of the two by Radlkofer on the basis of branchlets

3-angled versus 6-angled is, I believe, of little taxonomic significance

in this complex and variable species.

The form of cauliflory in this species as also in P. largifolia and

P. exalata, suggests an adaptation to either pollination or seed

dispersal by bats.

The specimens at hand do exhibit morphological variations that

roughly correlate with differences in geographical distribution.

Subspecies loretana has large, papery leaflets and is found mostly

in the Amazon Basin lowlands. Subspecies alata has smaller

leaflets that are subcoriaceous to coriaceous and occurs principally

in "ceja de la montana" forests on the Eastern Cordillera and, in

Ecuador, on the Western Cordillera and in the lowland forests

between that and the Pacific Coast. Material from the humid

lowland forests of Costa Rica, Panama, and the Choc6 of Colombia

more nearly fit subsp. loretana. In Peru, the two subspecies are

140 FIELDIANA: BOTANY, VOLUME 36

allopatric in the middle valley of the Huallaga River and near the

foothills at the western edge of the Basin (as, for example, at

Yurimaguas), but this zone of allopatry probably does not extend

above 800 or 900 m. altitude nor eastward more than 50 km. from

the foothills of the Cordillera.

There is considerable variation in presence and amount of

pubescence such that I am unable to see any justification for

retention of Macbride's variety pubens. Although this variation

occurs in both subspecies, the type of var. pubens can be assigned to

subspecies loretana. To demonstrate the variation in this feature,

the pubescent specimens in each subspecies are listed separately.

2a. subspecies alata

Leaves, leaf-bearing branchlets and mature, flowering stems smaller in most

dimensions than ssp. loretana. Leaves 12.5-22 cm. long; lateral leaflets (4.5) 6.5-8.5

(12) cm. long by (1.8) 2.3-3.6 (5.6) cm. wide; terminal leaflets (5) 6.5-10 (15) cm. long

by (1.7) 2.6-4.4 (7) cm. wide; petioles 3.0-4.7 (10) cm. long by 3-5 (9) mm. wide

(including wings); usually subcoriaceous.

Type: Tafalla s.n., from Chicoplaya, Peru.

PERU: [Dept. Loreto: Prov. Alto Amazonas;] Yurimaguas,

Poeppig 2239 (isotype F, of P. rhizantha P. & E.); Prov. Maynas;

Mishuyacu, near Iquitos, alt. 100 m., Klug 1034 (F, US). - Dept.

San Martin: [Prov. Lamas: Dist. Roque;] San Roque [on tributary

of Rio Sisa; not San Roque de Cumbasa], LI. Williams 7262 (F);

[Prov. San Martin:] Tarapoto, LI. Williams 6092 (F). - Dept.

Huanuco: [Prov. Huamalies; Dist. Monzon;] Chicoplaya,1 Tafalla

s.n. (isotype F; photo, ex B: F neg. 5589); Cachicoto, alt. 800 m.,

Woytkowski 7888 (US); [Prov. Huanuco; Dist. Chinchao;] Pampa-

yacu, 7,000 ft., Kanehira 46 (F). — Dept. Cuzco: Prov. Convencion;

Echarate, semi-tropical forest in deep woods, Goodspeed Exped.

(Stork, Norton, & Vargas) 10454 (F).

Pubescent phase: Dept. San Martin: [Prov. Moyobamba;]

Zepelacio [now Jepelacio], alt. 1,100-1,200 m., Klug 3263 (F, MO,

US). — Dept. Huanuco: [Prov. Pachitea;] Muna, in Andean forest,

[alt. ca. 2,300-2,400 m.] Woytkowski 5213 (F).

ECUADOR: [Prov. Bolivar:] Western Cordillera, lower valley

of [Rio] Chimbo, 1,000 m., Rimbach 196 (F-2).

'Type locality given in the text for volume IV of Ruiz & Pavon's, Flora Peruv. et

Chil. (see Ann. Inst. Bot. A. J. Cavanilles 12: 159. 1953.) as follows: "Habitat in

Peruviae Andium nemoribus Chicoplaya, ubi Joannes Tafalla hanc speciem delineavit

et descripsit." For the location of Chicoplaya, see Hodge, W. H., in Bot. Mus. Leafl.

14 (6): 143. 1950.

SIMPSON: REVISION OF PAULLINIA 141

BOLIVIA: [Dept. Beni; Prov. Ballivian;] Rurrenabaque, alt.

1,000 ft., Mulford Exped. (M. Cardenas} 1257 (US).

I have seen no material of the other collections cited by

Radlkofer under P. rhizantha but it should be noted that the R. S.

Williams collections he cites are all from Bolivia and not Colombia

as given in the "Pflanzenreich" (see below under subsp. loretana).

2b. subspecies loretana (Macbr.) D. Simpson stat. nov. P. alata (R.

& P.) G. Don var. loretana Macbr. Field Museum Nat. Hist., Bot.

Ser. 13, IIIA (2): 331. 1956. P. alata (R. & P.) G. Don var. pubens

Macbr. ibid. Plate la; plate 7/a.

Leaves larger in most dimensions (except petiole width) than subsp. alata, and

the texture usually chartaceous in contrast to a more coriaceous texture in ssp. alata.

Leaves averaging 25-35 cm. long; petioles usually 9-12 (15) cm. long; terminal leaflet

generally (10) 12-16 (20) cm. long by (4.5) 6.5-7 (9.5) cm. wide.

Type: LI. Williams 2339 from Caballo Cocha, Peru.

PERU: Dept. Loreto: [Prov. Maynas; Dist. Ramon Castilla;]

forest, Caballo Cocha on the Amazon River, LI. Williams 2339

(holotype F, isotype US); [Dist. Putumayo;] Florida, Rio Putumayo

at mouth of Rio Zubineta, alt. ca. 200 m., Klug 1993 (F, US); [Dist.

Iquitos;] Iquitos, alt. ca. 100 m., Killip & Smith 27421 (F, US); Villa

Olgita, Rio Itaya near Iquitos, U.C.B.P. (Mathias & Taylor} 3909

(F); [Prov. Alto Amazonas; Dist. Balsapuerto;] Balsapuerto, alt. ca.

220 m., Klug 2879 (F, US); [Prov. Coronel Portillo; Dist. Calleria;]

Pucallpa, on shores of the lake, Pau-Cocha, alt. 200 m., Woytkowski

6299 (US, MO); [Dist. Padre Abad;] Aguaytia, in forest, alt. 300 m.,

Woytkowski 5364 (F). — Dept. Amazonas: Prov. Bagua; [Dist.

Aramango;] left bank of Rio Maranon above Cascades de Mayasi

(opposite km. 278 of Maranon road), elev. 425-500 m., Wurdack

1977 (F-2, US). - Dept. San Martin: [Prov. Mariscal Caceres; Dist.

Uchiza;] Quebrada Huicte (Rio Huallaga), 529 m. alt., in tall forest,

U.C.B.P. (Jose Schunke V.) 6444 (F). - Dept. Huanuco: [Prov.

Leoncio Prado; Dist. Rupa Rupa;] in secondary forest, alt. 700 m.,

Tingo Maria, Woytkowski 5296 (F).

Pubescent phase: Dept. San Martin: Prov. Mariscal Caceres;

Dist. Tocache Nuevo; en bosque alto, desembocadura del Rfo

Mishollo, Jose Schunke V. 4708 (F). - Dept. Huanuco: Prov.

Huamalies [should be Prov. Leoncio Prado; Dist. Rupa Rupa;]

between Supte and Tulumayo Rivers, north of Tingo Maria, alt.

610 m., Goodspeed Exped. (Stork & Norton) 9564 (type of var.

pubens Macbr.; holotype F); [Dist. Rupa Rupa;] in forest, alt. 690

142 FIELDIANA: BOTANY, VOLUME 36

m., Tingo Maria, Woytkowski 5321 (F), 5325 (F, MO); Prov.

Pachitea; Dist. Honoria; (Bosque Nac. de Iparia: cerca del

campamento Miel de Abeja, 1 km. arriba de Tournavista) en monte

alto, alt. 300-400 m., Jose Schunke V. 1148 (F).

ECUADOR: Prov. Chimborazo: juncture of Provinces of

Guayas, Canar, Chimborazo, and Bolivar, foothills of the western

cordillera near the village of Bucay [= General Elizalde], 1,000-

1,250 ft. elev., W. H. Camp 3650 (F, US). - Prov. Guayas: in forest

near Pedro Carbo, alt. ca. 100 m., Haught 3121 (F, US); in forest

near Balao, Eggers 14384 (US). — Prov. Santiago-Zamora: primary

rain forest, Taisha, alt. 1,500 ft., Cazalet & Pennington 7773 (US).

BOLIVIA: Dept. La Paz (?): (Prov. Iturralde?;) Maquiqui, R.

S. Williams 689 (US). - Dept. Santa Cruz: Prov. Sara; bosque del

Rio Palometillas, 400 m., J. Steinbach 7608a (F, MO).

3. Paullinia exalata Radlk. Bot. Jahrb. 37(1): 150. 1905.

Pflanzenreich IV, 165: 244, no. 6. 1931.

Wood composite. Leaves large; petiole not winged, 7-22 cm. long; leaflets 5, (8)

16-22 cm. long by (3.2) 7.5-11 cm. wide, elliptic to obovate, cuspidate with a broad

acumen, entire or with a few remote, obtuse or rounded-sinuate teeth; petiolules

about 1 cm. long in the specimens and photograph cited below.

Syntypes. — Ule 5817a (B) from Brazil, and Weberbauer 1910

(B) from Peru.

PERU: Dept. Loreto: [Prov. Alto Amazonas; Dist. Yuri-

maguas;] up river from Yurimaguas, U.C.B.P. (Mathias & Taylor]

3919 (F-2); Prov. Coronel Portillo; Dist. Yarinacocha; Yarinacocha

[near Pucallpa], Sagastegui & Aldave 5665 (US); vicinity of

Pucallpa, southwest bank of Yarinacocha, U.C.B.P. (Mathias &

Taylor) 6023 (F); Dist. Iparia; Bosque Nac. de Iparia, a lo largo del

Rio Ucayali cerca del pueblo de Iparia, Jose Schunke V. 2636 (F).

-Dept. Junin: Prov. Tarma; [Dist. Chanchamayo;] La Merced in

the Chanchamayo Valley, Weberbauer 1910 (photo ex B: F neg.

5599).

Paullinia elegans Camb. in St. Hil. Fl. Bras. 1: 370, no. 1.

1825. Radlk. Monogr. Paull. 169 no. 9. 1895. Pflanzenreich IV, 165:

255, no. 11. 1931. Not included in "Flora of Peru."

This appears to be a wide-ranging, polymorphous assemblage

sparingly found in our area but more common toward the Atlantic

and Caribbean coasts of South America. On the basis of the

material I have examined there seem to be several regional forms

that warrant taxonomic recognition. The most distinctive of these is

SIMPSON: REVISION OF PAULLINIA 143

found in Paraguay, adjacent parts of Argentina and Brazil (Rio

Grande do Sul State). It is characterized by relatively small leaflets

having reticulate tertiary venation, and broadly marginal teeth.

Collections from Bahia State (Blanchet 1838, 1831, s.n.; Glocker 33:

Bondar 1665: all at F) are very near the Paraguayan material,

differing if at all, in that the distal leaflets are sessile and often the

leaves and leaflets somewhat larger. One herbarium sheet at F (no.

935984) contains fragments of a St. Hilaire collection that seem to

belong to this southern element. These fragments are from material

at P, and probably the type, and if so that element must be

considered typical when assigning epithets to the infraspecific taxa.

It should be noted that although Radlkofer included this

species among those of section Neurotoechus having composite

wood, this characteristic is not so marked as in species such as P.

alata (R. & P.) G. Don and P. exalata Radlk. In P. elegans there

seems to be, at least in some plants, a growth pattern in which the

central bundle develops normally while the peripheral bundles

remain small though separate.

4a. subspecies elegans. Plate Ic.

Syntypes: St. Hilaire s.n. from Rio Grande do Sul, and St.

Hilaire s.n. from Minas Gerais, Brazil.

Distribution (as represented by collections in herbarium F). —

Argentina (Provinces or Territories of: Formosa, Chaco, Santa Fe",

Missiones, and Corrientes), Paraguay, Brazil (States of Rio Grande

do Sul, Bahia; Radlk. cites collections from several other states as

well).

4b. subspecies neglecta (Radlk.) D. Simp. stat. nov. P. neglecta

Radlk. Monogr. Serjania 42, 71, no. 29. 1875. Monogr. Paull. 167, no.

8. 1895. Pflanzenreich IV, 165: 254, no. 10. 1931. Semarillaria nitida

Ruiz & Pavon, Fl. Peruv. et Chil. IV: t. 339. 1802.1 Plate Ib; plate

lib.

1 For the unpublished text intended to accompany the plates of volume IV of the

Flora Peruviensis et Chilensis, see Anal. Inst. Bot. Cavanilles 12: 157. 1954. Although

volume IV was published without text, the illustration is adequate to meet the

requirements of the Code (article 44) as an "illustration with analysis showing

essential characters." Thus Radlkofer's Paullinia neglecta, even though not

accompanied by a description, is not a nomen nudum since reference to the Ruiz &

Pavon illustration fulfills the requirements for validating it as a nomen novum

replacing Semarillaria nitida R. & P. (the epithet nitida is preempted in Paullinia by

P. nitida Kunth).

144 FIELDIANA: BOTANY, VOLUME 36

Wood composite. Leaves trifoliolate or an occasional leaf pinnately 5-foliolate;

petiole not winged; leaflets elliptic or rotund to obovate, subentire or usually with a

few blunt or rounded teeth on the distal half, venation reticulate but occasionally the

tertiary veins subclathrate. Inflorescence a spicate or racemose thyrse, the cincinni

subsessile. Fruit subglobose, stipitate; the valves sometimes slightly more thick-walled

than in subsp. elegans.

The occasional presence of pinnately 5-foliolate leaves that

Radlkofer mentioned is especially well illustrated in Schunke-V.

3877 where a few 5-foliolate leaves are interspersed among the

predominently 3-foliolate leaves! Since even on plants with 5-

foliolate leaves, 3-foliolate leaves predominate, this seems to be a

"good" (i.e., constant) and, therefore, useful feature for recognizing

this taxon.

Type: Tafalla s.n. from Vitoc, Peru.

PERU: Dept. Loreto: [Prov. Coronel Portillo; Dist. Calleria]

on the shore of the lake, Pau Cocha, Pucallpa, alt. 200 m.,

Woytkowski 6320 (MO). — Dept. San Martin: [Prov. Lamas;]

Rumizapa, near Tarapoto, LI. Williams 6752 (F); Prov. San Martin;

Tarapoto, LI. Williams 6626 (F); in forest, Juan Guerra, near

Tarapoto, LI. Williams 6851 (F), 6911 (F); Prov. Mariscal Caceres;

Dist. Tocache; Fundo Miramar 5 km. abajo de Tocache Nuevo,

Jose Schunke V. 3877 (F). - [Dept. Pasco (Dept. Junin on label):

Prov. Oxapampa;] low andean woodland, Yunguy, alt. 1,600 m.,

Woytkowski 6573 (MO), 6575 (F, MO, US, WIS). - Dept. Junin:

[Prov. Tarma; Dist. Chanchamayo;] along Rio Perene [now

considered part of Rio Chanchamayo], near "Hacienda 3," Colonia

Perene, alt. about 600 m., forest [at approximately 75° 15' W, 10°

55' S], Killip & Smith 25205 (US); cerca al Puente Herreria [bridge

over the Chanchamayo River near the village of Francia between

San Ramon and La Merced], alt. 809 m., U.C.B.P. (J. Schunke V.)

6200 (F-2). [Dist. Vitoc;] ad Vitoc arcem, Tafalla s.n.1 (photo ex B:

F neg. 5612). — Dept. Ayacucho: Prov. La Mar; Ayna, between

Huanta and Rio Apurimac, alt. 750-1,000 m., Killip & Smith 22743

(US). — Dept. Puno: Prov. Carabaya; [Dist. Sangaban;] puente

Inambari [near mouth of Rio Sangaban?], alt. 670 m., C. Vargas C.

16076 (US).

BOLIVIA: [Dept. La Paz: Prov. Larecaja;] Guanai, 2,000 ft.,

Rusby 530 (F, US), 626 (F). - [Dept. Beni: Prov. Ballivianes;] San

Pedro Hacienda, [near] Reyes, alt. 1,000 ft., Rusby 1330 (F, MO). -

Dept. Sta. Cruz: Prov. Cercado [= Prov. Ibanez?]; campos del

'The type has been cited as Pavon s.n., or Ruiz & Pavon s.n. but was actually a

Tafalla collection (see footnote 1, page 154).

SIMPSON: REVISION OF PAULLINIA 145

Cuchi, 450 m., J. Steinbach 7462 (F); Prov. Ibanez; Sta. Cruz de la

Sierra, Ismael Peredo (20/XI/1946) (F); Prov. Sara; bosques de

Buena Vista, 450 m. alt., J. Steinbach 6913 (F, MO) [the specimen

at MO has leaves trifoliolate only, that at F has both 3- and 5-

foliolate leaves in about equal number, all on the same branch].

I had earlier thought that P. tarapotensis Radlk., belonged in

subseries Alatae. Although I have seen no type material of P.

tarapotensis, a photograph of the holotype (bearing flowers but no

capsules) is available and in general appearance resembles P.

elegans var. neglecta, differing only in leaflet number. Later,

certain specimens were found that fit Radlkofer's description of P.

tarapotensis but have the capsules of subseries Pinnatae plus other

features that separate it from P. elegans. This discovery occurred

too late to revise the manuscript, instead, P. tarapotensis is now

deleted but will be treated in Part II of this revision.

Subseries 3. Pinnatae D. Simp, subser. nov.

Capsulae stipitatae, 2.5-3.5 mm. longae (stipite 3-12 mm. longo incluso), clavatae

vel plerumque pyriformes, interdum prominentis tribus, carinatis vel leviter cuculatis,

prope apicem instructis; valvis ca. 2 mm. incrassatis, ad maturitatem rigescentibus vel

rigidis.

Capsules stipitate, sometimes markedly so, clavate (especially in P. pinnata) or

more commonly pyriform (with the part above the stipe usually ellipsoidal),

sometimes with 3 carinate or slightly cuculate projections near the apex, mostly 2.5-

3.5 cm. long including the 3-12 mm. long stipe; the valves about 2 mm. thick, rigid at

maturity.

The species of this subsection found in the regions covered in

this work are P. clavigera Schlect., P. hemiptera D. Simp., P.

pinnata L., P. spicata Benth., and P. tumbesensis D. Simp. Other

species that probably belong to this subseries include P. sessiliflora

Radlk. and perhaps P. fibrigera Radlk., although I have seen no

material of this latter species (the several specimens so labeled in

hb. F belong rather to some species in Radlkofer's section III).

Type species: P. pinnata L.

5. Paullinia pinnata L. Sp. PL (ed. 1) 366, no. 7 pro partim. 1753.

Radlk. Monogr. Paull. 135, no. 6. 1895. Pflanzenreich IV, 165: 247,

no. 8. 1931.

Wood composite, sometimes the peripheral bundles remaining small and

indistinct, sometimes becoming simple by loss or sloughing of peripheral bundles.

Leaves pinnately 5-foliolate; leaflets remotely rounded-serrate; petiole and rachis

winged; stipule acicular, 4-7 mm. long, caducous. Inflorescence spicate, axial or

terminal; cincinni sessile or elongating in fruit to 5-8 mm. long, subtended by an

acicular bract usually not exceeding the flower buds. Capsule 25-30(35) mm. long, the

valve ca 2 mm. thick.

146 FIELDIANA: BOTANY, VOLUME 36

Said to be in Peru based on a Pavon collection in hb. Boiss. (G)

[fide Radlk.]. We have no duplicate of this collection at F, nor am I

able to find a photo of it in the "phototype" collection1. I have not

yet seen Peruvian material that definitely belongs to this species.

Nor have I seen any Ecuadorean or Bolivian collections of this

species, but it may possibly be found in the eastern and southern

parts of Santa Cruz Department of Bolivia.

This and the following species may be conspecific but the

matter is too complicated to attempt a solution here. There seem to

be three main forms of P. pinnata in the Western Hemisphere; one

with narrow elongate leaflets and narrowly claviform capsules

found in the Chaco forests of northern Argentina, Paraguay, and

adjacent parts of Brazil; one with leaflets broadly ovate to elliptic,

seldom more than 8 cm. long, and capsules pyriform or broadly

clavate, found from northeastern Brazil through northern Goias

State, Para State, the Guianas, and northeastern Venezuela; and

the third, a more heterogeneous element with larger leaflets that is

found from Mexico to Panama and in the West Indies. This third

form seems to intergrade with P. clavigera in certain parts of

Mexico and Central America. The narrow-leafleted "Chaco" form

mentioned above probably occurs in southeastern Bolivia and on

that basis the species is included in this treatment.

6. Paullinia clavigera Schlect. Linnaea 10: 239, no. 306. 1836.

Radlk. Monogr. Paull. 175, no. 16. 1895. Pflanzenreich IV, 165: 261,

no. 20. 1931. Not included in "Flora of Peru."

This species, although primarily from Central America and

Mexico, was said by Radlkofer to include a collection from

Amazonian Brazil; Ule 5954 from Fortaleza, on the lower Rio

Jurua. I have seen no material of the Ule collection, nor any

photograph of it. The abundant material from Mexico and Central

America in this herbarium (F) is obviously conspecific with several

Peruvian collections previously misidentified and filed under other

names. However, these collections do seem to represent two more or

less distinct elements which are here recognized as two varieties of

this species.

6a. var. clavigera.

Lianas; wood usually composite but the peripheral bundles remaining very small

relative to the central bundle, or occasionally simple. Leaves (in South American

material) very large, (16) 21-40 cm. long; petiole and rachis winged, petiole (4) 6-14

'See page 131 above for explanation of "phototype" collection.

SIMPSON: REVISION OF PAULLINIA 147

cm. long; leaflets not pellucid, mostly from 12 cm. long by 4 cm. wide to 18 cm. by 7

cm., subentire to remotely serrate-dentate, apically acuminate or acutely subcuspi-

date; stipules elongate, broadly ligulate or lanceolate, striate, 10-15 (25) mm. long.

Inflorescence borne singly in the leaf axil; the axis becoming robust at maturity;

cincinni sessile, subtended by acicular bracts about 4-6 mm. long; flowers pedicellate.

Capsules short stipitate, obovate or turbinate, 22-35 mm. long including the 3-5 mm.

stipe.

Type: Schiede 306 from Mexico (photo ex B: F neg. no. 5592).

PERU: Dept. Loreto: [Prov. Coronel Portillo;] Neshuya, 250

m. alt., U.C.B.P. (Jose Schunke V.) 6647 (F-2); Jose Schunke V. 912

(F).

BOLIVIA: [Dept. la Paz: Prov. Iturralde;] Tumupasa, R. S.

Williams 466 (US-2).

6b. var. bullata D. Simp. var. nov. Plate Hd.

A varietate typica differt foliis rugosis, stipulis aliquantum grandioribus, capsulis

pyriformibus, et stipitibus capsularum 6-12 mm. longis.

Like the typical variety except as follows: leaves rugose; stipules somewhat

larger; capsules pyriform; stipe 6-12 mm. long.

Type: Woytkowski 7149 from Dept. San Martin, Peru.

PERU: Dept. Loreto: [Prov. Maynas;] river bank, San Antonio,

Rio Momon near Iquitos, U.C.B.P. (Mathias & Taylor} 3880 (F,

US); [Prov. Requena;] vicinity of Requena, U.C.B.P. (Mathias &

Taylor) 5530 (F-2); [Prov. Coronel Portillo;] southwest bank of

Yarina Cocha, vicinity of Pucallpa, U.C.B.P. (Mathias & Taylor)

6020 (F). — Dept. San Martin: [Prov. Mariscal Caceres;] outskirts

of forest, Huinguilla [on Rio Huallaga between Campanilla and

Juan Jui], alt. 500 m. Woytkowski 7149 (holotype MO, hb. sheet no.

1,793,213; isotype US, hb. sheet no. 2,453,494). - Dept. Jum'n:

[Prov. Tarma; Dist. Chanchamayo;] in forest, Huatsiroke, alt. 1,800

m. [Rio Huatziroqui is a tributary entering the Rio Chanchamayo

(or Rio Perene") on the right bank, about 10 to 15 km. below the

confluence of the Rio Paucartambo], Woytkowski 5573 (F, MO-2).

7. Paullinia hemiptera D. Simp. sp. nov. Plate He; plate HI.

Frutex scandens; ligno simplici; ramis juvenilibus pubescentibus et leviter 3- ad

6-angularibus, deinde glabrescentibus et teretibus. Folium 5-foliolis, pinnatim

compositum; petiolo et rache alata; foliolis sessilibus, ellipticis vel oblongis vel

obovato-ellipticis, subintegeris vel in dimidio distali remote serrato-dentatis; foliolis,

rache, et petiolo tomento molli, velvetino obsiti; stipulis acicularibus, 4-8 mm. longis,

ad basirn 0.5-1.5 mm. latis. Inflorescentiae condensatae, ad nodos caulium veterum

effoliferorum fasciculatae. Capsulae clavatae, pubescentes, ad apicem per cristam

supra quoque valvam subalatae.

PLATE III. Paullinia hemiptera; from holotype, LI Williams 6517.

148

SIMPSON: REVISION OF PAULLINIA 149

Liana; wood simple; young stems weakly 3- to 6-angled, pubescent; older stems

mostly glabrous and terete. Leaves pinnately 5-foliolate; petiole and rachis winged;

petiole 3.5-8.0 cm. long; leaflets elliptic or oblong or obovate-elliptic, sessile, (4) 6-7.5

(10) cm. long by (2.0) 2.7-3.1 (4.7) cm. wide, subentire to remotely serrate-dentate in

the distal half; the leaflets, rachis and petiole pubescent beneath with soft, velvety

tomentum; stipules acicular, 4-8 mm. long, 0.5-1.5 mm. wide at base. Inflorescence of

2 - several condensed thyrses clustered at the nodes of mature, leafless stems. Pedicels

4-5 mm. long, subtended by a minute bract, jointed at midlength. Ovary densely

appressed pubescent, strongly 3-angled. Capsules clavate, appressed pubescent; the

hairs oriented toward the capsule apex; slightly winged in the apical part; the largest

capsules to 22 mm. long by about 8 mm. wide (may develop to larger size; none of

those seen had dehisced).

Type: LI Williams 6517 from Tarapoto, Peru.

PERU: Dept. San Martin: [Prov. San Martin;] Tarapoto, alt.

360-900 m., LI. Williams 6517 (holotype F, hb. no. 626,681; isotype

US, hb. no. 1,779,856); Rio Mayo near Tarapoto, LI. Williams 6273

(F); Juan Guerra near Tarapoto, LI. Williams 6856 (F).

This taxon is probably closely related to P. pinnata L., rather

than to P. alata (R. & P.) G. Don which it resembles in several

respects. The crest-like ridges near the apex of the capsule occur

occasionally in several species of section Paullinia, but seldom as

strikingly as in this species.

In choosing the epithet for this species I have followed

Radlkofer's practice of using the Latin adjective alata when

referring to winged petioles (e.g., P. exalata Radlk.), while retaining

the Greek pteron to indicate capsule wings (e.g., P. plagioptera

Radlk., P. isoptera Radlk.).

8. Paullinia tumbesensis D. Simp. sp. nov. Plate Id; plate IV.

Frutex scandens; ligno composite, fasciculis externis inconspicuis; ramulis, foliis

(stipulis petiolisque includentibus), inflorescentiis et sepalis puberulis vel tomentosis.

Folia 5-foliolato-pinnata; foliolis late ovatis vel ellipticis vel obovato-ellipticis, rigide

subcoriaceis, impunctatis; foliolis ca. 52-109 mm. longis, 29-49 mm. latis; rachibus et

pet ml is anguste alatis, alis 0.5-1.0 mm. latis utroque costarum; petioli plerumque 4-7

cm. longis; stipulis caducis, subulatis, 6-9 mm. longis. Inflorescentia thyrsum

spiciformem in axibus foliorum singulariter prodiens; cincinnis sessilibus; bracteis

plerumque 2-4 mm. longis, acicularis. Capsulae stipitatae, pyriformes, ad basim in

stipitem attenuatae, 25-32 mm. (stipite incluso) longae.

Climbing shrub or vine; wood composite, the peripheral bundles inconspicuous;

branchlets, leaves (including stipules and petiole), inflorescence branches, and sepals

puberulous or tomentose. Leaves pinnately 5-foliolate, 12-20 (25) cm. long; leaflets

broadly ovate or elliptic or obovate-elliptic, rigidly subcoriaceous, not pellucid

punctate, puberulous above on the midnerve and secondary veins, tomentose below

on the principle veins and sometimes sparsely between the veins; lateral leaflets

mostly 52 mm. long by 29 mm. wide to 91 by 49 mm., terminal leaflet 56 by 30 mm.

PLATE IV. Paullinia tumbesensis; from holotype, Simpson 407.

150

SIMPSON: REVISION OF PAULLINIA 151

to 109 by 49 mm.; rachis and petiole narrowly winged, the wing 0.5-1.0 mm. wide on

each side; petiole (2.4) 4.3-7.0 (9.4) cm. long; stipules mostly caducous, subulate, 6-9

mm. long. Inflorescence single, axillary, a spicate thyrse; the cincinni sessile; bracts

mostly 2-4 mm. long, acicular. Capsules stdpitate, pyriform, attenuate at base into the

stipe, 25-32 mm. long including stipe.

Type: Simpson 407 from Dept. Tumbes, Peru.

PERU: Dept. Tumbes: Prov. Zarumilla; Dist. Matapalo;

Bosque Nac. de Tumbes cerca de Campo Verde, alt. 600-800 m.,

Simpson 407 (holotype F-2, hb. sheets no. 1,716,731 fruiting

specimen; 1,716,730 flowering specimen; isotypes USM, US, G).

ECUADOR: Prov. Guayas; road form Guayaquil to Cuevedo,

km. 78, elev. 100 m., Dodson & Thien 1285 (F-2, WIS).

9. Paullinia spicata Benth. in Hook. Journ. Bot. 3: 193. 1851.

Radlk. Monogr. Paull. 104, no. 10. 1895. Pflanzenreich IV, 165: 242,

no. 12. 1931.

Wood composite, the branchlets strongly angled and the peripheral bundles small

but well defined, or the branchlets terete and the peripheral bundles inconspicuous to

absent. Leaves pinnately 5-foliolate; rachis and petiole wingless; leaflets elliptic to

obovate, remotely obtuse-crenate, acute-apiculate, tertiary nerves mostly clathrate;

stipules minute, covering the apical bud, promptly deciduous. Inflorescence a spicate

thyrse, borne singly in the leaf axil; cincinni condensed and sessile, each subtended by

an acicular bract 1.5-6.0 mm. long. Capsule stipitate, striate, 2.4-3.5 cm. long; stipe

0.6-1.0 cm. long; valves mostly 2-3 mm. thick; columella sometimes persistent.

Type: Spruce 524 from Para State, Brazil.

PERU: Dept. Cajamarca: [Prov. Jaen;] in forest, Jaen, alt. 500

m. [the town of Jaen is at an altitude of ca. 3,000 m., this collection

was probably made near Bellavista on the Rio Maranon],

Woytkowski 5602 (MO, US). - Dept. Loreto: [Prov. Alto

Amazonas;] dense forest, between Yurimaguas and Balsa-puerto,

alt. 135-150 m., Killip & Smith 28167 (US). - Dept. San Martin:

[Prov. Mariscal Caceres; Dist. Juanjui;] in the forest, Juanjui, alt.

400 m., Woytkowski 7075 (MO, US).

ECUADOR: [Prov. Guayas:] Balao, Eggers 14247 (US). This

specimen has flowering inflorescences but no capsules. It varies

from typical P. spicata in many respects and its assignment here is

tentative. A duplicate of Eggers 15846 at F appears to be markedly

different from the Eggers 14247 duplicate at US, although both

were cited by Radlkofer under P. spicata.

An excellent photo of the isotype at M is available (F neg.

5998) but I have otherwise seen no type material. The Killip &

152 FIELDIANA: BOTANY, VOLUME 36

Smith collection seems to agree well with the specimen in the

photograph, especially as to characters of the leaves, twigs, and

inflorescence. The two Woytkowski collections, though from widely

separated localities, are very similar, yet differ in many ways from

both the Killip & Smith collection and the type. Although their

assignment to this species is tentative, their capsules place them

definitely in Subseries Pinnatae.

Series B. Obovatae D. Simp. ser. nov.

Capsulae grandes, usque ad 7 cm. longae, stipitatae, pyriformes vel obovatae vel

fusiformes; valvis plerumque 5-8 mm. crassis, mesocarpio fibroso, plus minusve

elastico; placenta post dehiscentiam persistent!, longitudinaliter triquetro, per

angustam, filiformem basim ad receptaculum affixa.

Capsules large, to as much as 7 cm. long, stipitate, pyriform or obovate or

fusiform; valves of mature capsule mostly (3) 5-8 mm. thick, the fibrous mesocarp

slightly elastic; the placenta persistent after dehiscence, longitudinally triquetrous,

attached to the receptacle by a narrow thread-like base.

10. Paullinia obovata (R. & P.) Pers. Syn. PL 1: 443, no. 6. 1805.

Radlk. Monogr. Paull. 172, no. 13. 1895. Pflanzenreich IV, 165: 259,

no. 16. 1931. Semarillaria obovata Ruiz at Pavon, Fl. Peruv. et

Chil. Prodr. 54. 1794.

Liana; wood simple; branchlets weakly 3- to 5-angled. Leaves pinnately 5-

foliolate; rachis and petiole without wings; petiole mostly 4-12 cm. long; leaflets

ovate to elliptic to obovate, 6-18 cm. long and 3-9 cm. wide, acuminate, remotely

serrate-dentate, the teeth apices of a dense, gland-like tissue, base rounded to acute

(rarely cuneate), glabrous above except scabrous to puberulent along the main veins,

a tuft of rigid hairs in the axils of secondary veins beneath, sometimes in the axils of

tertiary veins as well, otherwise usually glabrous beneath or infrequently uniformly

puberulous, densely glandular to eglandular beneath; stipules caducous (no stipules

seen on any of the specimens cited). Inflorescence borne singly in the axil of the

leaves or sometimes terminal and axial; rachis usually rusty tomentose; cincinni

sessile, bracteate, the bract shorter than or rarely exceeding the flower buds; flowers

pedicellate. Fruit stipitate or rarely subsessile; obovate to pyriform to fusiform;

mostly 3-5 (7) cm. long including the 3-18 mm. long stipe; capsule valves 4-8 mm.

thick, fibrous, the fibers oriented longitudinally, sometimes very dense and woody,

sometimes corky; the valves dehiscent, usually falling away; the 3-angled placentum

usually persistent; seeds half enclosed in the aril; immature fruit green or greenish

orange (fide J. Schunke V.), red when mature (fide various collectors), the seed shiny

black (fide J. Steinbach) or chestnut- colored (fide Ruiz & Pavon) with a fleshy white

aril.

10a. var. obovata. Plate le.

Leaflets markedly serrate-toothed?, the serrations beginning near the base or at

least below the middle; at least some of the secondary nerves bifurcating near the

margin; a tuft of barbate hairs in the axils of the secondary nerves beneath, and in

the secondary nerve bifurcations, barbate hairs none in the axil of tertiary nerves or

rarely a few tertiary nerves with a small axillary tuft. Inflorescence mostly less than

SIMPSON: REVISION OF PA ULLINIA 153

8 cm. long; bracts subulate (fide Radlk.) or (in Waytkowski 7573) ligulate, broadly

acute, and exceeding the cincinni. Capsules mostly 3 (3.5) cm. long or less, obovate or

pyriform.

Type: Ruiz & Pavon s.n. from Pozuzo, Peru.

PERU: Dept. Loreto: [Prov. Coronel Portillo; Dist. Calleria;]

on the shores of the lake, Pau Cocha (Pucallpa) alt. 200 m.,

Woytkowski 6305 (MO, US); [Dist. Padre Abad;] in forest,

Aguaytia, alt. 300 m., Woytkowski 5375 (F, MO); river bank

opposite village of Aguaytia, U.C.B.P. (Mathias & Taylor) 3585 (F,

LA); on bank of stream, Previsto, alt. 420 m., Woytkowski 7573 (F,

MO, US). — Dept. San Martin: Prov. Mariscal Caceres; Dist.

Tocache Nuevo; en bosque alto, fundo Miramar 5 km. abajo de

Tocache Nuevo (margin izquierda del Rio Huallaga), J. Schunke V.

3879 (F). — Dept. Huanuco: [Prov. Leoncio Prado;] Camino

Jacintillo [a footpath beginning at the airport bridge opposite Tingo

Maria and leading up the left bank of the Rio Huallaga under a

precipitous mountainside, upstream to near the confluence of the

Rio Monzon], vicinity of Tingo Maria, U.C.B.P. (Mathias & Taylor)

5928 (F); [Prov. Pachitea; Dist. Pozuzo;] "habitat in Peruviae

Andium nemoribus Pozuzo,"1 Ruiz & Pavon s.n. (isotype F; photo

ex G: F neg. 23,655; photo ex F: F neg. 53,208); Dist. Honoria; al

borde del rio en bosque bajo, Isla de Pacanase a 5 km. arriba del

campamento (el campamento "Miel de Abeja" del Servicio Forestal,

a la orilla del Rio Pachitea), 1 km. arriba de Tournavista o unos 20

km. arriba de la confluencia con el Rio Ucayali, J. Schunke V. 2333

(F) [distributed as P. elongata].

lOb. var. subrotunda (R. & P.) D. Simp. stat. nov. Semarillaria

subrotunda Ruiz et Pavon, Fl. Peruv. et Chil. Prodr. 54. 1794. Ruiz,

Syst. Veg. 92. 1798. Paullinia subrotunda (R. & P.) Pers. Syn. PL 1:

443, no. 15. 1805. Radlk. Monogr. Paull. 174, no. 15. 1895.

Pflanzenreich IV, 165: 260, no. 19. 1931. Semarillaria nervosa Ruiz

et Pav6n, nomen ined., in scheda collectionis ex Vitoc a Tafalla

legit.

Young branches, inflorescences, and flower buds densely ferrugineous tomentose.

Leaflet margins with very shallow serrations, mostly limited to the apical half of the

leaflet; tertiary venation markedly densely clathrate; leaflets densely puberulous

beneath. Inflorescences usually 20-25 cm. long; bracts subulate, ca. 6 mm. long,

exceeding the flower buds, soon deciduous. Capsules unknown to me (illustrated in

Flora Peruv. et ChiL, vol. IV, t. 336).

1 From Anal. Jard. Bot. Madrid 12: 157. 1953.

154 FIELDIANA: BOTANY, VOLUME 36

Lectotype: Tafalla s.n. from Peru1 (lectotype MA, isolectotypes

US, fragment at F).

PERU: [Dept. Huanuco: Prov. Huanuco; Dist. Chinchao;]

Pampayacu and [San Juan de] Cochero, Poeppig 1327 (MO;

fragment at F). — [Dept. Junin: Prov. Tarma; Dist. Vitoc;] Vitoc,

Tafalla s.n. (isolectotype US; fragment at F; photo ex B: F neg.

5629).

Macbride commented in the "Flora of Peru" that this taxon

"seems too near P. faginea," but this and other comments on the

affinities of the Paullinia species that he treated should be

disregarded. While there are superficial resemblences, the two taxa

are, in reality, in separate and only distantly related sections of the

genus.

'No collection sites are given by Ruiz and Pavon in the Prodromus, but in the

Syst. Veg. (p. 93) they state: "habitat offatim in Peruviae nemoribus versus

Chinchao, Cuchero et Pozuzo vicos." I have seen none of their specimens from any of

those three locations, nor apparently had Radlkofer. In the unpublished text for vol.

IV of the Flora Peruv. et Chil. (Anal. Jard. Bot. Madrid 12: 154. 1953) the collection

data cited above was repeated and to it was added "..., et in Vitoc via prope Collam

et Pucara, ubi Johannes Tafalla." It seems likely that the collections from Chinchao,

Cuchero, and Pozuzo were lost in the fire at Macora, since Ruiz in his journal lists S.

obovata and S. subrotunda among those that he subsequently "succeeded in

describing in Huanuco, of those that were worked in Macora and were burned in that

fire." It is clear from his journal that neither he nor Pavon ever entered the

Chanchamayo Valley nor the tributary Rio Tulumayo Valley where the village of

Vitoc was located.

The collections from Vitoc and Pucara were made by Tafalla in 1794 six years

after Ruiz and Pavon's return to Spain, and Tafalla probably included them in a

shipment of specimens that he sent to Madrid in 1795 (see Steele, A. R., Flowers for

the King, p. 271. 1964). The Ruiz specimens cited by Radlkofer, as also the specimen

at US cited above, are undoubtedly duplicates of the Tafalla collection from Vitoc.

The US specimen has a label headed "Herb. Reg. Berolinense" and on the lower

parts of the label are; "Peruvia at Chili." and "Ruiz legit, ex herbario Lamberti."

This is perhaps from the Griefswald set sent to US in 1895 (see Taxon 19(4): 540.

1970). Although the label on the specimen at B did not so indicate, it too was

probably obtained from the Lambert herbarium. The duplicate at FI (cited as "hb.

Webb" by Radlk.) was surely part of the sets of specimens that Webb purchased

from Pavon in 1826 (see Steele, p. 314).

It seems probable that all the Ruiz and Pavon collections of this taxon were lost

and that of the specimens indicated by Ruiz as belonging to this taxon only the

Tafalla collection remains in existence. I, therefore, propose its designation as the

lectotype.

On the specimen at US (sheet no. 249,751) the label contains, in addition to the

information discussed above, a handwritten binomial which apparently is an

unpublished manuscript name. This binomial, "Semarillaria nervosa," is in a

distinctive hand, almost certainly that of Ruiz. The B specimen as shown in the

photograph has two handwritten binomials on the label; the first, "Semarillaria

nervosa Flor. Per," by the same hand as that on the US specimen, the second,

"Paullinia subrotunda Pers., Radlk.," seemingly in Radlkofer's hand. It may be that

when first received from Tafalla, Ruiz tentatively wrote this binomial on the label

and later, finding it to be the same as his description and illustration of S.

subrotunda so cited it in the manuscript of the Flora Peruv. et Chil., vol. IV, but

neglected to change the specimen labels.

SIMPSON: REVISION OF PAULLINIA 155

lOc. var. poly morpha D. Simp. var. nov. Plate If; plate V.

Varietas ab aliis varietatibus hujus specie! bracteis minutis vel vestigialibus,

capsulis plerumque fusiformibus et usque ad 7 cm. longis differt. A varietate

subrotunda pube sparsioribus in pagina inferiore foliolorum et in ramulis, alabastris

plerumque grandioribus, et ab varietate obovata capsulis longioribus, foliolis

plerumque grandioribus, et nervis tertiariis magis clathratis differt.

Differing from the other varieties of this species by the minute or vestigial bracts,

capsules that are fusiform and up to 7 cm. long, from the variety subrotunda by the

more sparse pubescence on the lower surface of the leaflets and on the branchlets, the

larger flower buds, and from the variety obovata by the longer capsules, somewhat

larger leaflets, and the more clathrate tertiary nerves.

Type: Jose Schunke V. 4580 from Tocache Nuevo, Peru.

PERU: Dept. Amazonas: Prov. Bagua; rainforest along Rio

Santiago 10-15 km. above mouth, elev. 250 m., Wurdack 2505 (F,

US); above Pongo de Manseriche, left bank of Rio Santiago,

overflow bank ["Dept. Loreto" on label is error], Mexia 6297 (F,

MO, US). — Dept. Loreto: [Prov. Alto Amazonas; Dist. Yuri-

maguas;] edge of river, Sapote Yacu, Sta. Rosa [ca. 15 km. b'near

distance downstream from Yurimaguas], alt. 115-210 m., LI.

Williams 4928 (F). - Dept. San Martin: Prov. Mariscal Caceres;

Dist. Tocache Nuevo; a orilla del rio en bosque alto, Quebrada de

Tananta (margin derecha del Rio Huallaga), Jose Schunke V. 4574

(F-2 [flowers and fruit]; duplicates not yet distributed), 4580

(holotype F-2, hb. sheets no. 1,716,732 and 1,716,733 [fruit only, no

flowers]; duplicates not yet distributed).

BOLIVIA: Dept. Sta. Cruz: Prov. Sara; quebrada humeda,

bosque, Buenavista, alt. 450 m., J. Steinbach 7102 (F, MO); bosque

del Rio Palometillo, alt. 400 m., J. Steinbach 6783 (F).

The vegetative morphology of this species seems to be

exceptionally variable compared to that of the other species in this

section. This was probably not apparent to Macbride when he

treated PaulUnia for the "Flora of Peru," for he cited only four

collections in all; two under P. obovata (the Ruiz and Pavon type

and Jose M. Schunke 118) and two under P. subrotunda (the type

collection and Poeppig 1327). For the type of P. subrotunda he had

only the photograph cited above, a fragmentary part of one leaflet

from a specimen at Madrid, and a fragment of the Poeppig

collection. It is now fairly obvious that Schunke 118 is not P.

obovata nor even closely related to it, but for the Ruiz and Pavon

type he had both a full specimen and a photograph of the duplicate

at B.

PLATE V. Paullinia obovata var. polymorpha; from holotype, J. Schunke V.

4580.

156

SIMPSON: REVISION OF PAULLINIA 157

Since Macbride's treatment was published new material has

accumulated in the herbarium that clearly demonstrates the

mutual affinities of P. obovata and P. subrotunda. At the same

time, these new materials raise problems that were not previously

apparent. The principal problem concerns devising a new taxonomic

treatment. Clearly P. obovata and P. subrotunda are conspecific,

but to what extent should infraspecific elements be given taxonomic

recognition and at what hierarchical levels? Unfortunately, even

with the additional collections of the last two decades the total of

material now available is still such a scanty sampling of this

widespread and complex species that it is not yet possible to devise

a satisfactory taxonomic treatment.

Even with these few collections there appear to be evident

some meaningful patterns of morphology and distribution, and

these I prefer to interpret as reflecting the recent evolutionary

history of the species. Some hypotheses concerning that evolution-

ary history are proposed as follows:

The species, perhaps originally a more homogeneous taxon

morphologically similar to Wurdack 2505 and LI. Williams 4928,

was geographically fragmented into three elements probably by no

later than the mid-Pleistocene. These three elements, var. obovata,

var. subrotunda, and var. polymorpha, evolved some differences in

morphology in response to ecological selection.

During the period of isolation and divergence the three

elements could have been distributed as follows: The var. obovata

occupied the middle Ucayali Valley centered around the

confluences of the Pachitea and Aguaytia Rivers; var. subrotunda

was confined to a narrow band of forest at middle elevations on the

eastern slopes of the Andes; and var. polymorpha was probably

endemic to a lowland forest refugium centered around the

confluence of the Santiago and Maranon Rivers and extending

south to the confluence of the Cainarache and Huallaga Rivers,

near present-day Yurimaguas.

Post-Pleistocene shifts in climate have brought these three

elements into contact again, with hybridization and considerable

gene flow between elements resulting in the blurring of

morphological differences.

Among the collections cited under var. polymorpha, Schunke

4574 and Mexia 6297 suggest a strong infusion of var. subrotunda

genes as also do the Steinbach collections from Bolivia.

158 FIELDIANA: BOTANY, VOLUME 36

11. Paullinia bracteosa Radlk. Bull. Herb. Boiss., ser. 2, 5: 321.

1905. Pflanzenreich IV, 165: 264, no. 24. 1931. Plate Ig.

The branches markedly 3-6-angled and sulcate, stout; wood composite but

inconspicuously so,1 the peripheral bundles remaining very small though distinct.

Leaves 5-foliolate pinnate; rachis and petiole broadly winged; leaflets oblong to

elliptic to obovate, remotely crenate to obtusely serrate, mostly 12-28 cm. long, (4) 7-

12 cm. wide, secondary nerves often plicate- sunken above; petiole (5) 10-26 cm. long;

stipules mostly 3-6 cm. long by 5-12 mm. wide, longitudinally striate, oblong or

narrowly elliptic, obtuse, usually persistent. Inflorescence stout, usually pubescent

throughout, borne singly in the leaf axil; bracts usually exceeding the cincinni, oblong

to elliptic, obtuse or broadly acute, 8-20 mm. long, mostly 4-8 mm. wide. Capsules

pyriform to broadly obovate, 3-7 cm. long; valves 3-8 mm. thick.

Type: Tonduz 11416 from Costa Rica.

PERU: Dept. Amazonas: [Prov. Bagua; Dist. Aramango;] in

forest at 300 m., Aramango, Woytkowski 5622 (US). — Dept. Loreto:

[Prov. Requena; Dist. Requena;] dense forest on high sandy ground

north of Requena, Chacra Canama [Camana?], e. side of river,

U.C.B.P. (Mathias & Taylor] 5545 (F); [Prov. Ucayali;] Quebrada

de Maquia (Contamana), in tall forest, 220 m. alt., U.C.B.P. (J.

Schunke V.} 6677 (F-2); J. Schunke V. 938 (F); [Prov. Coronel

Portillo; Dist. Padre Abad;] woods near the house of Don Diogenes

del Aguila, east of Aguaytia between Pucallpa road and Rio

Aguaytia, U.C.B.P. (Mathias & Taylor) 3545 (F), 5066 (F), 5358 (F),

5986 (F). - Dept. San Martin: [Prov. Huallaga; Dist. Saposoa;] in

purma [ = secondary vegetation], Saposoa, 400 m., Woytkowski 5070

(F, MO, US), 5404 (F, MO), 5487a (MO-2); in forest, Gramalote to

Saposoa, alt. 450 m., Woytkowski 5424 (MO); Prov. Mariscal

Caceres; Dist. Tocache; a orilla del rio en bosque alto, Fundo

Porvenir (margin derecha del Rio Huallaga), J. Schunke V. 4335

(F); [Dist. Campanilla;] in tall forest, 320 m. alt., at Campanilla,

U.C.B.P. (J. Schunke V.) 6314 (F-2). - Dept. Huanuco: [Prov.

Leoncio Prado;] at Monzon-Huallaga junction, U.C.B.P. (Mathias

& Taylor] 5339 (F).

BOLIVIA: Dept. La Paz: Prov. Larecaja; Tuiri (near Mapiri,

on left bank of Rio Mapiri), alt. 490-750 m., Krukoff 10756 (F, MO,

US) [distributed as P. ingaefolia Rusby].

The considerable number of collections of this species now in

herbaria convincingly demonstrate that the Central American

'Although the isotype at F seems to lack peripheral bundles, all the other Central

American collections (two from Costa Rica and 11 from Panama) have inconspic-

uously composite wood in which the peripheral bundles seem to remain very small

but distinct, as is also the case in the Amazonian materials.

SIMPSON: REVISION OF PAULLINIA 159

plants, though well separated from the Amazonian ones by

geography, differ to no significant degree in any of the critical

morphological features. It now seems that the Central American

plants are distributed throughout the rain forest of areas below 800

m. alt., from the Caribbean lowlands of Costa Rica and from the

Osa Peninsula on the Pacific side of Costa Rica southward, through

Panama into the Choco of Colombia. The Andean highland

separates that part of the distribution from the Amazonian

populations which occur principally in the western and south-

central parts of the Basin. Krukoff 6522 from Huamayta on the Rio

Madeira, Amazonas State, Brazil, represents the eastern edge of the

range as known to me at this time. I have seen no specimens from

Colombian Amazonia nor from any part of the Basin north of the

Maranon-Amazonas-Solimoes. P. naiguatensis Steyermark may

prove to be conspecific with P. bracteosa and, if so, would

considerably extend the range of the latter, for, in addition to the

type locality as cited by Steyermark (humid forest in the coastal

mountains of the Federal District, Venezuela), there are collections

identified by Steyermark as P. naiguatensis from Territorio do

Roraima, Brazil (Prance, et al. 4060) and an adjacent part of

Amazonas State (Prance et al. 10295).

12. Paullinia imberbis Radlk. Monogr. Paull. 177, no. 18. 1895.

Pflanzenreich IV, 165: 263, no. 23. 1931.

Branches 3-6-ridged and sulcate; wood simple. Leaves 5-foliolate pinnate; petiole

and rachis winged; leaflets oblong to elliptic to obovate or oblanceolate, remotely

serrate. Inflorescence single, axillary or terminal, spicate; cincinni sessile; bracts

small, subulate, 2-4 mm. long. Capsules pyriform or [fide Radlk.] ellipsoidal, 2-3 (-4?)

cm. long.

Syntypes: Martius s.n. from Barra [= Manaus] and Coara on

the Rio Negro, and at Para [Belem]; Melinon 59 from French

Guiana.

PERU: Dept. Loreto: [Prov. Maynas; Dist. Ramon Castilla;] in

forest at Caballo Cocha on the Amazon River [ca. 65 km. east of

Leticia, Colombia] LI. Williams 2363 (F-2).

There are two collections from Venezuela originally distributed

as P. pinnata but which are probably P. imberbis. These are LI.

Williams 15242 & 16160, both from Tatama on the upper Orinoco,

in Fed. Terr. Amazonas.

Aside from the three collections cited above I have seen no

other material belonging to this species and lacking any photograph

of type material must depend on Radlkofer's description alone.

160 FIELDIANA: BOTANY, VOLUME 36

Series C. Eriocarpae D. Simp. ser. nov.

Capsulae sessiles (estipitatae) vel subsessiles (turn stipites 5 mm. longae minores),

ovoidea vel ellipsoidea vel fusiformes, plerumque pubescentes; valvis 1-5 mm. crassis.

Lignum compositum. Folia 5-foliolato-pinnata; rachibus et petiolis alatis.

Capsules sessile (non-stipitate) or subsessile (then the stipe less than 5 mm. long),

ovoid to ellipsoid to fusiform, usually pubescent; the valves 1-5 mm. thick. Wood

composite. Leaves 5-foliolate pinnate; rachis and petiole winged.

This rather distinctive series includes P. eriocarpa Tr. & PL

(treated below), and P. leiocarpa Griseb., a species of the West

Indies and the Caribbean Coast of South America, extending

eastward through the Guianas to Brazil.

Type species: P. eriocarpa Triana & Planchon.

13. Paullinia eriocarpa Tr. & PI. Ann. Sci. Nat. Bot., ser. 4, 18:

353. 1862. Radlk., Monogr. Paull. 179, no. 20. 1895. Pflanzenreich

IV, 165: 265, no. 26. 1931. P. eriantha Benth. ex Radlk., Monogr.

Paull. 179, no. 20. 1895 (P. eriantha Benth. was a manuscript name

until Radlkofer's monograph provided an accompanying descrip-

tion). Plate Ih; plate lie.

Liana; wood composite, peripheral bundles smaller than the central although

remaining noticeably distinct into the older stems; young branches 5-6-angled,

densely pilose to lanate with golden yellow to tawny-colored hairs. Leaves pinnately

5- (rarely 3-) foliolate; rachis and petiole broadly winged; leaflets sessile, 7.5-20 cm.

long, 3.8-10 cm. wide, usually soft pubescent beneath, subcoriaceous, the tip broadly

acute or rounded and with an apiculus, the margin revolute, entire or few toothed,

thickened-cartilaginous and straw-colored; petiole 5 to 17 cm. long; stipules

lanceolate to narrowly ovate or elliptic, acute or short acuminate, 8-15 mm. long,

mostly about 3 mm. wide at base, sometimes to 9 mm. wide by 12 mm. long.

Inflorescence densely lanate pubescent, solitary and axial; cincinni sessile; flower

buds mostly 7-9 mm. diameter; bracts ovate or suborbicular, sometimes more narrow

near the inflorescence apex, broadly acute to rounded at the tip; capsules sessile,

fusiform or ellipsoid-fusiform, densely lanate pubescent; the hairs stiff, brittle, and

irritating to the skin.

Type: Triana 3452 from Villa vicencio, Colombia.

PERU: Dept. Amazonas: [Prov. Bagua;] forest at Aramango,

elev. 300 m., Woytkowski 5634 (MO). - Dept. Loreto: [Prov. Alto

Amazonas;] edge of forest, Fortaleza, Yurimaguas, LL Williams

4302 (F). — Dept. San Martin: [Prov. Lamas;] Rumizapa, near

Tarapoto, LI. Williams 6770 (F); [Prov. San Martin;] Tarapoto, LI.

Williams 5849 (F); [Prov. Huallaga;] secondary forest, Saposoa, alt.

400 m., Woytkowski 5069 (F, MO, US), 7260 (MO, US).

COLOMBIA: Comisaria del Putumayo: forest, Umbria, 0° 54'

N, 76° 10' W, alt. 325 m., Klug 1843 (F, US).

SIMPSON: REVISIN OF PA ULLINIA 161

LI. Williams 4302 from Fortaleza near Yurimaguas is a variant

form. It has very broad stipules, flower buds that are pedicellate

and smaller than usual in this species; the inflorescence bracts

exceed the buds, are only sparsely pubescent, and dark reddish-

brown in the dried specimen; the leaflets are nearly glabrous, the

apices are subcuspidate and bases are auriculate; the stipules are

broadly ovate, ca. 20 mm. wide by 22 mm. long. In all of these

features it contrasts with the other Peruvian collections and

because there are no fruits, its disposition is tentative.

The other material here cited agrees well with the photograph

of the type of P. eriantha Benth. ex Radlk., Spruce 4415 (photo ex

B: F neg. 5598) from near Tarapoto. I have seen neither a specimen

nor photograph of the type of P. eriocarpa Tr. & PL

REFERENCES

FAEGRI, K. and L. VAN DER PIJL

1966. The principles of pollination ecology. Pergamon Press, New York.

GOODSPEED, T. H. and H. E. STORK

1955. The University of California Botanical Garden expeditions to the Andes

(1935-1952). Univ. Calif. Publ. Bot, 28(3), pp. 79-142.

LANJOUW, J. and F. A. STAFLEU

1964. Index Herbariorum. Part 1. The herbaria of the world, ed. 5. Reg. Veg. vol.

31.

MACBRIDE, J. F.

1956. Sapindaceae, in Flora of Peru. Field Mus. Nat. Hist, Bot. Ser., 13 (part III

A, no. 2), pp. 291-391 (for PaulUnia, pp. 325-361).

MULLER, H.

1873. Die Befructung der Blumen durch Insekten. Verlag von Wilhelm Engelmann,

Leipzig, (see p. 154).

RADLKOFER, L.

1890. Uber die Gliederung der Familie der Sapindaceen. Sitzungsber. Math.-Phys.

Cl. Konigl. Bayer. Akad. Wiss. Miinchen, 20, pp. 105-379.

1896. Monographic der Sapindaceen Gattung Paullinia. Abh. Matk-Phys. Cl.

KonigL Bayer. Akad. Wiss. Miinchen 19, pp. 67-381. Separately printed and

issued in 1895.

1931-1934. Sapindaceae, in A. Engler, Das Pflanzenreich IV, no. 165 (for Paullinia,

pp. 219-320, 1931; and pp. 321-352, 1932).

THOMPSON, D'ARCY W. (translator and editor).

1883. Translation of Muller, Hermann, The Fertilization of Flowers, Macmillan &

Co., London.

162 FIELDIANA: BOTANY, VOLUME 36

ADDENDUM

Subsequent to writing the manuscript for this revision I

realized that some of Rusby's Bolivian species probably belonged to

section Paullinia. A loan from the New York Botanical Garden

herbarium enabled me to examine type specimens of those species.

Four of them proved to be members of this section and synonyms of

earlier names as listed here.

Paullinia ingaefolia Rusby, Mem. N. Y. Bot. Gard. 7 (3): 291. 1927.

Type: O. E. White 1275 from Rurrenabaque, Bolivia. = P.

imberbis Radlk.

P. pendulifolia Rusby, I.e., p. 291-2. Type: Rusby 1622 from

pampas near Lake Rogagua, Bolivia. = P. pinnata L.

P. ribesiaecarpa Rusby, I.e., p. 293. Type: Rusby 1730 from Reyes,

Bolivia. = P. elegans Camb. subsp. elegans.

P. tatei Rusby, Phytologia 1 (2): 64. 1934. Type: G. H. H. Tate 553

from Guanai, Bolivia. = P. elegans Camb. subsp. neglecta

(Radlk.) D. Simpson.

In the "Flora of Peru," Macbride (p. 343) commented under P.

imberbis Radlk. that P. ingaefolia Rusby "is apparently similar but

larger in all parts." Before seeing the type of P. ingaefolia Rusby, I

was inclined to treat it as probably conspecific with P. bracteosa

Radlk. However, the type specimen has not the distinctive large,

obtuse inflorescence bracts of P. bracteosa, but shows small,

acircular inflorescence bracts like those described by Radlkofer for

P. imberbis. Even though it seems to fit the description of P.

imberbis, having neither type material nor "phototype" of the

latter, the disposition of P. ingaefolia Rusby as a synonym of P.

imberbis Radlk, must be considered tentative.

SIMPSON: REVISION ON PAULLINIA

163

VI. INDEX TO EXSICCATAE

Collection

Blanchet, J. S.

Collection

Bondar, G.

Camp, W. H.

Cardenas, M.

Cazalet, P.C.D.

& T. D.

Pennington

Dodson & Thein

Eggers, H.F.A.

Glocker. E. F.

Goodspeed Exped.

(Stork & Morton)

(Stork, Horton,

& Vargas)

Haught, O.

Kanehira, R.

Killip & Smith

Klug, G.

Krukoff, B. A.

Martins

Mexia, Y.

Peredo, I.

Poeppig, E. F.

Prance, et aL

Rimbach, A.

Ruiz & Pavon

(also see Tafalla)

Rusby, H. H.

Sagastegui, A.

St. Hilaire, A. de

Schiede, C.J.W.

Schunke, J. M.

Schunke V., J.

Simpson, D. R.

Spruce, R.

Steinbach, J.

Tafalla, J.

Tate, G. H. H.

Tonduz, A.

Triana, J. J.

U.C.B.P.

(Mathias&

Taylor)

(J. Schunke V.)

Ule, E.

Vargas, C.

Weberbauer, A.

White, O. E.

Williams, L.

164 FIELDIANA: BOTANY, VOLUME 36

16160 159

Williams, R. S. 466 147

689 142

Woytkowski, F. 5069 160

5070 158

5213 140

5296 141

5321 142

5325 142

5364 141

5375 153

5404 158

5424 158

5487A 158

5573 147

5602 151

5622 158

5634 160

6299 141

6305 153

6320 144

6573 144

6575 144

7075 151

7149 147

7573 153

7888 140

Wurdack, J. J. 1977 141

2505 155, 157

UNIVERSITY OF ILLINOIS URBANA