Skip to main content

Full text of "Internal oral features of larvae from eight Anuran families : functional, systematic, evolutionary, and ecological considerations"

See other formats 


eae 
ae : ’ id hi a COMP. Zope 

NIVERSITY OF KANSAS Je eat MISCELLANEOUS 
USEUM OF NATURAL HISTORY Pua Or as ce 


Internal Oral Features of 
Larvae from Eight Anuran 
Families: 

Functional, Systematic, 
Evolutionary and Ecological 
Considerations 


By 
- Richard Wassersug 


UNIVERSITY OF KANSAS 
LAWRENCE 1980; 


Wassersug, E. 1980. 


UNIVERSITY OF KANSAS PUBLICATIONS 
MUSEUM OF NATURAL HISTORY 


The University of Kansas Publications, Mu: wu. .° Natural History, beginning 
with volume 1 in 1946, was discontinued with volunie 2u in 1971. Shorter research 
papers formerly published in the above series are now published as Occasional 
Papers, Museum of Natural History. The Miscellaneous Publications, Museum of 
Natural History, began with number 1 in 1946. Longer research papers are pub- 
lished in that series. Monographs of the Museum of Natural History were initiated 
in 1970. All manuscripts are subjected to critical review by intra- and extramural 
specialists; final acceptance is at the discretion of the Director. 


Institutional libraries interested in exchanging publications may obtain the Occa- 
sional Papers and Miscellaneous Publications by addressing the Exchange Librarian, 
University of Kansas Library, Lawrence, Kansas 66045. Individuals may purchase 
separate numbers of all series. Prices for all publications of the Museum may be 
obtained from the Publications Secretary, Museum of Natural History, University of 
Kansas, Lawrence, Kansas 66045. 


mus. COMP ZOO 


nf 


THE UNIVERSITY OF KANSAS 


MuSEUM OF NATURAL HISTORY 


MIsCcELLANEOUS PUBLICATION No. 68 
June 24, 1980.22 


Internal Oral Features of Larvae from Eight 
Anuran Families: 
Functional, Systematic, Evolutionary and 
Ecological Considerations 


By 


RICHARD WASSERSUG 


Department of Anatomy 
and 
Committee on Evolutionary Biology 
University of Chicago 
Chicago, Illinois 60637 


THE UNIVERSITY OF KANSAS 
LAWRENCE 
1980 


UNIVERSITY OF KANSAS PUBLICATIONS, MUSEUM OF NATURAL HISTORY 


Editor: E. O. Wiley 


Miscellaneous Publication No. 68 
pp. 1-146; 37 figures 
Published June 24, 1980 


MusEuM oF NATURAL HIsTOoRY 
THE UNIVERSITY OF KANSAS 
LAwreENce, Kansas 66045 
U.S.A. 


PRINTED BY 
UNIVERSITY OF KANSAS PRINTING SERVICE 
LAWRENCE, KANSAS 


CONTENTS 


NER © CANO Noe ee ke ee OL 1 
IN CKO WAC GL OCTETS a eee og ee, dO I, SR RS 2 
INA te ra a Sa aR or a 2 
IVA En CS se Bence foe a IE Re Re ee EY 5 

EGS GEM ONS 3 2a Pes oe oyre a teen eyo nn Be eR, BO AI ae 1 
PS CAD MCA Coe sk SA SO te at nw hs rape RRR il 

SSCA PIMUSSETUCT Sti Sh REIS UO ASTER SRT oe ares 7 
DISCO GIO SSE ASS en ee Sener Re re ee ee Sanne eee oe ae 10 
VV TLE SRODSTCETEC ONS fetes a ott ie coe PL Oe eee eR ee eee 10 
PAU ECSCISLCTIVUSUE Po I NEE ROT REE ERO De ee 14 
BOMMDINGOTICIL GUS cen te Pea ee a ede re ee eee ee 16 
IDISCOLLOSSUSIIICLIST wemmas Stet lees peri sweeties ors Seve cae t eee eee 19 
iT Op latyani Gaey tar ses oe eae a ke 21 
inopliny mus YdOrsalisne 22 ie sek se 7h eee ee cree 2 
Mircrolaylidae ye ai oe tee soe ae Sr re OE aes ce eee ae 24 
Microhylaiibend more mea ean ae a ee 24 
Microhiylathe ymonst ae Te a OH 
IVGCrON La ORIOL see ee cere RS =e ene ar Re 30 
J PCE OFE 6 Patina ter ser weet sa Oe one tees ot Soe 9 ee oh ll 32 
SOTO ODES DOM OG RODS ae 32 
WOLD O GUS OLD OL eae OR SE AI eo eS Oe eee 33 
IC DLObTAChitm@ NASSC big ses ee ate Sees Se ee 37 
NEC LOD TACT UAV TIVROS ive 1151S) eereeeeen es eae aaa 40 
Orcolalaxiping) oo ER NE nee Aero eerie 43 
sli Ge, seme: 20S IRS Aa te Pee ALES ee Eee EE oe 46 
ANOCCCU SPINOSG: 2 \e see ee TSE 2 ee ee 2 eee 46 
Gasirotheca tiobainbae =-t2. 222 a ee ee ee ee 48 
IPO LCD gh CNIUON LUGS) =e ceo tg a a8 ee oes ee ee 52 
EUG: AULPOCCUG 23 Sse a een aN tae ee Rca a te 55 
LLG OMAN OSCONLG. spss. crn DE Ota ee 59 
HGlagpnicbodes xxx. 2 ne. ee, = ee ee oe SP es 61 
ALU INA Ce eS te Nc ae Scere re eo eo 63 
LUGS CIT OCC GEG 3 ee ws ce we eg ec ne ee 67 
ELD ESQ COUCIISLS ox. cee Rs eas. oc ae Pann Mts eR pe esa ae 70 
TEAC OND, SOMO OUD, ake EE oe 72 
Rigchonlaniconhardschtltcet mess 2: = = aaa enna ae nen 76 
UNCTUS CREPUGIS® Sains Se eae na Ne ac ee 80 
STUAS CU SOT DUA sxe eet OO 82 
VNCGIU CIS COLA TH GS seas ae es 85 
Gentrolenid ae’ eee ee oe i ee as See 88 
Gentrolenclla fleischnann = ee eee 88 
Mendrobatidae 2 8 oat a om aie 91 
(ODUGSITA TIS STO OTIC ee ee 91 
Golostethis:-niutbicola ess 2 2-2 s Se eee cae eee 94 


DISCUSSION (eS eee eee 97 


Functionally ConsiGeratiOns) ee ree ee 97 


Kératinized structures 2.20.4... eee 97 
Infralabial papillae... 4.3. eee 98 
Lingual. papillae. ee See 99 
Buccal floor atena 2.22.5) 2 ee eee 99 
Buccal (pockets: 25...2. 2 ee 2 ee 100 
Prepocket papillae and other features of the buccal floor 101 
Ventral) velum 00.2 a ON ee 102 
The filter system. 2 105 
Branchial tood traps andusecretony dees) 22. saan 109 
Glottis ‘and laryngeal disc 0.00. ne 110 
Esophageal: funnel 2 2 2 eee TE 
Prenarial arena, 2.8.0 os he 111 
Inteimalonares: 22 2k ee 112 
PRostnanal arena. 2.8 22ers 113 
Kateral ridge papillae <2. 222 ee 114 
Buccal root arena’ 24.223 2 3.08 As 114 
Glandularzone and dorsal’secretory pits= 115 
Dorsaleve luiniyest2e es caesar ee PReenL eee ene 116 
Pressure .cushiOns)....2. 8 ee sate 
SystematicConsiderations 2.22.24 eee 117 
Ascaphidae)! 222 2000 22 ee 118 
Discoglossidac! (fi. 0 2 ae 118 
Rhinophrynidae 220 ee 118 
Microhylidaet 272. 2 et ee 118 
Relobatidae sre. ee Oe le 119 

13 G6 Y= es ets te Be La ON A PAM ery AT 119 
Dendrobatidae and’ Centrolenidac a= = 12] 
EVoluttonanya@onsiderations | 2s eee oe 12] 
The Ascaphus and Leiopelma lite cycles) eee 2h 
Evolutionanystrends inthe Discoglossidac. 122 
dhe Pipoidea and their relationships 2... 2 123 
the ‘microhylid* problemi 2.6 eee 124 
The origins of the “advanced” anurans (Type 4 larvae) 125 
Evolution of specific larval types in the genus, Hyla _-_-_-___- 126 
The evolution of ontogenies and its role in larval diversity 127 
Ecological’ Considerations: 22 2 TS Eee 128 
Tadpoleteedine ecology... 220 er 128 
Rarticle/sorting) the general mechanism 2 a 129 
Microhabit implications of morphological patterns — 130 
SUMMARY AND" CONCLUSIONS 02S 133 
LITERATURE CVRE D2: os eye ee ee 139 


APPENDIX: GLOSSARY OF TERMS 


INTRODUCTION 


It has been a full century since a 
biologist first studied tadpole morphol- 
ogy for clues to the evolution and sys- 
tematics of the Anura (Lataste, 1879). 
Generally recognized as the most signifi- 
cant work from these last hundred years 
is that of Orton (1953, 1957), who sorted 
all frog families into four groups based 
on larval characters, specifically external 
oral features and spiracle position. 

While herpetologists have largely ac- 
cepted Orton’s four superfamilial groups, 
there is continual disagreement about 
relationships between and within these 
groups. Controversies have centered on 
the question of how much weight should 
be given larval characters when larval 
morphologies suggest relationships dif- 
ferent from adult morphologies. Resolu- 
tion of this question has been hampered 
by the few larval characters which her- 
petologists have traditionally considered 
to have taxonomic value. With considera- 
tion of few characters, the chance of 
convergence is obviously high and con- 
fidence in derived systematic relation- 
ships is low. Until recently, most work 
concerning the higher relationships of 
anurans has involved only a few tadpole 
external characters. Starrett (1968, 
1973) and Sokol (1975, 1977a, 1977b) 
have now made much progress toward 
establishing the importance of tadpoles 
to anuran systematics. Although they 
have identified many new internal larval 
features of value to systematic discus- 
sions, the larger controversy, unfortu- 
nately, has not been resolved. Starrett 
and Sokol disagree on how to interpret 
relationships implied by many larval 
features. 

The present study began as a search 
for additional diagnostic characters to 
help clarify systematic problems in the 
Anura (Wassersug, 1976a). It has ex- 
panded into a more general, comparative 
study of the functional morphology and 
feeding ecology of anuran larvae. An 
effort is made here not only to identify 


characters with systematic import, but 
to discern patterns in oral features that 
can be correlated with our knowledge of 
tadpole ecology. 

I present here a comparative study 
of certain internal oral features of anu- 
ran larvae that have not been empha- 
sized by Starrett, Sokol, or other work- 
ers. The characters described are all 
surface features which lie between the 
opening of the mouth and the esopha- 
gus. I have emphasized those organs 
which come into direct contact with 
water and food in the mouths of tad- 
poles and are thus directly involved with 
the feeding process. 

I have chosen to examine morpholog- 
ical structures involved in feeding in 
part because they are convenient to 
study. The morphology of tadpoles is 
dominated by tissue related to feeding 
functions, especially ingestion, and it 
seems reasonable to assume that a tad- 
pole’s oral morphology will demonstrate 
adaptation to the environment in which 
a larva lives. If we understand the adap- 
tive significance of morphological fea- 
tures, we should be able to determine 
much about the ecology of the tadpole 
from its oral morphology. An ultimate 
goal of this comparative study is to un- 
derstand the morphology of anuran 
larvae in enough detail to be able to 
accurately predict a tadpole’s ecology 
from its morphology. 

Although this study is limited to sur- 
face features, some reference is made to 
underlying, cartilaginous elements such 
as the ceratohyal, which forms the pis- 
ton of the tadpole buccal pump, and the 
spicule, which supports part of the oral 
surfaces. A comparative study of the 
cartilaginous skeleton of the tadpole 
buccal pump is presented elsewhere 
(Wassersug and Hoff, 1979). 

Only free-living larvae are consid- 
ered. An ontogenetic series of one spe- 
cies, Hyla regilla, was examined in detail 
in a previous study so that morpho- 


2 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


logical features modified extensively 
through development could be elimi- 
nated from further consideration (Was- 
sersug, 1976b). It was necessary to do 
this because not all material available 
for study is of the same developmental 
stage. 

The core of this paper is descriptions 
of oral structures in the larvae of se- 
lected anuran species. This is followed 
by a four part Discussion. 


The first part of the Discussion 
(Functional Considerations) reviews the 
diversity of the structures, and an effort 
is made here to correlate known larval 
ecology with patterns in oral morphol- 
ogy. Functions for many of the oral 
structures presented in the Descriptions 
are deduced on the basis of what is 
known about larval ecology. The sec- 
ond part of the Discussion (Systematic 
Considerations) examines specific sys- 
tematic questions. The third part (Evo- 
lutionary Considerations) deals with 
questions of evolutionary history as well 
as the evolutionary mechanisms that 
could account for diversity among tad- 
poles. The last part of the Discussion 
(Ecological Considerations ) reverses the 
first part. Here, starting from a basic 
understanding of the morphology, an 
attempt is made to assess the ecology of 
tadpoles and correlate ecological pat- 
terns with known morphological pat- 
terns. 


ACKNOWLEDGMENTS 


This paper augments work presented 
in a doctoral thesis submitted to the 
University of Chicago in 1973. I want 
to thank the members of my committee, 
James Hopson, Robert Inger, George 
Rabb, David Wake and Rainer Zangerl, 
for their encouragement, advice, and 
most of all, their patience. 

I am grateful to William Duellman, 
Hymen Marx and David Wake for al- 
lowing me to dissect specimens in their 
care. Additional specimens were gener- 
ously provided by E. Crespo, M. Delsol, 


R. Demmer, W. R. Heyer, R. W. Mc- 
Diarmid and C. Richards. 

For logistical support I thank the Di- 
rectors and Department Heads of the 
Committee on Evolutionary Biology and 
the Department of Anatomy of the Uni- 
versity of Chicago; the Center for Grad- 
uate Studies and the Division of Rep- 
tiles, Field Museum of Natural His- 
tory, Chicago, and The Museum of 
Natural History at the University of 
Kansas. 

Financial support was provided by a 
fellowship from the Center for Graduate 
Studies, Field Museum of Natural His- 
tory; Block Fund and Hinds Fund, Uni- 
versity of Chicago; and the National 
Science Foundation (BMS 75-03447 and 
DEB 76-19275). 

I am extremely grateful to Marsha 
Greaves, who executed drawings for 
this paper. Ilse Hecht graciously pro- 
vided translations of several articles in 
foreign languages. Shirley Aumiller and 
Robert Kott helped with photography, 
portions of the manuscript were typed 
by Debra Randall and Karen Rosenberg. 
Special thanks are due Nancy Bradney, 
who helped in many aspects of produc- 
tion and editing. Steve Busack, William 
Duellman, Robert Inger, Dianne Seale, 
Otto Sokol, Linda Trueb and David 
Wake have read various portions of this 
manuscript; it has profited greatly from 
their constructive criticisms. 


MATERIALS 


Oral structures of larvae of 31 species 
from eight families are described. The 
tadpoles came from museum collections 
or the author’s private collection. An 
effort was made to examine specimens 
at or near Gosner (1960) stage 36. 

A few species (e.g. Acris crepitans) 
were chosen for study because they have 
the commonest type of anuran larvae 
(i.e., denticle pattern 2/3; inhabitants of 
small ponds and pools) and serve as 
reference forms for comparison with tad- 
poles of more exotic morphologies and 
ecologies. All other species were selected 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 3 


either because they came from families 
whose larvae have been associated with 
interesting taxonomic problems or be- 
cause they present unusual larval ecolo- 
gies. Particular emphasis was placed on 
the frogs assigned to the superorder Ar- 
chaeobatrachia by Duellman (1975); 
thus Ascaphus (Ascaphidae), Alytes, 
Bombina, Discoglossus (Discoglossidae ), 
Rhinophrynus_ (Rhinophrynidae), and 
several pelobatid larvae are described. 
Since the Microhylidae is a family of 
particularly problematic relationships, 
several Microhyla larvae are also de- 
scribed. 


Hylidae larvae were selected because 
of their great ecological diversity (see 
Duellman, 1970) while one member of 
the Centrolenidae and two species of 
Dendrobatidae are described solely be- 
cause of their unusual larval ecology. 

Certain families are not treated here 
because species from these families have 
been described or illustrated elsewhere: 
for example Pipidae (Sokol, 1975, 1977a 
and other references cited therein); Ran- 
idae [Rana agilis (Kratochwill, 1933), 
Rana temporaria (Savage, 1952; De- 
Jongh, 1968), Rana catesheiana (Grad- 
well, 1970, 1972a), Rana fuscigula 
(Gradwell, 1972c)], and the Pseudidae 
(Pseudis paradoxa larvae illustrated by 
W. Parker, 1881). One bufonid, Bufo 
bufo, has been partially described and 
illustrated by Savage, 1952. The avail- 
ability of specimens set some constraint 
on the families that could be studied; 
several major families with tropical dis- 
tributions (e.g., Leptodactylidae and 
Hyperolidae) are left for future work. 

Knowledge of the habitats and feed- 
ing ecology of the larvae studied here 
is summarized below. The species are 


loosely grouped by common features 
either of external morphology or ecol- 
ogy (references for most of these com- 
ments are given in the descriptions). 


BENTHIC LARVAE WITH ENLARGED 
SUCTORIAL MouTHS 


ASCAPHIDAE.—Ascaphus truei: ad- 
here to rocks in streams; well known for 
their large suctorial oral disc and adap- 
tation to torrential habitats. 

PELOPATIDAE. — Leptobrachium 
hasselti: inhabit quiet, clear regions of 
streams, where they graze on algae 
growing on rocks. In terms of body 
shape or tail length, L. hasselti larvae 
are little specialized for stream life, and 
are among the more generalized mego- 
phrynine tadpoles. Leptobrachium osha- 
nensis: more specialized to stream life 
than L. hasselti (Liu, 1950:191-201) 
since larvae have very long, strong tails 
and are good swimmers in running wa- 
ter; they stay on the bottom in shallow 
water and have an expanded oral disc, 
with a large denticle-free area.1 Oreo- 
lalax pingii: similar to Leptobrachium 
larvae in general appearance; described 
as “bottom feeders” (Liu, 1950). Their 
morphology suggests that they are inter- 
mediate in their tolerance for currents 
between L. hasselti and L. oshanensis. 

HYLIDAE.—Hyla mixe: among the 
most highly specialized larvae for stream 
life; have very large oral disc used for 
adhering to rocks in strong currents. 
Ptychohyla leonhardschultzei: inhabit 
small, quiet, peripheral pools in moun- 
tain streams. Smilisca sordida: inhabit 
streams, but only in regions of very 
gentle current; externally, they show 
few of the modifications characteristic of 
stream tadpoles. 


*Liu (1950:200) questioned the common interpretation (after Smith, 1926) that the mouth 
serves an adhesive function. He watched an L. oshanensis larva in his laboratory and noticed that 
when it rested on the bottom, only the tips of the marginal papillae touched the substrate. Water 
flowed into the mouth through the notch at the front and the back of the disc. Liu then concluded 
that the oral disc served the function of raising the head off the bottom to allow for respiration. 
His observations, however, remain inconclusive for the natural situation, because he described the 
action of the disc only in tadpoles confined to quiet water. The disc may still have an adhesive 
function in currents, not only for this species, but for all other species with enlarged oral discs. 


4 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


FUNNEL-MOUTHED TADPOLES 


MICROHYLIDAE.—Microhyla hey- 
monsi: inhabit quiet pools and feed on 
particles at the air-water interface; have 
rapidly vibrating, filamentous tail tip 
and can remain seemingly motionless at 
the surface film for long periods of time. 

PELOBATIDAE.—Megophrys — mi- 
nor: epitomize the surface film-feeding 
way of life; have a huge, upwardly di- 
rected, denticle-free oral disc;? occur in 
slow, flowing water and are, consequent- 
ly, equipped with a long, powerful tail 
which allows them to resist displacement 
downstream. 

HYLIDAE.—Ptychohyla — schmidto- 
rum: moderately specialized for stream 
existence; found in the quieter reaches 
of mountain pools; have expanded oral 
discs with a large denticle-free area. 

DENDROBATIDAE. — Colostethus 
nubicola: In aquaria they swim beneath 
the surface film (Savage, 1968). They 
have a large, anterior-directed oral disc 
with surface papillae which, according 
to Savage, aid in sorting particles from 
the water. Found under vegetation and 
rocks in side pools and rivulets of small 
streams; found in pools sufficiently small 
so as to be occupied rarely by fish. 


“FOSSORIAL’ STREAM FORMS 


CENTROLENIDAE.—Centrolenella 
fleischmanni: extremely elongated larvae 
which lack the suctorial mouth or fun- 
nel of other stream forms; found in 
cracks and crevices amongst the rocks 
and vegetation in shallow streams. 


HYLIDAE, Arboreal Larvae.—Ano- 
theca spinosa: live in the shallow water 
that collects in tree holes; known carni- 
vores with large beaks; specialized for 
feeding on mosquito larvae, other ar- 
thropods, and frogs’ eggs. Hyla dendro- 
scarta: extremely elongated tadpoles 
which live burrowed in the leaf axils of 


bromeliads; macrophagous (according to 
Orton, 1944) but lack the large mouth 
of Anotheca spinosa and presumably are 
more dietary generalists, feeding on 
small fragments of animal and _ plant 
matter that collect in their arboreal 
pond. 


MiIpWATER MACROPHAGOUS LARVAE 


MICROHYLIDAE.—Microhyla berd- 
morei: dispersed throughout the water 
column in quiet pools; lack keratinized 
mouth parts and are obligate feeders on 
fine, suspended matter. Microhyla or- 
nata: found suspended throughout the 
water column; presumably similar in 
their feeding habits to M. berdmorei. 

HYLIDAE.—Agalychnis _ callidryas: 
by rapidly vibrating the pointed tips of 
their tails these larvae can hang sus- 
pended in midwater in ponds; retain the 
typical hylid 2/3 denticle pattern and 
can facultatively graze on substrates be- 
sides feeding on microscopic particles 
midwater (see pers. comm. by McDiar- 
mid in Heyer, 1976:22). 


TEMPORARY PooL DWELLERS, Omnivores 


RHINOPHRYNIDAE. — Rhinophry- 
nus dorsalis: larvae lack keratinized 
mouth parts for biting or scraping, but 
are omnivores able to cannibalize small- 
er individuals and also efficient suspen- 
sion feeders of ultraplanktonic particles. 

PELOBATIDAE.—Scaphiopus bom- 
bifrons: active tadpole found in tempo- 
rary pools and known for rapid develop- 
ment and voracious feeding habits. Al- 
though carnivorous and cannibalistic, 
best considered omnivores; an oral disc 
of moderate size surrounding large and 
powerful beak; members of genus are 


efficient suspension feeders (Richmond, 
1947). 


ADDITIONAL Ponp LARVAE 
DISCOGLOSSIDAE.—Alytes obste- 


* According to Liu (1950:191) the disc folds shut when the tadpole is below the surface, but 
opens to form a funnel at the surface film when the tadpoles are feeding. The funnel may function 
in both feeding and respiration as a surface float; literature on the subject has been reviewed by 


Noble (1927) and Liu (1950). 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 5 


tricans: typical for quiet water tadpoles 
in general appearance and behavior; able 
to stay in midwater by gulping air inter- 
mittently, but feed on vegetation at or 
near the bottom. Alytes cisternasii: pre- 
sumably similar to A. obstetricans. Bom- 
bina orientalis: occur in assorted ponds 
or pools, often adjacent to streams; can 
be found in puddles lacking macroscopic 
vegetation; can be carnivorous. Perhaps 
appropriately viewed as omnivores with 
generalized external structures, capable 
of carnivory and suspension-feeding on 
microscopic particles. Discoglossus pic- 
tus: generalized pond larvae, presum- 
ably similar to Alytes and Bombina. 
HYLIDAE.—Gastrotheca riobambae: 
inhabit shallow, high elevation pools and 
exhibit no unusual behavior for anuran 
larvae. Hyla femoralis: midwater forms 
that live amongst vegetation in pools; 
have exceptionally high tail fin with 
terminal filament, but otherwise “typi- 
cal” in their external morphology and 
feeding habits. Hyla rufitela: live in 
small pools, often overgrown with mac- 
rophytes; have an unusual 2/4 denticle 
pattern but not known to have unusual 
feeding ecology. Hyla ebraccata and 
Hyla sarayacuensis: hide among the 
aquatic plants of the shallow parts of 
ponds, where the adults of the species 
breed; larvae characterized by a reduc- 
tion of the oral disc and the loss of den- 
ticle rows as compared with typical 
pond hylids; members of species group 
feed on submerged leaves or other de- 
bris (Starrett, 1973). Hyla phlebodes: 
found among vegetation in shallower 
parts of pools; have small mouth without 
oral disc and denticle rows; members of 
species group feed on the bottom (Star- 
rett, 1973). Acris crepitans: typical gen- 
eralist tadpole of small ponds and pools 
throughout much of eastern United 
States. Colostethus nubicola: inhabits 


small, rocky ponds; details of feeding 
ecology unknown. 


METHODS 


All tadpoles examined were meas- 
ured from snout to vent and staged ac- 
cording to Gosner, 1960. For each tad- 
pole the floor and the roof of the mouth 
were exposed following a simple dissec- 
tion procedure described by Wassersug 
(1976a). Tadpoles were pinned under 
water in a small tray of darkly stained 
paraffin and examined with a dissecting 
microscope. The larvae were lightly 
stained with either methy] blue or crys- 
tal violet to accentuate surface features, 
specifically mucous-secreting epithelia. 

Ventral and dorsal surfaces were pho- 
tographed on 4” x 5” format through a 
bellows-view camera using a 32 mm 
lens; a few photographs were taken with 
16 or 64 mm lenses. Tadpoles were usu- 
ally pinned to the paraffin at the tail 
only; care was taken not to distort or 
stretch oral surfaces. Camera lucida 
drawings were also made of many of 
the specimens, particularly the smaller 
ones. Either camera lucida drawings or 
photographs accompany the descriptions 
for most species; some descriptions have 
both for purposes of comparison. One 
species is described without photograph 
or illustration because it is nearly identi- 
cal to related forms that are illustrated. 

All illustrations in this descriptive 
section present the buccal floor and buc- 
cal roof as a single plate with the floor 
above and roof below; for drawings, dor- 
sal and ventral surfaces are magnified 
equally and a single scale line (=1 mm) 
is given. For the photographs, a single 
scale line is given if both halves of the 
plate are magnified equally; otherwise 
two separate scale lines are used. 

Just as there are optical limits on 
photographical qualities,? there are hu- 


* An effort was made to obtain scanning electron (SEM) photomicrographs of several small 
specimens because of the potential resolution and depth of field provided by this technique. SEM, 
however, had to be abandoned because it required pre-coating the specimens with a uniform, con- 
ductive metallic layer and the convoluted, intricate structure of the gill filters made such preparation 


impossible. 


6 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


man limits on how accurately certain 
ventral, pharyngeal features—specifical- 
ly, the gill filters—can be drawn. The 
gill filters in many of the drawings are 
slightly stylized; minor discrepancies be- 
tween the descriptions and the drawings 
remain. 

Species descriptions are grouped by 
family. Each description is provided 
with a reference or references that per- 
tain(s) to the species morphology, ecol- 
ogy, or both. The morphological descrip- 
tions follow Wassersug (1976a). Mer- 
istic features are presented for both the 
left and right sides of the buccal roof 
and floor; discrepancies between one 
side and the other, are noted. These 
data, which were originally collected to 
determine whether asymmetries in ex- 
ternal aspects of the respiratory system 
(i.e., position of the spiracle) were re- 
flected internally, give some indication of 
the amount of variation in any species. 
Meristic data from the pharynx are all 
taken from the left side. It can be noted 


Buccal 
Floor Arena 
Papillae 


Marginal 
Projection 


Filter eee pees 
Cavities I, 01,0 


4 ibe e ? Esophageal ——EE 


here that no consistent pattern of left or 
right handedness was found, either in- 
traspecifically (for Hyla regilla; Wasser- 
sug, 1976b) or interspecifically. 

Because of the complex, qualitative 
nature of many of the features described, 
I chose not to use a telegraphic form for 
the descriptive body of the study. Hyla 
regilla, illustrated and described in de- 
tail in Wassersug (1976b), is treated as 
the “typical” pond tadpole and consid- 
ered a reference for comparison with 
the other forms. 

Figures 1 and 2 show nearly all 
structures presented in the species de- 
scriptions. All morphological structures 
described are briefly defined in the 
Glossary. Terminology follows Wasser- 
sug (1976a). 

Two sources are followed for anuran 
systematics. I follow Orton (1957) in 
recognizing four larval types. Type 1 
includes the Pipidae and the Rhinophry- 
nidae; Type 2 consists of the Micro- 


Lower Beak 


Lingual Papillae 
Infralabial Papilla 
Buccal Floor Arena 


Prepocket 
Papillae 


Buccal 
Pocket 


Filter 
Rows 


SEF ete 


Funnel 


Glottis 


Fic. 1.—The floor of the mouth in a Hyla femoralis larva. Most morphological features discussed in 
the text are labelled in this Figure and Fig. 2. The scale line equals 1 mm. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 7 


Prenarial Arena 


Prenarial Papillae 


Postnarial Arena 


Median Ridge 


Buccal Roof 
Arena 


Lateral Pressure f NSS. 
Cushion ae 


Medial Pressure 
Cushion 


Esophagus oar acs REP ES 


Denticle Rows 


Upper Beak 


Narial Valve 
Projection 


Postnarial Papillae 


Lateral Ridge 
Papilla 


<r ——Lateral Roof 
5 Papillae 


Dorsal Velum 


Esophageal Funnel 


Fic. 2.—The roof of the mouth in a Hyla femoralis larva. Most of the morphological features dis- 
cussed in the text are labelled in this Figure or the previous one. The scale line equals 1 mm. 


hylidae only; Type 3 includes the As- 
caphidae and Discoglossidae; and Type 
4 includes all the remaining families. 
Starrett (1973) has given formal names 
to these groups. As I am not convinced 
that this series represents a phylogenetic 
sequence, I am reluctant to follow Star- 
rett's names. Unless otherwise noted I 
follow Duellman (1975) for the basic 


taxonomy of the Anura and recognize 
two broad groups of frogs, the archaic 
(Archaaeobatrachia) and advanced 
(Neobatrachia) forms. Which families 
belong to which group is highly contro- 
versial (cf. Sokol, 1977b) and I deal 
with this question to some extent in the 
discussion sections on systematics and 
evolution. 


DESCRIPTIONS 


ASCAPHIDAE 
Ascaphus truei Stejneger 
(Fig. 3) 


Material—_FMNH 166497 (stage 37, 
sv. 18.0 mm). Bird Tributary to St. Joe 
River, Shoshone Co., Idaho, U.S.A.; July 
22, 1965. 

Reference.—Stebbins, 1951 (p. 192). 

External.—Ascaphus_ truei larvae 
have a large suctorial oral disc, and a 
medial spiracle. This description is 


based on a typical individual with a 
3/11 denticle pattern, of which two up- 
per and three lower rows were multiple. 

Ventral buccal.—The floor of the oral 
cavity is slightly expanded anterolater- 
ally; bilateral infralabial papillae are 
absent. Immediately inside the mouth 
is a single flap-like fold of skin identi- 
fiable as a posteriorly-directed, bilobed, 
oral valve, the free edge of which sup- 
ports nine very small evenly spaced pa- 
pillae. The tongue, a transversely elon- 


MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Fic. 3.—Photographs of the floor (above) and roof (below) of the mouth in an Ascaphus truei 
larva. The scale line in this Figure and all that follow equals 1 mm. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 9 


gate dome, is covered with a dense field 
of several dozen, straight, blunt papillae. 
The majority of these lingual papillae 
are tall and subequal in size, but the 
anterior ones are smaller and thinner 
than the rest. The BFA papillae are 
organized in a broad, V-shaped row that 
begins anterolaterally as prepocket pa- 
pillae. The BFA papillae are eight per 
side, simple, straight and of modest 
size. They become progressively smaller 
posteromedially. Pustulations and pa- 
pillae of any sort are absent from the 
remainder of the buccal floor. Buccal 
pockets are long and shallow and un- 
perforated. The free velar surface is 
comparatively short. Its posterior mar- 
gin is shaped like a very broad “V” with 
a curved apex and it lacks posterior 
projections or a median notch. The velar 
margin forms a rim twice as thick as 
the free velar surface in front of it. The 
margin is curved upward; directed dor- 
sad rather than caudad. It is stiff, but 
no spicule support is evident; its rigid- 
ity results in part from the complete 
anchoring of the ventral velum to the 
dorsal margin of the filter plate on cb. 2. 
In fact, the velar surface extends pos- 
teriorly in thin, tapered bands for half 
the length of these filter plates. Well- 
formed secretory pits could not be re- 
solved, however, the velar margin is 
covered densely with extremely small, 
fine, irregular pittings. The anterior lim- 
its of this unusual mucosa could not be 
determined. 


Ventral pharynx.—The pharynx in 
Ascaphus truei is overall proportionally 
as large as in typical pond tadpoles, but 
the filter plates are rotated outward so 
that cb. 2 is more transversely oriented 
and filter cavity 3 is reduced in volume 
on each side. Filter plates of cb. 2 are 
straight while those of cb. 3 are strongly 
bowed, obscuring much of filter cavity 
3 from dorsal view. Counts for number 
of filter rows on cb. 1-4 are 9, 11, 10, 6 
respectively. The filter mesh of A. truei 
is much reduced. Secondary filter folds 
are short; tertiary folds are short and 


few. Neighboring filter rows do not 
abut, so filter canals are large, open 
channels, almost as wide as the filter 
row. The branchial food traps are small 
but well demarcated ventrally, where a 
sharp rim separates them from the filter 
tissue. Secretory ridges could not be 
resolved and must be extremely fine or 
absent. The glottis is a tiny, unper- 
forated slit under the velar margin. 
Glottal lips and a conspicuous laryn- 
geal disc are absent. The esophageal 
funnel has a narrow dorsal profile. 
Dorsal buccal.—The buccal roof is 
dwarfed in comparison to the huge up- 
per beak, and much narrowed anteriorly. 
The roof is transversely arched and has 
considerable depth compared to other 
species. The oral opening is small and 
positioned so far posterior that the pre- 
narial arena is also “preoral.” The arena 
is a slim, blind tunnel above the upper 
beak. It is devoid of papillae or pustu- 
lations. The internal nares are oblique, 
nearly longitudinally oriented slits and 
are on the lateral walls rather than on 
the roof of the buccal cavity. The an- 
terior narial walls make an anterolateral 
loop that bounds small, oval, oblique 
pits on each side. The walls surround- 
ing these pits are thick and pustulate. 
The remainder of the anterior walls are 
shallow and lack papillae. The posterior 
narial walls are tall and slightly pustu- 
late, but lack narial valve projections. 
All major landmarks of the buccal roof 
proper, such as the median ridge and 
papillae that could outline the post- 
narial and buccal roof arenas, are ab- 
sent. There are approximately ten small, 
simple, blunt papillae scattered about 
the buccal roof behind the nares; these 
show some tendency to be concentrated 
in the posterolateral corners of the buc- 
cal roof. The buccal roof sinks, then 
rises again just anterior to the dorsal 
velum, and forms a major, posteriorly 
directed V-shaped depression. The an- 
terior surface of this depression is lined 
with a few blunt, well-spaced pustula- 
tions. The glandular zone begins at the 


10 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


lowest point of this depression. The 
zone is of uniform anterior-posterior 
length except directly in front of the 
esophagus, where it is much abbreviated. 
The zone is made up of irregularly 
spaced, minuscule secretory pits that are 
extremely dense. The dorsal velum is 
short and thick with a smooth ventral 
margin. It is continuous across the mid- 
line, but its full extent in that region 
cannot be determined because of dam- 
age in dissection. 

Dorsal pharynx.—The two pairs of 
pressure cushions in A. truei are of typi- 
cal size and dimension. They are, how- 
ever, covered with a few faint, scattered 
pustulations along with a dense secre- 
tory epithelium. The ciliary groove is 
broad and shallow. 


Diagnostic summary.—The oral cav- 
ity of Ascaphus truei differs from that 
of all other tadpoles examined in the 
following, unique characters: presence 
of oral valve instead of paired infra- 
labial papillae; many papillae covering 
tongue anlage; prenarial arena reduced 
to short, blind tunnel; sensory pits an- 
terolaterally on the internal nares; large 
V-shaped depression extending across 
posterior buccal roof. A combination of 
the following features further differen- 
tiates the oral cavity of Ascaphus larvae 
from all other tadpoles: ventral velum 
fully attached to dorsal margin of filter 
plates; secretory tissue without pitted or 
ridged pattern; unperforated glottis; ab- 
sence of prenarial, postnarial and buc- 
cal roof arenas. 


DISCOGLOSSIDAE 
Alytes obstetricans Laurenti 
(Figs. 4, 5) 

Material—Uncatalogued, author's 
collection (stage 36, sv. 18.5 mm). Col- 
lected at “La barlieu,” 30 km SE of 
Lyon, Rhéne, France; no date. 

References.—Eibl-Eibesfeldt, 1953; 
Magnin, 1959. 

External_—Alytes obstetricans tad- 
poles have a 2/3 denticle pattern with 


the inner upper and two inner lower 
rows multiple. The spiracle is medial. 

Ventral buccal.—The floor of the oral 
cavity is relatively wide anteriorly and 
rounded rather than pointed. The buc- 
cal floor is flat but slopes forward so 
that the frontal plane of the mouth 
meets the frontal plane of the trunk at 
an angle; these planes are more nearly 
parallel in most tadpoles. Infralabial 
papillae are tall, narrow and have a fine 
serrated margin rather than secondary 
papillae. The transversely elongate 
tongue anlage lacks lingual papillae but 
is covered fully by a dense field of stout, 
blunt pustulations. The buccal floor 
arena is surrounded by eight BFA pa- 
pillae on one side and ten on the other. 
These BFA papillae are less numerous 
in front of the buccal pockets, more nu- 
merous posteriorly and slightly more 
attenuate and acutely pointed when 
compared to those of typical pond tad- 
poles. A couple of the largest papillae 
are laterally compressed and have ter- 
minal bifurcations. There are two dis- 
tinct conical prepocket papillae on each 
side. Some extremely tiny, pointed pa- 
pillae are clustered directly anterior to 
cb. 2 on the buccal floor, and there is a 
single small papilla above the lateral 
arm of the ceratohyal on each side in 
this specimen. Dozens of tiny, precise 
pustulations cover the buccal floor, with 
their greatest concentration anterolateral 
and posteromedial. Each buccal pocket 
has a large transverse pouch with a 
heavy fold rising off its posterior wall 
obscuring its floor. No open slit could 
be found within the pockets. The free 
velar surface is reduced. Spicular sup- 
port is lacking except for a tiny spur of 
cartilage at the points where the dorsal 
edge of the filter plates on cb. 3 meet 
the velum. The velar margin is fully 
fused to the top of the filter plates on 
cb. 2 and fused almost to its edge above 
cb. 3. The velum edge is curved up- 
ward above cb. 2. The velar margin is 
concave posteriorly between filter plates 
and when viewed from above appears 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 11 


Fic. 4.—Drawings of the floor (above) and roof (below) of the mouth in an Alytes obstetricans 
larva. 


12 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


as a series of crescents rather than a 
continuous posteriorly convex arc. Pro- 
jections of the velar margin are dis- 
placed medially and are pointed rather 
than curved. The projections associated 
with the third filter cavity and the me- 
dial notch are long, finger-like papillae. 
Secretory pits are restricted to the pos- 
terior projections of the velar margin. 
The pits are small and of low density. 
Ventral pharynx.—The branchial bas- 
kets of A. obstetricans are round, long, 
wide and deep. The first and second 
filter cavities are disproportionately 
large compared to the third. The dorsal 
margins of the filter plates of cb. 3 arch 
upward and blanket much of the third 
filter cavity. Although the filter plates 
are relatively long in A. obstetricans, 
the number of filter rows and density of 
the filter mesh is low. Counts for filter 
rows are 6, 9, 8, 5, for cb. 1-4. The filter 
mesh is comparatively reduced because 
the rows are narrow, and secondary 
and tertiary filter folds, while present, 
are very thin. The filter canals are 
large, nearly fully open channels. The 
short, flexible surface of the velum is 
covered ventrally by small, scattered se- 
cretory pits rather than by secretory 
ridges. These pits grade into weak 
“ridges” farther forward under the im- 
mobile portion of the velum. The 
“ridges,” however, are not well-formed 
and have a rather streaked appearance. 
The area covered by the secretory tissue 
is rather small and sharply separated 
from the filter epithelium by a rim 
around the food traps in each filter cav- 
ity. The glottis is 90% visible from above. 
It is long with large thick lips. The 
laryngeal disc, however, is not excep- 
tionally large. The esophageal funnel is 
large and broad in dorsal profile. 
Dorsal buccal—The buccal roof is 
broad anteriorly and has a “V” shaped 
posterior pharyngeal margin. The pre- 
narial arena is as large as that of the 
typical pond hylids examined, but ap- 
pears shorter in photographs because it 
is partially obscured by a posterior re- 


flection of the upper labial cartilage and 
beak. The prenarial arena is devoid of 
any papillae, pustulations, ridges or 
other projections. The internal nares of 
A. obstetricans are elongate, obliquely 
oriented slits. Their anterior walls each 
have a low, tiny, anteriorly-directed flap 
at their most anteromedial corners. Nor- 
mal prenarial papillae are absent, how- 
ever; the anterior narial walls abruptly 
expand into huge flaps that extend back 
the length of the nares. These flaps curl 
posteromedially under the narial open- 
ings. The tall, anteromedial edges of 
these flaps are serrated. The coiled, pos- 
terior margins of the flaps have a few 
small, pointed, irregular secondary pa- 
pillae on each side. The posterior narial 
walls have valves but lack any narial 
valve projections. Eight small, pointed 
cones of subequal size make up the post- 
narial papillae series. These papillae are 
clustered in a rather transverse patch in 
the middle of the postnarial arena. The 
median ridge is an anteriorly directed, 
triangular flap. The anterior surface of 
the median ridge is smooth but two tiny 
cusps line the lateral margins of the 
ridge on each side. Separate lateral ridge 
papillae are absent. The BRA is an 
elongate oval defined by approximately 
six, well-spaced papillae on one side and 
five on the other in this specimen. The 
BRA papillae are all small, pointed 
cones of subequal size. In addition to 
the main rows of papillae, there are two 
single papillae lateral to the arena on 
each side and an arc of tiny conical 
papillae at the posterolateral margins 
of the buccal roof. Randomly dispersed 
within the BRA are 50-60 tiny pustula- 
tions. Pustulations are absent outside of 
the arena. The glandular zone is nor- 
mal. The dorsal velum is short and not 
tightly coiled. It is completely and 
broadly interrupted medially and its me- 
dial edges are papillate on each side. 

Dorsal —pharynx.—Although — two 
broad, shallow waves of the velar sur- 
face are faintly visible in this specimen, 
pressure cushions are essentially absent. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 13 


Fic. 5.—Photographs of the floor (above) 


and roof (below) of the mouth in an Alytes obstetricans 
larva. 


14 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


The ciliary groove is similar to or slight- 
ly broader than that of a typical pond 
larvae. 

Diagnostic summary.—Tadpoles of 
the family Discoglossidae can be dis- 
tinguished from those of all other fami- 
lies by the presence of large flaps of the 
anterior narial walls which curl posteri- 
orly under the nares and their V-shaped 
posterior pharyngeal margin. All dis- 
coglossid tadpoles have the posterior, 
ventral velar margin ankylosed to the 
dorsal margins of the filter plates and 
appear to lack well organized secretory 
ridges in the branchial food traps. In- 
ternally Alytes tadpoles differ from other 
discoglossid larvae in the presence of 
four long projections of the ventral vel- 
um near the midline and pustulations 
covering the tongue anlage. Compared 
to A. cisternasii, A. obstetricans tadpoles 
lack paired lingual papillae, but have 
an oval-shaped buccal floor arena, short- 
er, fewer BFA papillae and _ slightly 
denser gill filters. A. cisternasii larvae 
tend to have more papillate and crenu- 
lated structures than the larvae of A. 
obstetricans; this can be seen in the in- 
fralabial papillae, the narial flaps and 
the median ridge. A. obstetricans, how- 
ever, is the only discoglossid examined 
with papillation of the median edge of 
the dorsal velum. 


Alytes cisternasii Bosca 
(Fig. 6) 


Material.—Author’s collection (stage 
34; sv. 15.5 mm). Collected in a perma- 
nently swampy area, in the province of 
Alto Alentejo, Portugal; no date. 

Reference.—Boulenger, 1891. 

External.—Alytes cisternasii have 
large robust larvae which in terms of 
size, shape and denticle pattern are vir- 
tually identical to A. obstetricans. The 
spiracle is medial. Boulenger (1891) 
could detect no characters that distin- 
guish the larvae of these two species but 
Crespo (pers. comm.) finds that in Por- 
tugal A. cisternasii larvae have a larger 
spiracle with a shorter, free terminal 


sleeve and generally fewer denticles than 
A. obstetricans larvae. The specimen de- 
scribed here had a 2/3 denticle pattern 
with all rows multiple. 

Ventral buccal.—(Unless otherwise 
noted A. cisternasii is identical to A. 
obstetricans). A. cisternasii has several 
pustulations inside the oral orifice an- 
terior to the infralabial papillae. Among 
the pustulations on the tongue anlage, 
two near the midline are particularly 
enlarged and, as such, resemble in shape 
and position the typical paired lingual 
papillae of most anuran larvae. Buccal 
floor arena papillae are taller and more 
numerous in this species. I count nine 
major papillae on one side, twelve on 
the other with an equal number of 
minor papillae as well as many small 
pointed pustulations in the posterior 
part of the arena and interspersed among 
the papillae forming the margins of the 
arena. There are several small conical 
papillae in front of the pockets on each 
side. Papillae posterior to the BFA are 
lacking. There appears to be less spicu- 
lar support than in A. obstetricans. Sin- 
gle secretory pits could not be discerned 
even at 100. 

Ventral Pharynx.—The branchial bas- 
kets are slightly more triangular in dor- 
sal view than those of A. obstetricans 
and perhaps a bit smaller in comparison 
to the size of the buccal cavity overall. 
Counts for filter rows run 9, 10, 8, and 
5 for cb. 1-4 respectively, which means 
that for individuals of comparable size 
and stage A. cisternasii has slightly more 
filters rows than A. obstetricans. In con- 
trast, filter mesh appears slightly denser 
than in A. obstetricans but this differ- 
ence may be due to preservation. The 
filter canals are necessarily less open in 
this specimen than in the specimen of 
A. obstetricans described above. No se- 
cretory ridges were evident in the 
branchial food traps. The glottis is less 
than 50% visible from above. The glottal 
lips are not as thick as in A. obstetricans. 


Dorsal buccal.—The flaps at the an- 
teromedial corners of internal nares in 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 15 


Fic. 6.—Drawings of the floor (above) and roof (below) of the mouth of an Alytes cisternasii larva. 


16 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


A. obstetricans are present in this species 
but much reduced. The large postero- 
medially directed flaps that arise from 
the anterior margin of the narial walls 
have a more distinctly papillate margin 
in this species compared with A. obstet- 
ricans. There are nine pointed cones 
within the postnarial arena. The median 
ridge in this species has a more crenu- 
late border than in A. obstetricans. The 
BRA is defined by approximately seven 
papillae on one side and five on the 
other in this specimen. The BRA papil- 
lae are more attenuate overall. The me- 
dial edges of the dorsal velum lack 
papillae but are pustulate in this speci- 
men. Most other differences illustrated 
in Figs. 4 and 6 reflect differences in dis- 
section. 

Dorsal pharynx.—As in A. obstetri- 
cans. 

Diagnostic summary.—See Alytes ob- 
stetricans (p. 14). 


Bombina orientalis Boulenger 
(Fig. 7) 


Material—Uncatalogued, author's 
collection (stage 34, sv. 14.1 mm). Lab- 
oratory raised population, Amphibian 
Facility, University of Michigan, Ann 
Arbor, Michigan. Initial stock from Ko- 
rea with no additional data. 

Reference.—Okada, 1931. 

External——Bombina orientalis larvae 
have a 2/3 denticle pattern with the in- 
ner rows double. The spiracle is medial, 
characteristic of discoglossid tadpoles. 

Ventral buccal—The floor of the 
mouth of B. orientalis is proportionally 
longer than that of A. obstetricans with 
the increase solely in the pharynx. As in 
A. obstetricans, the floor tips anteriorly 
downward from the frontal plane of the 
trunk. The mouth is anteriorly narrower 
than in A. obstetricans. A single, strong- 
ly compressed, dorsally projecting in- 
fralabial papilla is situated on each side 
in the typical position. These are not as 
tall as in A. obstetricans. Anteriorly, the 
papillae grade into a series of pustula- 
tions that are continuous across the mid- 


line. Posteriorly, the papillae degenerate 
into a jagged fringe. The free edges of 
the infralabial papillae are lined with 
short, attenuate, secondary papillae, 
eight on each side in this specimen. The 
single, median lingual papilla, situated 
far anterior, is anteriorly to posteriorly 
flattened and slightly concave posterad. 
It has a deep median notch and small, 
secondary, terminal subpapillate projec- 
tions. Posterolaterally on each side of 
the tongue anlage are single, large, coni- 
cal pustulations. The papillae of the 
BFA are much reduced in size and 
aligned in a “V” with the arms of the 
“V” forward. Anteriorly each arm turns 
laterally so the BFA papillae series is 
continuous with the prepocket papillate 
series. I counted six BFA papillae on 
one side and seven on the other, with 
an additional three papillae on each side 
in the prepocket papillae position. The 
prepocket papillae and the largest BFA 
papilla on each side have fine secondary, 
pointed, terminal papillae. Other papil- 
lae in the BFA row are simple, small, 
laterally compressed cones that grade 
down to pustulations. There are a few 
tiny pustulations near the prepocket pa- 
pillae, but the remainder of the buccal 
floor is smooth and free of any additional 
papillae or pustulations. The buccal 
pockets are unperforated and have the 
same orientation and length to width 
ratio as in A. obstetricans. They are, 
however, shallower. The short, free velar 
surface lacks spicular support but is 
quite inflexible due, in part, to the fact 
that it is fused almost to its margin at 
the top of the filter plate of cb. 2. The 
velar margin is posteriorly concave be- 
tween filter plates, giving the free velar 
surfaces the appearance of a series of 
crescents as in A. obstetricans. These 
projections are directly above the filter 
plates. The median portion of the velum 
forms a discreet series of waves and 
lacks a large median notch. The edge of 
the velum is curled up to form a thick- 
ened rim continuous from one side to the 
other similar to, but not as large as, the 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 17 


Fic. 7.—Drawings of the floor (above) and roof (below) of the mouth of a Bombina orientalis larva. 


18 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


raised velar margin in Ascaphus truei. 
The rim has the buffed texture charac- 
teristic of secretory tissue but distinct 
secretory pits could not be resolved. 

Ventral pharynx.—The branchial bas- 
kets are slightly taller and longer than 
those of A. obstetricans and much taller 
and longer than those of typical pond 
tadpoles. They are tallest posterolater- 
ally, where they extend well above the 
plane of the floor of the mouth. Com- 
pared with A. obstetricans, the third fil- 
ter cavities of B. orientalis show further 
reduction in volume at the expense of 
larger second filter cavities. The filter 
plates, particularly those of cb. 3, are 
imbricated to a great extent in B. ori- 
entalis. Counts for numbers of filter 
rows run 8-9, 11, 10, 6 for cb. 1-4. The 
filter mesh is similar to that of A. ob- 
stetricans. Filter rows are unusual in 
being very uniform in thickness and 
little expanded ventrally. The main fold 
is very straight and second folds are 
single, short crossbars on the main fold. 
Filter rows do not abut and the filter 
canals are large, open channels. Branch- 
ial food traps are extremely short and 
separated from the filter surfaces ven- 
trally by a large, distinct rim. Secretory 
ridges could not be resolved anywhere 
in the food traps even at 75x. They are 
presumably absent; nevertheless. the 
food traps are covered with a buff-tex- 
tured tissue that has the superficial char- 
acteristics of a glandular mucosa. The 
glottis of B. orientalis is larger than that 
of typical pond tadpoles, but smaller 
than that of A. obstetricans. The glottal 
lips are thickest in B. orientalis. The 
laryngeal disc is present, but not very 
distinctive. Fifty percent of the glottis 
is obscured from dorsal view by the mar- 
gin of the velum. The esophageal funnel 
has a narrow profile. The bore of the 
esophageal funnel, however, is large in 
B. orientalis and comparable to that of 
A. obstetricans. 

Dorsal buccal_—B. orientalis shares 
with A. obstetricans the V-shaped poste- 
rior margin of the pharynx. In B. orien- 


talis the pharynx is longer and narrow- 
er such that the roof of the mouth is 
more diamond-shaped rather than typi- 
cally triangular in ventral profile. The 
prenarial arena is like that of A. obstet- 
ricans, except that B. orientalis lacks the 
dorsal reflection of the beaks. The in- 
ternal nares are so similar in A. obstetri- 
cans and B. orientalis, that only differ- 
ences are presented here. In place of 
the anteriorly directed flaps of the an- 
terior walls are a cluster of minuscule 
attenuate papillae on each side in B. 
orientalis. The huge flaps of the anterior 
narial walls are larger in B. orientalis 
than in A. obstetricans, but are more 
uniform in width. They extend postero- 
medially over a greater portion of the 
postnarial arena. Two small, conical pa- 
pillae with fine apical irregularities are 
present in the small postnarial arena in 
a transverse line just anterior to the 
median ridge. The median ridge is iden- 
tical to that of A. obstetricans, except 
that fine secondary cusps are clustered 
on its ventral corner rather than along 
the lateral margins. Lateral ridge papil- 
lae and BRA papillae are absent. Ran- 
domly dispersed about the buccal roof 
between the median ridge and the glan- 
dular zone are a half dozen tiny, pointed 
papillae and an equal number of faint 
pustulations in this specimen. The glan- 
dular zone is of similar proportions to 
that of A. obstetricans but the anterior 
margin is V-shaped, paralleling the back 
of the pharynx. Well-defined secretory 
pits are visible only at the anterior mar- 
gin of the zone. The dorsal velum is 
slightly longer than that of A. obstetri- 
cans, but still shorter than that of pond 
hylids, at least laterally. The velum is 
broadly interrupted medially and has 
smooth medial margins reflected poste- 
riorly. 

Dorsal pharynx.—The pressure cush- 
ions are normal in terms of general size 
and depth, though the medial pressure 
cushions are slightly narrower. The cil- 
iary groove is similar to, but even broad- 
er than that of A. obstetricans. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 19 


Diagnostic summary.—The following 
set of oral characters readily distin- 
guishes Bombina orientalis larvae from 
all other discoglossid tadpoles studied: 
single, medial bifurcated, lingual papil- 
lae; buccal roof arena absent; secretory 
pits present only at the anterior margin 
of glandular zone. 


Discoglossus pictus Otth 
(Fig. 8) 


Material.—Author’s collection (stage 
33; sv. 7.7mm). Laboratory raised spec- 
imens, initially collected near Lyon, 
France; no date. 

Reference.—Boulenger, 1891. 

External.—Discoglossus pictus larvae 
have a single, medial spiracle and a rela- 
tively ventral mouth with two upper and 
three lower denticle rows. Typically all 
denticle rows are double but the outer 
rows may be single. 


Ventral buccal—D. pictus is de- 
scribed by comparison with Alytes and 
Bombina. The floor of the oral cavity 
is more pointed in D. pictus than in A. 
obstetricans and in general proportions 
more closely resembles Bombina than 
Alytes. Infralabial papillae are similar 
to those of Alytes only smaller. A few 
pustulations occur around the base of 
these papillae. The tongue anlage has 
six papillae with robust bases and 
pointed apices arranged in an arc with 
the most medial papillae most anterior. 
The medial two papillae are fused at 
their base and resemble the single lin- 
gual papilla seen in Bombina. A half 
dozen tiny pustulations occur anterior to 
the lingual papillae on the tongue an- 
lage. The BFA is surrounded by 8 major 
papillae on each side, aligned in straight 
rows beginning at the medial margin of 
the buccal pocket and converging pos- 
teriorly, and an equal number of minor 
papillae near their bases. The BFA 
papillae are simple, attenuate cones 
lacking any major secondary pustula- 
tions or papillae. The arena is particu- 
larly well demarcated anteriorly because 
of large pustulations that run on each 


side in a row from the lateral edge of 
the tongue anlage to the medial edge of 
the buccal pockets. There are no pustu- 
lations within the BFA and only a few 
very tiny pustulations on the buccal 
floor anterior to the pockets. Prepocket 
papillae are absent. Buccal pockets are 
shallow and unperforated, as in Bom- 
bina. The free velar surface has a thick 
glandular edge with a cusp projection 
over each filter cavity. The median 
notch is very weak. 

Ventral pharynx.—The branchial bas- 
kets have similar shape and proportions 
to those of Alytes. Counts for numbers 
of filter rows run 9, 10, 9, and 5 for cb. 
1-4. In all details of the filter rows and 
filter plates D. pictus is virtually indis- 
tinguishable from Alytes and Bombina. 
The branchial food traps are covered 
with small but distinct secretory pits, 
which do not appear to be organized as 
rows or ridges anywhere in the nharvnx. 
The glottis is not strongly elevated; its 
lips are thick, but not as thick: as taose 
of Bombina. Approximately one-third of 
the glottis lies under the ventral velum. 
In overall proportions the glottis of D. 
pictus is most similar to that of Alytes 
cisternasii among its near relatives. D. 
pictus shares a large esophageal funnel 
with other discoglossoid tadpoles. 

Dorsal buccal.—In shape and _ pro- 
portions the roof of the buccal cavity is 
similar to that of Bombina. The rostrum 
is not strongly turned ventrally and the 
prenarial arena is devoid of any surface 
irregularities, pustulations or papillae. 
The large flaps that arise from the an- 
terior wall of the nares in other disco- 
glossoids are present in D. pictus. They 
have a jagged, irregular free margin and 
are not as large as in Alytes or Bombina. 
The median ridge is similar to that of 
Bombina. There are a few pustulations 
in the postnarial arena but papillae in 
that region, as well as lateral to the me- 
dian ridge, are absent. Ten simple pa- 
pillae arranged roughly in an are (with 
the crown most posterior) define the 
posterior edge of the BRA. A few tiny 


20 


MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Fic. 8.—Drawings depicting D. pictus; floor of mouth (above) and roof 


(below). 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 21 


papillae on one side define the lateral 
margin of the BRA. Several dozen small 
pustulations occur within the BRA 
among the papillae that define it. Lat- 
eral to the BRA the buccal roof is 
smooth with the exception of a small 
longitudinally oriented pustulate ridge 
in the far lateral margin of the buccal 
roof on each side. The glandular zone 
has a V-shaped anterior margin and is 
of uniform length. Small tightly packed 
secretory pits are visible everywhere on 
the glandular zone. The dorsal velum is 
short, weakly curved and broadly inter- 
rupted on the midline. 

Dorsal pharynx.—The dorsal pharynx 
is indistinguishable from that of A. ob- 
stetricans. 

Diagnostic summary.—The great sim- 
ilarity of all the discoglossid tadpoles 
masks their minor differences. Only triv- 
ial features distinguish D. pictus from 
the Bombina and Alytes larvae exam- 
ined; e.g., papillation is generally greater 
than in Bombina, the arch of papillae on 
the tongue enlage is different than either 
Alytes examined, the sculpturing of the 
ventral velar margin is least in this 
genus, etc. Many of these differences 
may be size related. 


RHINOPHRYNIDAE 
Rhinophrynus dorsalis Dumeril & Bibron 
(Fig. 9) 


Material—Uncatalogued, author's 
collection (stage 36, sv. 15.5 mm). Col- 
lected in shallow pools of flooded drain- 
age ditch along Highway 190 approxi- 
mately 1.0 km north of Tehauntepec, 
Oaxaca Province, Mexico; July 17, 1970. 

References——Orton, 1943; Stuart, 
1961; Starrett, 1973; Wassersug, 1972. 

External.—Rhinophrynus larvae have 
a wide oral orifice that lacks an oral 
disc. Keratinized structures, both denti- 
cles and beaks, are absent. There is a 
single, short barbel that extends forward 
from the middle of the lower lip in the 
larvae from this population. Spiracular 
openings are paired and bilaterally sym- 
metrical in the genus. 


Ventral buccal.—The wide oral open- 
ing of this larva gives the front of the 
mouth a gently curved outline in dorsal 
view. On the lower lip, near the mid- 
line, are two pairs of tiny, thin papillae 
and some extremely small pustulations. 
Because the lower lip curls backward 
inside the mouth, the more ventral pair 
of papillae are obscured from view with- 
out stretching the lips. Lingual papillae 
are absent. BFA papillae are clustered 
along a broad arc, all well behind the 
buccal pockets; they do not circumscribe 
an “arena per se. These papillae, seven 
per side, are simple, anteriorly curved 
structures, large and of subequal size. 
The largest papillae are nearest the mid- 
line and displaced anterad. None of 
the papillae are bifurcated, although two 
on each side show some basal fusion. 
Within the arc of the BFA are two, 
small, attenuate papillae behind the buc- 
cal pockets near the midline. Prepocket 
papillae are absent. There are no papil- 
lae anywhere anterior to the buccal 
pockets, but above the massive cerato- 
hyals are numerous, randomly scattered 
pustulations and subpapillate projec- 
tions. The buccal pockets are very wide, 
deep and not long. They do not appear 
perforated. There is a large single fold 
of epithelium arising from the floor of 
the pockets. The free velar surface is 
not large. It is unusual in that it is 
fused medially with the dorsal margins 
of the fourth filter plates rather than 
continuous across the midline. The velar 
margin is also fused to the top of the 
filter plates on cb. 2 but is free for a 
short distance along cb. 3. Spicular sup- 
port for the velar margin is absent and 
the velar surface tends to arch upward 
above each filter cavity. Except for the 
points of attachment, the velar margin 
forms a rather smooth arch when viewed 
from above. Posterior projections of the 
velum above the filter cavities and in 
the median notch are absent. Secretory 
pits could not be resolved anywhere 
along the posterior velar margin. The 
division of the velum into two parts 


MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


22 


yo 


4 


below) of the mouth of a Rhinophrynus dor- 


salis larva. 


Fic. 9.—Photographs of the floor (above) and roof ( 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 23 


means that the glottis is fully exposed 
above the velum and is within the buc- 
cal cavity. The glottis has a predomi- 
nantly horizontal orientation, but is 
tipped slightly forward. A slight de- 
pression makes the anterior margin of 
the glottis stand out from the buccal 
floor. The glottal lips and laryngeal disc 
are comparable in size and shape to 
those of typical pond tadpoles. 


Ventral pharynx.—The branchial bas- 
kets have a “typical” length to width 
ratio, but encompass an atypically large 
proportion of the total oral volume. The 
first and second filter cavities are strong- 
ly bowed outward; the third filter cavity 
is round and very open when viewed 
from above. The extreme curvature of 
the filter plates make it difficult to esti- 
mate height to length ratios. The rela- 
tively tall filter plates are also thick. 
These plates are arched through the 
transverse plane in such a manner that 
they are more imbricated away from 
the midline than toward the midline. 
Counts for filter rows run 11], 13, 13, 8 
for cb. 1-4. The filter ruffles are very 
dense. Secondary and tertiary filter folds 
are very long and higher order folds 
are numerous. Full filter rows are very 
wide, particularly ventrally. The filter 
cavities are tall, nearly completely cano- 
pied passages. They were packed with 
fine silt in this specimen and other indi- 
viduals from this population. With care- 
ful manipulation it is possible to curl 
forward the velar margin and expose 
some of the secretory ridges of the 
branchial food traps. The food traps 
are quite tall and narrow, but their full 
area cannot be determined without sec- 
tioning. The secretory ridges are not 
continuous from one filter cavity to the 
next. The ridges are tall, thin, and 
highly regular. The esophageal funnel 
is of normal proportions. 

Buccal roof.—The roof of the oral 
cavity is far more squarish and less tri- 
angular in R. dorsalis than in any beaked 
tadpole. The buccal roof is basically flat, 
but has some depressions and ridges not 


observed in any other species. Smooth 
transversely oriented ridges occur on 
each side of the roof formed by an 
abrupt rise just posterior to the articu- 
lating surface of the ceratohyal with the 
palatoquadrate bars. These ridges, 
which do not cross the midline, align 
with buccal pockets below. More pos- 
teriorly within the glandular zone are 
three concavities of the roof on each 
side. These correspond to the upwell- 
ings of the free velar surface over each 
filter cavity. The middle concavity on 
each side is the largest and has a steep, 
transverse anterior wall. The lateral and 
medial pairs are more gentle, shallow 
hollows. The internal nares are tiny and 
far anterolateral; a prenarial arena is 
ill-defined. Anterior to the nares are a 
half dozen small, scattered pustulations. 
The narial openings are small transverse 
ovals, slightly larger medially than lat- 
erally. Their anterior and medial mar- 
gins are weakly defined and have a faint 
texture. Prenarial papillae are absent. 
The posterior narial walls are thin flaps, 
more horizontal than vertical, and about 
as wide as tall (vide long). Narial valve 
projections are absent. The median ridge 
and papillae of the buccal roof that 
demarcate postnarial and buccal roof 
arenas in other tadpoles are absent. Sev- 
eral dozen pustulations are dispersed 
over the buccal roof between the nares 
and the glandular zone. There are also 
two pairs of small, blunt papillae in the 
midportion of the buccal roof halfway 
between the midline and lateral margin; 
they are the only symmetrical projections 
of the buccal roof. The glandular zone 
is made up of dense, moderately large 
and conspicuous secretory pits that ex- 
tend back onto the pressure cushions. 
Although the anterior margin of the 
zone is quite wavy, the zone is of gen- 
erally uniform length except directly in 
front of the esophagus where it abruptly 
decreases to a negligible length. The 
average length of the zone is equal to or 
slightly greater than in typical pond 
larvae. The dorsal velum is shorter and 


24 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


not particularly coiled. It is rather stiff, 
projects directly anterad, and is absent 
on the midline. 

Dorsal pharynx.—As viewed from 
above, the dorsal pharynx on each side 
is shaped like a right, isosceles triangle 
with sides posterior and lateral, and the 
hypotenuse facing anteromedial. The 
pharynx is larger than in any beaked 
tadpole, with pressure cushions extend- 
ing back well past the posterior margin 
of the dorsal velum. There are three 
pairs of very large, distinct pressure 
cushions. The most lateral pairs are tall, 
oblique ridges with sharp ventral edges. 
The middle pair are oriented like the 
lateral pair, but are not as tall and have 
a posterior knob rather than a continu- 
ous sharp edge. The medial pair are 
large, round swellings. Behind the pres- 
sure cushions is a relatively narrow and 
shallow ciliary groove. 


Diagnostic summary.—The full at- 
tachment of the ventral velum to the 
tops of all filter plates, including those 
of cb. 4, immediately separates Rhi- 
nophrynus from all other species. The 
following is a partial list of other fea- 
tures which, in combination, diagnose 
Rhinophrynus: lingual papillae absent; 
massive prepocket buccal surface; buc- 
cal floor arena papillae in broad arch; 
glottis exposed on posterior buccal floor; 
absence of velar marginal projections; 
massive branchial baskets; well-devel- 
oped secretory ridges; absence of all 
features that would outline postnarial 
and buccal roof arenas; presence of lat- 
eral depressions of buccal roof above 
buccal pockets. 


MICROHYLIDAE 
Microhyla berdmorei (Blyth) 
(Fig. 10) 

Material—FMNH 187567 (stage 32, 
sv. 10.8 mm). Rock pool adjacent to 
flowing stream at Khao Yai, Nakhon 
Ratchasima, Thailand; January 5, 1969. 

References.—Inger, 1966 (p. 150); 
Heyer, 1973. 


External._—Microhyla berdmorei tad- 
poles are of medium size for the genus. 
As is characteristic of the family, they 
lack hard mouth parts and have a me- 
dial spiracle. In M. berdmorei the tail 
is pointed but lacks a vibrating filamen- 
tous tip. 

Ventral buccal—tThe floor of the 
mouth is shaped somewhat like a trape- 
zoid (base posterior) with the length 
and width approximately equal. The 
buccal floor behind the buccal pockets 
is extremely elongate. The ceratohyals 
are so short, compared to the length of 
the branchial baskets, that only a fifth 
of the mouth is anterior to the buccal 
pockets. The tiny lower lip is directed 
dorsally. There are three small, blunt 
infralabial papillae on each side con- 
centrated within the loop of the lower 
lip. The most anterodorsal papilla is 
a small, simple knob-like projection. 
There is an anteroventral papilla slightly 
more attenuate. The third papilla is pos- 
terior. It is a thick, transversely oriented 
flap that has a common ventral base 
with the anteroventral papilla. The pos- 
terior infralabial papillae from each side 
abut at the midline, obscuring much of 
the oral orifice. The tongue anlage is an 
elongate oval devoid of lingual papillae. 
The BFA papillae are arranged along a 
semicircle which is open anteriorly. The 
front edge of the arena is formed by the 
buccal pockets. The arena is substan- 
tially wider than the relatively short in- 
terpocket distance, but limited to the 
anterior half of the portion of the buccal 
floor behind the pockets. The papillae 
are small, attenuate cones, evenly spaced 
and lacking terminal irregularities. I 
counted five on one side and six on the 
other, with the largest ones most pos- 
terior. Arising from the center of the 
buccal floor at the back of the arena is 
the glottis—a remarkable anterad dis- 
placement from the typical mid-pharyn- 
geal position of this structure in all tad- 
poles with beaks. The elevated laryn- 
geal disc is tipped so the glottis is 
directed anterodorsally. The glottal slit 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 25 


Fic. 10.—Drawings of the floor (above) and roof (below) of the mouth of a Microhyla berdmorei 
we & larva. 


26 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


is of moderate size with very weak lips. 
Directly in front of the glottis is a single, 
small medial papilla. A single, conical, 
mediodorsally directed prepocket papilla 
can be found on each side. There are a 
few faint pustulations within the arena 
that disperse out over the lateral arms 
of the ceratohyals. A long trachea is 
visible in the buccal floor, dividing the 
BFA and the rest of the buccal floor, in- 
cluding the edge of the velum, into right 
and left halves. The buccal pockets are 
very large triangular pouches with their 
posterior margins far more obliquely 
oriented than in any hylid examined. 
The pockets are not deep compared to 
their unusual length, nor are they per- 
forated. The velar surface is astonish- 
ingly long—about half the length of the 
mouth. The freely movable portion of 
the velum is, however, only slightly long- 
er than in typical pond larvae. Spicular 
support is absent from this delicate sur- 
face. The velar margin lacks any pos- 
terior projections and forms a smooth 
arc as viewed from above. The margin 
is reflected dorsally along its medial edge 
where it joins the trachea. Secretory pits 
scattered along the posterior velar edge 
are limited to the absolute margin; these 
pits are few in number and quite small 
in size. 

Ventral pharynx.—The volume of the 
branchial baskets is over half of the 
total volume of the head and body; i.e. 
larger than in any beaked tadpole. The 
increase is primarily in length, with the 
baskets extending so far forward that 
almost half of the baskets are under the 
velum. Viewed from above, the baskets 
are basically shaped like elongate tri- 
angles and have their long axes oriented 
closer to the sagittal plane than in any 
beaked tadpole examined. The filter cav- 
ities decrease in size as one goes from 
the first to the third. The filter plates 
are almost vertical, except that a slight 
arch to the top portion of the third plate 
blankets the top of the third filter cavity. 
The extension of the filter plates under 
the velum makes the more anterior filter 


rows inaccessible without dissection of 
the ventral velum. Counts for filter rows 
run 22, 20, 18, 10 for cb. 1-4, exceeding 
the values for any beaked tadpole. The 
filter rows have normal proportions in 
terms of folding pattern, but are not 
quite as tall as in typical pond larvae. 
Neighboring filter rows fully abut and 
tertiary folds from neighboring rows in- 
terdigitate. The filter canals are very 
small, fully canopied tubes. The bran- 
chial food traps are unlike any seen in 
beaked species. They are slender, verti- 
cal crescents of a thick secretory tissue 
(hence the name crescentic organs; Sav- 
age, 1952), isolated at the anterior end 
of the filter cavities and inaccessible 
without dissection of the velar surface. 
The area covered by the organs is very 
small compared to the size of the filter 
cavities. Secretory ridges are very dense 
within the food traps and are not easily 
resolved in this specimen. Bound by the 
narrow crescentic organs are vertical re- 
gions of the filter cavities containing a 
few, loosely defined clusters of secretory 
cells. Most of the ventral surface of the 
velum is free of secretory tissue of any 
sort. The esophageal funnel is both nar- 
row in profile and small in diameter. 


Dorsal buccal—The roof of the 
mouth shares with the floor a trape- 
zoidal, almost rectangular profile. In de- 
tails, the roof differs radically from all 
non-microhylids examined. The upper 
lip is nearly horizontal so that the pre- 
narial arena is large and square. Far 
back in the arena, in line with the front 
of the internal nares is a single, medial, 
conical papilla, which is anteriorly 
curved and of medium size. Other sur- 
face features are absent. The most un- 
usual aspect of the roof is the internal 
nares. Each naris is a shallow, round 
cul-de-sac with all but the lateral third 
covered by a textured, presumably sen- 
sory, tissue patch. Thin, flap-like verti- 
cal walls surround the medial edges of 
these narial depressions. The antero- 
medial portions of the walls are low and 
coarsely serrate; the posteromedial por- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 27 


tions are relatively huge, tongue-like 
flaps that project anteriorly under the 
nares and touch on the midline. These 
projections are in a position correspond- 
ing to that of the narial valve projections 
in some beaked tadpoles. While they 
may be homologues to the narial valve 
projections, they are clearly not asso- 
ciated with any functional valves, for 
the nares are not perforated. These pro- 
jections are strongly concave anterad. A 
vertical ridge descends down the pos- 
terolateral surface of each projection and 
clearly stiffens it. The bases of these 
ridges align with a small triangular flap 
on each side of the buccal pocket just 
posterolateral to the nares. The nares 
lack any sort of rim or wall along their 
whole lateral margins. A postnarial are- 
na is absent. There are no postnarial pa- 
pillae, lateral ridge papillae, or median 
ridge to define this arena. Halfway back 
on the buccal roof are four simple, coni- 
cal BRA papillae. These papillae are too 
close to each other and to the midline 
to define a buccal roof arena. On each 
side of the midline, the smaller of the 
two papillae is displaced slightly antero- 
laterad. Beginning lateral to the BRA 
papillae and extending in a straight row 
anterolaterally across the buccal roof is 
a series of evenly spaced pustulations on 
each side. There are a few, very thin 
pustulations scattered over the buccal 
roof anterior to the BRA papillae; other- 
wise, papillae and pustulations of the 
buccal roof are absent. The glandular 
zone is of typical length. Its anterior 
margin is nearly transverse. The secre- 
tory pits are large and conspicuous, but 
of very low density. The dorsal velum 
at its maximum length is 25% longer than 
that of common pond hylid larvae. It is 
continuous across the midline, although 
rather abruptly constricted in that re- 
gion. 

Dorsal pharynx.—The two pressure 
cushions on each side are quite distinc- 
tive. The lateral one is, on the average, 
twice as long and tall, and four times 
as wide, as the medial cushion. Its tall- 


est point is at its posterolateral corner, 
where it is nearly twice as tall as at 
its midpoint. The medial cushion is 
more precisely defined and of more uni- 
form proportions. Laterally, the ciliary 
groove of M. berdmorei is shallow and 
very wide. Medially, the groove be- 
comes a narrow, deeply entrenched 
canal. 

Diagnostic summary. — Microhylid 
larvae differ from all other tadpoles in 
the attachment pattern for their ventral 
velum to their filter plates, the position 
of the glottis, the shape of the branchial 
food traps, and structure of the unper- 
forated internal nares. 

Microhyla berdmorei is readily dis- 
tinguishable from the other microhylids 
examined by the combination of: large, 
medially abutting, posterior infralabial 
papillae; a single preglottal papillae; 
high count of filter rows; narrow bran- 
chial food traps; four buccal roof pa- 
pillae; dorsal velum vestigial medially. 


Microhyla heymonsi Vogt 
(Fig. 11) 

Material—FMNH 187923 (stage 37, 
sv. 6.8 mm). Small pool above dammed 
stream at Sakearat Experimental Sta- 
tion, Amphoe Pak Thong Chai, Chang- 
wat Nakhon Ratchasima, Thailand; Feb- 
ruary 24, 1969. 

References.—Parker, 1934 (p. 135); 
Heyer, 1973. 

External.—Besides the beakless mouth 
and median spiracle typical of micro- 
hylid larvae, the tadpoles of Microhyla 
heymonsi are characterized by an oral 
disc expanded into a large, upwardly 
directed funnel. Only features which 
distinguish these larvae from the two 
previously described species of Micro- 
hyla are dealt with in the following de- 
scription. 

Buccal ventral.—Anteriorly the floor 
of the mouth is broad and gently curved, 
when viewed from above. Posteriorly, 
the branchial baskets are truncated. The 
prepocket portion of the buccal floor is 
much elongated compared to other spe- 


28 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Fic. 11.—Drawings of the floor (above) and roof (below) of the mouth of a Microhyla heymonsi 
larva. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 29 


cies of Microhyla. The posterior shift of 
the buccal pockets reflects the unusual 
shape of the underlying ceratohyals. 
These are extremely elongate medially 
and have posteriorly directed lateral 
arms. The infralabial papillae are as in 
M. berdmorei. BFA papillae are re- 
duced in size and number (3 on one 
side and 4 on the other), grading into 
pustulations. The papillae are blunt 
rather than conical. There is a single, 
blunt, medial papilla in front of the 
glottis. Prepocket papillae are large and 
strongly curved medially. The buccal 
pockets are narrow, nearly longitudinal- 
ly oriented slits. They are deep and 
their floors inaccessible without further 
dissection; whether or not they are per- 
forate was not determined. The glottis, 
which lacks elevated lips, sits on a huge 
laryngeal disc. It is directed dorsally 
and positioned about two-thirds of the 
distance back on the buccal floor. The 
trachea is much shorter in this species 
than in the other Microhyla. The velar 
margin is gently arched and lacks any 
projections. It is virtually continuous 
over the trachea, but left and right sides 
are interrupted by a small median notch. 
Secretory pits could not be resolved on 
the velar margin. 

Ventral pharynx.—The branchial bas- 
kets are shaped like equilateral triangles 
with more than half of the filters under 
the velar margin. The baskets are tiny, 
being proportionally smaller even than 
those of typical, Orton type 4 pond lar- 
vae. The third filter cavity is reduced 
in size and nearly inaccessible under the 
velum. All filter cavities contained fine, 
flocculent food matter in this specimen. 
Counts for filter rows on each filter plate 
were well below those for other Micro- 
hyla (and within a row of the counts 
for H. regilla). The filter mesh is not as 
dense as in the other Microhyla. Fine 
tertiary and higher order filter folds 
are lacking. Well-developed secretory 
ridges are present anteriorly under the 
ventral velum. Although the branchial 
food traps seem relatively larger than in 


the other two species, the full extent of 
the traps could not be determined be- 
cause of difficulty in staining and the 
small size of these tadpoles. The ridges 
were absent along most of the unat- 
tached velar surface. The esophageal 
funnel was slightly larger and broader 
than in the two other species. 

Dorsal buccal—There is a small, 
posteriorly directed V-shaped depression 
in the middle of the prenarial arena. A 
single, medial papilla of the prenarial 
arena arises at the posterior apex of this 
V-shaped concavity. This papilla is 
small, straight, and blunt. The extent of 
the presumed sensory patch in the in- 
ternal nares could not be determined. 
The narial walls are more similar to 
those of M. ornata than to M. berdmorei. 
The anteromedial portions of the narial 
walls are reflected backward under the 
narial openings. The bases of the large, 
posteromedial projections of the narial 
walls are widest in this species. The 
ridges on the backs of these projections 
are not as tall in M. heymonsi as in M. 
ornata or M. berdmorei. They are, how- 
ever, more sharply defined and com- 
pletely continuous with the small flaps 
posterolateral to the nares. In M. hey- 
monsi these flaps have arched rather 
than pointed apices. There are two tiny, 
blunt papillae far posterior on the buc- 
cal floor and several dozen blunt pustu- 
lations dispersed over the surface in a 
W-shaped pattern. I could resolve secre- 
tory pits, but only on the dorsal velum. 
The glandular zone, as a discreet region 
of the buccal roof, may be absent. The 
dorsal velum is shorter than in the other 
two species, at least in the neighborhood 
of the midline; nevertheless, the velum is 
still continuous across the midline. 


Dorsal pharynx.—Only lateral pres- 
sure cushions are well-defined and these 
are small, obliquely oriented ovals, much 
smaller than the medial pressure cush- 
ions in the other species. The ciliary 
groove is comparable to or slightly 
deeper than that of M. ornata. 


Diagnostic summary.—The mouth of 


30 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


M. heymonsi grossly differs from the 
mouth of the other Microhyla in the 
following features: longer prepocket 
buccal surface; ventral velum virtually 
continuous across the midline; branchial 
baskets highly reduced; very large 
esophageal funnel; V-shaped groove in 
prenarial arena; dorsal velum and pres- 
sure cushions reduced. 


Microhyla ornata Dumeril & Bibron 


Material_—Uncatalogued, author's col- 
lection (stage 36, sv. 5.9 mm). Collected 
in shallow cement drainage ditch on the 
grounds of Applied Scientific Research 
Corporation, 196 Ehahoyothin Rd., 
Bangkheng, Bangkok, Thailand; August 
22, 1968. 

Reference.—Lui, 1950 (p. 251-252). 

Microhyla ornata is so similar to M. 
berdmorei that only the differences are 
presented here. 

External.—Microhyla ornata is a 
small species with small tadpoles. Its 
larvae can be distinguished externally 
from those of previous species by their 
smaller size, rounder snouts and rela- 
tively longer, filamentous tails. 

Ventral buccal—The floor of the 
mouth is anteriorly rounder in M. ornata 
than in M. berdmorei. M. ornata lacks 
anterodorsal infralabial papillae and the 
posterior infralabial papillae are no more 
than gentle swellings incapable of abut- 
ting on the midline. The medial pre- 
glottal papilla is relatively taller in M. 
ornata, and in this species there is a 
second smaller, medial preglottal papilla 
in front of the first one. The prepocket 
papillae are relatively larger and curved 
medially. The margin of the velum has 
a single, broad posterior projection on 
each side above the filter plates of cb. 3. 
Numerous tiny secretory pits are present 
in a narrow row along the velar margin 
medial to the apices of the posterior pro- 
jections. The medial edge of the velum 
is not reflected posteriorly at the mar- 
gins of the trachea. 

Ventral pharynx.—Counts for filter 
rows run 14-15, 16, 13-14, 9 for cb. 1-4. 


Tertiary folds on the filter rows do not 
interdigitate as tightly as in M. berd- 
morei, but the difference is slight enough 
to be accountable to minor differences 
in preservation. The branchial food traps 
are wider and form nearly complete tori. 
The area bound by the food traps has 
faintly developed secretory ridges and, 
unlike M. berdmorei, these extend pos- 
teriorly onto the ventral surface of the 
ventral velum but are still absent along 
the posterior velar margin. 

Dorsal buccal.—Walls of the internal 
nares are not as tall as in M. berdmorei. 
The anteromedial portions lack serra- 
tions; the shorter, posteromedial projec- 
tions do not meet each other on the mid- 
line. The ridges on the back of these 
projections are very faint. The pre- 
sumed, sensory field in the internal nares 
covers a greater portion of the narial 
depressions. M. ornata has two rather 
than four BRA papillae. They are rela- 
tively taller than the BRA papillae of 
M. berdmorei. The rows of pustulations 
extending away from the BRA papillae 
are less conspicuous. The secretory pits 
of the glandular zone are even less dense 
than in M. berdmorei. The zone is frag- 
mented into four approximately equal 
portions with one division along the 
midline and two other divisions between 
the midline and the lateral buccal mar- 
gin. The glandular zone ends anterior 
to the base of the velum. The dorsal 
velum is more complete across the mid- 
line. Rather than being just a vestige, 
the mid-portion of this flap is equal to a 
fifth of the maximum length of the 
velum. 

Dorsal pharynx.—Differences are mi- 
nor and attributable to preservation. 

Diagnostic summary.—Microhyla or- 
nata is virtually identical to M. berd- 
morei. The few differences—viz., infra- 
labial papillae not abutting, double pre- 
glottal papillae, less filter rows, 2 buccal 
roof papillae, dorsal velum more com- 
plete across midline—may largely be 
ascribable to the differences in size. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 31 


Fic. 12.—Photographs of the floor (above) and roof (below) of the mouth of a Scaphiopus bombi- 
frons larva. 


32 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


PELOBATIDAE 
Scaphiopus bombifrons Cope 
(Fig. 12) 

Material—_FMNH 75020 (stage 36, 
sv. 15.5 mm). Collected from drainage 
ditch 4.6 km west of Fredrick, Tillman 
Co., Oklahoma; July 4, 1948. 

References.—Stebbins, 1951 (p. 205- 
207); Bragg, 1965 (p. 63-91). 

External.—The head is broad; the 
spiracle is sinistral. The oral disc is 
moderately small but surrounds a large 
and powerful beak. The most common 
denticle pattern is 4/4 but both the up- 
per and lower rows may range from 2 
to 6 in number (Bresler and Bragg, 
1954). 

Ventral buccal—The floor of the 
mouth in dorsal view closely approxi- 
mates an equilateral triangle. There are 
two pairs of moderate sized infralabial 
papillae. The first pair is far anterior 
within the arc of the infralabial carti- 
lage; these project anteriorly out of the 
mouth. The second pair is positioned 
more dorsolateral and posterior, and 
projects medially. All four papillae are 
subequal in size, slim, slightly com- 
pressed cylinders and lack marginal pa- 
pillae. The lingual papillae are typical 
in height and orientation, although they 
may have a slightly enlarged basal diam- 
eter. The apices of the papillae are 
covered with fine pustulations. The buc- 
cal floor arena is a poorly defined, egg- 
shaped patch bearing between 40 and 
50 papillae. Those papillae around the 
edge and within the patch are rather 
homogenous; they are all evenly spaced, 
medium small, straight and a few have 
fine terminal pustulations. Posteriorly, 
the papillae are replaced by pustula- 
tions. Papillae and pustulations are ab- 
sent directly in front of or behind the 
buccal pockets, which are very narrow, 
transverse depressions. The pockets are 
shallow and unperforated. The free 
velar surface has an average relative 
area but otherwise is quite unusual. The 
margin of the velum is a wide, smooth 
edged “V,” with each side being nearly 


straight or slightly convex anteriorly. 
There is a short, but relatively wide, me- 
dian notch. The velum is supported by 
only two pairs of spicules. One pair 
neighbors the midline and the other is 
nearby, directly over the filter plate of 
cb. 4. The spicules are wide and not 
very stiff. Small but dense and con- 
spicuous secretory pits form a band 
along the posterior velar margin. The 
band is widest around the median notch 
and thins out laterally so that secretory 
pits are absent along the lateral quarter 
of the velar margin. 

Ventral pharynx.—The branchial bas- 
kets are close to perfect ovals with the 
long axes of the ovals running antero- 
laterad to posteromesad. The baskets 
are moderately shallow. They are of 
average or slightly larger than average 
size. The filter plates of the second and 
third ceratobranchials do not project up- 
ward, consequently, each basket is a 
single common filter cavity. Counts for 
filter rows run 10, 14, 10-11, 9 for cb. 
1-4. These are all above the average; 
those for cb. 2 and 4 significantly so. 
The filter mesh on each filter row is very 
fine due to multitudinous tiny, tertiary 
filter folds on each secondary fold. The 
filter mesh, however, is not dense— 
rather wide gaps exist between neigh- 
boring filter rows. (The distance between 
rows may have been slightly accentu- 
ated by shrinkage from preservatives. ) 
Some of the rows, particularly those of 
cb. 2 and 3, are wider than normal. The 
filter canals are very tall and moderately 
wide. They are about 80% canopied by 
the filters. The ventral surface of the 
velum is covered by buff textured tissue, 
indicative of a secretory mucosa. At 
75 magnification no organized ridges 
or secretory pits could be resolved. This 
species presumably lacks secretory 
ridges. The glottis, glottal lips and 
laryngeal disc are typical in size and 
shape. The glottis is fully visible on the 
pharynx when viewed from directly 
above. The esophagus is large and the 
funnel is broad in dorsal profile. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 33 


Dorsal buccal.—S. bombifrons has a 
slightly larger prenarial arena and small- 
er postnarial arena than seen in a typical 
type 4 larva such as Hyla regilla. The 
smaller postnarial arena is the result of 
relative displacement on the median 
ridge anteriorly. The most unique fea- 
ture of S. bombifrons is a small, round, 
keratinized knob in the center of the 
prenarial arena. Other than this, the pre- 
narial arena is devoid of any papillae, 
pustulations or ridges. The internal 
nares are narrow, oblique slits. The 
anterior narial walls have slightly pustu- 
late anteromedial corners but lack pre- 
narial papillae. The narial valves have 
slightly arched margins, but lack narial 
valve projections. A deep groove behind 
each narial valve sets the valves off from 
the postnarial arena. The _ postnarial 
arena is bound laterally by two medium 
sized, robust papillae aligned directly 
behind the medial edges of the nares. 
These two papillae have pustulate 
crowns and a secondary, short, conical 
papilla at their lateral bases. A single, 
tiny, pointed papilla occurs posterolat- 
eral to the two postnarial papillae. The 
median ridge is a moderately tall, tri- 
angular flap with a relatively shorter 
base. The ridge is thick and stiff. There 
is a single, large, pointed, asymmetric, 
marginal projection on the median ridge 
in this specimen. The anterior surface 
has pustulations, but they are concen- 
trated in a rim around the edge of the 
flap. There are two small, pointed pus- 
tulations within the prenarial arena. A 
buccal roof arena is absent. Scattered 
about the buccal roof behind the median 
ridge are a few dozen pointed pustula- 
tions. Two or three of the more anterior 
ones are slightly larger than the others. 
The glandular zone is overall very wide, 
only narrowing slightly near the midline; 
its anterior margin is a distinct, broad 
“V.” The secretory pits are uniformly 
large and conspicuous, but not particu- 
larly dense. The dorsal velum is very 
long and broadly interrupted on the mid- 
line. Each side is anterodorsally coiled. 


Dorsal pharynx.—A single, massive 
pressure cushion covers the dorsal 
pharynx on each side. These cushions 
are deep, elongate ovals with their ma- 
jor axes running anterolateral to postero- 
dorsal. The ciliary groove was damaged 
in dissection, but appears very broad 
medially. 

Diagnostic Summary. — Scaphiopus 
bombifrons larvae are unique among the 
species examined in the following fea- 
tures: pattern and shape of infralabial 
papillae; shape of lingual papillae; shape 
of ventral velar margin; presence of 
keratinized knob in prenarial arena; size 
and shape of pressure cushions. 


Megophrys minor Stejneger 
(Fig. 13) 


Material—FMNH 49857 (stage 34, 
sv. 16.5 mm). Collected from pools un- 
der cascades of small mountain streams 
or in side pools of large mountain 
streams where current is slow, at ap- 
proximately 9100 m elevation, near 
Taosze, Mount Omei, Szechwan, China; 
June 15, 1946. 

Reference.—Liu, 1950 (p. 188-191). 

External—Megophrys minor larvae 
have a huge, upwardly directed, denticle 
free oral disc. They are generally 
thought to lack keratinized beaks, but 
the fine, hair-like structures at the mar- 
gin of the oral orifice (noted by Liu, 
1950) are undoubtedly vestiges of oral 
beak serrations. The tadpoles have a 
long, powerful tail; the spiracle is sinis- 
tral. 


Ventral buccal.—The anterodorsad 
direction of the lower beak results from 
a general, dorsal curvature of the an- 
terior portion of the mouth. Because 
the prepocket portion of the oral floor 
is relatively large and expanded antero- 
laterally, when the mouth is pinned flat 
the oral floor is in the shape of an elon- 
gate trapezoid rather than a triangle. 
M. minor has four infralabial papillae 
within the loop of infralabia and visible 
in the oral orifice of an undissected tad- 
pole. These papillae were illustrated 


34 


Fic. 


MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


13.—Photographs of the floor (above) and roof (below) of the mouth of a Megophrys minor 
larva. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 35 


and described in detail by Liu, 1950 
(Fig. 382 and p. 190). In addition 
there are five other infralabial papillae, 
one on the midline, posterior to the visi- 
ble four, and two others on each side, 
dorsolateral to the loop of the infralabial 
cartilage. All the infralabial papillae are 
thick, fleshy lobes that lack terminal ser- 
rations or papillae. The five most mesial 
papillae curve slightly anterad. The four 
most lateral papillae are elevated knobs. 
The most anterior and dorsal pair of 
lateral infralabial papillae are oval (an- 
teromedial to posterolateral major axis), 
palp-like projections on the posterior 
corners of the infralabial cartilage. The 
more dorsal pair of lateral infralabial 
papillae are also oval (transverse major 
axis) but are smaller than the pair just 
described. The medial papilla is a tri- 
angularly shaped projection that fits in 
tightly between the smaller and more 
posterior pair of external visible papillae. 
Lingual anlage and lingual papillae were 
not observed in the normal position. The 
medial infralabial papilla could be a 
homologue of the lingual papillae; how- 
ever, the far anterior placement of this 
papilla argues against such a hypothesis. 
Buccal floor arena papillae are absent. 
Instead of a BFA M. minor has two 
massive ridges that run the length of 
the prepocket surface of the buccal floor. 
Each ridge has a predominantly anterior 
to posterior orientation. They are slight- 
ly bowed inward and have elevated ter- 
minal knobs. The ridges are quite resil- 
ient, but not supported by cartilage. 
They are set off from the rest of the 
buccal floor by deep creases both in 
front and in back of them. Immediately 
medial to the ends of the buccal pockets 
are small, anteriorly concave flaps that 
surround the terminal knobs of each 
ridge. There is one blunt, rather antero- 
medial, prepocket papilla on each side, 
plus a small series of prepocket pustu- 
lations. A half-dozen stubby pustula- 
tions are scattered over the dorsomedial 
portion of the buccal floor. A few small 
pustulations are also on the buccal floor 


in a narrow, transverse arch behind the 
buccal pockets and in an anterior-pos- 
terior line bowed outward slightly, above 
the lateral portions of the ceratohyals. 
The buccal pockets are very short and 
strongly arched anteriorly. Despite the 
fact that they are quite shallow, folds 
at the bottom of each pocket make it 
difficult to determine if the pockets are 
perforated. If the pockets are perfo- 
rated, the openings must be extremely 
small. The free velar surface is moder- 
ately large and supported by long, thin 
spicules. The posterior margin is un- 
usual in having a U-shaped medial mar- 
gin and a straight, nearly transverse lat- 
eral margin. There is a sharp inflection 
point between the straight and curved 
portions. Posterior projections are lack- 
ing; the median notch is small. The 
curved portion of the posterior margin 
is slightly thickened and his a buffed 
texture; however, individual secretory 
pits could not be resolved, even at 75 
magnification, anywhere along the velar 
margin. 

Ventral pharynx.—The branchial bas- 
kets are similar in size, shape and propor- 
tion to those of other megophrynines, 
particularly Leptobrachium. As viewed 
from above, the baskets are oval, nearly 
round, and cover an area slightly less 
than they do in H. regilla. They are very 
shallow, and the filter plates of cb. 2 
and 3 are very low so that there is only 
a single common filter cavity on each 
side. Counts for filter rows run 8, 9, 9, 
7 for cb. 1-4. The first three values are 
insignificantly different from those of H. 
regilla while the count for cb. 4 is sig- 
nificantly above the H. regilla mean. 
The filter rows are quite wide, but the 
filter nitches are larger than in other 
megophrynine tadpoles because the fil- 
ter folds are very thin. This may be, in 
part, a preservational artifact. On many 
of the filter rows tertiary filter folds are 
so long that they abut with the filter 
folds of neighboring rows. Filter rows 
from opposing filter plates meet end to 
end above each gill slit rather than inter- 


36 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


digitating, and thus they obscure the 
gill slits. The filter canals are small but 
most are fully canopied. Nevertheless, 
the general low density of the filter mesh 
makes the canals rather open, diffuse 
corridors. The ventral surface of the 
ventral velum is covered with a fine, buff 
textured tissue, suggesting a secretory 
mucosa. Organized secretory tissue, as 
either pits or ridges, could not be re- 
solved at even 75 magnification and 
is presumed absent from the branchial 
food traps. The glottis is small, with a 
typical amount of exposure posterior to 
the ventral velum. It has elevated lips, 
sits on a transversely oriented laryngeal 
disc, and is oriented slightly anteriorly 
rather than dorsally. The esophageal 
funnel has a moderately large bore, but 
a generally narrower profile than that 
of other megophrynine larvae. 


Dorsal buccal.——The upper beak is 
nearly straight rather than curved, thus, 
the prenarial arena is very wide and the 
buccal roof more trapezoidal than tri- 
angular in shape. Individual papillae 
are absent from the buccal roof; instead 
rows of papillae are replaced by ridges. 
Projecting down from the middle third 
of the prenarial arena is a large, V- 
shaped ridge. The anterior corners of 
the ridge are the thickest and tallest 
portions and are reflected medially. 
There is a deep sulcus within the apex 
of the “V.” Behind the apex a medium 
sized, round “knob” descends from the 
roof. On each side paralleling the arms 
of the V-shaped ridge, but far lateral, 
is a short, thin ridge. Posterior to the 
ends of these ridges and directly lateral 
to the apex of the V-shaped ridge are 
clusters of medium small, blunt pustu- 
lations. The nares are of moderate size 
and transversely oriented. The inter- 
narial distauce is large. The anterior 
narial walls are slightly pustulate along 
their more medial halves and possess a 
stiff, medium tall, conical papilla along 
their more lateral halves. The narial 
valves are tall and each possesses a 
distinct, pointed projection. Although 


the prenarial papillae and the narial 
valve projections are relatively tall, they 
are still far shorter and less attenuate 
than in other megophrynine larvae ex- 
amined. The postnarial arena is a mod- 
erately small, smooth, triangular recess 
in the buccal roof. The arena is filled 
by the body of the median ridge, which 
is a massive, nipple-shaped, anteriorly 
directed projection. This projection ex- 
tends far forward, anterior to the nares 
and under the posteromedial “knob” in 
the prenarial arena. Despite its unusual 
size and shape, the median ridge in M. 
minor is similar to the typical flap-like 
median ridge in most other species, be- 
ing joined to the buccal roof along a 
narrow, transverse line and free to move 
ventrally in the sagittal plane. The post- 
narial arena is bound laterally by wide, 
low ridges that extend far posterior and 
are major features on the buccal roof. 
These ridges are the presumed homo- 
logues of the postnarial papillae and pos- 
sibly the BRA papillae collectively. The 
ridges begin as anterad projections that 
extend almost as far forward as the tips 
of the median ridge. The ridges become 
the medial walls of the internal nares 
and bow outward along most of their 
length. The end of each ridge is re- 
curved. The ridges are tallest and thin- 
nest anteriorly, and lowest and widest 
posteriorly. They end just behind the 
middle of the buccal floor and well be- 
hind the base of the median ridge. The 
lateral ridge papillae are represented by 
large, thick, obliquely oriented flaps, far 
lateral to the middle of the ridges just 
described. On each side anterolateral to 
these flaps is a similar subsidiary pro- 
jection. These subsidiary lateral ridge 
“papillae” are only a third to a half the 
size of the main ones. Posteromedial 
to the main lateral ridges are small 
ridges each with a gently crenulate mar- 
gin and a predominantly anterior-pos- 
terior orientation. These ridges are in a 
position comparable to the lateral roof 
papillae seen in many tadpoles and are 
presumed homologues of these struc- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 37 


tures. They blend posteriorly into a 
cluster of pustulations. A buccal roof 
arena is absent. Pustulations are present 
behind the median ridge and mesial to 
the posterior arms of the long ridges 
that bound the postnarial arena. There 
is also a loose cluster of faint, blunt 
pustulations in the posteromedial por- 
tion of the buccal roof. The glandular 
zone is narrow and thins out toward the 
midline; it is absent in front of the 
esophagus. The secretory pits at the 
anterior margin of the glandular zone 
are elongate in an anterior-posterior 
direction. The remaining secretory pits 
are relatively heterogenous in size and 
shape, but the larger and rounder pits 
tend to be along the posterior margin 
of the zone. It is difficult to separate the 
dorsal velum from the pressure cushions 
for the velum seems very short and di- 
rected ventrally rather than anteriorly 
or anteroventrally. This may, however, 
be an artifact from shrinkage. The 
velum is broadly interrupted medially 
and the medial margins on each side are 
reflected dorsally into the esophageal 
funnel. 

Dorsal pharynx.—M. minor has only 
one large pressure cushion per side. The 
cushion is tallest and widest far lateral 
and triangular in shape. The ciliary 
groove was damaged in dissection and 
its proportions could not be determined 
in this specimen. 

Diagnostic summary.—Megophrys, as 
represented by Megophrys minor, differs 
from all other tadpoles examined in the 
structure of the infralabial papillae; 
shape of the ventral velar margin; large, 
nipple-shaped median ridge and large 
lateral ridge papillae. 


Leptobrachium hasselti Tschudi 
(Fig. 14) 


Material —FMNH 14828 (stage 35, 
sv. 27.8 mm). Swampy area adjacent to 
stream at Labang Camp on Sungei 
Seran, Bintulu District, Fourth Div., 
Sarawac, Borneo; November 5, 1963. 

Reference.—Inger, 1966 (p. 33-36). 


External. — Leptobrachium hasselti 
tadpoles commonly have a 6/6 or 7/6 
denticle pattern but there is both zoo- 
geographic and ontogenetic variation in 
the number of denticle rows (see Inger, 
1966, Table 3). The spiracle is sinistral. 
The body is deep and the tail is shorter 
than in most megophrynine tadpoles. 


Ventral buccal_—The floor of the 
mouth is broad despite the fact that 
anteriorly the oral orifice is compara- 
tively small. There is a single, major 
hand-shaped infralabial papilla on each 
side which is typical in size, shape and 
position but with marginal “fingers” long 
and attenuate. Between the lateral mar- 
gin of the keratinized beak and the in- 
fralabial papilla is a series of three 
papillae or pustulations in a straight 
anterior to posterior row. The largest is 
the most anterior one. Ventral and me- 
dial to the hand-like infralabial papillae 
are two moderately tall, attenuate pa- 
pillae. These have pustulate surfaces. 
Lingual papillae are not evident on the 
tongue anlage but the two papillae just 
mentioned could be possible analogues 
or even homologues of the lingual pa- 
pillae. The buccal floor arena is well 
defined by 10-12 moderately large, at- 
tenuate, conical BFA papillae on each 
side and a field of 20 or more smaller, 
but similarly shaped papillae posterior- 
ly. The lateral BFA papillae are aligned 
parallel to the sagittal plane. The tall- 
est ones show some slight, marginal 
pustulation but none are bifurcate. 
There is evidence of preservational 
artifact in papillae; the BFA _ papil- 
lae are compressed flat against the 
buccal floor, with the lateral papillae 
pointing away from the midline. Pre- 
pocket papillae are absent, but there 
are well developed pustulations in the 
prepocket position on each side of 
the buccal floor. Anterior to these, 
above the lateral arms of the cera- 
tohyal, are a few dozen or so tiny 
papillae and associated pustulations in 
a diffuse patch. Tiny conical papillae 
and pustulations are also present in 


38 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Fic. 14.—Photographs of the floor (above) and roof (below) of the mouth of a Leptobrachium 
hasselti larva. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 39 


abundance on the buccal surface in a 
triangular area between the BFA, the 
back of the buccal pockets and the velar 
surface. The buccal pockets have a nearly 
transverse orientation: they are wide 
and deep. They are clearly perforated, 
although the perforations are not visible 
without retracting the large folds on the 
posterior wall of each pocket. The free 
velar surface is very long; it is supported 
by long, thin spicules that extend back 
to the posterior velar margin. The velar 
margin is nearly straight on each side 
of the midline, its medial portion is 
coarsely serrated. There is a well de- 
veloped median notch bound by two 
papillae-like projections on each side. 
The more medial projections are two to 
three times the size of those immediately 
lateral. The velum possesses a secondary 
margin between the branchial baskets 
ventral to the main edge. This secon- 
dary edge also possesses a pair of pos- 
terior projections. Large, conspicuous 
secretory pits of moderate to low density 
can be found along the middle half of 
the velar margin. The pits are concen- 
trated on the medial posterior projection. 


Ventral pharynx.—The branchial bas- 
kets are oval, elongate in the transverse 
plane, and extremely shallow. All filter 
cavities are of nearly equal volume, but 
the filter plates are so short that they 
effectively form a single common cavity 
for each branchial basket. The shorten- 
ing of the filter plates is reflected in an 
increase in length to height ratios for 
all the filter plates, particularly those of 
the second and third ceratobranchials. 
Counts for filter rows run 5, 8, 9, 6 for 
cb. 1-4. The values for the first two 
arches are extremely low while the two 
other values are within the typical tad- 
pole range. The filter mesh is not dense 
although the filter rows are all very 
wide. The wide rows result from longer 
secondary and tertiary filter folds. Nev- 
ertheless, neighboring filter rows do not 
abut and the filter canals are shallow, 
60% canopied by the filters, and about a 
third as wide as the rows that bound 


them. Branchial food traps are com- 
pletely confined to the horizontal free 
velar surface; however these surfaces are 
covered by large, dense secretory pits 
instead of secretory ridges. These pits 
completely cover the food traps and can 
be found on the ventral surface of the 
ventral velum directly in front of the 
glottis where they tend to align in trans- 
verse rows. The glottis is small, with 
weakly developed lips; a laryngeal disc 
is not visible. Ninety per cent or more 
of the glottis is fully visible when viewed 
from above but this large exposure may 
have resulted in part from tissue shrink- 
age. The esophageal funnel is anteriorly 
broad but not large. 

Dorsal buccal.—The buccal roof is 
broad overall, but the prenarial arena is 
narrow. Six pustulations within the nar- 
ial arena show no particular symmetry. 
The nares are wide and have a strong 
transverse orientation. A small pustulate 
ridge projects from the anteromedial 
corner of each naris toward the center 
of the prenarial arena. In the middle of 
the otherwise low anterior narial walls 
arise single, tall, attenuate prenarial pa- 
pillae on each side, matched by equally 
tall and attenuate narial valve projec- 
tions of the posterior wall. Both the 
prenarial papillae and the narial valve 
projections point away from the narial 
openings, presumably because of shrink- 
age. The median ridge is near the nares 
and the postnarial papillae are in a 
transverse line between the median 
ridge and the laterally displaced lateral 
ridge papillae; consequently, the post- 
narial arena is shaped much like a trans- 
versely elongate oval, two to three times 
as wide as long. The postnarial papillae 
are more attenuate than those of H. 
regilla but equally numerous. The me- 
dian ridge, which lacks a secondary 
ridge, is small and shaped like a half- 
circle with fine, evenly spaced, marginal 
serrations. Lateral ridge papillae are 
twice as tall as the median ridge. They 
are narrow, transversely compressed 
flaps with two attenuate peaks. The 


40 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


medial peaks are two to three times as 
tall as the lateral peaks. The buccal roof 
arena is a well-defined oval bounded by 
eleven attenuate conical papillae on one 
side and twelve on the other. Far lateral 
to the middle of the BRA are clusters of 
two or three smaller, secondary papillae. 

The buccal roof is evenly covered 
with fine pustulations. In addition, 
there are larger, more robust pustula- 
tions scattered within the BRA. Later- 
ally the glandular zone is of typical 
length along the anterior-posterior axis 
but medially it is reduced to a length 
of only two or three secretory pits. The 
dorsal velum has been badly distorted 
from alcohol storage and its full extent 
cannot be determined. There is a large 
gap between the left and right halves 
of the velum and the medial edges—un- 
usual in being papillate—are reflected 
dorsomedially. I counted seven papillae 
on each side and roughly an equal num- 
ber of pustulations. 


Dorsal pharynx.—The nature of the 
pressure cushions could not be deter- 
mined because of shrinkage. It seems, 
however, that there is a single cushion 
that fits into the common filter cavity 
rather than multiple pressure cushions 
on each side. The ciliary groove is un- 
usually wide, particularly laterally, but 
this again may be a preservational arti- 
fact. 

Diagnostic summary.—See Oreolalax 
pingii (p. 46). 


Leptobrachium oshanensis (Liu) 
(Fig. 15) 


Material—FMNH 49589 (stage 34, 
sv. 12.2 mm). Collected from side pools 
and pools beneath cascades of small 
streams at approximately 1-1.1 km eleva- 
tion, Mt. Omei, Szechwan, China; June 
15, 1946. 

Reference.—Liu, 1950 (p. 199-201). 

External_—Leptobrachium — oshanen- 
sis larvae have an enlarged oral disc 
(like many stream tadpoles) but much 
of it is denticle free. The usual pattern 
for denticle rows is 4/3, but it is not 


uncommon for there to be one row more 
or less above or below. The spiracle is 
sinistral; the body and tail are elongate. 


Ventral buccal_—The lower beak is 
larger and positioned more dorsally in 
L. oshanensis than in L. hasselti. The 
floor of the mouth is of typical width 
and not as broad as in L. hasselti. Very 
large, hand-like infralabial papillae are 
present; like many other oral papillae, 
they show a tendency to be divided into 
long, secondary papillae. The infralabial 
papillae are tall structures with a half- 
dozen or more extremely attenuate, 
pointed “fingers” that arch anterad. 
There are two rather than three papillae 
on each side anterior to the infralabial 
papillae. The more anterior papillae are 
the larger. These are flattened into the 
transverse plane, bifurcated and are 
pointing out of the mouth. The papillae 
between these and the hand-like papil- 
lae are small, simple projections. The 
two medial papillae, which were noted 
as possible anteriorly displaced lingual 
papillae in L. hasselti, are taller in L. 
oshanensis and also project anteriorly. 
Papillae are absent from the tongue an- 
lage. On each side, laterally adjoining 
the tongue anlage, are small, oval swell- 
ings; on the top of these swellings are 
one or two moderate sized attenuate 
papillae. The buccal floor arena differs 
from that of L. hasselti in being fully 
open posteriorly. The BFA papillae 
rows converge slightly both anteriorly 
and posteriorly, giving the arena a typi- 
cal oval design. While BFA papillae 
are as numerous as in L. hasselti, they 
are taller and even more attenuate. 
Three BFA papillae on one side and 
four on the other have terminal divisions. 
The papillae in the middle of the series 
are much larger than those in the front 
or back of the rows. There are one or 
two small, conical spikes in the pre- 
pocket positions. These are at the back 
of a short row of similarly shaped pa- 
pillae that extend in a line anterome- 
dially over the lateral arms of the cera- 
tohyals. L. oshanensis lacks the papil- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 4] 


Fic. 15.—Photographs of the floor (above) and roof (below) of the mouth of a Leptobrachium 
oshanensis larva. 


42 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


late patches seen in L. hasselti in front 
and behind the buccal pockets. There 
is, however, a relatively dense field of 
stiff, conical pustulations within the pos- 
terior half of the buccal floor arena not 
seen in the other species. These pustu- 
lations extend farthest forward on the 
midline to about the level of the buccal 
pockets. Posteriorly, they reach the 
velar margin and generally increase in 
size as one goes backwards over the 
buccal floor. The buccal pockets are 
longer and not as wide as in L. hasselti. 
The bottoms of the pockets are fully 
obscured by massive folds from the pos- 
terior walls. I could not find clear nat- 
ural perforations. The free velar sur- 
face differs from that of L. hasselti in 
having weak posterior projections rather 
than medial crenulations over the filter 
cavities and in having a uniform, nar- 
row, but thickened band of secretory 
pits along its posterior margin. A deep 
median notch is bound by massive 
brushes of extremely attenuate papillae. 
These arise from a double—dorsal and 
ventral—medial edge on the ventral 
velum, as in L. hasselti. The numerous 
papillae are preserved in an anteriorly 
folded direction in this specimen. 
Ventral pharynx.—The branchial bas- 
kets are smaller than in L. hasselti. The 
filter plates have even less height. 
Counts for filter rows run 7, 10, 8, 5 for 
cb. 1-4. These values are below the 
mean for H. regilla, but none signifi- 
cantly. The values for cb. 1 and 2 are 
actually above the values for L. hasselti; 
this is possible within the smaller bran- 
chial baskets of this species because 
the filter rows are narrower. Filter ca- 
nals are slightly more open in L. osha- 
nensis than in L. hasselti. As with L. 
hasselti, the branchial food traps are 
covered with secretory pits rather than 
ridges and the pits are continuous across 
the ventral surface of the ventral velum 
and in front of the glottis. The glottal 
lips are larger in this species than in L. 
hasselti and the glottis is fully exposed 
within the median notch. There are two 


tiny anterior papillae on the glottal lips. 
The laryngeal disc is visible as a very 
wide, transversely oriented, oval ridge. 
The esophageal funnel is of comparable 
size and proportions in the two species. 


Dorsal buccal—The roof of the 
mouth has roughly the same proportions 
as in L. hasselti. The upper beak is re- 
flected posteriorly in L. oshanensis, ob- 
scuring about 40% of the prenarial arena 
from ventral view. The number of pus- 
tulations in the prenarial arena is re- 
duced. The prenarial papillae and narial 
valve projections are not quite as tall 
as in L. hasselti, but are still far taller 
than in any hylid. The narial valve pro- 
jections are displaced slightly laterally. 
The median ridge is farther posterior 
and the postnarial arena is larger than 
in L. hasselti but not as large as in a 
typical tadpole. The positions of the 
postnarial papillae are as in H. regilla, 
not as in L. hasselti. The median ridge 
and lateral ridge papillae are similar 
to those of L. hasselti except for being 
slightly taller and wider. The papillae 
that define the buccal roof arena are 
like those of L. hasselti in number and 
position, but more attenuate in L. 
oshanensis. The largest papillae are an- 
terior. The papillae in the clusters lat- 
eral to the BRA are more numerous in 
this species. The BRA encloses a field 
of stout conical pustulations, which are 
very large, stiff, subpapillate structures 
of only moderate density at the front of 
the arena. Posteriorly, the pustulations 
become both smaller and more numer- 
ous. The pustulate field ends in a trans- 
verse line that laterally abuts with the 
glandular zone; the glandular zone is 
broadly interrupted medially. The dor- 
sal velum is also interrupted in front of 
the esophagus but the gap between the 
left and right portions of the glandular 
zone is twice the width of the gap be- 
tween the left and right halves of the 
velum. Secretory pits within the glan- 
dular zone are similar in both species. 
The dorsal velum is generally short; it 
is anchored on each side of the roof just 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 43 


behind the medial limits of the glandu- 
lar zone. Although it is a free flap both 
lateral and medial to these points, the 
medial portions of the dorsal velum are 
relatively immobile. The medial por- 
tions of the velum are turned backwards 
and are lined with long, attenuate pa- 
pillae, twice as many as in L. hasselti. 
Papillae in this general region are not 
restricted to the margin of the velum; 
they are also prevalent on the pharyngeal 
surface of the velum and on the buccal 
roof above the medial edges of the 
velum. 

Dorsal pharynx.—tThe lateral and me- 
dial pressure cushions are completely 
confluent. The ciliary groove is very 
wide and shallow. 

Diagnostic summary.—See Oreolalax 


pingii (p. 46). 


Oreolalax pingii (Liu) 
(Fig. 16) 


Material—_FMNH 50988 (stage 36, 
sv. 25.8 mm). Collected among stones 
in slow-running streams at approximate- 
ly 2.7-3.4 km elevation, between Hsuan- 
shenpa and Lolokou (N. 28°0’, W. 102° 
30’) on the road between Sichang and 
Chaochiao, Sikang, China; May, 1942. 

Reference.—Liu, 1950 (p. 149-150). 

External.—O. pingii tadpoles have an 
invariant 5/5 denticle pattern. The spir- 
acle is sinistral. In general appearance 
they are similar to Leptobrachium lar- 
vae. In terms of tail length O. pingii 
larvae are intermediate between L. has- 
selti and L. oshanensis. Internally O. 
pingii larvae are sufficiently similar to 
the two species of Leptobrachium de- 
scribed above that comparison with 
those species is emphasized in the fol- 
lowing description. 

Ventral buccal.—The buccal floor is 
more triangular than in either species of 
Leptobrachium and the prepocket por- 
tion is relatively longer. The infralabial 
papillae are most similar to those of L. 
oshanensis. They differ in being overall 
smaller and in being covered by large, 
pointed pustulations. All but the small- 


est buccal papillae in O. pingii have this 
unusual pustulate surface. The most an- 
terior infralabial papillae are very stout 
and are divided into four or more pro- 
jections. Smaller papillae between the 
main infralabial papillae and the anterior 
pair are absent. The medial pair of in- 
fralabial papillae, between the main 
infralabial papillae, are very tall and 
have two or more terminal subdivisions. 
As in Leptobrachium, the lingual anlage 
lacks papillae. The buccal floor arena 
is similar to that of L. oshanensis; the 
BFA papillae are a bit larger. The 
largest BFA papillae are fused basally 
with their neighbors. Papillae in the 
prepocket position and over the lateral 
arms of the ceratohyals are like those of 
L. oshanensis in position and shape. Pa- 
pillate patches are present directly be- 
hind the buccal pockets as in L. hasselti, 
although they are not as numerous as in 
that species. The pustulate field within 
the BFA is more similar to that of L. 
oshanensis than that of L. hasselti, but 
differs from the former in extending far- 
ther forward on the midline and being 
made up of overall smaller, blunter, and 
less dense pustulations. The far poste- 
rior portion of this pustulate field has 
the smallest pustulations and the great- 
est density. The buccal pockets are in- 
termediate between those of L. hasselti 
and L. oshanensis in all aspects that have 
been considered; there is a definite small 
perforation of the pockets in this species. 
The free velar margin is supported 
by thicker spicules in O. pingii than in 
either species of Leptobrachium. There 
are three distinct small, round, posterior 
projections of the velar margin on each 
side. These are similar in position to 
those of H. regilla but more distinct. 
The median notch and the two projec- 
tions that immediately bound it are a 
bit larger than in L. hasselti, but other- 
wise similar. The odd, secondary, ven- 
tral edge to the medial portion of the 
velar margin—characteristic of both spe- 
cies of Leptobrachium—is rudimentary 
in O. pingii. On the ventral side of the 


44 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Fic. 16.—Photographs of the floor (above) and roof (below) of the mouth of an Oreolalax pingit 
larva. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 45 


medial portion of the velar margin are 
four small, simple, posteriorly directed 
papillae. The lateral pair are the largest, 
but are still too small to reach the velar 
margin and be seen from above. Secre- 
tory pits on the velar margin are as in 
L. oshanensis. 


Ventral pharynx.—The branchial bas- 
kets of O. pingii resemble those of L. 
hasselti in shape; they are intermediate 
between those of the two species of 
Leptobrachium in relative size. Essen- 
tially there is only one common filter 
cavity in O. pingii larvae on each side; 
for it is only far forward, under the 
velum, that the filter plates of any cerato- 
branchial is as tall as its filter rows. 
Counts for filter rows are identical to 
those of L. oshanensis. No characters of 
the filter themselves serve to distinguish 
the filters of O. pingii from those of L. 
oshanensis. Secretory pits could be found 
only on the ventral surface of the ventral 
velum. In that regard, O. pingii is no 
different from the species of Leptobrach- 
ium. In terms of the pattern of the pits 
in front of the glottis, the glottis itself, 
and the esophageal funnel, O. pingii is 
indistinguishable from L. hasselti. 

Dorsal buccal.—The buccal roof is 
more triangular in O. pingii than in 
either species of Leptobatrachium. In 
size and proportions the prenarial arena 
is intermediate between that of the two 
species of Leptobrachium. There are 
two tall, but simple, cylindrical papillae 
positioned far laterally within the pre- 
narial arena. These project ventrally be- 
tween the main infralabial papillae and 
the anterior infralabial pair. The inter- 
nal nares are most similar to those of 
L. hasselti. They differ in the following 
ways: the anterior walls are pustulate 
(and there is a second prenarial papillae 
on one side in this specimen); the narial 
valve projections are not as tall. The 
median ridge is more posterior in O. 
pingii than in either species of Lepto- 
brachium, but still semicircular in shape. 
Marginal serrations are very coarse. The 
median ridge has a pustulate anterior 


surface, but lacks a secondary ridge. The 
postnarial arena is average in size. Post- 
narial papillae (four per side) are posi- 
tioned in oblique clusters as in H. regilla 
and L. oshanensis. These papillae are 
larger than in L. oshanensis, the largest 
being the most posterolateral. All have 
serrated anterior margins. The lateral 
ridge papillae are huge, flap-like struc- 
tures positioned far lateral as in Lepto- 
brachium, but much larger than in that 
genus, and with three, rather than two, 
terminal projections. The BRA of O. 
pingii is anteriorly narrower, but poste- 
riorly wider, than that of L. oshanensis. 
Within the arena proper is a field of two 
to three dozen small to medium sized 
papillae, most of which have terminal 
pustulations. A small, triangular cluster 
of papillae extends laterally from the 
posterior portion of the BRA. The di- 
mensions of the glandular zone are most 
comparable to those of L. hasselti. The 
zone differs, however, from either species 
of Leptobrachium in having extremely 
large and conspicuous secretory pits 
along its anterior margin. The dorsal 
velum is shorter than that of L. oshanen- 
sis and interrupted medially. The an- 
terior margin of the velum is nearly 
transverse except in the neighborhood 
of the midline, where it is reflected 
posteriorly. Papillae are present along 
the medial half of the velar margin. 
These are as numerous as in L. hasselti 
and intermediate between those of L. 
hasselti and L. oshanensis in size. 


Dorsal pharynx.—A_ single, trans- 
versely elongate pressure cushion is pres- 
ent on each side. These cushions extend 
from the lateral margin of the pharynx 
halfway to the medial margins of the 
dorsal velum. Their shorter, anterior- 
posterior axes are a quarter or less the 
length of their longer, transverse axes. 
The cushions have indented anterior and 
medial margins, which give them an 
autonomy from the dorsal velum not 
seen in any other tadpoles. The ciliary 
groove is similar to that of Leptobrach- 
ium. 


46 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Diagnostic summary—The  non- 
Megophrys megophrynines form a tight 
cluster of species that share the unique 
character of having multiple, layered 
papillation of the medial portion of the 
ventral velar margin. They also tend to 
have extremely tall prenarial and post- 
narial papillae and papillation of the 
medial margin of the dorsal velum. Lep- 
tobrachium oshanensis can be distin- 
guished from L. hasselti by an assort- 
ment of minor characters. The most 
conspicuous of these includes the size 
of the infralabial papillae and extent of 
papillation of the medial margins of the 
dorsal and ventral vela. Oreolalax can 
readily be distinguished from the species 
of Leptobrachium examined by the un- 
usual pustulation of its buccal papillae. 


HYLIDAE 
Anotheca spinosa Steindachner 
(Figs. 17, 18) 


Material—KU 60017 (stage 37, sv. 
14.5 mm). Collected in tree hole con- 
taining approximately 60 ml of water at 
1600 m elevation, Vista Hermosa, Oa- 
xaca, Mexico; August 12, 1960. 


Reference.—Duellman, 1970 (p. 149- 
150). 

External—Anotheca spinosa has 
stout, muscular larvae with very broad 
beaks and a 2/2 denticle pattern in early 
stages. The upper rows of denticles tend 
to disintegrate in later stages. The spir- 
acle is sinistral. 

Ventral buccal—The floor of the 
mouth in the frontal plane is roughly an 
equilateral triangle. Two distinct pustu- 
lations are aligned directly behind the 
keratinized beak on each side. Behind 
these and about halfway between the 
symphysis and the articulation of the 
infralabial cartilage with Meckel’s carti- 
lage is an infralabial papilla on each 
side; the papillae are shaped like slightly 
compressed cylinders and point antero- 
dorsally. Secondary, stubby, terminal 
projections on the papillae number five 
on one side and four on the other. There 


is only one lingual papilla in this speci- 
men. It is terminally bifurcated and 
laterally expanded, and looks much like 
two normal papillae that have fused on 
the midline. The buccal floor arena is 
very broad. The lateral row of papillae 
outlining the arena converge little to- 
ward the midline at either the front or 
the back of the arena. The BFA papillae 
are average in number, but shorter and 
blunter than in typical hylids. The 
larger papillae, i.e., those in the pos- 
terior half of the arena, are flattened 
against the floor of the cavity and di- 
rected out away from the midline rather 
than medial as in most other tadpoles 
examined. This unusual orientation may, 
in part, result from shrinkage following 
preservation. The two largest BFA pa- 
pillae on each side have terminal bifur- 
cations. Some three dozen small, blunt 
pustulations and tiny papillae are scat- 
tered about in the posterior half of the 
BFA. A half dozen similar projections 
are above the lateral arms of the cerato- 
hyal outside of the arena. The buccal 
pockets are short, slender, and more 
obliquely oriented than those of a typi- 
cal hylid; they are not perforated. The 
free velar surface is very much reduced. 
Supporting spicules are thin and short. 
The posterior margin of the velum is 
predominantly transverse with very faint 
crenulations projecting posteriorly over 
the first two filter cavities. The shallow 
median notch is so broad that it covers 
no less than half of the velar edge. Se- 
cretory pits could not be found on the 
margin and are likely absent. 

Ventral pharynx.—The branchial bas- 
kets are greatly reduced. The filter plates 
are reduced in all directions, although 
they still retain some height. Vestiges of 
filter rows are present as thin, irregular 
folds of epithelium. These number 3, 3, 
3, 2 for cb. 1-4. Only the main folds are 
present. The filter canals are completely 
open and two to four times as wide as 
the vestiges of the filter rows. Branchial 
food traps are absent; that is, no orga- 
nized secretory ridges could be resolved 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES AT 


Fic. 17.—Drawings of the floor (above) and roof (below) of the mouth of an Anotheca spinosa 
r larva. 


48 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


anywhere in the pharynx. The glottal 
slit and lips are large and _ vertically 
oriented. They are 50% under the velar 
margin. The glottis sits on a_ large, 
gently domed, laryngeal disc. The 
esophageal funnel is of enormous width 
and extremely broad profile. 


Dorsal buccal.—The roof of the 
mouth shares the equilateral triangular 
shape of the floor of the mouth. The 
lateral margins of the roof, however, 
curve downward more than in other tad- 
poles examined, giving the roof consid- 
erable depth. The internal nares and 
median ridge are far forward on the 
buccal roof. The prenarial arena is twice 
as broad as long. Two tall, knobby pro- 
jections with blunt apices arise from 
the middle of the prenarial arena. The 
tips of these projections are turned pos- 
teriorly. The stout common pedestals 
for these projections bow forward on the 
midline and extends back on each side 
as ridges that become continuous with 
the posterior wall of the internal nares. 
The internal nares are shallow slits, 
more elongate and oblique than in a 
typical hylid larva. The far anteromedial 
corners are pustulate flaps. Anterior 
walls are effectively absent except lat- 
erally where there are some blunt pus- 
tulations and a distinct, pustulate, pre- 
narial papilla on each side. The posterior 
narial walls each have a slight to mod- 
erate narial valve projection, but the 
projections are in the middle rather than 
at the anteromedial end of the posterior 
wall. There is a deep groove that sep- 
arates the posterolateral portion of the 
posterior narial wall from the postnarial 
arena on each side. It is not clear 
whether the posterior narial walls of A. 
spinosa tadpoles would be efficient as 
valves. The postnarial arena contains 
two tall, straight papillae that project 
anterolaterally under the narial valve 
projections. These have pustulate apices 
and some sign of terminal bifurcation. 
The median ridge is reduced to a tall 
papilla with an expanded base and a 
pustulate anterior margin. It is com- 


pressed in the sagittal plane and col- 
lapsed asymmetrically against the buccal 
roof in this specimen. Anterolateral to 
the median (ridge) papilla and within 
the postnarial arena are fifteen to twenty 
pustulations and a single, simple pa- 
pilla. Posterolateral to the median 
(ridge) papilla are two typical lateral 
ridge papillae. The buccal roof is devoid 
of any papillae that could outline a buc- 
cal roof arena. A hundred or more blunt 
pustulations, including a few stubby sub- 
papillate projections, cover the buccal 
roof proper. I could not resolve a glan- 
dular zone. The dorsal velum is com- 
pletely absent. 

Dorsal pharynx.—Without a dorsal 
velum the buccal and pharyngeal roofs 
are completely confluent. There are no 
pressure cushions. 

Diagnostic summary.—Anotheca lar- 
vae are uniquely different from all other 
tadpoles examined in the following fea- 
tures: single medial lingual papillae; 
gigantic median notch and ventral velar 
margin; filter rows reduced to thin, sin- 
gle folds; dorsal velum and_ dorsal 
pharyngeal features absent. 


Gastrotheca riobambae (Fowler) 
(Fig. 19) 


Material.—Author’s collection (stage 
37, sv. 20.6 mm). Laboratory raised; 
Mus. Nat. Hist. Univ. Kansas. 

References.—Hoogmoed, 1967; Duell- 
man, 1974. 


External.—G. riobambae larvae are 
large, with dorsolateral to nearly dor- 
sal eyes, a 2/3 denticle formula, and 
sinistral spiracle. Tadpoles of this species 
are often known under the name Gastro- 
theca marsupiata (see Duellman and 
Fritts, 1972 and Duellman, 1974, for the 
most recent discussion of the systematics 
of the genus). 

Ventral buccal.—The width to length 
ratio for the mouth is typical for a hylid 
tadpole but the floor is noticeably round- 
er than in most species examined. There 
are four infralabial papillae in a trans- 
verse line. The medial pair consists of 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 


Fic. 18.—Photographs of the floor (above) and roof (below) of the mouth of an Anotheca 
larva. 


49 


spinosa 


50 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


— 


Fic. 19.—Photographs of the floor (above) and roof (below) of the mouth of a Gastrotheca rio- 
bambae larva. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 51 


moderately tall, finger-like projections. 
The lateral two are shorter and may be 
considered bipartite. On each side dor- 
sal to the infralabial papillae, in the 
angle between the upper and lower 
beaks, is a small, but distinct kerati- 
nized spur. These spurs are directed 
anteriorly and separate from the main 
keratinized beaks. They are just far 
enough back inside the mouth so that 
they cannot contact a smooth substrate 
when the lower jaw is fully depressed. 
G. riobambae has four lingual papillae. 
The two tallest are near the midline; the 
other two are slightly shorter and pos- 
terolateral to the medial pair. The buc- 
cal floor arena is wide. The BFA pa- 
pillae series is continuous posteriorly 
across the midline. I counted 17 BFA 
papillae on one side, 19 on the other; 
only nine rather small BFA papillae are 
anterior to the buccal pockets. The 
papillae behind the pockets are not in 
tight rows but spread out laterally to 
cover the region of the buccal floor be- 
tween the median ends of the buccal 
pockets and the front of the second filter 
cavity. The four largest papillae, viz. 
those just medial to the buccal pockets, 
show basal bifurcation; a few of the 
other larger papillae have minor pustu- 
lations along their sides. Posterome- 
dially within the arena is a pair of small 
papillae. There are also about sixty tiny, 
faint pustulations randomly dispersed 
within the caudal two-thirds of the 
arena. G. riobambae has typical pre- 
pocket papillae. The buccal pockets are 
wide, with a strongly oblique orienta- 
tion. The pockets are shallow and per- 
forated. The free velar surface is long, 
supported by spicules; the posterior pro- 
jections of the velar margin are also long. 
The projections associated with the third 
filter cavity are displaced medially. All 
pairs of projections are directed postero- 
medial. The single pair of projections 
surrounding the median notch are thick- 
er than the rest of the velar surface; they 
are directed dorsad because the mid- 
portion of the velar margin is forced 


forward by a relatively gigantic laryn- 
geal disc. The projections are covered 
with dense, conspicuous secretory pits. 
Ventral pharynx.—tThe branchial bas- 
kets are long and deep. The baskets 
overall have a more circular, rather than 
typically oval, dorsal profile. This shape 
seems to reflect modifications of the 
third filter cavity. The tiny, third filter 
cavity, as viewed from above, is a nearly 
closed chamber capped by the imbricat- 
ing, curved dorsal margin of the filter 
plate on cb. 3. Filter plates of cb. 1 and 
cb. 2 are relatively long; plates of cb. 3 
are exceptionally tall. Counts for filter 
rows are very near the mean for H. 
regilla, but filter rows are wider with 
longer secondary folds. Filter canals are 
of typical proportions. Filter rows at 
the bottom of the plates lining filter 
cavities 1 and 2 are fused, not only with 
neighboring rows on each plate, but with 
the abutting filter rows of the opposing 
plate. Thus, the gill slits are completely 
obscured from dorsal view and water 
that enters those two filter cavities must 
pass through individual filter niches in 
order to reach the atrial chamber. The 
branchial food traps of G. riobambae 
cover a disproportionately large area in 
the anterior portions of the filter cavities. 
Secretory ridges are of greater absolute 
size and height than in a typical hylid 
tadpole, but whether this difference is 
significant could not be determined. 
While the glottal slit and glottal lip pro- 
portions are not unusual, the glottis in 
G. riobambae differs from most other 
tadpoles examined in having a predomi- 
nantly vertical orientation. The glottis 
sits on a huge, but not very conspicuous 
laryngeal disc. G. riobambae has esoph- 
ageal funnel of very broad profile. 
Dorsal buccal.—The roof of the 
mouth of G. riobambae is broad ante- 
riorly and quite round. The nares and 
median ridge are positioned far back on 
the buccal roof, resulting in relatively 
large pre- and postnarial arenas. The 
caudal half of the large prenarial arena 
is filled by a gentle, upside-down, U- 


52 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


shaped depression of the rostrum. Pus- 
tulations are absent within this arena. 
The internal nares are transversely ori- 
ented, quite small, and nearly as wide as 
long. The walls, particularly the medial 
portion of the anterior wall, are very 
thick. Pustulations are few and faint on 
the anterior wall; prenarial papillae are 
absent. Narial valve projections are very 
slight. A deep groove separates the poste- 
rior narial wall from the postnarial arena. 
Postnarial papillae are typical in posi- 
tion and number, although with pustu- 
late anterior margins. The median ridge 
is not unusual in shape or marginal 
sculpturing, but lacks pustulations or a 
secondary ridge on its anterior surface. 
Within the postnarial arena are three 
or four tiny, blunt papillae. One is lo- 
cated medially in the anterior half of the 
arena; the remaining ones are aligned 
transversely just in front of the median 
ridge. The lateral ridge papillae are 
slightly compressed into the sagittal 
plane. They are bifurcate, with shorter 
anterior and taller posterior projections, 
and have very tiny pustulations on their 
anterior surfaces. The BRA is well de- 
fined—oval in shape with a truncated 
anterior margin—by a continuous arc of 
no less than 16 papillae. These papillae 
are simple and moderate in size. In 
general, BRA papillae are evenly spaced 
in single file; however, in the postero- 
lateral margins of the arena there are 
One or two small secondary papillae 
neighboring the main row. Fifty to one 
hundred small, pointed pustulations are 
randomly dispersed within the arena. 
Three or four small, pointed papillae 
may be found clustered along the far 
lateral margins of the buccal roof. The 
glandular zone of G. riobambae is nar- 
row and tends to be of a uniform width 
across the buccal roof. Secretory pits 
are very conspicuous and large, as were 
the secretory ridges in the ventral 
pharynx; the apparent accentuation of 
larval secretory tissue in this species, 
however, may not be significant when 
overall size is taken into account. The 


dorsal velum is average or a bit larger 
than average for a hylid tadpole; the 
gap between the left and right sides of 
the dorsal velum is great. 

Dorsal pharynx.—G. riobambae has 
two large and conspicuous pressure 
cushions, obliquely oriented on each 
side. The lateral cushions are the larger 
and rounder, and the medial cushions 
are of subequal width to the lateral pair. 
The ciliary groove is not noticeably dif- 
ferent from that of any typical hylid 
larva. 

Diagnostic summary.—Gastrotheca 
riobambae has: ventral fusion over the 
gill slits of the filter rows from neigh- 
boring filter plates; a medially directed 
spur on each side at the front of the 
mouth and four lingual papillae. Cur- 
sory examination of larvae from other 
Gastrotheca species (G. monticola, KU 
142847, N = 1, stage 35; G. marsupiata, 
KU 139442, N = 4, all stage 36) sug- 
gest that these features characterize the 
genus not the species. In G. marsupiata 
the fusion of the filter rows, however, is 
incomplete in the posterior portion of 
the filter cavity. 


Hyla femoralis Sonnini & Latreille 
(Fig. 20) 

Material. — Uncatalogued, author's 
collection (stage 36, sv. 16.0 mm). Col- 
lected from small pond, 1.25 km NE 
Florida Technical Univ., Orange Co. 
Florida; August 12, 1975. 

Reference.—Wright, 1932. 

External.—This is possibly the most 
beautiful tadpole in North America. This 
specimen has a tall tail fin, which is al- 
most as tall as it is long, and ends in a 
pointed flagellum. The denticle pattern 
is 2/3; the spiracle is sinistral. 

Ventral buccal.—H. femoralis is so 
similar to H. regilla that only differences 
are emphasized in this description. The 
floor of the mouth is slightly broader 
posteriorly in H. femoralis than in H. 
regilla. The infralabial papillae are 
smaller and narrower. The lingual pa- 
pillae are positioned far forward and 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 53 


Fic. 20.—Drawings of the floor (above) and roof (below) of the mouth of a Hyla femoralis larva. 


54 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


almost touch the infralabial papillae. The 
buccal floor arena is an oval area well 
defined by 11 papillae on one side, eight 
on the other. The two to three largest 
papillae on each side, those directly me- 
dial to the buccal pockets, arise from a 
common ridge-like base. Approximately 
20 pustulations occur in the posterior 
half of the BFA. Four to six large pustu- 
lations/subpapillate structures are dis- 
persed on the buccal floor anterior to the 
buccal pockets. The buccal pockets are 
shallow. The floor of each pocket is 
completely obscured by a large fleshy 
fold from its dorsal wall. Any perfora- 
tion of the floor of the buccal pockets 
must be relatively small and very me- 
dial. The free velar surface is slightly 
longer than in H. regilla. 


Ventral pharynx.—The branchial bas- 
kets are larger than those of H. regilla. 
The posterior margin of the baskets are 
elongated, reflecting elongation of the 
second and third filter plates. The second 
and particularly the third filter plates are 
noticeably taller in this species. The 
third filter plate has an arched dorsal 
margin which curls medially and nearly 
completely covers the third filter cavity. 
Filter counts cb. 1-4 were: 11, 11, 14, 
17; these are all well above the counts 
for H. regilla at any stage. Filter rows 
abut completely; the filter canals are 
narrow and hidden from view. The fil- 
ters have a fully developed mesh made 
up of secondary, tertiary and higher 
order filter folds. Among the hylids ex- 
amined only Agalychnis had a filter mesh 
that was denser. Branchial food traps 
are necessarily large, covering the ven- 
tral surface of the ventral velum and 
much of the anterior region of each 
filter cavity. A pattern of well organized 
secretory ridges could be observed on 
the underside of the ventral velum. The 
glottis is of average size and only a 
quarter covered by the margin of the 
ventral velum. The lips are slightly 
heavier than those of H. regilla. The 
esophageal funnel is narrow. 

Dorsal buccal.——The buccal roof is 


indistinguishable in shape from that of 
H. regilla, except that the dorsal pharyn- 
geal region is longer. In the center of 
the prenarial arena descends a bulge 
approximately twice as wide as the mini- 
mum distance between the internal 
nares. The bulge has a distinctive an- 
terior arched margin which is lined with 
a half-dozen blunt pustulations (similar 
structures are seen in Smilisca sordida). 
The internal nares are identical to those 
of H. regilla in proportion, shape, orien- 
tation and valvular structure. The pre- 
narial papillae however are smaller than 
those of the average H. regilla. The 
median ridge is small and has a pustu- 
late irregular margin. Two large papil- 
lae with strongly pustulate edges lie 
halfway between the median ridge and 
the internal nares. Smaller papillae arise 
from the base of these large papillae on 
their anteromedial side. There are two 
large pustulations in the anterior portion 
of the postnarial arena and a large bi- 
cusped subpapillate projection in the 
middle of the posterior part of the post- 
narial arena. There is a large, flap-like, 
laterally compressed, irregularly shaped 
projection in the typical position of the 
lateral ridge papillae on each side. The 
buccal roof has three distinct papillae 
on each side arranged in an oblique row 
(anterolateral to posteromedial). These 
define the lateral bounds of the BRA. 
There is a single tiny papilla located far 
laterally on the buccal roof. Approxi- 
mately fifty pustulations are distributed 
between the median ridge and the pos- 
terior quarter of the BRA. The more 
posterior pustulations are smaller and 
more numerous. Three to four pustula- 
tions occur on the buccal roof lateral to 
the BRA. The glandular zone is rela- 
tively narrow and slightly longer later- 
ally than medially. It is made up of 
large distinct secretory pits. The dorsal 
velum is of comparable size to that of 
H. regilla. It is barely continuous across 
the midline. 

Dorsal pharynx.—The pressure cush- 
ions are large obliquely oriented swell- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 55 


ings. The lateral pressure cushion is the 
larger. The medial pressure cushion has 
itself a medial swelling presumably a 
pressure cushion for the third filter cav- 
ity. The ciliary groove is a large open 
trough. 

Diagnostic summary.—The oral cay- 
ity of H. femoralis is very similar to that 
of H. regilla but can be distinguished 
from the latter species by larger bran- 
chial baskets, taller filter plates and 
denser gill filters. Papillation and pustu- 
lation is slightly more extensive in H. 
femoralis than in H. regilla. 


Hyla rufitela Fouquette 
(Gisigst 2b 22)) 


Material. — Uncatalogued; author's 
collection (stage 36, sv. 15.4 mm). Col- 
lected in grassy pools on the edge of the 
town Rincon de Osa, Puntarenas, Costa 
Rica; March 3, 1970. 

Reference.—Duellman, 1970 (p. 243- 
244). 

External—tThis tadpole is propor- 
tioned like most “generalized” hylid lar- 
vae. The spiracle is sinistral. These tad- 
poles are otherwise unusual in having a 
2/4 denticle pattern. 

Ventral buccal.—The interior of the 
mouth has a rounded appearance though 
the width to length ratio for the oral 
floor is average. This species has super- 
numerary infralabial papillae, eight per 
side in the specimen at hand. Of the 
eight, three are far anterior, stubby, 
blunt and with auxiliary pustulations; 
these form a transverse row on each 
side. The largest of the three is nearest 
to the midline. Immediately behind the 
middle of these is a wide, blunt papillae 
of medium size. Posterior to this are 
three compressed papillae with jagged 
margins which form a second transverse 
row. The middle papilla of this poste- 
rior row is of the same size as the pa- 
pillae in the front row; the other two 
are much smaller. The last infralabial 
papilla on each side is behind the second 
row, and it is the largest of all eight 
papillae. Topographically the most pos- 


terior papilla on each side appears to be 
the homologue of the infralabial papilla 
in species that have only one per side. 
This papilla is a large, transverse, slight- 
ly conical fan. It has a constricted base 
and strongly papillate margin. The sec- 
ondary marginal papillae, numbering 
four and six per side, are further deco- 
rated with yet finer jagged projections. 
The BFA is demarcated by many small 
papillae and pointed pustulations. In ad- 
dition to having papillae outlining an 
ovoid BFA, a transverse row of papillae 
above the junction of the ceratohyals 
and hypobranchial plates splits the arena 
into anterior and posterior portions. I 
counted thirteen BFA papillae on one 
side and fifteen on the other. Little 
weight can be given these counts, how- 
ever, because of great difficulty in as- 
sessing what is a “miniscule papilla” 
versus a “subpapillate pustulation” in 
this particular species. Larger projec- 
tions that are unquestionably of “pa- 
pilla” proportions tend to be small, com- 
pressed cylinders; three had terminal bi- 
furcations. The largest papillae are in 
the transverse row and pointed ante- 
riorly. The concentration of pustulations 
and papillae elsewhere on the buccal 
floor, including lingual papillae and pre- 
pocket papillae are typical of hylid lar- 
vae. The buccal pockets are long and 
very obliquely oriented, with a predomi- 
nant angle of 30° from the transverse 
plane. The pockets are definitely per- 
forated, but the slits are obscured from 
dorsal view by strong forward projec- 
tions of the mediodorsal portion of the 
posterior pocket walls. The extensive 
free velar surface is supported by long, 
thin spicules that nearly reach the velar 
margin, which is a relatively smooth 
broad arch. Marginal projections over 
the filter cavities are single, faint peaks. 
The medial portion of the velar margin 
is the smoothest segment, with only the 
tiniest median notch. A very narrow 
band of moderately large and distinctive 
secretory pits lines the dorsal surface of 
the velar margin. The pits are absent 


MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


56 


Fic. 21.—Drawings of the floor (above) and roof (below) of the mouth of a Hyla rufitela larva. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 57 


lateral to the tip of the spicule over cb. 
2 and in the immediate neighborhood of 
the median notch. 


Ventral pharynx.—The branchial bas- 
kets in toto, and the filter cavities in 
particular, are large, despite the fact 
that the width to length ratio for the 
branchial baskets is not unusual. The 
increased size is reflected in counts of 
filter rows, viz. 9, 11, 11, 9 for cb. 1-4. 
These are higher than for most hylids 
examined, the count for cb. 4 significant- 
ly so. Ratios of length to height for the 
filter plates are not great. Thus, al- 
though the plates are longer, they are 
not taller. The filter mesh is dense; filter 
rows are wide and secondary filter folds 
are long. There are few partial filter 
rows. The filter canals are fully cano- 
pied. The branchial food traps are long 
because the velar surface is large. The 
traps, however, do not cover a great ver- 
tical area and are presumably limited by 
the shallowness of the branchial baskets. 
Secretory ridges are visible through the 
velum. These are extremely uniform in 
spacing and shape and continuous above 
all the filter cavities on each side. The 
glottis is 80% blanketed by the ventral 
velum. The glottal lips are of uneven 
thickness (thickest posteriorly) but of 
typical elevation. The esophageal funnel 
has a slightly narrower dorsal profile 
than in typical hylid larvae. 

Dorsal buccal.—The buccal roof has 
the same general positioning of major 
features, such as nares and median 
ridge, of most hylids; specific regions, 
however, differ grossly. In the middle 
of the prenarial arena is a small trans- 
verse pustulate row, or low ridge, no 
wider than the internarial distance. The 
anterior narial wall makes a full loop 
into the prenarial arena on each side, 
forming huge vacuities with high, thick, 
pustulate walls. Although these pockets 
are attached to the internal nares, they 
are cul de sacs with smooth round bot- 
toms. These structures may be Jacob- 
son’s organs, but they have not been 
seen in any other hylids. Lateral to these 


pockets the anterior narial wall has one 
to three tiny, pointed prenarial papillae 
and pustulations. The narial opening is 
a large elongate oval, oriented a bit 
more obliquely than in typical hylid lar- 
vae. The posterior narial walls are tall 
with an arched ventral margin. Narial 
valve projections have pustulate apices 
that extend forward part way under the 
unusual prenarial vacuities. There are 
six to eight postnarial papillae, with 
additional pustulations; these loosely 
aligned in a broad, anteriorly directed 
“V.” The papillae are all rather small, 
subequal, blunt cylinders. The larger 
are the more posterolateral and tend to 
have terminal pustulations. The median 
ridge is more rectangular than triangu- 
lar. Its free margin is slightly sculptured 
with faint serrations; a few pustulations 
occur on the ventral half of the anterior 
surface. Lateral ridge papillae are 
shaped like halves of crescents with their 
peaks pointed medially. Their longer, 
more anterior margins are strongly ser- 
rated. Two or three tiny, pointed pa- 
pillae are found in the lateral margins 
of the buccal roof, but they do not de- 
fine a BRA. Instead, there are 200 or 
more small pustulations evenly spaced 
in the middle third of the buccal roof. 
These thin out laterally. The glandular 
zone is made up of large, distinctive 
secretory pits. The zone, overall, is short 
and of uneven length. The zone is long- 
est laterally and nearly absent on the 
midline. The dorsal velum is of normal 
proportions or of slightly greater lateral 
length than in other hylids examined. It 
extends across the midline as a low, but 
movable flap-like ridge. 

Dorsal pharynx.—Pressure cushions 
are present as two large, obliquely ori- 
ented bulges on the posterior surface 
of the dorsal velum on each side. These 
are more distinctive than those seen in 
other species of the genus. The lateral 
cushion is the longer; the medial cushion 
is the rounder. Medial to the inner cush- 
ion on each side is a small roll in the 
dorsal velum which may be a vestige of 


58 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Fic. 22.—Photographs of the floor (above) and roof (below) of the mouth of a Hyla rufitela larva. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 59 


a third cushion. Behind the pressure 
cushions are comparatively broad, shal- 
low ciliary grooves. 

Diagnostic summary.—Hyla rufitela 
differs from all other tadpoles examined 
by the development of large cul de sac 
vacuities extending anteriorly from the 
anterior narial walls. 


Hyla dendroscarta Taylor 
(Fig. 23) 

Material—MVZ uncatalogued (stage 
26, sv. 12.0 mm). Collected from large 
bromeliad by the road at 1400 m eleva- 
tion (Highway 150), approximately 5 km 
W Orizada, Vera Cruz, Mexico; Sept. 16, 
1972. 

Reference. — Duellman, 1970 (p. 
436). 

External.—Hyla dendroscarta larvae 
are extremely elongate, with a dorsoven- 
trally flattened body and a sinistral spir- 
acle. The denticle pattern is 3/4. 

Ventral buccal_—The buccal cavity 
is squarish, rather broad anteriorly. The 
infralabial papillae are simple pads that 
lack secondary papillae and pustulations. 
They are of moderate size and just touch 
on the midline. There are two, typical, 
lingual papillae. The BFA is well de- 
marcated by two nearly straight rows of 
papillae that converge posteriorly to 
form a “V.” I counted eight on one side 
and nine on the other, with an equal 
number of pustulations arranged along 
the same general line as the papillae. 
Some small additional papillae occur 
lateral to the posterior margin of the 
arena. All the papillae have sharp 
apices. The central region of the arena 
is bare. The prepocket region is bare. 
Careful manipulation failed to reveal 
open buccal pockets at 50 magnifica- 
tion. The free velar surface is of average 
length for typical pond larva, but has a 
relatively dense glandular margin. Indi- 
vidual secretory pits could not be re- 
solved at 125. Supporting spicules ap- 
pear to be relatively short but the mar- 
gin appears quite stiff. There are a few 
minor crenulations and a deep median 


notch over the glottis, but no distinctive 
marginal papillae. 

Ventral pharynx.—The branchial bas- 
kets are oval, almost round, in dorsal 
view with a length to width ratio of 
about 1. They are of average depth, 
compared to other hylids of similar size. 
The first and second filter cavities are 
quite open and fully exposed from 
above. The third filter cavity, however, 
is small and largely hidden by the im- 
brications of the filter plates on the third 
ceratobranchial. The dorsal margin is 
straight on the second filter plate, but 
has a distinctive apex on the third filter 
plate. I counted: 10 filter rows on cb. 
1; 10 on cb. 2; 10 on cb. 3; 7 on cb. 4. 
The filter rows are particularly dense 
and thick at the base of the branchial 
baskets. They are packed such that 
without subjecting them to extreme ma- 
nipulation they completely obscure the 
gill slits in dorsal view. Partial filter 
rows are short. The density of the 
filters at the base of the filter rows, is an 
unusual feature; the filter density overall 
is slightly less that of typical pond hy- 
lids, largely due to more open filter 
canals. The ventral surface of the ven- 
tral velum is covered by secretory tissue, 
evidenced by its buffed texture and 
staining properties. The branchial food 
traps are relatively short and well or- 
ganized secretory ridges were not ob- 
served. The glottis is well developed 
particularly for a tadpole of this stage. 
The glottis lies more than half way under 
the margin of the ventral velum but is 
still fully visible in dorsal view because 
of a large notch above it. The pharyn- 
geal disc is not particularly elevated. A 
tiny papilla projects up from the front 
of the glottal lips where they meet on 
the midline. This is completely hidden 
in the undistorted specimen by the ven- 
tral velum. 

Dorsal buccal—The roof of the 
mouth shares with the floor a generally 
squarish shape. The nares and median 
ridge have a typical position in the 
mouth for hylid larvae. There is a small 


60 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Fic. 23.—Drawings of the floor (above) and roof (below) of the mouth of a Hyla dendroscarta 
larva. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 61 


medial swelling in the prenarial arena. 
In the middle of this swelling are two, 
short, blunt papillae oriented on a trans- 
verse line and adjacent to each other at 
the midline. The nares are obliquely 
oriented with a slightly swollen rim sur- 
rounding them both anteriorly and pos- 
teriorly. A few minor pustulations arise 
from the lateral portions of the narial 
wall, but prenarial papillae proper are 
absent. The narial valves are large free 
flaps with a distinctive projection aris- 
ing between the midpoint and the me- 
dial margin of the posterior wall. The 
postnarial arena is triangularly shaped 
and is defined anteriorly by a short row 
of papillae, two on one side and three 
on the other, plus a relatively small me- 
dian ridge posteriorly. The short median 
ridge has three distinctive projections 
with the single medial one most distinc- 
tive. Lateral to the median ridge are 
two attenuate lateral ridge papillae, one 
of which has a secondary pustulation 
near its base. These lateral ridge pa- 
pillae are in line with two rows of buc- 
cal roof papillae that extend backward 
in almost parallel lines. Pustulations are 
lacking in the postnarial arena but are 
quite evident in the BRA. The BRA is 
a tall trapezoid with the median ridge 
and the lateral ridge papillae forming 
the wider base. Seven tall papillae, all 
lacking secondary pustulations, form the 
BRA papillate rows on each side. Pa- 
pillae and pustulations are lacking else- 
where on the buccal roof. The glandular 
zone is short with a relatively uniform 
length from the lateral margin of the 
roof of the mouth to the midline. 
Dorsal pharynx.—The dorsal velum 
is also short with two, distinctive, round 
pressure cushions of subequal size. The 
dorsal velum is absent on the midline. 
The esophageal funnel has a narrow pro- 
file. The ciliary groove is barely visible 
laterally but it is a distinct, open trough 
as it leads into the esophagus. 
Diagnostic summary.—The varying 
density of gill filters—very dense ven- 
trally and much less dense dorsally— 


readily characterizes this larvae. This 
feature, however, may simply reflect the 
early ontogenetic stage of this specimen 
(H. dendroscarta larvae are rare in col- 
lections and this was the largest speci- 
men available for study. The specimen 
was 40% longer than a free-living, feed- 
ing, stage 25 individual also examined). 
Several less questionable, but also less 
obvious features collectively distinguish 
this larvae from others examined. These 
include: simple, blunt infralabial pa- 
pillae in association with tall, attenuate 
BFA, BRA and lingual papillae; absence 
of obvious secretory pits or ridges in 
branchial food traps, presence of papilla 
at anterior edge of glottis; thickened rim 
around naries. 


Hyla phlebodes Stejneger 
(Fig. 24) 


Material—KU 68405 (stage 31, sv. 
7.3 mm). Collected in forest pool at 
100 m elevation, Puerto Viejo, Heredia, 
Costa Rica; July 15, 1961. 

Reference.—Duellman, 1970 (p. 222). 

External. — Hyla phlebodes (like 
other members of the Hyla microceph- 
ala species group) is characterized by a 
tail that terminates in a long pointed 
filament, tiny recessed beaks, and the 
absence of denticles and oral disc pa- 
pillae. The spiracle is sinistral. 

Ventral buccal.—The trend toward 
reduction of oral structures seen exter- 
nally is continued internally; this reduc- 
tion is greater in H. phlebodes than in 
other Hyla larvae examined. The floor 
of the mouth is disproportionately elon- 
gate anterior to the buccal pockets. Be- 
cause of the tiny beaks the mouth comes 
to an acute anterior termination. Single, 
relatively huge, infralabial papillae on 
each side overlap one another at the 
midline. These are roundish, slightly 
antero-posteriorly compressed structures, 
which lack any secondary projections. 
Each is larger than the oral orifice. 
There are no other papillae anywhere 
inside the mouth. The only surface 
features arising from the buccal floor 


62 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Fic. 24.—Drawings of the floor (above) and roof (below) of the mouth of a Hyla phlebodes larva. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 63 


are five or six tiny pustulations above 
the lateral portion of the cerato- 
hyal on each side. The buccal pockets 
are tiny, shallow and unperforated. They 
extend back from the midline at a great 
angle. The free velar surface is very 
short and unsupported by spicules. The 
trailing edge is a smooth curve without 
any posteriorly directed projections al- 
though a large, deep, median notch is 
present. Secretory pits could not be re- 
solved on the velar margin and are prob- 
ably absent. 


Ventral pharynx.—The branchial bas- 
kets are reduced in all planes. Filter 
plates are practically absent. Knobs on 
the gill bars, vestiges of filter rows, line 
the gill slits. These “filter rows” num- 
ber 4, 5, 5, 3 for ceratobranchials 1-4 
respectively. Filter folds are absent. 
There is little space left for branchial 
food traps in the shallow pharynx, and 
they may be absent. Secretory ridges, if 
present, would have to be restricted to 
the ventral surface of the velum, but 
none could be resolved on that surface. 
A large glottal slit fills the median notch 
of the velum. The slit sits on well- 
developed, elevated lips and is fully 
exposed when viewed from above. There 
is a large, unelevated laryngeal disc. 
Behind this extends a broad esophageal 
funnel. 


Dorsal buccal.—The buccal roof, like 
the buccal floor, is elongate. The nares 
are far anterior and the prenarial arena 
is proportionally small. Surface struc- 
tures are absent in the prenarial arena. 
The internal nares are small slits oriented 
at an angle 35° from the transverse 
plane. The narial walls are not tall and 
lack any sculpturing such as prenarial 
papillae or narial valve projections. Not 
quite halfway back on the buccal roof 
are two single pustulations, one each 
aligned directly behind the lateral mar- 
gins of the nares. With intense staining 
some faint pustulations may be seen on 
the middle posterior portion of the roof. 
The buccal roof is otherwise devoid of 
any ridges, pustulations, papillae, etc. 


A discrete anterior margin for the secre- 
tory zone could not be resolved, al- 
though tiny secretory pits are present 
posteriorly in the esophageal funnel. 
Remnants of a dorsal velum are present 
as two tiny, asymmetric, anteriorly di- 
rected flaps. These flaps are completely 
absent both in front of the large esopha- 
geal funnel and laterally; in total they 
traverse only a quarter of the mouth. 

Dorsal pharynx.—The two small flaps 
identified as remnants of the dorsal 
velum are far posterior and a dorsal 
pharynx is essentially absent. Pressure 
cushions are not present and the ciliary 
groove is a tiny crease behind the ves- 
tigeal dorsal velum. 

Diagnostic summary.—Some material 
of Hyla microcephala has been available 
for study; in that no major differences 
were found between H. microcephala 
and Hyla phlebodes, the following diag- 
nosis can stand as a diagnosis for the 
H. microcephala species group as a 
whole. These larvae can be distin- 
guished from all other tadpoles exam- 
ined by the unique knob-like vestiges of 
the filter rows on the diminutive cerato- 
branchials. In addition, H. phlebodes 
has the most massive, and unusually 
shaped, ceratohyals (and associated 
musculature) of any Hyla examined. 
The following selected features in com- 
bination further diagnose this species: 
large, medially, overlapping infralabial 
papillae; buccal papillae absent; glottis 
large; branchial baskets very small and 
filter row remnants few in number; se- 
cretory ridges absent; esophageal funnel 
large. 


Hyla mixe Duellman 
(Fig. 25) 


Material—KU 104183 (stage 28, sv. 
11.8 mm). Collected attached to small 
stones in gravel-bottom pools in cascad- 
ing small stream on N slope of the Sierra 
de Juarez, 4.2 km S of Compamento 
Vista Hermosa, Oaxaca, Mexico; Feb- 
ruary 16, 1966. 


64 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Fic. 25.—Drawings of the floor (above) and roof (below) of the mouth of a Hyla mixe larva. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 65 


Reference.—Duellman 1970 (p. 426- 
497). 

External—Tadpoles of this genus 
have a relatively huge, ventral, oral disc 
with a 7/10 to 7/11 denticle pattern. 
The spiracle is sinistral. 


Ventral buccal.—The width to length 
ratio of the whole oral cavity reveals a 
relatively short oral cavity in this spe- 
cies. These tadpoles have the most ex- 
treme proliferation of internal oral pa- 
pillae seen in any Hyla. Five small 
posteriorly directed papillae are present 
at the anterior limit of the infralabial 
region. These are evenly spaced imme- 
diately behind the keratinized beak. 
Two, major, fringed flaps are behind this 
row and quite similar in shape and posi- 
tion to typical hand-like infralabial pa- 
pillae. These later papillae are wide, 
and their secondary marginal papillae 
(four per side) are long and attenuated. 
Fine tertiary divisions of the marginal 
papillae are present on both sides. Two 
pustulations are present adjacent to the 
midline between the major pair of in- 
fralabial papillae. No other symmetrical 
projections neighbor the midline in the 
immediate region behind these pustula- 
tions, so they may be the homologues of 
the lingual papillae despite their rela- 
tively anterior position and small size. 
The buccal floor arena is sharply de- 
marcated posteriorly and laterally by a 
continuous fringe of 70 small to medium- 
sized, pointed papillae. The fringe be- 
gins far laterally in front of the buccal 
pockets, curves mesad around the end 
of the pockets, and continues obliquely 
to a point on the midline just anterior 
to where the free velar surface begins. 
The BFA papillae and prepocket pa- 
pillae series are thus connected. This 
line of papillae is occasionally two pa- 
pillae wide along its length. The papillae 
in the anterolateral and posteromedial 
portions of the BFA are the shortest. 
The tallest papillae are immediately me- 
dial to the buccal pockets. A very few 
of the larger papillae exhibit bifurca- 
tions near their bases. Many of the 


larger papillae have bumps or pustula- 
tions about halfway from their base. The 
papillae are virtually all curved; those in 
the typical BFA position are curved an- 
teromedially; those in the prepocket 
position arch directly posterior over the 
buccal pockets. Anterior to the fringe 
the buccal floor is void of pustulations 
or projections. Behind the fringe are 
secondary transverse rows of very small, 
anteriorly curved papillae. These rows 
are made up of five papillae on one side 
and four on the other. They are re- 
stricted to the areas between the base of 
the spicules of cb. 2 and cb. 3 on the 
most posterior part of the non-velar buc- 
cal floor. The buccal pockets are slender 
and transversely elongate. They are deep 
and each is occluded by a large, ante- 
riorly directed fold arising from the pos- 
terior dorsal surface of the pocket on 
each side. The overhanging papillae 
from the prepocket portion of the pa- 
pillae wall on the buccal floor further 
obscures the entrance to the buccal 
pockets. It is doubtful that the pockets 
are perforated in this species. Although 
the ratio of the length of the mouth to 
the maximum length of the velum is 
within the range for more typical hylids; 
the total velar area may actually be a 
bit larger. The supporting spicules are 
relatively heavy and long. They reach 
the posterior margin of the velum, caus- 
ing protuberances in the velar margin 
visible from above. Four undulating 
peaks on the velar margin are distinctly 
leptokurtic; they point posteromedially 
rather than posteriorly. Two auxillary 
peaks neighbor the median notch. A 
few small secretory pits were seen on 
the dorsal surface of the posterior edge 
of the velum, but the overall density or 
anterior extent of the secretory pits 
could not be assessed. 

Ventral pharynx.—The branchial bas- 
kets are relatively short, and this may 
account for the appearance of shorten- 
ing in the whole oral cavity. The most 
extensive reduction is in the fourth cera- 
tobranchial. Compared to other mem- 


66 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


bers of the genus, the branchial baskets 
are shaped as if they were pushed out- 
ward along their posteromedial margin. 
The third filter cavity is a slim, deep 
pocket concealed by the recumbent dor- 
sal margin of the filter plates on cb. 3. 
Because the branchial baskets are short 
while the velum is not, the filter cavities 
are blanketed more fully by the velum 
than in most species examined. Less 
than 10% of filter cavity 3 lies behind the 
trailing edge of the ventral velum. In 
terms of relative height of the filter 
plates and numbers of filter rows (i.e., 
Co, 1 == 10, Go, O = NOY, Go,’ S == GIO); 
cb. 4 = 4-5), Hyla mixe is not signifi- 
cantly different from more typical hylid 
larvae. The filter mesh is less dense. 
Secondary and finer filter folds are thin 
and the space between neighboring filter 
rows is large. Thus, the filter canals 
may be slightly more open in H. mixe 
than in other congeners. The large, stiff 
spicules made exploration of the ventral 
velar surface particularly difficult and 
I could not resolve secretory ridges any- 
where under the velum. However, in 
this specimen broad bands of light, floc- 
culent food matter were adhering to the 
ventral surface of the velum parallel to 
the posterior velar margin, suggesting 
the presence of some organized secre- 
tory tissue in this region. The glottis is 
open and slightly elevated but glottal 
lips are absent and the laryngeal disc 
is not visible; this may in part reflect 
the early developmental stage of this 
specimen. The glottis is more than 80% 
hidden from dorsal view by the edge of 
the velum. Behind the glottis is a mod- 
erately broad esophageal funnel. 
Dorsal buccal.—A massive soft, fleshy 
cone extends down from the prenarial 
arena into the medial space between the 
infralabial papillae. The base of the 
structure is oval in shape, about half as 
long as wide, and is 50% wider than the 
internarial distance. The height of the 
cone is equal to its width. As the cone 
descends it gently curls anteriorly. A 
half dozen pustulations cover the sides 


of the cone near its pointed tip. Directly 
in front of the internal nares are single, 
small, thin papillae, one on each side. 
No other structures are present in the 
prenarial arena. The internal nares are 
3 1/2 times wider than long (length 
measured along longer axis), significant- 
ly longer than in more typical hylid lar- 
vae. The angle made by the long axis 
of the internal nares from the transverse 
plane (45°) is very large. The anterior 
narial walls are characterized by a few 
knobby pustulations and a single tiny pa- 
pilla on each side very near the antero- 
medial corners on each side. The poste- 
rior wall is not very tall and completely 
lacks a narial valve projection. Each 
side of the postnarial arena is bound by 
a straight row of three attenuate postnar- 
ial papillae. These rows are oriented 
nearly parallel to the sagittal plane, but 
with their most anterior papillae slightly 
closer than their posterior papillae. The 
three papillae on each side increase in 
size from front to back. The largest one 
is as tall as the conical protuberance in 
the prenarial arena. The two other pos- 
terior papillae in each row arch slightly 
medially and have secondary pustula- 
tions on their anterior margins. In the 
position of the median ridge is a single, 
tall, extremely attenuate papilla which 
curves slightly anteriorly and has a ru- 
gose anterior surface. It is subequal to 
the tallest postnarial papillae. There are 
no pustulations within the postnarial 
arena. Two projections are evenly spaced 
directly behind the median papilla on 
the midline. The anterior one is a plain, 
small papilla; the posterior one is but a 
pustulation. The lateral ridge papillae 
are developed into relatively huge, lon- 
gitudinally oriented, papillate flaps. The 
flaps are as tall as long and point me- 
dially. They are displaced slightly pos- 
terior in relation to the median (ridge) 
papilla. Distinct attenuate marginal pa- 
pillae, four on one flap and five on the 
other, account for more than half of the 
maximum height of these projections. 
The marginal “papillae” have pustulate 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 67 


sides in addition to one or two minor 
basal bifurcations. The median (ridge) 
papilla and the flap-like ridge papillae 
outline the anterior end of a large, nearly 
rectangular, buccal roof arena. The pos- 
terior portion of this BRA is defined by 
lateral rows of five simple, thin papillae. 
The first four in each row are in a con- 
tinuous line running back from the flap- 
like lateral ridge papillae. The last and 
smallest papilla in each row is displaced 
medially. The tallest papillae in the 
BRA series are half the height of the 
lateral ridge papillae. Some half dozen 
faint pustulations are scattered in the 
anterior end of the BRA. Lateral to the 
BRA and two-thirds of the distance back 
on the buccal roof are single, small pa- 
pillae, one per side. The only other fea- 
tures on the buccal roof are some pustu- 
lations directly behind these lateral roof 
papillae. These pustulations contact the 
anterior edge of the glandular zone. The 
glandular zone has a very uniform, buff 
texture. It was virtually impossible to 
resolve the individual miniscule secre- 
tory pits of the glandular zone even at 
75>. The anterior margin of the secre- 
tory zone is arched posteriorly, so that 
the zone has a 50% greater length later- 
ally than on the midline. The dorsal 
velum is strongly coiled, but did not 
appear to be particularly long in this 
specimen (some of the dorsal velum was 
destroyed in dissection). If the velum 
is continuous across the midline, it is not 
exceptionally tall in that region. 


Dorsal pharynx.—A single, shallow, 
mediolaterally elongated pressure cush- 
ion could be resolved on one side but 
was removed from the specimen along 
with portions of the dorsal velum in 
order to gain access to the filter cavities 
during the study of ventral features. 
Details of the ciliary groove could not 
be determined. 

Diagnostic summary.—Hyla mixe 
tadpoles are unique among the tadpoles 
examined in having the papillae of the 
buccal floor arena so numerous and close 
to each other that a continuous fringe is 


formed. The larva are also distinguished 
by having: reduced third filter cavities; 
a massive cone-like projection of the 
prenarial arena; a median ridge reduced 
to a papillae; extremely large lateral 
ridge papillae. 


Hyla ebraccata Cope 
(Fig. 26) 


Material—KU 104130 (stage 35, sv. 
10.8 mm). Collected amidst emergent 
weedy, vegetation in shallow pond in 
clearing at edge of forest, at 100 meters 
elevation, Puerto Viejo, Heredia, Costa 
Rica; June 21, 1966. 

Reference.—Duellman, 1970 (p. 230- 
W3')) 

External. — Hyla ebraccata larvae 
have a long, pointed filamentous tail. 
They have a small, anteriorly directed 
mouth without denticle rows and a 
highly reduced oral disc. The eyes are 
far lateral; the spiracle is sinistral. 

Ventral buccal.—These larvae are 
characterized by a series of reductions 
in the oral cavity from what may be con- 
sidered the typical anuran condition. 
Single anterior to posterior compressed 
infralabial papillae are present on both 
sides. These small, flap-like papillae are 
approximately twice as wide as tall and 
have a slightly rugose dorsal margin. 
On the midline are a pair of small pa- 
pillae which may be homologues of the 
lingual papillae. Their far forward posi- 
tion, on the infralabial cartilage directly 
between the infralabial papillae, speaks 
against homologizing these projections 
with the lingual papillae; on the other 
hand, there are no other papillae di- 
rectly behind them and the anlage of the 
tongue is absent. Five or six very fine, 
simple papillae are dispersed laterally 
about the buccal floor and in front of 
the buccal pockets. A buccal floor arena 
per se is absent. Pustulations present on 
the buccal floor are scattered largely in 
front of the buccal pockets and are 
about as numerous as the buccal floor 
papillae. The buccal pockets are shal- 
low and clearly not perforated. The 


68 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Fic. 26.—Drawings of the floor (above) and roof (below) of the mouth of a Hyla ebraccata larva. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 69 


trailing edge of the ventral velum is a 
rather smooth semicircle. The undula- 
tions in the edge normally present above 
each filter cavity in other hylid larvae 
are reduced greatly in amplitude. The 
lateral peaks over filter cavity 1 are more 
distinct than those over filter cavity 3. 
The medial portion of the velar margin 
has a weak notch but is, otherwise, free 
of any abrupt contouring. Tiny secre- 
tory pits are on the dorsal velar margin. 
These occur quite far forward and cover 
much, if not all, of the free dorsal velar 
surface. The pits are densest near the 
midline. The free velar surface is com- 
paratively short considering that the 
spicules are quite long. This makes the 
edge rather stiff and immobile medially. 
The free velar surface also tends to be 
thick; secretory ridges cannot be seen 
without turning the edge over. 


Ventral pharynx.—The branchial bas- 
kets are about as long and wide as in 
typical members of the genus, but not 
nearly as deep. The filter plates are so 
short, and imbricated to such an extent, 
that filter tissue on cb. 2 and 3 is vir- 
tually restricted to the lateral faces. The 
filter plates of the first three ceratobran- 
chials are four to five times as long 
as high. Those of the fourth cerato- 
branchial are about twice as long as 
high. The number of full filter rows on 
the filter plates are: cb. 1 = 8, cb. 2 = 
8-9, cb. 3 = 11, cb. 4 = 8-9. Except for 
the high number of filter rows on cb. 4, 
these are well within the range for more 
typical hylid larvae. There are no par- 
tial filter rows. The filter mesh is very 
loose and, although secondary filter folds 
are numerous, they are quite thin and 
tertiary folds are lacking. Neighboring 
filter rows on each filter plate do not 
abut, consequently, filter canals are open 
channels along their whole length. Well 
defined branchial food traps are asso- 
ciated only with the third filter cavity; 
all secretory ridges are concentrated 
laterally and restricted almost complete- 
ly to the horizontal surface of the velum. 
Where secretory ridges appear, they are 


moderately wide and not dense. The 
trough between the secretory ridges is 
shallow. As viewed from above about 
half of the glottis is under the velum. 
The laryngeal disc is transversely elon- 
gated. The posterior margin of the disc 
is turned up so that the glottal slit is 
oriented more vertically than horizon- 
tally. Glottal lips are well developed. 
The esophageal funnel is broad and 
the lumen of the esophagus extremely 
large. The funnel is tightly juxtaposed 
to the laryngeal disc and by its mere 
size seems responsible for the distortion 
in the laryngeal disc. 


Dorsal buccal.—The prenarial arena 
is short. A single, relatively large bulge, 
which is convex anteriorly, arises in this 
arena and takes up most of its area. 
There are between one and two dozen 
tiny pustulations in the prenarial arena, 
mostly on the margins of the “bulge.” 
The internal nares are small; narial walls 
are not tall. Prenarial papillae are ab- 
sent from the anterior walls, although the 
walls’ margins are slightly rugose. The 
narial valves are two to three times as 
wide as tall and lack narial valve projec- 
tions. Both postnarial and buccal roof 
arenas are absent. A short, transverse 
crease halfway back on the buccal roof 
may be a remnant of the median ridge. 
It is not particularly wide, nor straight, 
and is relatively farther posterior than 
the median ridge in most Orton type 4 
tadpoles. In front of this crease are a 
dozen or so very tiny, pointed pustula- 
tions. These are absent in the back half 
of the buccal roof. The only papillae on 
the buccal roof are two, lateral to the 
median (ridge) crease and displaced 
slightly forward. These papillae are 
small, simple and may be considered 
either lateral ridge or buccal roof pa- 
pillae, depending on the faith one has 
in the homology of the “crease” with the 
median ridge. The tiny secretory pits, 
comprising the glandular zone, are suf- 
ficiently difficult to resolve that the an- 
terior limits of the zone cannot be 
determined. Laterally, the dorsal velum 


70 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


is short; however, the velum is continu- 
ous across the midline. Whereas the 
lateral portions of the dorsal velum 
project anteriorly, the velum_ twists 
sharply to project directly caudad on 
the midline. This middle portion of the 
velum is half as long as the velum at its 
widest lateral point. 

Dorsal pharynx.—Both medial and 
lateral pressure cushions are fused into 
a single, relatively large but shallow, 
roundish pad. Consistent with the large 
esophageal funnel, the ciliary groove is 
wide and not very deeply entrenched. 

Diagnostic summary.—See Hyla sa- 
rayacuensis (p. 72). 


Hyla sarayacuensis Shreve 
(Fig. 27) 
Material—KU 121413 (stage 34, sv. 


11.1 mm). Collected in grass marsh at 
1260 meter elevation, Rio Negro, Tungu- 
rahua, Ecuador; July 25, 1968. 

Reference.—None. 

External—The larvae of Hyla sa- 
rayacuensis have not been described in 
detail. It is sufficient to say that H. 
sarayacuensis tadpoles possess the assort- 
ment of larval features characteristic of 
the Hyla leucophyllata group: a tiny 
inset terminal mouth, absence of a pa- 
pillate labial disc and denticle rows, 
xiphiceral tail filament with tall equal 
or subequal dorsal and ventral fins. The 
spiracle is sinistral. Compared to H. 
ebraccata, H. sarayacuensis tadpoles 
have a relatively tall tail. 


Ventral buccal—The interior of the 
mouth of H. sarayacuensis is so similar 
to H. ebraccata that all features which 
unify the two species and contrast them 
with more typical Hyla larvae are not 
repeated. Instead, emphasis is placed 
on the few differences between H. sa- 
rayacuensis and H. ebraccata. A single 
anterior to posterior compressed infra- 
labial papilla is present on each side in 
H. sarayacuensis. These have blunt dor- 
sal margins and are smaller and about 
half as wide as the infralabial papillae 
in H. ebraccata. Medial papillae of any 


sort, such as those suggested as possible 
lingual papillae in H. ebraccata, are 
absent. The tiny secretory pits on the 
dorsal edge of the ventral velum are just 
as dense in H. ebraccata, but thin out 
anteriorly and may not cover the whole 
free velar surface. 


Ventral pharynx.—H. sarayacuensis 
exhibits some slight further reduction in 
the pharyngeal region compared with 
H. ebraccata. Counts for filter rows 
om the filter’ plates) run: cbaeiae— iG: 
Coy == 6a, Coy 2 = & co. 4b = &. 
The values for the first three cerato- 
branchials can be considered low; 
those for cb. 2 significantly so. Again, 
the number of filter rows on cb. 4 is 
significantly above the average for more 
typical larvae of the genus. Only with 
the greatest difficulty and repetitive 
staining could secretory ridges be re- 
solved under the ventral velum. Bran- 
chial food traps are not well demarcated. 
The greatest concentration of secretory 
ridges seems to be lateral and anterior 
and even here the topography is quite 
faint. 


Dorsal buccal.—As in H. ebraccata, 
in the prenarial arena a faint, anteriorly 
convex bulge is seen, but pustulations 
are absent. Small prenarial papillae are 
present, three on one side and one one 
on the other. These projections arise 
from the medial half of the anterior 
narial walls. Other pustulations or ru- 
gosities are absent from the narial walls. 
The two lateral roof papillae, seen in 
H. ebraccata, are present in H. saraya- 
cuensis. They are slightly behind the 
median papilla and half the distance 
laterally. In this specimen the anterior 
margins of the secretory zone could be 
resolved. The zone is of uniform length 
and has a smooth semicircular anterior 
margin. The ratio of the length of the 
buccal roof to the length of the glandu- 
lar zone was 4:1, which means that the 
zone is unusually long. The dorsal velum 
is similar to H. ebraccata but has slight- 
ly shorter median and lateral lengths. 

Dorsal pharynx.—Details of the dor- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 71 


Fic. 27.—Drawings of the floor (above) and roof (below) of the mouth of a Hyla sarayacuensis 
larva. 


72 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


sal pharynx region could not be deter- 
mined because of damage during dissec- 
tion. 

Diagnostic summary.—Hyla_ ebrac- 
cata and Hyla sarayacuensis as repre- 
sentatives of the Hyla leucophyllata spe- 
cies group are characterized by the 
following oral features: medially elon- 
gate prepocket buccal surface; reduction 
of the oral papillation and projections 
that define arenas; reduced marginal 
projections of ventral velum; proportion- 
ally large third filter cavity with com- 
paratively high number of filter rows in 
ch. 4: less dense filter mesh than typical 
(2/3 denticle patterns) hylid tadpoles; 
reduction of area of secretory ridges in 
the posteromedial portion of the bran- 
chial food traps; secretory pits covering 
extensive areas of dorsal surface of ven- 
tral velum and posterior buccal roof; 
large esophageal funnel. Both Hyla 
ebraccata and Hyla sarayacuensis tad- 
poles are definable by the combination 
of characters given above; the differ- 
ences between these two species are 
minor and given in the descriptions. 


Ptychohyla schmidtorum chamulae 
Duellman 


(Figs. 28, 29) 


Material—KU 75409 (stage 35, sv. 
12.7 mm). Collected attached to stones 
in pools in montane stream at 1600 me- 
ters elevation, 17.6 km NW of Pueblo 
Nuevo Solistahuacan, Chiapas, Mexico; 
June 14, 1963. 

Reference.—Duellman, 1970 (p. 529- 
Ba2)e 

External—aA slightly elongate tad- 
pole, with a 3/3 denticle pattern. The 
circumference of the oral disc is ex- 
panded into a denticle-free funnel of 
modest size. The spiracle is sinistral. 

Ventral buccal—The buccal floor is 
comparatively short and broad. The 
most anterior features in the mouth are 
eight medium sized, attenuate papillae, 
four per side. These line the base of 
the keratinized beak and arch forward, 
pointing out of the oral orifice. A bit 


more posterolateral on each side are 
single, large fleshy infralabial papillae. 
These are slightly compressed and bear 
faint, terminal pustulations and pointed 
apices. Dorsolateral to these papillae, 
over Meckel’s cartilage, are soft palps 
or cushion-like bulges of the buccal 
lining. These are oval and obliquely 
oriented. The lingual papillae are ex- 
ceptionally long and thin, and arch for- 
ward. The tongue anlage is narrow and 
extending obliquely back from its base 
are two wide but low elevations of the 
buccal surface. These curve slightly to- 
ward the medial margins of the buccal 
pockets before fading into the buccal 
floor. The BFA papillae are arranged 
in a “U” which is open anteriorly. The 
top of the “U” flares laterally such that 
the BFA papillae are in a continuous - 
row with the prepocket papillae. These 
papillae are all attenuate, of small to 
medium size, with curved pointed tips. 
I counted 18 on one side and 20 on the 
other. Two or three of these on each 
side can be considered prepocket pa- 
pillae. Only three BFA papillae were 
bifurcated. The center of the arena is 
devoid of pustulations. The only pustu- 
lations present on the buccal floor are 
at the posterior margin of the BFA and 
anterolateral above the body of the 
ceratohyals. Two or three medially di- 
rected papillae are among these antero- 
lateral patches of pustulations. The buc- 
cal pockets are wide in P. schmidtorum; 
they are also clearly perforated in this 
species. P. schmidtorum has a relatively 
large free velar surface supported by 4 
pairs of long spicules, the more medial 
of which fully reach the velar margin. 
The posterior edge of the velum has 
moderately acute peaks associated with 
each filter cavity. The peaks of the third 
filter cavity are displaced medially and 
are the largest. In addition to these un- 
dulations the velar margin is slightly 
arched around the tips of the spicules. 
The middle portion of the velum is 
strongly emarginated, with four or five 
papilla-like projections (preserved fold- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 73 


Fic. 28.—Drawings of the floor (above) and roof (below) of the mouth of a Ptychohyla schmid- 
torum larva. 


74 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


ed up and forward in this specimen). 
Individual secretory pits could not be 
resolved on this surface, although the 
apices of the trailing velar projections 
tended to have a buffed texture. 
Ventral pharynx.—The branchial bas- 
kets of P. schmidtorum are of average 
size for a hylid, but have the appearance 
of being displaced laterally along their 
posteromedial border. The baskets are 
relatively shallow, yet the filter plates 
are imbricated to an extreme and, con- 
sequently, are still quite tall. Ratios for 
maximum height to maximum length 
show the filter plates of P. schmidtorum 
to be as tall or taller than typical hylid 
tadpoles. Numbers of filter rows on each 
plate fall well within the normal hylid 
range; the filter mesh, however, is not 
dense. The filter rows are low, narrow 
creases with short secondary branches. 
Higher degree folds are rare and, when 
present, short. They occur predomi- 
nately along the ventral margin, where 
the filter rows are their widest. The 
filter canals are wide, open channels as 
wide or wider than the filter rows. Se- 
cretory ridges are visible through the 
dorsal surface of the velum; their size 
and proportions are typical for a hylid 
tadpole. The branchial food traps are 
not unusual. The glottis would not be 
visible in dorsal view, if the mid-portion 
of the velum had been preserved in the 
normal position. The glottis has a small 
and rather inconspicuous laryngeal disc. 
The esophageal funnel is not large but 
has a broad profile in dorsal view. 
Dorsal buccal.—Anteriorly the roof 
of the mouth is slightly expanded, other- 
wise gross dorsal proportions are typical 
for a hylid. There is a central trough 
that runs the length of the prenarial 
arena. It is surrounded posteriorly by 
a narrow V-shaped ridge (open ante- 
riorly) that increases progressively in 
height toward the internal nares. The 
ridge terminates posteriorly in a single 
projection. Other papillae and pustu- 
lations in the prenarial arena are absent. 
The internarial distance is great. The 


nares have a nearly transverse orienta- 
tion and are laterally elongate. Their 
anterior walls are thin and low; they 
lack any pustulations or prenarial pa- 
pillae. Their posterior walls are long and 
low and narial valve projections are 
faint or not present. P. schmidtorum 
lacks a well defined postnarial arena. In 
the position of the median ridge is a 
single, stiff, tall papilla. Lateral to it 
are obliquely oriented (anteromedial to 
posterodorsal) ridges on each side. 
These ridges have thick, wide bases and 
curved, bumpy ventral margins. They 
are tallest and _ thickest anteriorly, 
abruptly descending from the buccal 
roof at a position where one would ex- 
pect to find the most anterior postnarial 
papillae in other species. They continue 
back an equal distance behind the me- - 
dian (ridge) papilla and appear to be 
homologues of the postnarial papillae. 
Continuing posteriorly from the end of 
these oblique ridges are BRA papillae 
in two lines that converge slightly to- 
ward the midline; seven and eight to a 
side. The first and last papillae in the 
rows are diminutive; others are moder- 
ately large, and the five largest (three 
on one side, two on the other) have 
terminal bifurcations. The most poste- 
rior BRA papillae are displaced slightly 
toward the middle, reinforcing the im- 
age of a circumscribed “arena.” Within 
the middle of the BRA are approximately 
50 pustulations, the ten largest concen- 
trated anteriorly near the base of the 
median (ridge) papilla. Directly lateral 
to the end of the oblique ridges are two 
moderately large, laterally compressed 
papillae, one per side. Both of these 
have terminal bifurcations. They also 
are possible homologues of the lateral 
ridge papillae. Directly behind these, 
lateral to the BRA and in a line with 
the middle of the arena, are tight clus- 
ters of small pustulations, including one 
or two simple, small papillae. Still fur- 
ther posterior, lateral to the end of the 
BRA, are two small, tranversely oriented 
knolls. The glandular zone is very short 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 75 


Fic. 29.—Photographs of the floor (above) and roof (below) of the mouth of a Ptychohyla schmid- 
torum larva. 


76 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


and of uneven thickness. It is nearly 
absent on the midline but wider later- 
ally. Secretory pits are similar in size 
and dens:ty to many other hylids exam- 
ined but are not especially conspicuous, 
presumably because of differences in 
preservation. The dorsal velum is slight- 
ly shortened; it is completely absent on 
the midline. The free velar margin is 
moderately crenulate, particularly me- 
dially. 

Dorsal pharynx.—Pressure cushions 
are absent in P. schmidtorum. The cil- 
iary groove is very broad and shallow. 

Diagnostic summary.—Tadpoles of 
the genus Ptychohyla can be distin- 
guished from larvae of the other hylids 
examined by a combination of the fol- 
lowing characters: attenuate lingual pa- 
pillae; buccal floor and roof arenas elon- 
gate; median projections of the posterior 
velar margin elongate; third filter cavity 
truncated; filter mesh of low density; 
median ridge attenuate. 

Ptychohyla schmidtorum was_ the 
only funnel-mouthed hylid larva I ex- 
amined. In its oral cavity, the larvae 
exhibits the assorted modification asso- 
ciated with the funnel-mouthed condi- 
tion; principally, the fusion of papillae 
series into ridges and the posteriorly di- 
rected V-shaped depression of the pre- 
narial arena. 


Ptychohyla leonhardschultzei (Ahl) 
(Figs. 30, 31) 


Material—KU 139924 (stage 26, sv. 
14.4 mm). Collected on bottom of riffles 
in streams at 720 m elevation, 18.1 km N 
of San Pedro Mixtepec, Oaxaca, Mexico; 
March 20, 1970. 

Reference. — Duellman, 1970 (p. 
543). 

External.--The body is slightly elon- 
gate. The denticle formula is 4/6. The 
oral disc is relatively large, but does not 
have the expanded, denticle-free margin 
seen in P. schmidtorum. 

Ventral buccal.—The floor of the 
mouth in P. leonhardschultzei is rela- 
tively short; it has a squarish shape. In 


a transverse row at the base of the kera- 
tinized beaks are two pairs of small, 
jagged papillae. The smaller of the pairs 
is more lateral. Several subpapillate pus- 
tulations are still more lateral. Behind 
these pustulations on each side is a 
major infralabial papilla, shaped like a 
cupped hand with long, pointed “fingers” 
projecting forward. The fingers, four 
and five per side, have scattered, pointed 
pustulations on their margins. The two 
lingual papillae are very attenuate, as 
are most other oral papillae. The BRA 
is an elongate oval. The arena is sur- 
rounded by 86 moderately tall, thin, 
pointed papillae in this specimen. These 
papillae are strongly curved so as to 
point medially. Clusters of papillae near 
the lateral ends of the buccal pockets 
are fused at their base to form multi- - 
papillate structures which look like mini- 
ature deer antlers, one per side. These 
have six “points” on one side and eight 
on the other. Only two other papillae 
on the buccal floor show any sign of 
bifurcation. There are three and four 
small, conical prepocket papillae on op- 
posite sides. Between the anterior end 
of the BFA papilla series and the pre- 
pocket papillae are twelve to twenty pa- 
pillae of the same shape and size as 
those in the BFA series. These fully 
cover the surfaces above the body of the 
ceratohyals. Papillae from the BFA se- 
ries also extend laterally behind the buc- 
cal pockets to cover a small region of 
the buccal floor anterior to the second 
filter cavity. There are some 60 or so 
very tiny pustulations dispersed within 
the buccal arena. The buccal pockets 
are long, shallow, and perhaps a bit 
more transversely oriented than in typi- 
cal hylid larvae. They are conspicuously 
perforated. The free velar surface is 
large. It is supported by stiff spicules; 
the more medial pairs reach the velar 
margin. Single, symmetrical, leptokurtic 
projections of the velar margin are 
aligned over filter cavities 1 and 2; the 
more medial of these projections point 
posteromedially. Eight other distinct 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 77 


o ari é, apake ‘mn, 
gg al eee Akan aT > Mims 
aS une! n hy 
re SEE ee ae 


* (ion IM 


Fic. 30.—Drawings of the floor (above) and roof (below) of the mouth of a Ptychohyla leonhard- 
schultzei larva. 


78 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


projections are clumped together on the 
velar margin above the glottis. Secretory 
pits are mostly limited to the surface 
of the posterior projections. Pits are 
nearly absent from the velar surface it- 
self except for a very narrow band near 
the midline. 

Ventral pharynx.—The branchial bas- 
kets are of typical width to length ratio 
but have the appearance of being dis- 
placed laterally along their posterome- 
dial border. The third filter cavities are 
reduced in size and partially obscured 
from dorsal view by the medially imbri- 
cating, curved margins of the third filter 
plates. Filter plates of cb. 1 and cb. 2 
are relatively short, while those of cb. 3 
are comparatively tall. Counts for num- 
ber of filter rows fell within the normal 
range for typical hylid larvae, but those 
of cb. 1 and cb. 4 were slightly low, 
while those of cb. 2 and cb. 3 were 
slightly high. The filter mesh of P. leon- 
hardschultzei is quite open, but not as 
spacious as in Ptychohyla schmidtorum. 
P. leonhardschultzei has a_ typical 
branching pattern for the filter folds, but 
all the folds are exceptionally slender, 
resulting in large filter niches. Branchial 
food traps and their secretory ridges are 
not unusual. The glottis is 80% under 
the velum. It is small and has thin, 
elevated lips. The laryngeal disc is not 
visible. The esophageal funnel is of com- 
parable size to that of the more typical 
hylid larvae examined. 


Dorsal buccal.—The buccal roof ap- 
pears wide anteriorly. The central por- 
tion of the prenarial arena is displaced 
ventrally along the edge of an anteriorly 
convex arch. This arch is slightly wider 
than the internarial distance. The edge 
of the arch is lined with a dozen or so 
large, blunt pustulations or subpapillar 
projections and has its greatest depth 
anteriorly. The internal nares of P. leon- 
hardschultzei are elongate. Prenarial pa- 
pillae are lacking, but about a_ half 
dozen pointed pustulations are dispersed 
over the medial half of the anterior 
narial wall; the posterior narial walls 


lack narial valve projections. Postnarial 
papillae are located on narrow, obliquely 
oriented (anteromedial to posterolat- 
eral) ridges of moderate size between 
the anterior end of the internal nares 
and the median ridge. The tallest post- 
narial papillae are in the middle of their 
supporting ridges, where the ridges 
themselves are tallest. I counted seven 
postnarial papillae per side. The smaller 
papillae grade into pustulations at the 
end of the ridges; the larger are of mod- 
est proportion with jagged anterior mar- 
gins. The median ridge is a small isos- 
celes triangle, with a basal width equal 
to its height, and with slightly concave 
sides. It has four pustulations scattered 
over its anterior surface. There are two 
or three tiny, pointed papillae in front 
of the median ridge in the postnarial . 
arena. Laterally, at the posterior end 
of the ridges supporting the postnarial 
papillae, are two relatively giant, multi- 
papillate flaps. These flaps, presumed 
homologues of the lateral ridge papillae, 
are compressed into the sagittal plane. 
They have five, attenuate pointed papil- 
lae along their lateral margins. Two or 
three of these papillae have minor bi- 
furcations. Extending directly back from 
the base of the flaps on each side is a 
row of tall, thin, pointed papillae that 
define the lateral limits of the BRA. 
These rows converge only slightly until 
they are far back on the buccal roof, 
where they turn sharply toward the mid- 
line. The BRA is thus roughly a rec- 
tangle, with rounded corners. It is bound 
by a total of 34 BRA papillae in this 
specimen. Three or four of these pa- 
pillae have a tiny, pointed pustulation 
near their apices but in general, buccal 
roof papillae are little ornamented. A 
hundred or so small blunt pustulations 
are dispersed within the arena. Ten 
to twelve tiny papillae, not quite as tall 
as those in the BRA series, form a sec- 
ondary line or cluster of papillae lateral 
to the middle of the arena. Lateral to 
the posterior end of the arena are small 
pustulate fields also including a couple 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 79 


ING 


Fic. 31.—Photographs of the floor (above) and roof (below) of the mouth of a Ptychohyla leon- 
hardschultzei larva. 


80 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


of tiny papillae on each side. The glan- 
dular zone is not wide. Secretory pits 
are small in P. leonhardschultzei, al- 
though they are not atypically dense. 
Overall, the dorsal velum is of typical 
size and shape, but it is interrupted me- 
dially and its free medial edges are 
sculptured into a series of small, simple 
papillae on each side. 

Dorsal pharynx.—The pressure cush- 
ions are equal to, or larger than, those 
of more typical hylid larvae. The me- 
dial pressure cushions are the larger and 
rounder of the pairs. The ciliary groove 
is deep and narrow. 

Diagnostic summary.—Tadpoles of 
this species have the greatest number of 
buccal floor and buccal roof arena pa- 
pillae of any species examined and, as 
such, can be distinguished readily from 
all other hylid larvae examined. 


Acris crepitans Baird 
(Fig. 32) 


Material—FMNH 11916 (stage 36, 
sv. 16.8 mm). Collected 6.4 km W of 
Fort Worth, Tarrant Co., Texas; March 
29, 1931. 

Reference.—Stebbins, 1951 (p. 301- 
304); Heyer, 1976. 

External.—Tadpoles of the genus 
Acris have a 2/3 denticle pattern and a 
sinistral spiracle. Their body form and 
habits typify the “common” pond tad- 
pole. 

Ventral buccal.—Acris crepitans lar- 
vae are indistinguishable from H. regilla 
larvae in the shape of their mouths and 
general oral proportions. The few dif- 
ferences between the two species large- 
ly correlate with the greater size of A. 
crepitans. The position and shape of 
papillae on the buccal floor, including 
the infrarostral, lingual, BFA and the 
buccal pocket papillae, are identical to 
H. regilla. The same situation holds for 
pustulations on the buccal floor. The 
BFA papillae may be slightly taller and 
more numerous than those of H. regilla, 
but not significantly so. (In the speci- 
men on which this description is based, 


there are three lingual papillae. Exami- 
nation of other A. crepitans, however, 
shows that two lingual papillae are typi- 
cal and that this individual is abnormal 
in that feature). The buccal pockets of 
A. crepitans are clearly perforated. The 
secretory pits on the dorsal margin of 
the velum are possibly a bit larger than 
those of H. regilla, but the difference is 
slight and may be accounted for by pres- 
ervational differences. 


Ventral pharynx.—The filter rows fall 
within the normal range for H. regilla 
of comparable stage; filter rows are a bit 
wider in A. crepitans than in H. regilla. 
Overall, the filter mesh appears slightly 
less dense in A. crepitans. Again this 
difference is small enough to be ascrib- 
able to differences in preservation. 


Dorsal buccal.—All major features © 
are as in H. regilla; the few distinctions 
between A. crepitans and H. regilla are 
almost all comprehendable as simple 
augmentations associated with overall 
increase in size. There are twice as many 
prenarial papillae in A. crepitans as in 
the average H. regilla; the narial valve 
projections are more attenuate. I count- 
ed seven postnarial papillae in oblong 
(anterolateral to posteromedial) clusters 
on each side of the midline in this speci- 
men; this is significantly more than in 
H. regilla. At least one of these was bi- 
furcated. A couple of small papillae 
and a half dozen or so pustulations were 
within the postnarial arena. The median 
ridge and lateral ridge papillae are simi- 
lar to those of H. regilla. Additional 
papillae, anterolateral to the lateral 
ridge papillae, are present. There are 
twice as many BRA papillae in A. crepi- 
tans as there are in H. regilla. Ten or 
more papillae with some associated pus- 
tulations occur lateral to the BRA on 
each side. Evenly dispersed within the 
arena are well over a hundred medium 
to small pustulations. The glandular 
zone is identical to that of H. regilla, 
although it is a bit easier to resolve 
secretory pits in the one specimen on 


hand. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 81 


Fic. 32.—Photographs of the floor (above) and roof (below) of the mouth of an Acris crepitans 
larva. 


82 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Dorsal pharynx.—Pressure cushions 
can be discerned as two round faint 
swellings on the posterior surfaces of 
the dorsal velum on each side. These 
cushions and the ciliary groove are in- 
distinguishable from those of H. regilla. 

Diagnostic summary.—Tadpoles of 
the genus Acris are so similar externally, 
that it is reasonable to consider Acris 
crepitans as representative of the genus. 

Acris crepitans tadpoles have a high- 
er number of postnarial papillae than 
any other larvae examined. There are 
significantly more buccal roof arena pa- 
pillae in Acris crepitans tadpoles than in 
Hyla regilla tadpoles. Otherwise, the 
larvae of these two species are virtually 
identical. 


Smilisca sordida (Peters ) 
(Fig. 33) 


Material. — Uncatalogued, author's 
collection (stage 37, sv. 11.5 mm). Col- 
lected singly in clear shallow, but slow 
flowing water of the Rio Rincon near 
ten mile mark on the Pacifica Road from 
Rincon de Osa, Puntarenas, Costa Rica; 
March 4, 1970. 

Reference. — Duellman, 1970 (p. 
617). 

External.—These tadpoles have a 2/3 
denticle pattern and a sinistral spiracle; 
only a slight ventral shift of the mouth 
and elongation of the tail distinguishes 
these tadpoles from typical pond polli- 
wogs. Externally they show few of the 
modifications characteristic of stream 
tadpoles. 


Ventral buccal.—Anteriorly the floor 
of the mouth is slightly expanded lat- 
erally. The infralabial papillae are 
cupped flaps of skin no different in size 
and shape than those of typical pond 
hylid larvae, but with finely serrated 
margins rather than papillate “fingers.” 
Lingual papillae are tall. The buccal 
floor arena is defined by a modest num- 
ber of papillae, some eight per side. The 
BFA papillae do not converge strongly 
toward the midline at the anterior limit 
of the BFA. The three or four largest 


BFA papillae, those directly medial to 
the ends of the buccal pockets, are fused 
to their neighbors at their bases. There 
are single, prepocket papillae on each 
side, which are relatively tiny and sharp- 
ly pointed. A dozen or so large pustu- 
lations are in the posterior half of the 
BFA. Fine pustulations are also scat- 
tered above the lateral arms of the 
ceratohyals, among the BFA papillae, 
and posteriorly behind the arena. This 
last posterior patch extends laterally to 
the front of the second filter cavity. 
Single, tiny papillae are among the pus- 
tulations on the ceratohyal arms in front 
of the second filter cavity. The buccal 
pockets of S. sordida are large, long and 
more transversely oriented than those of 
H. regilla. They are conspicuously per- 
forated. The free velar surface is large 
and supported by stiff, long spicules that 
come close to reaching the velar margin. 
Posterior projections from the velar sur- 
face are long and leptokurtic. The pro- 
jections associated with the third filter 
cavities are displaced medially. The pro- 
jections of both the second and third 
filter cavities are pointed posteromedi- 
ally. One or two tall secondary projec- 
tions are developed next to the relatively 
deep median notch. Large, conspicuous 
secretory pits form a thin band on the 
posteromedial portion of the velar mar- 
gin. This band is diminished laterally; 
distal to the second filter cavity the pits 
are restricted to the marginal projections 
of the velum. 

Ventral pharynx.—The branchial bas- 
kets viewed from above are typical in 
size and shape. They are, however, not 
particularly deep and individual filter 
plates are imbricated to a great extent. 
The third filter cavity is reduced in 
width. In terms of other characters of 
the filters, viz. the length to height ratio 
of the filter plates, number of filter plates 
on the rows, pattern of the filter folding, 
density of filter mesh, etc., S. sordida is 
indistinguishable from more typical hy- 
lid larvae. Branchial food traps are 
necessarily shallow because of the shal- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 83 


g o wi zt 
ae 8 i pa % 


ZHU op 


Fic. 33.—Drawings of the floor (above) and roof (below) of the mouth of a Smilisca sordida larva. 


84 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


lowness of the branchial baskets, but, 
because of the larger free velar surface, 
the actual area covered by secretory 
ridges is not reduced. The secretory 
ridges are similar to those of typical 
pond larvae. The glottis is typical. The 
laryngeal disc is not visible. The esoph- 
ageal funnel is a bit broader but no 
larger in S. sordida than in the reference 
hylid larva. 

Dorsal buccal.—The buccal roof has 
the same length to width ratio as H. 
regilla, but a slightly more square, less 
triangular shape. The central portion of 
the prenarial arena bulges ventrally. On 
the surface of this shallow bulge is an 
anteriorly directed pustulate arch. The 
arch is twice as wide as the internarial 
distance. The internal nares of S. sor- 
dida are more elongate and obliquely 
oriented than in a typical pond larva. 
Prenarial papillae are lacking, but many 
tiny, pointed pustulations line the whole 
length of the low, anterior narial walls. 
The posterior narial walls lack narial 
valve projections. Postnarial papillae, 
two on one side, four on the other are 
small, simple, blunt projections. These 
are in obliquely oriented (anteromedia] 
to posterolateral) lines that include an 
equal number of evenly spaced pustu- 
lations. The reduced median ridge is as 
tall as it is wide and has a peaked rather 
than horizontal ventral margin. This 
ventral margin is lined with many fine 
serrations, and there are single isolated 
pustulations on the anterior surface of 
the median ridge. Extending back from 
the ends of the postnarial papillae row, 
posterolateral to the median ridge, are 
large, nearly longitudinally oriented 
flaps. These are presumed homologues 
of the lateral ridge papillae. Each flap 
is tallest posteriorly and deeply notched, 
so that it has at least five sharp peaks 
along its margin plus a couple of minor, 
pointed pustulations. Continuing pos- 
teriorly in line with the long axes of the 
lateral flaps are single rows of BRA pa- 
pillae on each side. These are slender, 
pointed papillae of modest size, num- 


bering five per side. There is a gap 
equal to a seventh of the length of the 
buccal roof separating the first BRA 
papillae from the posterior edge of the 
lateral flaps. The BRA papillae rows 
converge only slightly towards the mid- 
line and outline a rather elongate buccal 
roof arena. There are sixty or more pus- 
tulations within the arena; their greatest 
concentration is posterior. Directly lat- 
eral to the gap between the lateral flaps 
and the most anterior BRA papillae are 
small clusters of four or five pustulations 
and one tiny papilla per side. The glan- 
dular zone is slightly shortened both 
medially and laterally compared with 
more typical pond larvae. The secretory 
pits are not unusual in any way. The 
dorsal velum is short and has a broad 
gap across the midline. The more me- 
dial limits of the velum have a pustulate 
margin. 

Dorsal pharynx.—Pressure cushions 
are of comparable length to those of H. 
regilla but are necessarily smaller be- 
cause of the decrease in free velar sur- 
face. The ciliary groove is narrow. 

Diagnostic summary.—Because only 
one species of Smilisca was examined in- 
ternally, it would be premature to at- 
tempt to diagnose the genus. The Smi- 
lisca sordida group of Smilisca includes 
stream breeders; it is a reasonable hy- 
pothesis, however, that the oral cavity 
of S. sordida larvae will resemble that 
of other stream Smilisca tadpoles more 
than it will resemble that of the larvae 
from the non-stream breeding Smilisca 
baudinii group. On the other hand, it 
is possible that with such slight differ- 
ences between typical pond larvae, such 
as Hyla regilla, and Smilisca sordida, 
other species of Smilisca may not be 
distinguishable from S. sordida in larval 
oral anatomy. The following are a few 
of the oral features in which Smilisca 
sordida larvae differ from Hyla regilla 
tadpoles: lingual papillae taller; spic- 
ules stiffer and longer; velar marginal 
projections longer; filter plates more im- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 85 


bricated; branchial food traps shallower; 
pressure cushions smaller. 


Agalychnis callidryas (Cope) 
(Fig. 34) 


Material. — Author’s collection (3 
specimens: stage 35, sv. 19.9 mm; stage 
37, sv. 18.6 mm; stage 39, sv. 19.2 mm). 
Collected in massive swarming schools 
in the “Caiman” pond, by Pacifica Road, 
9.7 km south of Rincon de Osa, Punta- 
renas, Costa Rica; March 4, 1970. 

Reference.—Duellman, 1970 (p. 107- 
108). 

External—tTadpoles of the genus 
Agalychnis all have an elongate, fila- 
mentous tail. Their eyes are far lateral, 
their mouths are directed anteriorly or 
anteroventrally; the spiracle is just left 
of the midline. The denticle pattern is, 
however, the common 2/3 arrangement. 


Ventral buccal——The floor of the 
mouth is anteriorly broader and _ poste- 
riorly more elongate than in typical hy- 
lid larvae. A single, tiny, soft spur is 
present on the infrarostral cartilage 
posterior to the edge of the beak but in 
line with the keratinized cutting edge 
on each side. The infralabial papillae 
all lie well medial to these spurs. Three 
to five infralabial papillae per side are 
present. The largest papilla on each side 
is always the most medial. These more 
medial papillae are tall, tapered, and 
round to slightly oval in cross section. 
They curve directly anterad. The small- 
er, more lateral papillae are round in 
section and do not curve anterad. These 
more lateral infrarostral papillae form 
an anterior to posterior line or cluster, 
with the most anterior papilla slightly 
larger than the more posterior ones. 
The largest medial papillae are 1.5 to 
2 times taller than any of the lateral 
papillae. All papillae are unbifurcated. 
The two lingual papillae are so close to 
each other and to the midline that they 
effectively share a common base. The 
buccal floor arena is outlined by two 
posteriorly converging lines of papillae. 
Anteriorly these rows flare out laterally 


into a dense field of numerous tiny pus- 
tulations and subpapillae above the body 
of the large ceratohyal on each side. 
Six to nine (X = 7.5) BFA papillae 
make up the row on each side; none are 
bifurcated. The posterior four to five 
papillae are equal or subequal in size 
and shape. These are in a perfectly 
straight row. The papillae at the begin- 
ning of the series are similar, but smaller 
and not in such straight lines. Anteriorly 
two small papillae lie inside the arena. 
These are next to and parallel the main 
papillae row. Sixty or more tiny pustu- 
lations cover the buccal floor. They are 
most concentrated posteromedially in 
the BFA and laterally above the cerato- 
hyals. There are no isolated prepocket 
papillae per se but a few small papillae, 
not separable from the pustulate lateral 
field at the anterior end of the BFA 
papillae rows, may be homologous to 
the prepocket papillae in other species. 
The buccal pockets may be perforated 
in the stage 37 specimen, but are 
not evidently open in either the stage 
35 or 39 specimens; a consistent spe- 
cies pattern cannot be resolved with 
the small sample at hand. The anterior 
to posterior length of the free velar 
margin over the filter cavities is larger, 
and the supporting spicules longer, than 
in most other hylids. The margin of the 
ventral velum has four posteriorly di- 
rected acute peaks. The peaks, which 
are most obviously homologous with the 
velar crenulations normally above the 
third filter cavity in other hylids, are 
displaced medially. Several paired auxil- 
lary projections in this region effectively 
produce a field of papillae trialing off 
the edge of the velum above the glottis. 
Single, broad crenulations are present on 
the velar margin between the peaks 
above filter cavities 1 and 2; these are 
also posteriorly directed. The pitted 
glandular zone on the dorsal surface of 
the ventral velum has a sharply defined 
anterior margin. The glandular tissue is 
densest on or near the major peaks and 
tends to thicken the edge of the velum. 


MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


86 


Fic. 34.—Photographs of the floor (above) and roof (below) of the mouth of an Agalychnis calli- 


dryas larva. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 87 


Ventral pharynx.—Agalychnis is char- 
acterized by substantially larger bran- 
chial baskets, a greater amount of filter 
surface, and a finer filter mesh than any 
other hylid examined. Filter plates are 
longer and particularly taller than in 
H. regilla, although the length to width 
ratio for the branchial baskets in toto 
do not differ; the additional filter surface 
in Agalychnis is taken up in increased 
curvature and imbrication of the filter 
plates. The major axis of the second 
filter cavity projects back from the mid- 
line at a very acute angle. The cerato- 
branchials and the filter plates that they 
support are strongly bowed outward. 
Filter plate 1 is so curved that in cross 
section it circumscribes half a circle. 
The dorsal margins of the filter plates 
on cb. 2 and cb. 3 are not straight, but 
arch upward to a peak just behind the 
edge of the ventral velum. These plates 
are strongly imbricated toward the mid- 
line causing the filter cavities, particu- 
larly cavity 3, to be almost completely 
encircled by filter epithelium. The num- 
ber of full filter rows on each plate are 
exceptionally high; they range as fol- 
lowseco les loch: 2) = 19-15. cb: 
3 = 12-13, cb. 4 = 8-9. The filter mesh 
is far tighter than in other hylids. Sec- 
ondary filter folds are numerous and 
dense. Filter rows do not simply termi- 
nate at the ventral borders of the filter 
plates, but interdigitate with the rows on 
the opposite filter plates, completely ob- 
scuring the gill slits. Partial, tapered 
filter rows project deeply ventrad_ be- 
tween normal full rows and fill any 
spaces between the full filter rows. Fil- 
ter canals are virtually complete tubes, 
rarely visible without displacement of 
the filter rows. Narrow and _ closely 
packed secretory ridges are readily visi- 
ble on the ventral surface of the ventral 
velum. The ridges appear discontinuous 
directly under the spicule of cb. 2. 
Branchial food traps cover a large pro- 
portion of the anterior surface in the 
filter cavities extending down the medial 
anterior wall of each cavity. The glottis 


appears fully developed, although not 
large. The glottal lips are raised on a 
strongly developed anterior to posterior- 
ly directed ridge. The papillate mid- 
portion of the ventral velum margin 
makes it impossible to estimate, in any 
consistent way, the amount of glottis 
obscured by the velum. The esophageal 
funnel is of relatively narrow profile. 


Dorsal buccal.—The roof of the 
mouth is quite wide anteriorly. In the 
center of the large prenarial arena is a 
single, darkly staining, semicircular 
(concaved posteriorly ) low ridge. Even- 
ly spaced on the top of the ridge are 
short, knobby projections, 4 to 7 per 
side. Posterolaterally these are mere 
pustulations, but anteromedially they de- 
velop into unusually shaped, low nodu- 
lar papillae. There is a faint medial 
gap in the ridge. Medially within the 
semicircle, are one or two more typical, 
prenarial pustulations. The internal 
nares are not unusual in shape, except 
that the openings to the olfactory cap- 
sule are small and far lateral. There are 
no distinct prenarial papillae, although 
the raised anterior wall of each naris is 
rugose in texture, particularly medially. 
The narial valve projection is shorter 
than in more typical hylid larvae. Be- 
tween the nares and the median ridge 
lie three distinct papillae on each side. 
These papillae all arise from a common 
straight ridge oriented anteromedially to 
posterolaterally. On each side the most 
anterior of the three papillae lies direct- 
ly behind the narial valve projection. 
This is a very large, conical, acutely 
pointed papilla. It is taller than the 
narial wall and projects anteriorly under 
the narial valve projection in one speci- 
men but a bit more medially in the 
other two. The middle papilla in the 
series is the smallest. It is a simple, tiny, 
finger-shaped papilla aligned behind the 
middle of the nares and about halfway 
back in the narial arena. This papilla 
is directed anteriorly. The last papilla 
on the ridge is a simple, stout, cone one- 
half to one-third the size of the largest 


88 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


anterior papilla. This last papilla aligns 
itself behind the lateral end of the nares 
and two-thirds of the way back in the 
narial arena. It is oriented medially. In 
two of the three specimens, postnarial 
papillae show some very fine pustula- 
tions along their most anterior margins. 
A single, basal bifurcation was found 
asymmetrically on the third papilla in 
one of the specimens; elaborate bifurca- 
tions of the posterior narial papillae are 
probably rare in this species. The last 
papilla in the posterior narial papillae 
row may be homologous to the lateral 
ridge papillae; however, they are dis- 
placed a bit forward and because they 
sit on a common ridge with the other 
papillae within the narial arena they are 
considered here part of the postnarial 
papillae series. There are no other pa- 
pillae directly lateral to the median 
ridge. The median ridge is a transverse 
flap, shaped like an isosceles triangle. 
It is positioned slightly anterior to the 
middle of the buccal roof. The ridge has 
a basal width one to two times its height 
and a single or occasionally bifurcated 
apex. No secondary anterior fringe is 
present. Between seven and ten pustu- 
lations are scattered within the post- 
narial arena. The buccal roof arena 
covers the flat, medial one-third of the 
buccal roof. The arena is bound later- 
ally by two patches of simple, moder- 
ately small, finger-like papillae, seven to 
nine per side. The patches trend antero- 
laterally to posteromedially. Posteriorly, 
the patches tend to narrow into single 
rows. The papillae do not vary much in 
size, but the largest ones are generally 
in the middle portion of the patches. 
None of the papillae are bifurcated. 
Within the arena are approximately 
eighty randomly dispersed, darkly stain- 
ing spots, almost too small to be called 
pustulations. These are absent outside 
of the arena. A secondary cluster of 
lateral roof papillae appears on each 
side about halfway between the main 
BRA papillae patches and the lateral 
edge of the mouth. The one to six pa- 


pillae in these lateral “clusters” are, on 
the average, one-fourth the height of 
the main BRA papillae. The glandular 
zone is well defined and made up of 
conspicuous secretory pits, which are 
large, round and abutting. These extend 
well over the dorsal velum and onto the 
pressure cushions. They are densest on 
the edge of the velum. The anterior 
margin of the glandular zone is strongly 
scalloped. Two lateral swells on each 
side match perfectly the anterior limits 
of the free edge of the ventral velum 
over filter cavities 2 and 3 below. The 
dorsal velum is divided on the midline 
by a substantial gap. The medial mar- 
gins of the dorsal velum may point to- 
wards each other or the edges may turn 
posterior to form a funnel that parallels 
the esophageal funnel. Laterally, the © 
edge of the dorsal velum is thick, glan- 
dular, and not curled. The velum is 
relatively and absolutely smaller than 
in more typical hylid larvae. 

Dorsal pharynx.—The pressure cush- 
ions are comparatively huge, obliquely 
oriented ovals, extending well down 
into filter cavities 1 and 2. The me- 
dial cushions are about three times as 
wide as the lateral cushions and twice 
as tall. The ciliary groove is relatively 
narrow and deeply entrenched. 

Diagnostic swmmary.—Of the spe- 
cies examined in detail, Agalychnis calli- 
dryas differs from all other hylids (sensu 
lato) in having massive branchial bas- 
kets, dorsally arched filter plates, and an 
extremely dense filter mesh. Cursory 
examination of other species in the genus 
of Phyllomedusa suggest that the diag- 
noses will not distinguish A. callidryas 
from other middle American phyllome- 
dusine tadpoles. 


CENTROLENIDAE 
Centrolenella fleischmanni (Boettger) 
(Fig. 35) 
Material. — Uncatalogued, author's 


collection (R. W. McDiarmid Field No. 
6865 “clutch #3”) (stage 30, sv. 6.9 
mm). Raised from eggs. Monteverde, 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 89 


Fic. 35.—Drawings of the floor (above) and roof (below) of the mouth of a Centrolenella fleisch- 
manni Jarva. 


90 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Puntarenas Prov., Costa Rica; collected 
July 4, 1971; preserved Aug. 20, 1971. 

Reference. — Starrett, 1960; Villa, 
1971; Heyer, 1976. 

External_—The tadpole of Centrole- 
nella fleischmanni is extremely elongate, 
with a tail more than twice as long as 
its body. The denticle pattern is 2/3 
and a single, sinistral spiracle is located 
3/4th the way back on the body. The 
external nares are tiny pores; the eyes 
are small, deepset and directed dorsally. 


Ventral buccal.—Internally the floor 
of the mouth is of the overall triangular 
shape of most tadpoles; however the 
oral orifice itself is narrow and the buc- 
cal surface behind the buccal pockets is 
exceptionally elongate. There are two 
pairs of infralabial papillae; a medial 
pair adjacent to the midline and a lateral 
pair placed more typically over Meckel’s 
cartilage. The medial papillae are sim- 
ple, tall, but blunt structures that over- 
lap on the midline. The lateral pair are 
small, anteroposteriorly flattened flaps, 
which are slightly concave anteriorly; 
one has slight marginal pustulations, but 
they both lack secondary papillae. The 
base of the lingual papillae is a trans- 
versely oriented ridge and on its apex 
is a transverse row of four small lingual 
papillae. There is but one BFA papillae 
on one side, four on the other; these 
are displaced far laterally and barely 
define an arena. The papillae lack sec- 
ondary pustulations; they are flattened 
against the buccal floor and have slight- 
ly knobby crowns. Most of the buccal 
floor lies behind these papillae and com- 
pletely lacks topographic relief. Papillae 
and pustulations are also absent from 
within the BFA or in the region anterior 
to the buccal pockets. The prepocket 
region is small. The buccal pockets are 
shallow creases with a strongly oblique 
orientation. Buccal slits appear (?) pat- 
ent. The free velar surface is gently 
curved, lacking projections. The median 
notch is very weak and shallow. A thin 
zone of glandular tissue is evident over 
the margin. 


Ventral pharynx.—Seen from above 
the ventral pharynx is long laterally and 
narrows sharply toward the midline. The 
result is that the branchial baskets ap- 
pear relatively triangular in shape. The 
baskets are extremely shallow. The filter 
plates are short and only slightly imbri- 
cated. The tallest plate is the second. 
The fourth filter plate is strongly arched 
in the transverse plane. Filter rows are 
well organized along the dorsal margins 
of the filter plate particularly the second 
filter plate; ventrally the filters become 
less organized and have a puffy, spongy- 
like texture. It is possible to discern sec- 
ondary and occasionally tertiary filter 
folds on some of the filter rows but 
others appear relatively disorganized, 
lacking an obvious branching pattern. 
Despite the fact that the higher order - 
filter folds are not always visible, there 
are no large spaces between filter rows; 
filter canals appear to be relatively nar- 
row, shallow slits. Counts for filter rows 
run: ¢b. 1, 9: cb. 2) 15:chaowlleichees 
10. These should be considered mini- 
mum estimates as it was very difficult 
to collect these data on such a small 
specimen. The gill filters are unusual in 
that they show such a broad range of 
structural organization. The light grainy 
texture characteristic of secretory tissue 
is evident on the short surface of the 
branchial food traps; however, even at 
100, well organized secretory ridges 
could not be discerned in this region 
and may be lacking in the species. The 
glottis is distinct with heavy lips. An 
elevated laryngeal disc is absent at this 
stage. The esophageal funnel is of com- 
parable shape to that of most other anu- 
rans examined. 

Dorsal buccal.—The buccal roof has 
the basic triangular shape typical of 
other anurans. The prenarial arena is 
narrow, long and nearly vertically ori- 
ented, a reflection of the ventral position 
of the oral orifice. There is one stiff, 
enlarged triangular projection which ex- 
tends down from the posterior portion 
of the prenarial arena. It has a slightly 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 91 


compressed, transversely oriented base. 
There are no other papillae or pustula- 
tions in the prenarial arena. The inter- 
nal nares are large, longitudinally ori- 
ented, open pits. The long median walls 
of these vacuities virtually meet on the 
midline. The prenarial papillae are 
points that extend forward from the 
front of the tall narial walls. Narial 
valve projections are lacking. The narial 
walls on each side circumscribe a large 
shallow narial depression (Jacobson’s 
organs?) at the base of which is a small 
slit whose margins completely abut. Be- 
cause of the small size of the specimen, 
complete exploration was impossible, but 
these slits appeared to be patent internal 
narial openings. The postnarial arena is 
extremely small and contains a single 
sharp papilla located on the midline in 
the middle of the arena. It is hidden 
from ventral view by a simple, triangu- 
larly shaped median ridge that is ad- 
pressed against the buccal roof. Al- 
though the median ridge is not large, 
the narial depressions are so long that 
the apex of the median ridge projects 
between the posterior portions of the in- 
ternal nares. Immediately to the sides 
and slightly anterior to the median ridge 
lie two pointed papillae, presumed hom- 
ologues of the lateral ridge papillae. 
They fit in the space between the median 
ridge and the posterior wall of the in- 
ternal nares. A short distance lateral 
and slightly posterior to each lateral 
ridge papilla lies a single obliquely ori- 
ented, posteromedial to anterolateral 
compressed flap on each side. The hom- 
ologues of these projections are uncer- 
tain. Whereas they are too far posterior 
to be associated with the median ridge, 
they are too far anterior to be obvious 
buccal roof papillae. Each flap has a 
terminal, anterolateral, fingerlike projec- 
tion that points anteriorly, and a more 
posteriomedial cusp similarly directed 
along the same edge. All of the struc- 
tures so far described lie in the anterior 
half of the buccal roof. Papillae and 
pustulations of any sort are lacking from 


the remainder of the buccal roof. The 
boundaries and general proportions of 
the glandular zone could not be dis- 
cerned in this specimen and a distinct 
glandular zone may be lacking in this 
species. The dorsal velum is short, un- 
curled, anteriorly directed, and continu- 
ous across the midline. 

Dorsal pharynx.—On the posterior 
surface of the dorsal velum are two 
extremely slight swellings on each side, 
the only evidence of pressure cushions. 
The medial swellings are larger and 
more obvious. The ciliary groove is a 
shallow broad channel. 

Diagnostic summary. — Many fine 
features of papillae pattern and shape 
are unique in this form compared to the 
other larvae described. The combina- 
tion of the following major features, 
however, readily serve to distinguish the 
oral cavity of this species from all oth- 
ers: long buccal floor and roof behind 
buccal slits and median ridge respec- 
tively, with both areas lacking papillae 
and pustulations; shallow branchial bas- 
kets; very large, longitudinally oriented 
internal narial depressions; four lingual 
papillae. 


DENDROBATIDAE 


Colostethus subpunctatus (Cope) 
(Fig. 36) 

Material—MVZ 63199 (stage 34, sv. 
11.7 mm). Collected from small rain 
pool at 2650 m elevation Bogota, Cundi- 
namarca, Colombia; Oct. 15, 1950. 

External.—Colostethus subpunctatus 
larvae have flattened venters and dor- 
sally directed eyes. The spiracle is sinis- 
tral and denticle formula is 2/3. 

Reference.—Stebbins and Hendrick- 
son, 1959. 

Ventral buccal—The floor of the 
mouth is generally similar to that of 
many pond tadpoles. In addition to 
weak spurs, directed anteriorly at the 
very margins at the oral orifice, there 
are three major papillae associated with 
the infralabial region. The smallest in- 
fralabial papillae are two blunt projec- 


92 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Fic. 36.—Drawings of the floor (above) and roof (below) of the mouth of a Colostethus subpunc- 
tatus larva. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 93 


tions arising from the posterior floor of 
the infralabial region. Their bases abut 
on the midline. In the typical position 
for infralabial papillae in other species, 
arise transversely compress large bifur- 
cate palps with secondary, small, mar- 
ginal pustulations. The tips of these 
palps touch on the midline. The third 
pair of infralabial papillae arise from 
the anterolateral edge of the large pair 
just described. These are flaps com- 
pressed in the transverse plane. They are 
nearly as wide as the previously de- 
scribed pair, but not as tall. They have 
several marginal serrations, but lack sec- 
ondary papillae. The two lingual pa- 
pillae are tall. The buccal floor arena is 
well defined by nine papillae on one side 
and ten on the other. The arena is rela- 
tively narrow and the BFA papillar rows 
have a slightly oblique (anterolateral 
to posteromedial) orientation. BFA pa- 
pillae are similar to those in most pond 
tadpoles; however, they lack extensive 
marginal papillations and are noticeably 
taller and more attenuate. One or two 
small papillae lie anterior to the medial 
margin of the buccal pockets. There is 
a weak field of pustulations in the pos- 
terior two thirds of the BFA. Buccal 
pockets are more obliquely oriented than 
in H. regilla. Large mucosal folds from 
the posterior walls obscure the base of 
each buccal pocket. The pockets are of 
average depth and may be perforated on 
one side in this specimen. The free 
velar margin is of average length, but 
relatively thick and inflexible. Tiny, but 
dense, secretory pits can be resolved at 
50> lining the free edge of the velum. 
The free edge is strongly sculptured with 
posteromedially directed cusps above 
the second and third filter plates and 
additional posteriorly directed cusps 
above the fourth filter plates. There is 
a very deep, distinct median notch. 
Ventral pharynx.—The branchial bas- 
kets are of comparable size and shape 
to those of typical pond tadpoles. The 
filter plates, however, are strongly imbri- 
cated medially and the medial gap be- 


tween the left and right branchial bas- 
kets is very large. Filter rows count run 
8, 9, 9, 5—similar to that of typical pond 
tadpoles. The filter mesh is low in densi- 
ty. Neighboring filter rows do not abut, 
except at their ventral extremes, and par- 
tial filter rows are short; thus leaving the 
filter canals as large open channels. 
Many of the filter rows lack tertiary or 
higher order folds, particularly in the 
small third filter cavity. The branchial 
food traps are relatively shallow. At 
50 magnification a weak pattern of 
transversely oriented ridges can be seen 
on the ventral surface of the ventral 
velum. The glottis is completely exposed 
in the median notch. It is slightly ele- 
vated and has distinctive lips; however, 
it is quite small in all dimensions com- 
pared to the overall size of the buccal 
cavity in this tadpole. The esophageal 
funnel is very large. 

Dorsal buccal.—In general propor- 
tions the roof of the mouth is similar 
to that of pond tadpoles. There is large 
prenarial arena; from the middle of the 
arena descends a broad, anteriorly di- 
rected, pustulate arch. Halfway between 
this structure and the internal nares lies 
a single large pustulation on the mid- 
line. The internal nares have a relative- 
ly horizontal orientation with a heavy, 
thick anterior wall, that lacks prenarial 
papillae. The posterior narial valve has 
a weak narial valve projection. Aligned 
in a row in the postnarial arena, parallel 
to the internal nares, run 4 robust pa- 
pillae on each side. The most posterior 
of these lies lateral to the median ridge 
and may be the homologue to the lateral 
ridge papilla. These are transversely 
flattened bifid structures. All of the 
papillae associated with the postnarial 
arena including the median ridge have 
pustulate anterior margins. The smallest 
postnarial papillae is most anterior. The 
median ridge is a simple, triangular flap 
of average size. The buccal roof arena 
is defined by five, extremely attenuate 
papillae on one side and four on the 
other. A few subpapillate projections 


94 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


arise far lateral to the buccal roof. There 
are several dozen barely visible pustu- 
lations distributed within the BRA. The 
glandular zone has a very distinctive 
“Vv” shaped anterior margin with large 
conspicuous secretory pits. Posteriorly 
the secretory pits tend to become much 
smaller, denser, and less distinct, until 
one reaches the regions of the pressure 
cushions, where they again become large 
and conspicuous. The free velar edge is 
short, discontinuous on the midline. The 
medial limits of the velar margin have 
two or three small, simple papillae. 
Dorsal pharynx.—The dorsal pharynx 
was damaged in dissection, but enough 
of the pressure cushions were preserved 
to indicate that these are relatively shal- 
low bulges with the lateral cushion far 
more distinct than the medial. Details 
of the ciliary groove were not preserved. 
Diagnostic summary.—No single fea- 
ture distinguishes these larvae from oth- 
ers examined; however, the following 
characters, in combination easily diag- 
nose the oral cavity of C. suwbpunctatus 
larvae: 3 pairs of infralabial papillae, 
BFA and BRA papillae very tall, bran- 
chial baskets widely separated, filter 
plates strongly imbricated but filter mesh 
weak, glottis small, secretory pits large. 


Colostethus nubicola (Dunn) 
(Fig. 37) 
Material.—Author’s collection, uncat- 
alogued (stage 34.5, sv. 8.3 mm). Col- 
lected under leaves in small side pools 
off of a tributary to the Rio Aquabiena, 
behind Tropical Science Center, Rincon, 
Osa Penn., Costa Rica; March 8, 1970. 
Reference. — Dunn, 1924; Savage, 
1968; Heyer, 1976. 
External—Colostethus nubicola lar- 
vae have an umbelliform, denticle-free, 
oral disk, directed anteriorly in my speci- 
mens. The anterior surface of the disk 
is pustulate; the nonmuscular portion of 
the tail is slight; the spiracle is lateral. 
Ventral buccal.—The floor of the 
mouth is broad; the oral orifice wide. 
The lower beak is directed dorsally and 


displaced posteriorly in relation to the 
tongue anlage. As a result the lingual 
papillae lie directly medial, rather than 
posterior, to the major infralabial pa- 
pillae. The more medial infralabial pa- 
pillae are small, robust, blunt projections 
arising from the base of the keratinized 
beak at the anterior limit of the mouth. 
The larger, infralabial pair, in a more 
typical position for infralabial papillae, 
are complex bifurcated structures. They 
are directed medially but do not come 
into contact or touch the interceding 
lingual papillae. These infralabial pa- 
pillae have a cupped dorsal portion with 
the hollow of the cup pointing posterior- 
ly. The lower portion is a single blunt 
projection, directed medially. The in- 
fralabial papillae are fleshy, stiff struc- 
tures that lack additional surface pustu- - 
lations or papillae. The buccal floor 
arena is exceedingly broad. Two, soft, 
longitudinally oriented swellings arise 
from the buccal floor anterior to the buc- 
cal pockets. These run approximately 
from the back of the infralabial region 
to the medial corner of the buccal pock- 
ets and define lateral boundaries for the 
BFA. The posterior half of the BFA is 
defined by conventional rows of typical 
papillae; however, these rows continue 
laterally as small papillae in front of 
the buccal pockets. I counted five BFA 
papillae on each side. Assorted pustu- 
lations occur in the posterior portion of 
the BFA and on the buccal floor direct- 
ly anterior to the buccal pockets. The 
buccal pockets have a large, fleshy pos- 
terior flap. They are deep and appear 
perforated in this specimen. The free 
velar surface is long and supported by 
long, conspicuous spicules. The free 
edge is similar to that of Colostethus 
subpunctatus; there are small posteriorly 
directed peaks on the free velar edge 
over the second filter plate and medially 
directed peaks over the third filter plate. 
A tiny cusp, pointed posteriorly on each 
side, is aligned over the top of the fourth 
filter plate. There is a well-developed 
medial notch. Secretory tissue exposed 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 95 


Fic. 37.—Drawings of the floor (above) and roof (below) of the mouth of a Colostethus nubicola 
larva. 


96 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


on the dorsal side is similar to that of 
C. subpunctatus. 

Ventral pharynx.—Branchial baskets 
are similarly shaped to those of C. sub- 
punctatus; however, the second filter 
plate is not as tall while the third is 
slightly taller and has a curved dorsal 
margin which virtually covers the third 
filter cavity. Counts for filter rows are 
7, 8, 8, 5 for cb. 1-4; these are slightly 
less than those of Hyla regilia at a com- 
parable stage. The filter density and 
fold pattern is almost indistinguishable 
from that of C. subpunctatus and less 
dense than that of typical pond tadpoles. 
Tertiary filter folds are less well devel- 
oped in this dendrobatid than in the 
one previously described. Filter canals 
are very open channels. The branchial 
food traps are shallow, but cover a large 
area due to the large free surface of the 
ventral velum. Secretory ridges could 
barely be discerned at 50 magnifica- 
tion. The glottis is two thirds covered 
by the ventral velum. In size and shape 
it is similar to C. swbpunctatus. The 
esophageal funnel is relatively broad but 
not as large as in C. swbpunctatus. 


Dorsal buccal.—The buccal roof is 
elongate. The prenarial arena is flat and 
very conspicuous. A narrow sharp ridge 
runs from the anteromedial margin of 
the dorsal beak two-thirds of the way 
back in the prenarial arena. A papilla 
descends on each side of the prenarial 
arena about halfway between the upper 
beak and the nares. Each is transversely 
compressed with a gently curved, free, 
ventral edge. Medially the ventral edge 
for each papilla continues along the 
buccal roof as a posteriorly directed 
ridge that converge on a tiny papilla 
midway between the internal nares. 
There are no pustulations or additional 
structures in the prenarial arena. The 
internal nares are oriented transversely. 
Their anterior walls are shallow and 
lack projections or papillations of any 
sort. There is a short, but distinct, narial 
valve projection far medial on the pos- 
terior edge of the narial valve. A glo- 


bous projection arises from the lateral 
wall of the buccal roof on each side. 
Each fills the space between the narial 
valve projection and the lateral margins 
of the nares. The postnarial arena is 
extremely small and is bounded poste- 
riorly by a small, truncated papilla 
which on topographic grounds may be 
the homologue of the median ridge. No 
other papillae are present in the post- 
narial region except for a giant fleshy 
flap behind each narial valve projection 
and a pustulation (homologue of lateral 
ridge papilla) immediately lateral to the 
median ridge on each side. The flaps 
just described is the largest structures 
on the buccal roof. They are transverse- 
ly compressed with a rounded, free edge 
that covers the medial half of the inter- 
nal nares on each side. The next largest - 
structures associated with the buccal 
roof are simple, soft, oval projections 
arising from the far lateral walls of the 
buccal roof directly posterolateral to the 
large flaps just described. These more 
lateral structures may be homologues of 
lateral roof papillae in more typical anu- 
ran larvae. A faint, longitudinal ridge 
runs the length of the middle third of 
the buccal roof on each side. These 
structures vaguely define the lateral mar- 
gins of a narrow BRA. There is a field 
of some twenty to thirty small pustula- 
tions in the region directly behind and 
medial to the ridges, but no papillae of 
any sort are present on the buccal roof 
behind the median (ridge) papillae. The 
glandular zone is large, but its anterior 
margin is poorly defined in this speci- 
men. Individual secretory pits could 
not be discerned. The dorsal velum is 
of comparable size to that of more typi- 
cal pond tadpoles. It is barely continu- 
ous across the midline. 

Dorsal pharynx.—Pressure cushions 
are distinct. The lateral cushion is small- 
er and more elongate along an anterior- 
posterior axis. The medial cushion is a 
larger, more oval bulge. The ciliary 
groove was not preserved in dissection. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES oF 


Diagnostic summary. — Colostethus 
nubicola is very different from C. sub- 
punctatus in its buccal anatomy but it 
shares many features with other funnel- 
mouthed larvae from other families. 
However, its particular pattern of infra- 
labial papillae, large flaps behind the 
nostrils, and parallel, pustulate ridges in 


place of the BFA papillae, were ob- 
served in no other species. The great 
difference between C. nubicola and C. 
subpunctatus, plus the fact that only two 
species have been examined, means that 
it is not possible to provide a single 
diagnosis for the family Dendrobatidae 
at this time. 


DISCUSSION 


FUNCTIONAL CONSIDERATIONS 


From structural relationships and 
knowledge of the habits of various tad- 
poles, functions can be hypothesized for 
many of the structures presented in the 
Descriptions. In the following section 
variation in internal oral features is re- 
viewed and possible functions for these 
structures are discussed. In many cases 
additional comparative and experimen- 
tal work will be necessary to verify func- 
tions proposed here. 


KERATINIZED STRUCTURES 


Two types of internal keratinized 
structures have been observed. The first 
is a medially directed spur located at 
the posterodorsal corner of the infrala- 
bial cartilage, just posterior to the lower 
beak. These spurs are well developed 
in Gastrotheca riobambae and show 
weak development in Agalychnis calli- 
dryas and Colostethus subpunctatus. 
They are too far back in the junction of 
the jaws to be applied to the substrate 
and the fact that they are directed me- 
dially prevents their occlusion with 
either the upper or lower beaks. They 
have not been discussed in the litera- 
ture, but have been illustrated in Rana 
rugulosa by Noble (1927, Fig. 10). Their 
occurrence in tadpoles of such diverse 
morphology, ecology, and relationship 
makes it difficult to infer a function for 
these poorly known structures. The 
spurs simply may be associated with 
relatively large size in free-swimming 
tadpoles of certain of the more advanced 
Type 4 families. It is possible that they 


assist in shredding food that is spit in 
and out of the mouth. 


The second type of internal kerati- 
nized structure is the median cornified 
knob in the prenarial arena of Scaphio- 
pus bombifrons. This knob has not been 
described previously, although it was 
mentioned as a feature that could sepa- 
rate Spea from Scaphiopus in Altig’s 
“Key to U.S. tadpoles” (1970). The 
knob is ideally positioned to assist in 
cutting long, firm plant material. Elon- 
gate matter drawn into the mouth would 
be held in position by the medially di- 
rected pair of infralabial papillae. The 
lower beak would close between the 
knob and the upper beak, thereby shear- 
ing the material. The presence of such 
a preparative aid to ingestion in S. bom- 
bifrons is consistent with the known 
voraciousness of these larvae. 

Ueck (1967) reports keratinized 
structures within the oral tube of a Hy- 
menochirus boettgeri larva. Examina- 
tion of a photomicrograph (Ueck, 1967, 
Fig. 53), indicates identical morphology 
with my own sectioned material of this 
species. What Ueck calls “die verhorn- 
ten Zellen” appear to be cornified squa- 
mous cells in a thickened layer identi- 
cal to what is seen in the mouth of adult 
Hymenochirus (O. Sokol, pers. comm.). 
Griffiths (1963) said that the mouth of 
Pseudhymenochirus is “armed” with 
minute denticles. Although he cited an 
unpublished observation for proof of 
this, I find no evidence for such denti- 
cles in my own sectioned material of 
this genus or any other pipid larvae. 


98 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


INFRALABIAL PAPILLAE 


There are normally two major infra- 
labial papillae in all free-swimming tad- 
poles of Types 3 and 4. These may be 
augmented by as many as six minor 
papillae on each side (e.g., Hyla rufitela, 
Agalychnis callidryas). The two major 
infralabial papillae can be long and sec- 
ondarily papillate (e.g., Ptychohyla leon- 
hardschultzei, Leptobrachium oshanen- 
sis, Oreolalax) or short and globose (e.g., 
Hyla phlebodes, Hyla dendroscarta). 

In Ascaphus the individual papillae 
are replaced by a large, bilobed oral 
valve (Gradwell 197la, 1973). This 
structure can be derived easily from two 
infralabial papillae that have fused on 
the midline. This valve allows Ascaphus 
larvae to cling to substrata without con- 
tinuous pumping. However, such a valve 
does not occur in other suctorial tad- 
poles (e.g., Hyla mixe; see also Grad- 
well, 1975b, for Litoria lesueuri, L. 
booroolongensis, L. glandulosa, L. citro- 
pa, and Mixophyes balbus) and is not 
essential for suctorial clinging. 

Rhinophrynus has four tiny papillae 
unlike any seen in other tadpoles. 

Judging from the drawings by Weisz 
(1945), Sterba (1950), Sokol (1962, 
1977a) and Ueck (1967), infralabial pa- 
pillae are absent in the Pipidae. 

Microhyla has two or three knob-like 
infralabial papillae per side. 


There are three possible functions for 
the infralabial papillae: they may func- 
tion as respiratory structures, sensory 
structures, or mechanical interactors 
with food or water currents. Respiration 
can hardly be a major function of these 
structures; for even in the species with 
highly dendritic infralabial papillae, 
their surface area is only a small fraction 
of the surface area of the remainder of 
the buccal floor and pharynx. Accord- 
ing to Gradwell (1972a, for R. cates- 
beiana), “Relative to the pharynx and 
gill cavity the buccal lining is poorly 
vascularized and is probably of little 
significance for blood oxygenization by 
the buccal water currents.” 


The infralabial papillae extend into 
the prenarial arena. The common com- 
plex, hand-like infralabial papillae are 
aligned directly in front of the internal 
nares. In this position the papillae could 
very easily serve the function of direct- 
ing particulate matter medially and 
away from the nares, where large parti- 
cles could cause obstruction. 

A possible function for the simple 
globose papillae of some species is 
chemoreception; however, the papillae 
have a poor surface-to-volume ratio for 
this purpose. The papillae may be me- 
chanical sensors; in species where they 
are in medial contact, objects entering 
the mouth will collide with them. In 
Centrolenella, Microhyla and Mego- 
phrys, the papillae are arranged so that 
large particles can be prevented from - 
accidentally entering the mouth cavity, 
a function proposed by Liu (1950, p. 
191). Indeed, in funnel-mouthed forms 
(e.g., Ptychohyla schmidtorum, Mego- 
phrys minor, Microhyla heymonsi, Colo- 
stethus nubicola) the marginal palps fit 
into the lateral margins of the posterior- 
ly directed “V” of the prenarial arena. 
A tight tongue-in-groove configuration is 
formed, making it impossible for large 
particles to enter the corners of the 
mouth. 


Abundant medium size plant frag- 
ments have been observed in the guts 
of Hyla microcephala, which is similar 
to Hyla phlebodes and a close rela- 
tive of it (Wassersug and Rosenberg, 
1979). The globose, medially abutting, 
infralabial papillae of these species may 
serve a function of providing informa- 
tion on the position of plant fragments, 
informing the tadpole when filamentous 
plant matter is far enough back in the 
mouth to be bitten off. 


The globose infralabial papillae of 
Hyla dendroscarta are an enigma. They 
may simply be embryonic, considering 
the early stage of the tadpole examined. 

Large, elongate infralabial papillae 
are seen in all stream-adapted tadpoles 
that feed on the bottom (e.g., Ptychohyla 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 99 


leonhardschultzei, Oreolalax, Lepto- 
brachium oshanensis, Hyla mixe) except 
Ascaphus. Duellman (1970, p. 412) il- 
lustrated large, complex infralabial pa- 
pillae in Hyla smaragdina, a species of 
the stream-adapted Hyla_ sumichrasti 
group. The papillae point out of the 
mouth in these stream forms and may 
serve as taste or pressure receptors when 
the tadpoles contact a hard surface. 
Support for this hypothesis must await 
experimental or neurohistological stud- 
ies. 
LINGUAL PAPILLAE 


Ascaphus has a dense field of several 
dozen lingual papillae, whereas lingual 
papillae are absent in microhylids, Rhi- 
nophrynus, and, by definition, in the 
aglossal pipids. Discoglossid larvae have 
a variety of patterns for the surface anat- 
omy of the lingual anlage. At one ex- 
treme is Alytes obstetricans, with a dense 
field of pustulations similar to the field 
of papillae in Ascaphus; at the other 
extreme is Bombina orientalis with a 
single medial projection. Alytes cister- 
nasii has two lingual papillae and many 
pustulations; Discoglossus pictus has an 
arch of six enlarged papillae of which 
the medial ones are fused. These latter 
two species demonstrate morphological 
patterns intermediately between the ex- 
tremes of the family. 

A single medial projection occurs in 
Anotheca, but it is grossly dissimilar to 
that of Bombina. Duellman (1970, p. 
326) illustrated a single medial projec- 
tion in Hyla zeteki. If this is a homo- 
logue to the lingual papilla, it is a re- 
markable convergence with Anotheca. 
(Both H. zeteki and Anotheca are large- 
mouthed, carnivorous, arboreal larvae.) 

The remaining tadpoles examined 
normally had either zero, two, or four 
lingual papillae. The papillae were ab- 
sent in H. phlebodes, H. sarayacuensis, 
H. mixe and the megophrynine pelobat- 
ids. In some of these forms, however, 
anterior papillae or pustulations near the 
infralabial papillae may be homologues 
of the lingual papillae. Among the 


hylids examined, only Gastrotheca has 
four lingual papillae. Rana esculenta, 
R. pipiens, Pelobates fuscus and Bufo 
cinerus (= B. viridis) have two lingual 
papillae (Schultze, 1870; Hammerman, 
1964), whereas Rana sylvatica, R. cates- 
beiana and R. clamitans have four 
(Helff and Mellicker, 1941; Hammer- 
man, 1964). Hammerman (1964) 
claimed that R. temporaria has two lin- 
gual papillae, but Savage (1952) illu- 
strates and DeJongh (1968) reports four 
in this species. Kratochwill (1933) re- 
ported four lingual papillae in Rana 
agilis (= R. dalmatina). Kenny (1969a) 
illustrates two lingual papillae in Phyl- 
lomedusa trinitatis. Lingual papillae 
have a chemoreceptive function (Ham- 
merman, 1967, 1969; Helff and Mel- 
licker, 1941). 


BuccAL FLOoR ARENA 


The BFA is nonexistent in certain 
species (e.g., Megophrys minor, Hyla 
phlebodes, H. ebraccata, H. sarayacuen- 
sis) or ill-defined (e.g., Scaphiopus bom- 
bifrons, Centrolenella fleischmanni) be- 
cause of the absence of papillae. In all 
other species studied papillae of the 
buccal floor define some sort of bilateral- 
ly symmetrical area. The arena is usu- 
ally oval or egg-shaped in the hylids, 
pelobatids and Alytes. [Illustrations in 
the literature indicate that the oval pat- 
tern also occurs in Rana and Bufo (see 
Savage, 1952). In Ascaphus, Bombina, 
Discoglossus, Colostethus, Agalychnis 
(and its relative Phyllomedusa trinitatis, 
Kenny, 1969a), Hyla dendroscarta, and 
the stream-adapted Hyla mixe and 
Ptychohyla schmidtorum, the papillae 
at the front of the arena diverge to form 
a “V” or “U” shaped arena. In Rhino- 
phrynus and the microhylids the papillae 
are limited to a broad arc behind the 
buccal pockets. Papillae of the buccal 
floor are evidently absent in the Pipidae 
(Sokol, 1962, 1977a; Ueck, 1967; Sterba, 
1950). 


The BFA papillae are most often 
slightly compressed, conical structures, 


100 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


which if large enough, curve medially. 
Certain tadpoles have rather specific 
types of papillae unlike those seen in 
any other species (e.g., the blunt, out- 
turned cylindrical papillae of Anotheca; 
the pustulate papillae of Oreolalax). The 
papillae are most numerous and attenu- 
ate in the stream species, the extremes 
being Ptychohyla leonhardschultzei (89 
papillae) and Hyla mixe (70 papillae). 
Gradwell (1972c) independently noted 
that papillae of the buccal floor are 
“better developed” in the stream-adapted 
tadpoles of Rana fuscigula than in lar- 
val R. catesbeiana. 


Bifurcation of BFA papillae seems to 
be associated with the size of individual 
papillae rather than with an arena of 
some specific shape or with the number 
of papillae. Terminal bifurcations are 
commonest on the larger papillae in spe- 
cies which have tall papillae. 


Possible functions for the BFA pa- 
pillae are similar to those proposed for 
the infralabial papillae. Again, respira- 
tion may be ruled out as a primary func- 
tion, although in some of the stream- 
adapted species the surface of the BFA 
papillae is substantial. A sensory func- 
tion is most commonly proposed (Kra- 
tochwill, 1933; Kenny, 1969a and 1969b). 
Gradwell (1972a) wrote that “the shape 
and orientation (of these papillae) sug- 
gest a role as detectors of particulate 
material in the respiratory system.” 
While this is plausible, I suspect that 
the papillae may do more than just 
sense particles; they could be important 
both in sorting and directing particulate 
matter in the mouth. In most species 
the buccal roof arena and buccal floor 
arena intermesh; thus, the spaces be- 
tween the papillae are not large (0.1- 
1.0 mm on the average). When a parti- 
cle larger than the distance between two 
buccal floor or roof papillae comes into 
the mouth it should be retained medially. 
Because the arenas narrow posteriorly, 
particles would then be funneled poste- 
riorly and into the esophagus. Of course, 
if a particle is so large that it cannot 


pass through the BFA and BRA papillae 
at the back of the arena, it could be 
coughed out of the mouth. Such cough- 
ing motions are well known in tadpoles 
subjected to irritating substances in the 
water. One reason for having a buccal 
straining mechanism is that it prevents 
large particles from entering the pharyn- 
geal cavity where they might clog the 
filters or the food traps. A similar siev- 
ing function has been proposed by Se- 
vertzov (1969). 

Stream-adapted tadpoles that scrape 
food from rocks have the densest mesh 
of buccal floor and roof papillae. These 
are tadpoles which ingest a_ rather 
coarse, self-generated suspension of 
foods, as opposed to the free suspension 
of microscopic plankton found in ponds 
or pools. The stream tadpoles have re- 
duced gill filters, and instead of using 
pharyngeal structures to strain their food 
the major job may be done by the BFA 
and BRA papillae. 


The pond-adapted tadpoles of the 
Hyla leucophyllata and H. microcephala 
groups lack buccal arenas, and must use 
their small orifices as the primary food 
sieve. The reduced BFA of Scaphiopus 
bombifrons is difficult to interpret. It is 
possible that with its relatively fine filter 
mesh and single large, filter cavity on 
each side, Scaphiopus faces little risk 
of clogging the filters. 


Megophrys minor and, to a lesser 
extent, Ptychohyla schmidtorum and 
Colostethus nubicola, have replaced 
rows of papillae with ridges. Both of 
these funnel-mouthed forms have two 
ridges that extend posteriorly from the 
tongue anlage and interlock with ridges 
on the roof of the mouth. These massive 
structures probably act as baffles, chan- 
neling particles backward in the mouth. 


BuccaL PocKETS 


The shape and size of the buccal 
pockets depend on the shape of the 
posterior margin of the ceratohyals and 
the anterior margin of the branchial bas- 
kets. The pockets may be very large in 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 101 


all dimensions, as in Leptobrachium, or 
so small as to be nearly absent, as in 
Hyla phlebodes. The pockets are evi- 
dently perforated only in Centrolenella 
fleischmanni, Colostethus nubicola, Pty- 
chohyla, Similisca sordida, Hyla rufitela, 
Gastrotheca riobambae, Hyla regilla (6 
out of 9 individuals), Oreolalax pingii 
and Leptobrachium hasselti. Some in- 
dividuals examined were too small to 
determine the state of this character. 
Perforation does not seem to be associa- 
ated consistently with any particular 
habitat or behavior and is a variable 
character within genera and species. 


The buccal pockets, as folds of skin, 
allow some autonomous movement of 
the buccal pump relative to the bran- 
chial baskets. One might expect that 
species with much rotation in the buc- 
cal pump (see Wassersug and Hoff, 
1979) would have the deepest pockets, 
but this does not appear to be so. 


The significance of perforation of the 
pockets is not clear. Gradwell and Pasz- 
tor (1968) discovered that the buccal 
pockets were open gill slits in Rana 
catesbeiana and constructed an elaborate 
hypothesis for the function of “this pha- 
ryngeal by-pass.” They suggested that 
the open slits may be a common feature 
which escaped the notice of earlier work- 
ers because of concealment by the pos- 
terior walls of the ceratohyals. In a 
later article Gradwell (1972a) viewed 
the condition of Rana catesbeiana as ex- 
ceptional, differing from “Ascaphus, 
Xenopus, Scaphiopus, Phyllomedusa, and 
many other genera.” The Australian lep- 
todactylid Mixophyes balbus lacks per- 
forated buccal pockets (Gradwell, 
1975b). I doubt that the pharyngeal by- 
pass is of much significance in terms of 
passage of water. Even in species with 
large buccal pockets the slits are small 
compared to the pharyngeal gill open- 
ings. Epithelial folds from the walls of 
the pockets clearly act as valves, retard- 
ing passage of fluid. The strange ecolog- 
ical and taxonomic distribution of this 
character makes it difficult to identify a 


function for it. Perhaps perforation is 
simply an embryonic feature occasionally 
retained in Type 4 tadpoles. 


PREPOCKET PAPILLAE AND OTHER 
FEATURES OF THE BUCCAL FLOOR 


Prepocket papillae, or at least papillae 
or pustulations over the body of the 
ceratohyals, are present in virtually all 
species examined except Scaphiopus 
bombifrons. Typically there are one to 
three papillae that project backward 
over the buccal pockets. These papillae 
are most numerous in species with a 
“U” or “V” shaped buccal floor arena, 
where the BFA papillae series becomes 
continuous with the prepocket papillae 
series. In species which have few BFA 
papillae, there are few papillae on the 
arms of the ceratohyals. Microhyla, Cal- 
luella, and Glyphoglossus (see Savage, 
1952, Figs. 21 and 26) and, presumably, 
other microhylids have only one very 
large prepocket papilla. From the liter- 
ature one can conclude that such papil- 
lae are absent in the pipids. 

The prepocket papillae may prevent 
particles from entering the buccal pock- 
ets. Together with the papillae of the 
buccal arenas they could act as sieves. 
The lateral roof papillae, in species 
where they occur, are aligned over the 
buccal pockets and may assist the pre- 
pocket papillae in some protective func- 
tion. In Rhinophrynus, where prepocket 
and lateral roof papillae are absent, 
there are transverse ridges on the buccal 
roof that insert into the buccal pockets; 
these may accomplish the same protec- 
tive task. 


In many species with numerous buc- 
cal floor papillae, a distinctive row or 
patch of papillae and pustulations may 
develop behind as well as in front of 
the buccal pockets. In at least one genus 
(viz., Ptychohyla) these are opposed 
dorsally by a similar patch. Again, a 
straining or sieving function may be in- 
ferred for these structures. 

In species of Microhyla there are one 
or more papillae aligned directly in front 


102 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


of the glottis. These could be very im- 
portant in sensing particles entering the 
mouth and also in directing particles 
away from the glottis. Rhinophrynus, 
which, like the microhylids, has its glot- 
tis arising from the middle of the buccal 
floor, also has anteromedial papillae in 
front of the glottis on the buccal floor 
arena. In Rhinophrynus, however, these 
are paired and lateral to, rather than 
centered on, the midline. 


Other tadpoles have unique arrange- 
ments of buccal floor features for which 
it is not so easy to hypothesize functions. 
Examples are the transverse row of buc- 
cal floor papillae seen in Hyla rufitela 
and the lateral papillate lobes of the 
tongue anlage in Leptobrachium osha- 
nensis. The singular occurrence of these 
features and our limited knowledge of 
the ecology and behavior of these tad- 
poles make functional speculations in- 
appropriate at this time. 


VENTRAL VELUM 


This is an important valve which 
separates the anterior buccal cavity from 
the posterior pharyngeal cavity. Kenny, 
Gradwell, DeJongh and others have pre- 
sented evidence to show that the margin 
of the ventral velum can press against 
the buccal roof and prevent regurgita- 
tion of water from the pharynx when 
the buccal floor is depressed. The ven- 
tral velum also supports the mucosa of 
the branchial food traps. Only the val- 
vular function of the velum is considered 
here. Although secretory pits on the 
dorsoposterior margin of the velum are 
discussed in this section, comments on 
the ventral secretory tissue are reserved 
for the section on the food traps. 


Six major configurations of the ven- 
tral velum are noted. Two of these have 
the ventral velum divided into right and 
left halves and interrupted on the mid- 
line. Of these two, one is the condition 
found in Rhinophrynus; the other is the 
condition in the microhylids (with the 
technical exception of Microhyla hey- 
monsi, see Description and below). In 


Rhinophrynus the velum is fully (to its 
posterior margin) attached to the tops 
of the filter plates of cb. 2 and 3. In the 
microhylids it is attached only partially 
(anteriorly) to the tops of these arches. 
In both cases major projections of the 
velar margin are lacking, the margin is 
gently curved, and the glottis arises from 
the buccal floor anterior to the velar 
margin. This condition, which I ob- 
served in Microhyla berdmorei and M. 
ornata, was observed by Savage (1952, 
1955) in Glyphoglossus molossus, Chap- 
erina fusca, Calluella guttulata, Kaloula 
pulchra, and Hypopachus barberi (= 
aquae), and by Gradwell (1974) in 
Phrynomerus annectens. Clearly the pat- 
tern is typical of the family. 

Of the four other configurations for 
the velum, one is shared by the Orton = 
Type 3 tadpoles, Ascaphus, Bombina, 
Discoglossus and Alytes. This is the con- 
dition in which the velum is continuous 
across the midline, but anchored to the 
tops of cb. 2 and 3, and the margin is 
either straight or convex anteriorly be- 
tween the filter cavities. These four gen- 
era are listed here in the order of de- 
creasing thickness of the velar margin. 

The most common configuration is 
for the velum to be continuous across 
the midline, and partially attached to 
the tops of the filter rows. Posterior pro- 
jections of the velar margin are usually 
present and conspicuous. This is the 
condition for the dendrobatids, centro- 
lenids, and most hylids examined. Judg- 
ing from the literature, this pattern is 
typical of other advanced Type 4 tad- 
poles. Scaphiopus bombifrons and the 
megophrynines seem to be exceptions. 
In those tadpoles a fifth configuration is 
seen. These species have the velum an- 
chored anteriorly but virtually lack any 
attachment of the velum to the tops of 
cb. 2 and cb. 3. Schultze (1892) illus- 
trated Pelobates fuscus with very tall 
filter plates on cb. 2 and 3, so it is possi- 
ble the velum in that species is more 
like that in the other Orton Type 4 tad- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 103 


poles than that in the pelobatids I ex- 
amined. 

The last configuration is the absence 
of a valvular velum in the pipids (Sokol, 
1962, for Hymenochirus; Menzies, 1967, 
for Pseudhymenochirus; Weisz, 1945; 
Sterba, 1950; and Gradwell, 197la, for 
Xenopus; Sokol, 1977a, for Pipa). Not 
all pipids are identical. In Xenopus the 
floor of the common pharyngeobranchial 
tract is covered by secretory ridges mor- 
phologically similar and presumably ho- 
mologous (Kenny, 1969b) to those 
found on the ventral surface of the ven- 
tral velum in Type 4 tadpoles. In Hy- 
menochirus and Pseudhymenochirus the 
surface is devoid of any organized secre- 
tory tissue. 


In species with a valvular velum 
there is always positive pressure in the 
atrial chamber surrounding the gill fila- 
ments and one-way flow out of the spir- 
acle. The division of the ventral velum 
into right and left halves in Rhino- 
phrynus and the microhylids should not 
affect the valvular function of the velum. 
Even for species with an undivided vel- 
um, Kratochwill (1933) and Kenny 
(1969a) argued that the flow of water 
from the buccal cavity into the pharynx 
will be by separate right and _ left 
streams. Gradwell (1970) verified this 
and showed experimentally that each 
side of the velum can act independently 
in R. catesbeiana. This independence is 
obligatory in Rhinophrynus and the mi- 
crohylids. In the pipids, which lack a 
valvular velum, one-way flow is achieved 
by the opercular flaps acting like valves, 
opening and closing with each cyclic 
depression of the buccal floor. For these 
species atrial pressures must be lower 
than ambient pressures during part of 
the cycle. 

For any species the mobility of the 
velar margin and the strength of the 
valvular seal are determined by the velar 
length and thickness, the extent of at- 
tachment of the velum to the underlying 
filter plate, and the stiffness and length 
of the spicules. In most species the 


velum can close passively from back 
pressure. This is the same mechanism 
that closes similarly shaped pocket 
valves in the vertebrate circulatory sys- 
tem. Because of the long attachment of 
the velum to the tops of the filter plates 
in Ascaphus, discoglossids and Rhino- 
phrynus, not only the velum but the pos- 
terior portion of the hypobranchial plate 
must rotate upward in order for the 
buccal floor to be sealed off from the 
filter chambers. In all tadpoles except 
pipids (Sokol, 1977a) the anterior edges 
of the hypobranchial plates are over- 
lapped by the ceratohyals. As the cera- 
tohyals are depressed the hypobranchial 
plate rotates on a transverse axis. This 
mechanism aids in closing the velar 
valve and is reviewed in Wassersug and 
Hoff (1979). (As a functional com- 
plex, this mechanism provides insight 
into why pipids, which lack the hypo- 
branchial/ceratohyal articulation, also 
lack a valvular velum. ) 

In Ascaphus truei there is a posterior- 
ly directed “V”-shaped trench on the 
buccal roof above the ventral velar mar- 
gin. The thickened velar margin insets 
into the trench when the valve is shut. 
Such a structural arrangement is remi- 
niscent of a gasket on a high pressure 
seal. That such a system should appear 
in this torrent-adapted larva is no sur- 
prise. Gradwell (197la) found a mean 
maximum oral disc suction of — 125 torr 
in five live, stage 32 Ascaphus tadpoles. 

The discoglossids show a thickened 
velar margin similar to that of Ascaphus 
but not as extensive. They also lack a 
trench on their buccal roof. 


Although Rhinophrynus has little mo- 
bility of the velar margin, the roof of 
the mouth above the free velar surface 
is contoured to assure sufficiently large 
contact area to close the valve. 


All the stream tadpoles with suctorial 
mouths except Ascaphus have spicules 
that are longer, stouter, or both, than the 
spicules of Hyla regilla or Acris crepi- 
tans. These stream forms must produce 
relatively great negative buccal pressures 


104 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


in order to adhere to rocks with their 
oral discs. Presumably, the enlarged 
spicules serve to strengthen the valve 
and prevent prolapse. The tadpoles of 
the Hyla leucophyllata species group 
have wide spicules and a thickened free 
velar surface. One can hypothesize that 
these tadpoles produce relatively great 
suction during inspiration. This may be 
important for drawing rather large frag- 
ments into the mouth. Hyla phlebodes 
has a very short ventral velum unsup- 
ported by spicules. The shortness of the 
velum insures some protection from col- 
lapse when the valve is elevated. Micro- 
hyla heymonsi, which appears to feed on 
large particles, has the thickest and pre- 
sumably the strongest ventral velum in 
the genus. Microhyla berdmorei and M. 
ornata, in contrast, have very delicate 
ventral vela and it is clear that they 
could not endure a high pressure differ- 
ential between the buccal and pharyn- 
geal cavities without a collapse of the 
valve. 

Two arboreal species, Hyla dendro- 
scarta and Anotheca spinosa, have re- 
duced ventral vela with little or no spic- 
ular support. The velum may be a func- 
tional valve in H. dendroscarta, but is so 
reduced in A. spinosa that it is hard to 
imagine how it could function. In Ano- 
theca the pharyngeal cavity itself is also 
extremely reduced, so that it may not 
matter whether the velum is valvular or 
not. The most unusual vela were seen 
in Megophrys minor and Scaphiopus 
bombifrons. At the present, the func- 
tional implications of their velar shapes 
are obscure. 


Two general patterns are evident for 
sculpturing of the velar margins as seen 
in dorsal view. This edge is most likely 
to appear as a smooth, gentle arc in tad- 
poles associated with standing water. In 
tadpoles associated with moving water 
marginal projections are distinct and 
numerous along the velar margin. There 
are, however, some outstanding excep- 
tions. Marginal projections are lacking 
in Ascaphus and quite distinctive in 


Agalychnis. When present, projections 
usually number one each for each filter 
cavity, with two or more additional pro- 
jections in front of the esophagus. Su- 
pernumerary projections are present in 
Leptobrachium and Oreolalax. In these, 
general projections may arise from the 
ventral medial surface of the velum as 
well as from the posterior medial edge. 
The most extreme development of super- 
numerary projections is seen in Lepto- 
brachium oshanensis, which has a multi- 
layered papillate fringe on the medial 
edge of its ventral velum. When the 
branchial baskets are an unusual shape, 
such as with the reduction of filter cav- 
ity 3 in certain stream tadpoles, there is 
a concomitant shift in the position of the 
projections (medial in the example of 
the stream tadpoles; see below). 


There is no literature concerning the 
possible function of the velar projections, 
although they must affect the valvular 
workings of the velum. With long and 
numerous projections such as are pres- 
ent in stream tadpoles, it becomes im- 
possible for the velar seal to open or 
close abruptly. The projections may play 
a role in regulating laminar flow and 
directing currents over the edge of the 
velum when the buccal floor is elevated 
and water is pushed over the velar edge 
into the pharynx. 


The medial portion of the ventral 
velum is displaced dorsally and anterior- 
ly in Anotheca and Gastrotheca due to 
a large laryngeal disc. The same portion 
of the velum is turned upward and for- 
ward in Ptychohyla schmidtorum and 
Leptobrachium oshanensis, but neither 
of these species has a large, early devel- 
oping glottis. Although the configura- 
tion in these two species may be a pres- 
ervational artifact, it is strange that it 
should occur in larvae of such similar 
overall ecology. 

In a general way, the extent of de- 
velopment of the median notch reflects 
the size and development of the glottis. 
At one extreme is Anotheca spinosa with 
a median notch half the width of the 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 105 


ventral velum and a large, functional 
glottis, which develops early, exposed 
within. At the other extreme is Ascaphus 
truei, which lacks a median notch and 
functional lungs as a larva. 


The secretory pits on the velar mar- 
gin have been discussed previously by 
Kenny (1969b) for one species (Pseudis 
paradoxa) and more recently and exten- 
sively for a large variety of species 
by Wassersug and Rosenberg (1979). 
No particular pattern of density or size 
is evident other than that the pits are 
most common on the posterior projec- 
tions of the velar margin and conse- 
quently most numerous in species with 
well developed posterior velar projec- 
tions. In Ascaphus and Bombina a fine- 
grained secretory tissue is seen under 
the dissecting microscope instead of well 
organized secretory pits. In Alytes a few 
tiny pits can be seen on the margin, 
while in Rhinophrynus secretory pits per 
sé are absent from this region. In spe- 
cies with a reduced velum (e.g., Hyla 
phlebodes) secretory tissue along the 
ventral velar margin cannot be resolved 
with light or scanning electron micro- 
scope (Wassersug and _ Rosenberg, 
1979). 

The pits of the ventral velum oppose 
the secretory tissue of the glandular zone 
and presumably function in concert with 
that tissue. As noted by Kratochwill 
(1933), the edge of the free velar sur- 
face matches the shape of the glandular 
zone. Gradwell (1970) suggested that 
friction of the velar edge meeting the 
buccal roof may be reduced by mucous 
secretion. This mucus could be extruded 
from the margin of the ventral velum as 
well as the glandular zone and, as dis- 
cussed by Wassersug and Rosenberg, 
could act both to lubricate and seal the 
valve. Kenny (1969b) and Wassersug 
and Rosenberg (1979) have suggest- 
ed that the secretory pits may be in- 
volved in the food trapping process. 
Mucous strands excreted by the secre- 
tory pits of the roof and floor could 
catch and aggregate plankton much as 


has been proposed for the mucous strand 
generated by the secretory ridges under 
the velum (Wassersug, 1972). 


THE FILTER SYSTEM 


Different species vary enormously in 
the extent of the development of the fil- 
ters. This variation is most evident in 
the size and shape of the branchial bas- 
kets. Massive branchial baskets charac- 
terize Agalychnis callidryas, Phyllome- 
dusa trinitatis (Kenny, 1969a), the mi- 
crohylids excluding Microhyla heymonsi 
(see Savage, 1952, 1955, for genera other 
than Microhyla), Xenopus (Sterba, 1950; 
Ueck, 1967) and free-swimming Pipa 
(Sokol, 1977a). Among the species ex- 
amined, Bombina orientalis, Scaphiopus 
bombifrons, Hyla femoralis, Hyla rufi- 
tela, and Gastrotheca riobambae have 
baskets slightly larger than “typical” 
pond larvae, such as Hyla regilla and 
Acris crepitans, but not nearly as large 
as in the species just mentioned. In 
those species with an enlarged pharynx 
the greatest increase in size is in the 
length of the filter plate. Also common- 
ly exhibited is an increase in height. 
The ceratobranchials tend to have a 
more longitudinal, less oblique, orienta- 
tion in these forms. Scaphiopus is some- 
what an exception in that its branchial 
baskets are bowl-shaped and, although 
larger overall than those of H. regilla, 
the filter plates of cb. 2 and 3 are much 
reduced in height. 

Large branchial baskets are asso- 
ciated with specialization for suspension- 
feeding, particularly in species which 
lack keratinized mouth parts (Savage, 
1952). However, one must be cautious 
of this generalization (see Kenny, 
1969a); pharyngeal size alone cannot be 
the determining factor for the size range 
or volume of particles removed from the 
water by a tadpole. The density of the 
filter mesh, and other oral features such 
as buccal floor and roof papillae, bran- 
chial food traps (Wassersug and Rosen- 
berg, 1979), volume of the buccal 
pump (Wassersug and Hoff, 1979) and 


106 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


size of the orifice will have a_pro- 
found effect on the quality and quantity, 
including size range, of food ingested. 

At the other extreme, the filter bas- 
kets are much reduced in Hyla phle- 
bodes and Anotheca spinosa, and slightly 
reduced in Hyla sarayacuensis, Colo- 
stethus nubicola, Microhyla heymonsi 
and Megophrys minor. Filter surfaces 
are totally lacking in Hymenochirus (So- 
kol, 1962; Ueck, 1967) and in Pseud- 
hymenochirus (pers. obs.). Reduction 
is exhibited in both the height and 
length of the filter plates. 

Reduction of the branchial baskets 
occurs in three types of tadpoles. The 
first includes known carnivores such as 
Hymenochirus and Anotheca. These 
tadpoles lack any ability to extract fine 
plankton from suspension and instead 
select relatively large individual prey 
items. From stomach contents (Parker, 
1931; Cei, 1968) it would seem that the 
aberrant leptodactylid Lepidobatrachus 
fits into this class; indeed, serial sections 
of Lepidobatrachus llanensis larvae re- 
veal extreme reduction of the branchial 
baskets (Wassersug, unpublished). 

The second type includes tadpoles 
which feed at the surface using upturned 
funnel mouths. Examples are Colo- 
stethus nubicola, Microhyla heymonsi 
and Megophrys minor. This reduction 
suggests that funnel-mouthed tadpoles 
are not as dependent on microscopic 
particles for food as is implied by some 
of the older literature. Pope (1931) 
commented on the sparsity of micro- 
scopic organisms in the clear mountain 
pools inhabited by funnel-mouthed lar- 
vae. Heyer (1973) found small arthro- 
pods in the guts of larval M. heymonsi, 
and Smith (1917) showed surprise at 
the relatively large particles that Micro- 
hyla achatina could ingest. 

The last type of tadpole with reduced 
branchial baskets includes pond species 
which lack denticle rows, exemplified by 
Hyla phlebodes and, to a lesser extent, 
by Hyla sarayacuensis. Wassersug (1973) 
suggested that hylid larvae with fila- 


mentous tails and reduced denticle rows 
may be convergent with truly special- 
ized suspension-feeders such as Xenop- 
us, but this now appears to be an over- 
generalization, at least for H. phlebodes. 
In an earlier part of this discussion, it 
was suggested that H. phlebodes tad- 
poles may ingest vegetative strands that 
they cut into fragments of moderate 
size. The reduction of the branchial 
baskets is consistent with the view that 
larger fragments and not the smaller 
plankton are the important food for this 
species. Larvae of the ranid genus 
Ooeidozyga have mouths which are sim- 
ilar in gross features to the hylids of the 
H. leucophyllata and H. microcephala 
species groups. Serial sections of the 
pharynx and digestive tract of an Ooei- 
dozyga laevis larva reveal reduced bran- - 
chial baskets as well as coarse plant 
matter in the intestines (Wassersug, un- 
published). While this further demon- 
strates convergence with certain hylid 
larvae, the observation does not support 
the view that Ooeidozyga larvae are 
carnivorous (Alcala, 1962). Heyer con- 
cluded from his study of larval gut con- 
tents that Ooeidozyga tadpoles are omni- 
vores, or if carnivorous, only facultative- 
ly so. 

Stream-adapted tadpoles do not show 
any strong trend toward either expand- 
ed or reduced branchial baskets. How- 
ever, a common feature of bottom-dwell- 
ing forms with suctorial mouths is an 
anterolateral displacement of the pos- 
teromedial edge of the branchial bas- 
kets. This gives the third filter cavity a 
truncated appearance. Also, there is a 
tendency (except in the stream mego- 
phrynines) toward extensive imbrication 
of the filter plates. These modifications, 
amounting to a broadening and dorsal- 
ventral shortening of the pharynx, are 
consistent with the low, broad, profile 
characteristic of stream-adapted orga- 
nisms (Hora, 1930). Centrolenella lar- 
vae, in contrast, exhibit elongation of the 
filter cavity. Their branchial profile ap- 
pears to reflect the general elongation of 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 107 


the tadpole and can be understood as an 
adaptation to burrowing (Gans, 1975). 

Although relative size only is dis- 
cussed above (i.e. size of the pharynx 
compared to size of the buccal cavity), 
it should be noted that there seems to 
be a natural upper limit on the absolute 
size of the branchial baskets. Despite 
their giantism, larvae of Pseudis para- 
doxa have branchial baskets no larger 
than those of bullfrog tadpoles one third 
to one half their total length (Parker, 
1881). 

Directly correlated with the overall 
size and shape of the branchial baskets 
is the number of filter rows on each filter 
plate. The number of filter rows, how- 
ever, is somewhat affected by the abso- 
lute size of the pharynx, so comparison 
of species with very large larvae with 
species with very small larvae may give 
slightly different counts even though the 
tadpoles have the same shape and pro- 
portions. 

Of the species examined, the highest 
counts for filter rows were in Microhyla 
berdmorei, M. ornata, Rhinophrynus dor- 
salis and Agalychnis callidryas, whereas 
the lowest counts were in Anotheca spi- 
nosa and Hyla phlebodes. In the above 
species, counts were significantly (p < 
.05) different from Hyla regilla for all 
four filter plates (cf. Table 1; Wasser- 
sug, 1976b). In certain other species, 
one or two, rather than all four, filter 
plates had significantly fewer or more 
filter rows than H. regilla. Thus, Hyla 
sarayacuensis shows a tendency toward 
reducing the number of filter rows as 
does Leptobrachium hasselti, whereas 
Scaphiopus bombifrons and Hyla femo- 
ralis show a tendency toward increasing 
the number of rows. 

Kratochwill (1933) gave counts for 
filter rows in Rana dalmatina that were 
within the range for Hyla regilla, except 
for cb. 4, which had a higher count. 
Such a difference may reflect differences 
in the shape of the branchial baskets or 
in counting procedure. There are no filter 
rows to count in Hymenochirus (Sokol, 


1962), whereas Pipa carvalhoi has counts 
of approximately 14, 19, 25 and 16 (So- 
kol, pers. comm.), all significantly above 
the range for H. regilla. From Sterba’s 
(1950) illustration it is clear that Xeno- 
pus laevis has significantly more filter 
rows on all arches than does H. regilla. 
The number of filter rows seems like a 
rough, but good, indicator of the amount 
of effort any species puts into micropha- 
gous suspension feeding. 

Substantial variations in the filter 
mesh between species provide a clear 
indication of a species’ ability to extract 
small particles from suspension. Filter 
porosity, however, may not be the abso- 
lute determinant of the lower size limit 
of the particles a filter feeder can extract 
(Kenny, 1969a; Wassersug, 1972; see 
LaBarbera, 1978). Of the species ex- 
amined, Rhinophrynus dorsalis, Agalych- 
nis callidryas, Microhyla berdmorei, and 
M. ornata have a particularly tight filter 
mesh, whereas Anotheca and Hyla phle- 
bodes have no filter mesh. Species with 
a filter mesh slightly denser than that of 
Hyla regilla include Hyla femoralis, 
Hyla rufitela and Gastrotheca riobam- 
bae. Species with a slightly less dense 
mesh are Ascaphus truei, the discoglos- 
sids, megophrynine pelobatids, dendro- 
batids, Microhyla heymonsi, and the re- 
maining hylids with the exception of 
Acris crepitans and Smilisca sordida. 
Scaphiopus bombifrons has a denser fil- 
ter mesh than H. regilla on the rows 
proper but greater spaces between the 
rows. 

A tight filter mesh is most commonly 
achieved by longer tertiary branches on 
the filter folds and by an increase in the 
number of higher order folds. It is pos- 
sible to achieve a tighter mesh without 
changes in the branching pattern by 
widening the individual side folds (e.g., 
Agalychnis) or by widening the rows in 
toto (e.g. Gastrotheca). The continua- 
tion of filter rows from one filter plate to 
the next over the gill slits at the bottom 
of the filter cavities is a unique feature 
of Gastrotheca. This assures that large 


108 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


particulate matter in the filter cavities 
cannot accidently escape through the gill 
slits into the atrial chamber. Sokol 
(1977a) reports a somewhat similar sit- 
uation in free-swimming Pipa larvae. In 
these tadpoles, however, filter ruffles are 
fused with the ruffles on neighboring 
rows of each filter plate rather than ver- 
tically over the gill slit. In Pipa, filter 
canals are present but overlaid with a 
continuous filter mesh that must be cut 
in order to expose the canals. I have 
verified this unusual anatomy in my own 
material of Pipa parva. 


Reduction of filter mesh is most com- 
monly achieved by deletion of higher 
order folds and shortening of secondary 
folds. Slight reduction can be accom- 
plished without a change in the branch- 
ing pattern by a simple thinning of all 
the folds. Wispy ribbons of tissue re- 
main as the vestiges of filter rows in 
Anotheca spinosa. In the extreme reduc- 
tion of the filter rows of Hyla phlebodes 
the main fold is retained as a raised knob 
on the gill bar. 


A few species have filter rows that 
are distinctive for other reasons than the 
tightness of their filter mesh. Ascaphus 
truei and Bombina orientalis share un- 
usually straight main filter folds with 
short, robust, side branches. 


Kenny (1969a) labelled the tops of 
the filter folds as filter ridges in Phyllo- 
medusa trinitatis and called the edges 
of the filter ridges “filter shelves.” How- 
ever, Gradwell (1972a) states that he 
found no filter shelves or filter crevices 
in Rana catesbeiana, whereas he does 
identify “pointed cells” at the crests of 
the ridges. I suspect that the discrep- 
ancy observed here is more the result of 
semantics and preservational artifacts 
than true differences between species. 
Both Kenny’s and Gradwell’s Fig. 9— 
their independent illustrations of the fil- 
ter rows—are somewhat ambiguous. I 
interpret the “tops of the filter ridges” of 
Kenny to be the same as the “pointed 
cells” of Gradwell. Under this designa- 
tion, Rana catesbeiana must necessarily 


have filter shelves. The acuteness of the 
ridge tips will vary with preservation. 
Gradwell preserved his specimens in 
Bouin’s Reagent, which tends to shrink 
tadpoles more than preservatives used 
by Kenny. Shrinkage would accentuate 
the ridge tips, narrow the filter shelves, 
and open the filter niches in such a way 
that individual filter crevices may not be 
obvious. 


Gradwell (1972a) has discussed the 
filters as respiratory structures. They are 
well vascularized, but their actual im- 
portance to gas exchange has not been 
measured and must vary greatly between 
species. In tadpoles such as Xenopus 
which lack gill filaments, the gill filters 
are obviously important respiratory 
structures (Millard, 1943). 


In many of the specimens examined, 
flocculent matter covers the filter sur- 
faces or packs the anterior ends of the 
filter cavities. There can be little ques- 
tion that the main function of the filters 
is to retain in the pharyngeal cavity par- 
ticulate material which enters the mouth 
with the respiratory current. The cur- 
rent discussion is limited to a considera- 
tion of the role of the filters in feeding. 


Kenny (1969a) set the lower limit 
for the dimensions of the filter niches at 
5 pm for Phyllomedusa trinitatis. Be- 
cause P. trinitatis has large, dense filters, 
this may be near the lower limit of the 
pore size for anurans in general. This 
measurement of 5 yum, however, was 
made on fixed material and, as noted by 
Kratochwill (1933), under positive pha- 
ryngeal pressure the filters should 
flatten and expand to form a tighter 
mesh. The strange, truncated shape of 
the filters and their resilience may relate 
to this ability to vary surface area and 
porosity with pressure. Under actual 
feeding conditions the filter niches may 
be smaller than 5 pm in some of the 
species specialized for suspension-feed- 
ing. Nevertheless, when very fine par- 
ticulate matter is found on filter surfaces, 
the particles are invaribly clumped in 
mucus; yet the filters themselves are not 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 109 


secretory. Thus, despite fluctuating filter 
niche porosity, ultraplanktonic entrap- 
ment must be initiated elsewhere in the 
mouth than on the filter surfaces. 


Questions remain as to how food 
matter is removed from the filter, and 
why mucus does not adhere more tena- 
ciously to these surfaces. Kenny (1969a) 
has suggested that there is some back- 
wash through the filters during part of 
every pumping cycle. In Rana cates- 
beiana, however, the pharyngeal pres- 
sure is only briefly and very slightly 
negative during the pumping cycle 
(Gradwell, 1972b), so that backwashing 
is unlikely. Reorientation of the filter 
plates may actually dislodge food parti- 
cles from the filter plates during each 
pumping stroke. Because of rotation 
in the hypobranchial plate the filter 
plates should be stretched every time 
the buccal floor is depressed. In most 
species, the pressure cushions are shaped 
to fit the filter cavities closely. The se- 
cretory pits found on the cushions might 
help lift material from the filters. The 
answer to why mucus does not clog the 
filters could conceivably lie in some 
aspect of the histochemistry of the mu- 
cus and the cells that make up the filter 
epithelium. Kenny (1969b) noted that 
the cells which line the filter niches can 
be distinguished from other cells in the 
filters because they stain unusually heay- 
ily with Orange G. 


BRANCHIAL Foop TRAPS AND 
SECRETORY RIDGES 


In most species the entire free velar 
surface is covered ventrally by secretory 
ridges. Thus, the parameter most rele- 
vant to the size of the branchial food 
traps is the size of the free velar surface. 
The height of the branchial baskets may 
also affect the size of the collecting sur- 
face because the secretory ridges com- 
monly cover a portion of the anterior 
and lateral surfaces of the jfilter plate. 
When the branchial baskets are very 
shallow, the ventral extent of the secre- 
tory area must be limited. Of the spe- 


cies examined, the largest collecting sur- 
faces were observed in Gastrotheca rio- 
bambae and Agalychnis callidryas. The 
smallest were seen in Hyla phlebodes. 
Branchial food traps are absent in Ano- 
theca and we may conclude from the 
literature that they are absent in Hymen- 
chirus and Pseudhymenochirus (Sokol, 
iets) 


The shape of the area covered by 
secretory tissue is largely determined by 
the shape of the ventral velum. Where 
the velum is completely anchored to the 
top of the filter plate (in Ascaphidae, 
Discoglossidae, and Rhinophrynidae), 
each filter cavity has a separate collect- 
ing area. In Ascaphus and the discoglos- 
sids the branchial food traps are further 
distinguished by a large, elevated rim 
that separates them ventrally from the 
filter surfaces on the filter plates. Lepto- 
brachium has an expansion of the area 
covered by secretory tissue onto the dor- 
sally facing surface between the fourth 
ceratobranchials anterior to the glottis. 

The microhyids and Rhinophrynus 
have collecting surfaces which are simi- 
lar in their gross shape. The surface is 
more (e.g., Microhyla berdmorei or less 
(e.g., Microhyla heymonsi), restricted 
to a dense, narrow, vertical crescent at 
the anterior end of the filter cavity. In 
Microhyla berdmorei much of the velar 
surface is devoid of secretory tissue. 


Perhaps the most unusual collecting 
surfaces occur in Xenopus. Here the 
secretory ridges are transposed to the 
ventral surface of the common bucco- 
pharyngeal cavity along the top of the 
filter plates. The gross shape of this 
secretory surface has been described and 
illustrated by Weisz (1945), Sterba 
(1950), Gradwell (197la, 1975a) and 
Sokol (1977a). Pipa and Hymenochirus 
lack these collecting surfaces (Sokol, 
1975). 

The histology of the secretory tissue 
has received extensive discussion by 
Kenny (1969b) and Wassersug and Ro- 
senberg (1979). Kenny reported se- 
cretory pits in Pseudis paradoxa along 


110 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


the margin of the ventral velum and 
thought them unique to P. paradoxa. 
Secretory pits are in fact common on the 
most posterior edge of the velum in 
many species, particularly those with 
long velar projections. Organized secre- 
tory ridges (Wassersug, 1972) are ab- 
sent in the Ascaphidae, Discoglossidae 
and Pelobatidae. In these forms the 
collecting surfaces have a fine-grained 
texture under light microscopy, indicat- 
ing the presence of secretory cells but 
without the ridge organization. Rhino- 
phrynus shows some development of 
ridges on part of the branchial food trap 
surface. Secretory ridges degenerate in 
Hyla regilla near metamorphic climax. 

Kenny (1969b) found no evidence 
that secretory cells were arranged either 
in pits or rows in the tadpoles of the 
hylid Amphodus auratus. Amphodus 
larvae are arboreal, and Kenny associated 
the absence of a ridged pattern with 
the unusual larval habitat of this species. 
I could not perceive secretory tissue un- 
der the velum in either Anotheca spinosa 
or Hyla phlebodes, nor could I find any 
ridges in Hyla dendroscarta. In members 
of the Hyla leucophyllata group, secre- 
tory ridges appear concentrated in the 
more anterolateral region of the pharynx 
near the ciliary groove. Whereas the 
above tadpoles are a rather heterogen- 
eous assemblage, they do offer some sup- 
port for the idea that absence of secre- 
tory tissue in tadpoles of advanced frogs 
is associated with abandonment of mi- 
crophagous suspension-feeding as the 
sole source of nourishment. 

The amplitude and frequency of se- 
cretory ridges may be important in de- 
termining the potential of the secretory 
surface to trap particles of a particular 
size or density. The pattern of the ridges 
seen in Xenopus (Kenny, 1969b) is 
unique for that genus. The ridged pat- 
tern in the microhylids is also unique for 
that family. In a general way, Kenny 
associates the features of the ridges in 
Xenopus and the microhylids with a spe- 
cialized, suspension-feeding way of life. 


Although I have concurred with this 
view (Wassersug, 1972), the question of 
the functional meaning of the textural 
patterns of secretory tissue remains un- 
answered. 


GLOTTIS AND LARYNGEAL Disc 


There seems to be considerable range 
in the size of the glottis and laryngeal 
discs, although most species do not vary 
much from the proportions observed in 
Hyla regilla. The extremes were seen in 
the discoglossids, Anotheca and Hyla 
phlebodes at the upper limit and As- 
caphus truei at the lower. To some ex- 
tent an enlarged glottis seems to char- 
acterize large individuals. A large glot- 
tis is clearly associated with early func- 
tioning of the lungs for respiration or 
hydrostasis. The small glottis seen in. 
larval Ascaphus truei is consistent with 
the fact that these tadpoles live in well- 
oxygenated water where stability in a 
current could be jeopardized by air- 
filled lungs. Although Ascaphus truei is 
the extreme, other stream-bottom forms 
also show some reduction in the size of 
the glottis compared with typical pond 
larvae. Bufo larvae which have a small 
nonfunctional glottis develop their lungs 
just before metamorphosis (Savage, 1950 
and 1961; Starrett, 1973) and it is no 
surprise that they are relatively restrict- 
ed to a bottom habitat in water of low 
turbidity (Wassersug, 1973; see also 
Wassersug and Seibert, 1975). 

The position of the glottis in relation 
to the velar margin again reflects larval 
lung utilization. Rhinophrynus, micro- 
hylid, and pipid larvae all have the glot- 
tis fully exposed on the buccal floor. 
These are tadpoles which have the abil- 
ity to stay in the water column con- 
tinuously, and their lungs must have a 
hydrostatic function. Although the re- 
maining tadpoles have the glottis behind 
the ventral velum, many have it fully or 
nearly fully exposed when viewed from 
above. These include the larvae of 
Scaphiopus bombifrons, Leptobrachium 
oshanensis, Hyla phlebodes, Hyla den- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES MULL 


droscarta, and Colostethus subpunctatus. 
On the other hand, the nonfunctional 
glottis of Ascaphus tadpoles is fully 
blanketed by the ventral velar margin. 
Evidently, accessibility to the glottis 
may be associated with early function of 
the lungs for respiratory as well as hy- 
drostatic regulation. 


ESOPHAGEAL FUNNEL 


The lumen of the esophageal funnel 
will limit the size of the largest food 
item that can enter the alimentary canal. 
In this regard the data in the descrip- 
tions, although subjective and only com- 
parative, are valuable as one index of 
feeding potential. Ueck (1967) noted a 
great discrepancy between the sizes of 
the esophagus in Xenopus and Hymeno- 
chirus. Xenopus has a strictly micropha- 
gous diet and consequently a slender 
esophagus. Hymenochirus is strictly a 
carnivore on large prey; it has a com- 
paratively huge esophagus. 

Among the tadpoles examined, the 
largest esophageal bore is seen in Ano- 
theca larvae. The discoglossids, Hyla 
phlebodes, H. ebraccata, H. sarayacuen- 
sis, and the funnel-mouthed microhylids, 
all have relatively large esophagi, 
suggesting an ability to ingest compara- 
tively large particles. The non funnel- 
mouthed microhylid larvae have esoph- 
agi with the smallest diameter. This 
observation is consistent with the view 
that these tadpoles are specialized for 
microphagy. 

The profile of the esophageal funnel 
reflects the shape of the posterior margin 
of the branchial baskets. Because of the 
lateral displacement of the third filter 
cavity in tadpoles associated with the 
stream habitat (with the exception of 
Ascaphus truei), these forms tend to 
have broad esophageal funnels. Dis- 
tinctly broad esophageal funnels can also 
be seen in the discoglossids, certain den- 
drobatids (Colostethus subpunctatus), 
amphignathodontine hylids, and _ the 
Hyla microcephala and H. leucophyllata 
species groups. A conspicuously narrow 


lumen characterizes the non funnel- 
mouthed microhylids and Agalychnis 
callidryas. The narrow esophageal fun- 
nel in the larvae of these species is 
ascribable to the relatively elongate 
branchial baskets that bound the esoph- 
agus. 
PRENARIAL ARENA 


For any tadpole, the size and shape 
of the prenarial arena is determined by 
the position of the internal nares, the 
breadth of the upper beak, and the 
curvature of the rostrum. The nares are 
far forward in Rhinophrynus tadpoles 
and larvae of the Hyla microcephala and 
Hyla leucophyllata species groups; con- 
sequently, these species have small pre- 
narial arenas. Another larva with a 
greatly reduced prenarial arena is As- 
caphus truei; this is also the species with 
the strongest posteroventral curvature of 
the rostrum. Wide prenarial arenas char- 
acterize the amphignathodontine hylids, 
microhylids, and funnel-mouthed tad- 
poles in general, although a compara- 
tively narrow prenarial region is found 
in the megophrynine pelobatids, exclud- 
ing Megophrys. 

Many species have specific structures 
within the prenarial arena. The func- 
tional significance of only a few of these 
are comprehensible given the sample at 
hand. A distinct, posteriorly directed 
“V” shaped ridge occurs in Ptychohyla 
schmidtorum, Megophrys minor, Colo- 
stethus nubicola and Microhyla hey- 
monsi, but in no other species examined. 
The ridges interlock with a_palp-like 
infralabial papillae in these funnel- 
mouthed forms and, as has been sug- 
gested earlier, could prevent large par- 
ticles from entering the corners of the 
mouth. We may expect to find this 
structural arrangement in other funnel- 
mouthed species. 


The huge, fleshy, pustulate cone de- 
scending from the prenarial arena in 
Hyla mixe is of sufficient size to occlude 
the oral opening. In the absence of an 
oral valve derived from the infralabial 
papillae (cf. Ascaphus), this dorsal 


112 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


structure may serve a valvular function. 

A pustulate or papillate, anteriorly 
curved crest is present in the prenarial 
arena of Hyla femoralis, Colostethus sub- 
punctatus, Ptychohyla leonhardschultzei, 
Agalychnis callidryas, Anotheca spinosa, 
and Smilisca sordida. This is a group 
of larvae of such ecological and taxo- 
nomic diversity that a single function 
may be unlikely for this crest. Centro- 
lenella fleischmanni has a short trans- 
verse ridge in the prenarial area; several 
other tadpoles have single, knob-like 
structures in this region (e.g. Hyla ru- 
fitela). Comparing pictures of the dorsal 
and ventral surfaces of the various tad- 
poles shows that all these structures lie 
immediately above and behind the lower 
beak when the mouth is closed. They 
may have a mechanical function related 
to maintaining beak occlusion or posi- 
tioning and anchoring macroscopic food 
items during mastication. Alternatively, 
future histological work may reveal that 
this prenarial feature serves a sensory 
role. Orton Type 1, 2 and 3 larvae lack 
these structures. 

The keratinized knob in the prenarial 
arena of Scaphiopus bombifrons has 
been discussed already under the head- 
ing Keratinized Structures. 


INTERNAL NARES 


The nares are perforated in all but 
microhylid tadpoles. They may be large 
open vacuities, as in Centrolenella and 
the discoglossids, or diminutive struc- 
tures, as in Hyla phlebodes and Rhi- 
nophrynus. When the nares are larger 
than in typical pond larvae, such as 
Hyla regilla, they also tend to be more 
elongate and more obliquely oriented. 
A few species have internal nares more 
transversely oriented than H. regilla; 
these include the megophrynine pelo- 
batids, Rhinophrynus, the dendrobatids, 
and Ptychohyla schmidtorum. In the 
elongate Centrolenella fleischmanni tad- 
poles the nares are, understandably, 
longitudinally oriented. 

The fact that the nares do not per- 


forate until late in development in the 
microhylids may relate to the efficiency 
of their buccal pump. Presumably, the 
risk of leakage at a valve is reduced with 
fewer orifices that open and close with 
each pumping cycle. 

In nonmicrohylid larvae the anterior 
narial walls are commonly thickened 
and pustulate, with one or two small, 
prenarial papillae. While it is possible 
that such prenarial projections are sen- 
sory, their position directly in front of 
or under the narial openings suggests 
that at least the larger papillae serve a 
protective function. They could keep 
larger particles from moving up into the 
nares, where they might disrupt valvu- 
lar action or might detrimentally coat 
olfactory surfaces. In species where gen- 
eral oral papillation is reduced, pustu- . 
lations or papillae of the narial walls 
are invariably absent. On the other 
hand, proliferation of floor and roof pa- 
pillae does not necessarily mean an in- 
crease in the number or size of prenarial 
papillae. 

Megophrynine pelobatids tend to 
have a short ridge projecting forward 
from the anteromedial edge of the narial 
wall into the prenarial arena. In approx- 
imately the same region anterior to the 
nares, Hyla rufitela and Centrolenella 
fleischmanni develop a large cul de sac 
not of the same shape but presumably 
having the same chemoreceptive func- 
tion. Ascaphus truei has somewhat simi- 
lar, but smaller and more lateral, pits 
adjoining the anterolateral wall on each 
side. The “ciliated epithelial bands” re- 
ported in this region by Van Eeden 
(1951) may be important in cleansing 
these organs. Leptobrachium and Oreo- 
lalax have the most attenuate prenarial 
papillae of any of the tadpoles examined; 
however, the most extreme modification 
of anteronarial walls are the flaps found 
in the discoglossids. These flaps are 
most likely protective. The fact that 
they should develop in the discoglossids 
and no other taxa may be related to the 
plane of the buccal floor and roof in 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 113 


discoglossids compared to other families. 
Discoglossids have a buccal floor that 
tips gently downward anterior to the 
pharynx and they lack a strong rostral 
curvature. Most other species examined 
have a comparatively horizontal buccal 
cavity and abrupt curvature in the ros- 
tral region. With this latter orientation 
feeding currents impinge on the pre- 
narial arena before being deflected back- 
ward. In the discoglossids, however, 
feeding currents must flow into the 
mouth along a more posteriorly directed 
course toward the internal nares. If not 
for the flaps on the anterior narial walls, 
olfactory tissues could be abraded by 
coarse particles in the feeding currents. 

The posterior walls of the nares are 
sufficiently flexible and properly posi- 
tioned to act as valves in all the tadpoles 
examined except the microhylids and 
possibly the arboreal hylids. The di- 
mensions of the valves strictly reflect 
the dimensions of the nares. The narial 
valve projections, however, vary sub- 
stantially in size. In approximately half 
the species examined (megophrynine 
pelobatids, microhylids, Hyla_ rufitela, 
H. femoralis, H. dendroscarta, Anotheca 
spinosa, Gastrotheca riobambae, Aga- 
lychnis callidryas, and Acris crepitans) 
have narial valve projections equal to or 
larger than those of H. regilla. Dendro- 
batids have smaller valve projections. 
Of the remaining forms many, such as 
all Orton Type 1 and 3 larvae, Centrole- 
nella, Hyla phlebodes, H. ebraccata, and 
H. sarayacuensis tadpoles, lack the pro- 
jections. The narial valve projections of 
Anotheca spinosa are in the middle of 
the posterior walls on each side rather 
than at the medial corners. The most 
attenuate narial valve projections occur 
in Leptobrachium and Oreolalax larvae. 
Although narial valve projections may 
be sensory structures, it is equally likely 
that, in conjunction with prenarial pa- 
pillae and pustulations, they block large 
particles from entering the internal 
nares. 

In the microhylids that I examined 


the narial valves are abbreviated trans- 
versely but expanded forward and down- 
ward as large, cup-shaped projections. 
Judging from the diagram of Savage 
(1952, 1961), and Gradwell (1974), this 
is typical of the family. These projec- 
tions are stiffened by a ridge that runs 
down their posterior margin. Noble 
(1927) illustrates these projections in 
Microhyla pulchra, noting that they 
would effectively isolate the olfactory 
chambers from the buccal cavity on each 
side. He also speculates that these flaps 
could control and even direct water into 
the higher portions of the olfactory 
chambers. Gradwell offered the same 
hypothesis for the function of these 
structures in Phrynomerus; it seems like 
a very reasonable hypothesis. 


POSTNARIAL ARENA 


Superposed photographic prints and 
drawings of the floor and the roof of 
the mouth reveal that the tongue anlage 
makes a near-perfect fit into the post- 
narial arena. Species which have a poor- 
ly outlined postnarial arena (e.g., micro- 
hylids, Rhinophrynus, Hyla phlebodes, 
H. ebraccata, and H. sarayacuensis) also 
tend to show little development of the 
tongue during the stages that I exam- 
ined. Similarly, it is reasonable that 
postnarial arena structures are absent in 
the aglossal pipids. Other than to form 
a receptacle for the growing tongue, the 
roles of the postnarial papillae and me- 
dian ridge are not evident. Although 
they have been variously illustrated 
(Goette, 1874; Schulze, 1889; Kenny, 
1969a and 1969b), and DeJongh (1968) 
offered a fairly complete description for 
Rana temporaria, no author has dis- 
cussed these or any of the other buccal 
roof structures in a systematic fashion. 
Except for Kenny (1969a and 1969b), 
who labelled these projections along 
with the lateral ridge papillae as “sen- 
sory, no function has ever been sug- 
gested for them. 

When a postnarial arena is present, 
it is bound laterally by an average of 2 


114 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


or 3 postnarial papillae and assorted pus- 
tulations. The greatest number of post- 
narial papillae that I counted in any 
specimen was 7 per side (Acris crepi- 
tans). In funnel-mouthed larvae, these 
papillae are consolidated into a massive 
ridge on each side. A smaller ridge is 
developed in Ptychohyla leonhardschult- 
zei and to a still lesser extent in Centro- 
lenella fleischmanni and Agalychnis cal- 
lidryas. 

The median ridge defines the back 
of the postnarial arena. The ridge has a 
curved, pointed, or straight margin in 
most tadpoles examined and in ones il- 
lustrated by other authors; although the 
pattern is rather species-specific, from 
looking across the samples it is clear 
that there is a continuous spectrum of 
possible shapes for this structure. In 
Hyla mixe, Centrolenella fleischmanni, 
Anotheca spinosa, and _ Ptychohyla 
schmidtorum the median ridge is re- 
duced to a tall papilla. In Megophrys 
minor, the ridge is replaced by a large, 
nipple-shaped, projection; a similar but 
much smaller projection occurs in Colo- 
stethus nubicola. 


Serrations or fine papillae often oc- 
cur on the free ventral margin of the 
median ridge in tadpoles with a wide 
ridge. When the median ridge is rela- 
tively large it is often covered with pus- 
tulations or a secondary serrated ridge 
on its anterior surface. 


Because of the medial position of the 
postnarial arena, this collection of pro- 
jections, together with the tongue an- 
lage, may be important in dividing up 
the respiratory current into right and 
left streams. This is not to say that the 
papillae are without additional possible 
functions such as the sensory role im- 
plied by Kenny. 


LATERAL RIDGE PAPILLAE 


The structures which I have called 
the lateral ridge papillae are develop- 
mentally associated with the postnarial 
and buccal roof arenas. When these 
arenas are reduced or absent, the lateral 


ridge papillae invariably are reduced or 
absent. Increases in the number and 
size of arena papillae vary directly with 
increases in the size of the lateral ridge 
papillae. However, the papillae are ab- 
sent in all Orton Types 1, 2, and 3 lar- 
vae that I examined. 

The papillae may be simple and cy- 
lindrical with minor terminal and an- 
terior pustulations or secondary papillae 
(e.g., Hyla regilla, H. dendroscarta, Acris 
crepitans, and Anotheca spinosa); or 
they may be rather huge flaps with long, 
finger-like, marginal papillae (e.g., Hyla 
mixe, Ptychohyla leonhardschultzei, 
Oreolalax). There are all grades be- 
tween these extremes. In funnel-mouthed 
larvae the lateral ridge papillae are 
thick, stout projections lacking secon- 
dary papillation. 

Kenny (1969b) illustrated the lateral 
ridge papillae of Hyla geographica as 
being rather small flaps with jagged mar- 
gins. Savages (1952) illustration of 
these structures shows them to be of the 
rather huge “hand-like” variety in Bufo 
bufo. In general, large lateral ridge pa- 
pillae seem to be associated with the 
stream habitat. 


These papillae could have any or all 
of the functions suggested for oral pa- 
pillae elsewhere in the mouth—sensory, 
respiratory, or mechanical (as intercep- 
tors of feeding currents). Their position 
in front of the buccal pockets and their 
relatively large size in tadpoles that in- 
gest coarse material (e.g., stream forms; 
see Ecological Considerations) offer 
some support to a mechanical intercep- 
tor hypothesis. This does not rule out 
other functions. 


BuccaL Roor ARENA 


The buccal roof arena has already 
received attention under the heading of 
Buccal Floor Arena and only a few 
comments are added here. 

Anterior BRA papillae are positioned 
medially to the anterior BFA papillae; 
thus, buccal roof arenas tend to be more 
elongate and rectangular than buccal 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 115 


floor arenas. The number of BRA pa- 
pillae correlates directly with the num- 
ber of BFA papilla although there are 
never as many roof as floor papillae 
(maximum number observed was 34 in 
Ptychohyla_ leonhardschultzei). When 
the BFA is reduced, the BRA generally 
diminishes. In Colostethus nubicola the 
BRA papillae are reduced in height and 
coalesced into ridges. 

BRA papillae are, typically, simple 
cones. They are rarely bifurcated or dec- 
orated with terminal pustulations. 

Although all the authors mentioned 
above as providing illustrations of post- 
narial structures included some or all of 
the buccal roof proper in their drawings, 
their illustrations appear incomplete in 
regard to post-median ridge features. 
As to size, shape, number, and position 
of the accuracy of the representations of 
buccal roof arena papillae cannot be de- 
termined, because independent descrip- 
tions of these structures are not pro- 


vided. 


The possible functions for the BRA 
papillae are treated above with those of 
the BFA papillae. 


GLANDULAR ZONE AND DorsAL 
SECRETORY Pits 


Few generalizations can be made 
about the size and shape of the glandu- 
lar zone. Some tadpoles which clearly 
are not specialized for microphagy (e.g., 
Microhyla heymonsi, Anotheca) com- 
pletely lack a visible glandular zone in 
the stages that I examined. In other 
tadpoles the zone is interrupted medially 
(e.g., Rhinophrynus, Microhyla ornata, 
Leptobrachium oshanensis) or extreme- 
ly reduced in that region (e.g., Mego- 
phrys minor, Oreolalax, Leptobrachium 
hasselti, Hyla rufitela). The glandular 
zone was, on the average, proportion- 
ately wider in Hyla mixe, Hyla saraya- 
cuensis, and the megophrynine pelo- 
batids than in Hyla regilla. It was 
narrower in Ptychohyla, Smilisca, and 
Gastrotheca. The anterior margin of the 
zone was so indistinct in Hyla phlebodes, 


H. ebraccata, and Centrolenella that its 
length could not be determined. The 
posterior extension of secretory pits onto 
the dorsal velum and pressure pads is 
highly variable. 

Kenny (1969b) illustrates the glan- 
dular zone of Hyla geographica as di- 
vided medially. In Xenopus laevis and 
Pipa (Sokol, 1977a), the glandular zone 
is largely restricted to the individual 
pressure pads (Wassersug, unpublished 
scanning electron micrographs) but may 
extend far anterior to the pressure pads 
in Xenopus tropicalis (Sokol, pers. 
comm ). 

Whereas the shape and size of the 
total glandular zone fails to reveal pat- 
terns that can be easily associated with 
function, phylogeny, or ecology, the se- 
cretory pits themselves do reveal such 
patterns. Ascapthus truei has irregular- 
ly shaped, poorly defined pits that are 
much smaller than those of Hyla regilla. 
Bombina is similar to Ascaphus, but has 
well defined pits, at least along the an- 
terior margin of the glandular zone. 
Other tadpoles with smaller pits than 
those of Hyla regilla include Centrole- 
nella fleischmanni, Hyla phlebodes, H. 
mixe, H. ebraccata, and Ptychohyla leon- 
hardschultzei. None of these species is 
believed to be particularly micropha- 
gous. On the other hand, a list of tad- 
poles with secretory pits much larger 
than those of Hyla regilla include Hyla 
rufitela, Agalychnis callidryas, Gastro- 
theca riobambae, Rhinophrynus dorsalis, 
Microhyla berdmorei, Scaphiopus bom- 
bifrons, and Xenopus laevis. These are 
all tadpoles that, based on the totality 
of characters examined and their known 
behavior and ecology, should be as effi- 
cient as, if not more efficient than, Hyla 
regilla in microphagous suspension-feed- 
ing. Thus there seems to be a weak 
inverse relationship between the size of 
the secretory pits and the typical size of 
the particles upon which a larva feeds. 
There also seems to be a loose inverse 
correlation between the size of secretory 
pits and their density; the density of the 


116 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


secretory pits is reduced in forms with 
good microphagous abilities such as 
Rhinophrynus, Microhyla berdmorei, M. 
ornata, and Scaphiopus bombifrons, 
compared to more typical pond larvae. 


Feeding currents are directed against 
the glandular zone by the ventral velum, 
particularly when the velum is in a par- 
tially elevated position. Kenny (1969a) 
called this region the “dorsal food traps” 
and considered the glandular zone a pri- 
mary site for mucous entrapment of food 
particles. In the earlier discussion of 
secretory pits on the ventral velar pro- 
jections, I hypothesized a possible mech- 
anism for the generation of mucous 
strands from the glandular zone. Kenny 
has suggested that the ventral velum 
may actually push particles onto the 
zone during valvular closure. There can 
be little question that food is in one way 
or another trapped in mucus along the 
zone, for in some specimens flocculent 
matter can be found adhering directly 
to this surface. 

Whether the glandular zone is a pri- 
mary region for food entrapment is 
another question. In specimens where 
particles were adhering to the glandular 
zone, they were also adhering to the 
branchial food traps and were in far 
greater abundance on the latter surfaces. 
Massive accumulations of particulate 
matter on the glandular zone would pre- 
vent normal valvular closure. For the 
valve to be operational when there is 
substantial mucous entrapment on the 
glandular zone, mucus and food aggre- 
gates must be swept off the surface with 
efficiency and regularity. 

Hypothetically, there are other pos- 
sible functions for the mucous secretion 
of the glandular zone besides food en- 
trapment. Mucus may be important for 
proper valvular function of the ventral 
velum. Both lubricant (Gradwell, 1970) 
and sealant (Kratochwill, 1933) roles 
have been proposed for this secretion in 
relation to the valvular operation of the 
ventral velum. 


DorsAL VELUM 


Of the species I examined, the only 
one which completely lacks a dorsal 
velum is Anotheca spinosa. In a few 
others, the dorsal velum appears vesti- 
gial (e.g., Hyla phlebodes with its two, 
tiny, lateral flaps). In pipids, which lack 
a valvular ventral velum, the dorsal 
velum is also absent. 


In most species the dorsal velum is 
distinct and is divided into right and 
left halves. The few species in which 
the velum is continuous across the mid- 
line include Ascaphus truei, the micro- 
hylids, Centrolenella, Hyla rufitela, H. 
femoralis and tadpoles of the Hyla leu- 
cophyllata species group. In_ species 
where the velum is broadly interrupted 
medially, the free medial edges are often 
reflected posteriorly, leading into the 
esophagus. 


In Orton Type 3 tadpoles, the velum 
is shorter than in Hyla regilla. The vel- 
um is also distinctly shorter than that of 
H. regilla in Rhinophrynus, Centrole- 
nella, Agalychnis, Hyla ebraccata, Hyla 
sarayacuensis, Ptychohyla schmidtorum, 
Smilisca sordida, and the megophrynine 
pelobatids (with the possible exception 
of Leptobrachium hasselti). The only 
species in which the velum is unques- 
tionably longer than that of H. regilla 
are Microhyla berdmorei, Microhyla or- 
nata, and Scaphiopus bombifrons. 

Papillae occur along the free medial 
margin of the dorsal velum in Ptycho- 
hyla and the megophrynine pelobatids. 
Although they were not seen in Hyla 
mixe, these papillae may be common in 
suctorial, stream larvae of Orton Type 4. 
Alytes shows some papillation in this 
region of the velum. These papillae have 
not been discussed or illustrated previ- 
ously; presumably, they serve a sensory 
or mechanical straining function, but 
these are mere speculations. 

At present it is not possible to asso- 
ciate the size of the dorsal velum with 
other aspects of the biology of any group 
of tadpoles. 

There are two schools of thought 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES a, 


concerning the primary function of the 
dorsal velum. These schools are identi- 
fiable with the two names given this 
structure, viz., “dorsal velum” and “pos- 
terior filter valve.” Among the latter 
school, Kratochwill (1933) holds that 
the dorsal velum, in conjunction with the 
pressure cushions, functions as the pos- 
terior half of the valve system that seals 
off the filter chambers from the buccal 
cavities. Kenny (1969a) has essentially 
the same view, although he does not be- 
lieve that closure is complete. Since 
microhylids have a ventral velum par- 
tially anchored to tops of the filter plates 
(fully anchored at the top of cb. 4), 
Gradwell (1974) has argued that a val- 
vular function is essential for the dorsal 
velum in these larvae. Savage (1961), 
as a proponent of the “velum” school, 
proposes that the major function of this 
fold is to deflect currents down into the 
filter cavities in such a way that vortices 
form in the branchial baskets between 
the individual filter plates. DeJongh 
(1968), who also used the term “velum” 
sees this structure as having a protective 
function. In his view that when the 
ventral velum is in its down position, 
the dorsal velum is pulled taut and pre- 
vents water from flowing directly along 
the ciliary groove where it could displace 
mucous strands. Although it is obvious 
(and was recognized by Kratochwill) 
that the dorsal velum could act both as 
a valve and a velum, topographically at 
least, a velar function is obligatory for 
this structure. In accepting a velar func- 
tion, I am not accepting all the details 
of Savage's centrifugal feeding mechan- 
ism. Indeed, the objections that Kenny 
(1969a) raised to Savage’s centrifugal 
feeding theory seem completely valid. 


PRESSURE CUSHIONS 


Pressure cushions were regularly 
damaged in dissection and will not be 
discussed in detail. Overall, the cushions 
show a very strong correlation in size 
and shape with the size and shape of 
the branchial baskets. For example, 


when the branchial baskets are elongate, 
the pressure cushions are elongate (e.g., 
microhylids, Agalychnis, Rhinophrynus);: 
when the filter plates of the second and 
third ceratobranchials are reduced so 
that there is a single filter cavity on each 
side, there is also a single pressure cush- 
ion on each side (pelobatids). 

When the filter cavities are extremely 
large, as in Rhinophrynus, Xenopus 
(Sterba, 1950), and Pipa (Sokol, 1977a), 
there are three rather than two distinct 
cushions on each side. Sokol (1977a) 
reports another oddity in Pipa parva; 
this is the occurrence of papillae at the 
anteroventral corners of the huge pres- 
sure cushions of this genus. 

The function of the pressure cushions 
is not clear. Kratochwill (1933) as- 
sumed that they could abut tightly to 
the filter surfaces and actually push 
water through the filters on the expira- 
tion half of each pumping cycle. This 
seems doubtful, considering the delicate 
nature of these pads, although I concede 
that in living tadpoles fluid pressure may 
make these structures stiffer than they 
appear in preserved larvae. A stronger 
objection is that a single cushion rarely 
fills a whole filter cavity. 

The pressure cushions have also been 
thought to serve as a protective rim for 
the ciliary groove. Although the poste- 
rior margin of the cushions makes up 
the anterior margin of the ciliary groove, 
far more tissue is involved in the cush- 
ions than is necessary to demarcate the 
groove. Thus, I doubt that this is the 
sole function of the cushions. 


Although I cannot offer a completely 
satisfying hypothesis for the function of 
the cushions, I suspect from the com- 
mon occurrence of scattered secretory 
pits on these surfaces that they are in- 
volved in direct contact with food mat- 
ter. A possible role in lifting food from 
the filters was mentioned above. 


SYSTEMATIC CONSIDERATIONS 


“Tadpole morphology represents per- 
haps one of the most useful and most 


118 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


misused of the available character com- 
plexes to be used in frog classification.” 
(Lynch, 1971:200). 

An effort is made here to discern 
shared, derived characters among inter- 
nal oral features that either substantiate 
or refute traditional taxonomic schemes. 
No formal effort is made to reassign any 
taxa, since such activities cannot be 
justified solely on the basis of larval 
characters. Species are discussed by 
family in the same order in which they 
are presented in the Descriptions. 

Results from this study generally sup- 
port traditional systematic assignments 
at the familial and lower taxonomic 
level. Suprafamilial relationships are 
discussed separately in the section titled 
Phylogenetic Considerations. 


ASCAPHIDAE 


Ascaphus larvae have many features 
which distinguish them from all other 
tadpoles, including the direct-develop- 
ing Leiopelma embryos. Although I 
have not personally examined Leiopelma 
ontogenetic series from the genus have 
been described by Stephenson (1950a, 
1950b, 1955). In comparing Stephen- 
son’s figures and description with my 
figures and description of Ascaphus I 
can find no common, clearly derived fea- 
tures in these genera. In the absence of 
shared, derived characters it may be con- 
cluded that early developmental stages 
do not provide any evidence for group- 
ing Ascaphus and Leiopelma in the 
same family. I support Savage (1973) 
and Goin et al. (1978) in recogniz- 
ing two families, the Ascaphidae and 
the Leiopelmatidae (contra Duellman, 
1975). 


DISCOGLOSSIDAE 


Lanza, Cei and Crespo (1975, 1976) 
have argued on the basis of immunolog- 
ical data for partitioning the Discoglossi- 
dae into the Bombinatidae (with Bom- 
bina) and the Discoglossidae (with Dis- 
coglossus and Alytes). Except in some 
fine points of lingual papillae pattern, 


the three genera have very similar lar- 
val oral structures. 

Certain of their shared features, such 
as expanded internal narial flaps and 
the pharyngeal shape, warrant including 
these genera in the same family on 
strictly phenetic grounds. I consider it 
unlikely that these similarities are due 
to convergence. The features in ques- 
tion may be primitive ones for anurans 
and of little use in discerning sister 
groups. If much weight, however, is 
given immunological data, we may find 
grounds for splitting the Archaeoba- 
trachia as currently recognized into sev- 
eral additional families. 


RHINOPHRYNIDAE 


Although Rhinophrynus dorsalis \ar- 
vae share certain features of other fami- . 
lies, such as an exposed glottis on the 
floor of the mouth (a la Microhylidae) 
and the full attachment of the ventral 
velum to the dorsal margins of the sec- 
ond and third filter plates (a la Disco- 
glossidae), they are overall unlike any 
other tadpoles examined and completely 
justify the monotypic assignment of this 
species. 


MICROHYLIDAE 


At the familial level it is clear that 
the microhylids are tightly united and, 
at the same time, separate from all other 
frog families by the uniqueness of their 
larvae (Orton, 1957). Although Savage 
(1952, 1955), Nelson and Cuellar 
(1968), and Gradwell (1974) have all 
described larval oral features in a variety 
of microhylid genera, their descriptions 
are neither detailed enough nor com- 
parative enough to help determine 
whether internal oral features can be 
useful in discerning relationships within 
this family. Of the three species of 
Microhyla I examined, M. ornata and 
M. berdmorei are extremely similar. Con- 
sidering the large number of superficially 
similar species in the genus (Parker, 
1934) it is doubtful that the minor oral 
features used to distinguish these two 
species would separate either of them 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES LIL) 


from all other species in the genus. The 
funnel-mouthed form, M. heymonsi, 
however, is immediately separable from 
the rest of the genus (with the possible 
exception of the other known funnel- 
mouthed species, M. achatina) in both 
external and internal features. It is 
doubtful that any other free-swimming 
microhylid larvae are much like M. hey- 
monsi in oral features; M. achatina and 
M. heymonsi may be justifiably removed 
from the genus Microhyla, if any weight 
is given to their larval characters. 


PELOBATIDAE 


The shallow filter plates of cb. 2 and 
cb. 3 and the bowl shape of the bran- 
chial baskets are unique features shared 
by all pelobatid larvae examined and 
the family is easily justified on these 
features. Three major subfamilial groups 
can be seen in the material at hand, and 
the European pelobatids may form a 
fourth division. Megophrys stands apart 
from the other pelobatids because of 
adaptations associated with the funnel- 
mouth. It forms the first division. The 
second division consists of the nonfun- 
nel-mouthed megophrynines represented 
here by Leptobrachium and Oreolalax. 
Tadpoles of these genera are sufficiently 
similar in larval morphology to justify 
their separation from Megophrys (In- 
ger, 1966). The third division consists 
of Scaphiopus, which is grossly different 
from the other pelobatids. Many of 
these differences may relate to the fact 
that these are temporary pond dwellers 
rather than inhabitants of streams. 


HYLIDAE 


Although the hylid larvae examined 
were few compared to the number of 
species in this family, the sample was 
originally selected because of its ecologi- 
cal diversity, and consequently demon- 
strated much correlated morphological 
diversity (see Functional Considerations 
and Ecolog’cal Considerations). Given 
this great morphological diversity, few 
unique features emerge which might 


define the family. Whether there are 
any larval features unique to the Hyli- 
dae cannot be determined until more 
tadpoles of other advanced families are 
studied. At present it seems that all 
hylids have a basically similar pattern 
of attachment for the ventral velum. 
Moreover, all have perforated, slit-like 
or oval internal naries. But comparison 
with descriptions in the literature for 
other families, such as the Ranidae and 
Bufonidae, and my own preliminary ob- 
servation of leptodactylid larvae, sug- 
gest that these are not unique features; 
alone they will not distinguish the hylids 
from most other neobatrachian families. 

Recently there has been much work 
on the subfamilial systematics of the 
Hylidae. Duellman (1970) associated 
the genera Anotheca and Gastrotheca 
in the subfamily Amphignathodontinae, 
while Maxson (1977) has argued from 
immunological data that Anotheca more 
appropriately belongs in the hylinine 
subfamily. Both Anotheca spinosa and 
Gastrotheca riobambae share a generally 
robust habitus and have the medial por- 
tion of the ventral velum deflected by a 
large laryngeal disc. The large laryngeal 
disc in both species probably results 
from independent convergence on early 
lung use. The larva of Anotheca spinosa 
is otherwise so bizarre that oral struc- 
tures are of little use in establishing its 
subfamilial status. 

Maxson (1976) has also used immu- 
nological data to argue for elevation of 
the Phyllomedusinae to familial status. 
The one phyllomedusine larva that I 
examined, Agalychnis callidryas, differs 
from all the other hylids in many fea- 
tures (e.g., the shape of the prepocket 
buccal surface, buccal floor arena, velar 
margin, and filter apparatus), but is vir- 
tually identical to Phyllomedusa trini- 
tatis as illustrated by Kenny (1969a). 
Considering the extreme _ similarity 
among phyllomedusine tadpoles in gross 
external morphology (for Central Amer- 
ican species, see Duellman, 1970), be- 
havior, and ecology (Starrett, 1973; 


120 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Wassersug, 1973), we may safely con- 
clude that internal oral features of A. 
callidryas are typical for the subfamily. 
Phyllomedusine tadpoles as a group 
seem to be more specialized for suspen- 
sion feeding (contra Kenny, 1969a) than 
other hylid larvae. Whether this war- 
rants their separation from the Hylidae 
at the family level cannot be determined 
at this time. 

At least three species of Phyllome- 
dusa, P. guttata, P. cochranae and P. 
jandaia are funnel-mouthed forms (Bo- 
kermann and Sazima, 1978) with a par- 
ticular external morphology regularly 
associated with a complex internal mor- 
phology unlike that seen in P. trinitatis 
or A. callidryas. If further examination 
of funnel-mouthed phyllomedusines re- 
veals the typical internal specializations 
of other funnel-mouthed larvae, this may 
warrant placing them in their own genus. 
If, however, they show the specializa- 
tions for microphagy characteristic of 
other phyllomedusine larvae, then, col- 
lectively, the uniformity and uniqueness 
of the phyllomedusine larvae may sup- 
port the elevation of the subfamily to 
family status. 

Distinguishing features of the larvae 
of the four genera of Hylinae frogs, 
Ptychohyla, Hyla, Smilisca, and Acris, 
are all of a quantitative rather than qual- 
itative nature. There is as much differ- 
ence between species within the genera 
Ptychohyla and Hyla as there is between 
members of all four genera. Hyla re- 
gilla, Acris crepitans, and Smilisca sor- 
dida are all very similar and show the 
generalized “pond” tadpole pattern for 
internal oral structures. 

Ptychohyla schmidtorum and P. leon- 
hardschultzei differ from one another at 
least as much internally as externally, 
completely supporting the separation of 
these forms into the two species groups, 
P. schmidtorum and P. euthysanota, re- 
spectively (Duellman, 1970). 

All species groups within Hyla rec- 
ognized by Duellman (1970) appear 
valid according to the features I have 
examined. A case can be made, how- 


ever, for elevating the rank of some of 
these groups (see below). 

H. rufitela is the northernmost mem- 
ber of the Hyla albomarginata species 
group, and has a larva externally like 
typical pond Hyla such as H. regilla (H. 
eximia species group) except for its 2/4 
denticle pattern. But internally, certain 
features such as the narial vacuities and 
the transverse row of papillae on the 
buccal floor, justify supraspecific sepa- 
ration of this Hyla from the other Hyla 
in this study. Unfortunately, not enough 
is known about other members of the 
albomarginata species group to say 
whether these features are diagnostic of 
the group in general. Hyla dendroscarta 
(H. bromeliacia group) has a variety of 
oral features which again would sepa- 
rate it from more typical pond Hyla lar- - 
vae. Likewise, the similarity between 
Hyla ebraccata and Hyla sarayacuensis 
justifies their separation from the other 
forms. H. ebraccata and H. sarayacuen- 
sis form a closely united group readily 
distinguishable in their oral anatomy 
from all other tadpoles examined; simi- 
larities in their larval morphologies com- 
pletely justify their association in the 
Hyla leucophyllata species group. Equal- 
ly divergent from the basic pond hyli- 
nine type are H. mixe (H. mixomaculata 
group) and H. phlebodes (Hyla micro- 
cephala species group). In the shape of 
its branchial baskets and gill filter den- 
sity, Hyla femoralis likewise stands apart 
from the other Hyla examined, including 
the Holarctic H. regilla. Maxson and 
Wilson (1975, Table 4) reach the same 
opinion based on albumin similarities. 

If one general conclusion can be 
reached from the diversity of larval oral 
structures in Hyla, it is that the genus 
is an artificial group that warrants frag- 
mentation into several genera. There is 
no question that, if adult frogs showed 
as much diversity in oral structures as 
these tadpoles do, herpetologists would 
have partitioned them into separate gen- 
era decades ago. 

On the other hand, there are no 
unique features of Smilisca or Acris tad- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 121 


poles that would help to define these 
genera as distinct from Hyla. Maxson 
and Wilson (1975) considered Acris as 
very different from Hyla at the organis- 
mal level (based on karyotypic and 
adult osteological characters; cf. Duell- 
man, 1970 and Gaudin, 1974) but wrote, 
“A protein-based classification would in- 
clude [Acris] in the same taxonomic 
category as the North American Hyla 
species but exclude many South Ameri- 
can species currently placed in the genus 
Hyla.” In this conflict of organismal 
versus molecular evolution the tadpoles 
appear to support the systematic ar- 
rangement derived from protein struc- 
ture rather than the one based on adult 
characters. According to Maxson and 
Wilson’s data for two Holarctic Hyla 
that I examined, Hyla regilla is more 
similar to Acris than it is to Hyla femo- 
ralis. Again, this similarity is borne out 
by tadpole oral characters. In terms of 
both molecular structure and larval anat- 
omy, Acris is more similar to certain spe- 
cies of Hyla than are many species of 
Hyla to each other. The oral morphol- 
ogy of hylinine tadpoles provides the 
first anatomical support for systematic 
patterns formulated from molecular 
data. For this reason tadpole structures 
deserve consideration in future work on 
anuran taxonomy. 


DENDROBATIDAE AND CENTROLENIDAE 


Of the two families remaining, there 
are no internal oral features that readily 
distinguish the dendrobatid larvae from 
other advanced families. The differences 
between Colostethus subpunctatus and 
Colostethus nubicola are massive and, 
again, if internal oral features of the 
larvae are taken into consideration, the 
funnel-mouthed form could justifiably 
be separated into a different genus. 

Features that are unique to Centro- 
lenella fleischmanni are virtually all as- 
sociated with the extreme elongation of 
the tadpole (e.g., elongated internal 
naries, elongation of the second and 
third ceratobranchials). | Considering 


that elongation is a general character- 
istic of centrolenid larvae, these internal 
features probably characterize the fam- 
ily and serve to distinguish these larvae 
from tadpoles of other families. 


EVOLUTIONARY 
CONSIDERATIONS 


THE Ascaphus anv Leiopelma 
Lire CYCLEs 


Despite their extreme specialization 
for life in torrential streams, Ascaphus 
larvae have all the essential, anatomical 
features for suspension-feeding (Was- 
sersug, 1975). They have, for example, 
well developed gill filters. Although the 
filter mesh in Ascaphus is not nearly as 
dense as in typical pond forms, the fil- 
ters are more fully developed than in 
certain other stream forms that clearly 
had pond-inhabiting ancestors (e.g., 
Ptychohyla). Ascaphus larvae have ex- 
panded ceratohyals that can serve as a 
buccal pump and a ventral velum that 
can act as a valve to maintain one-way 
flow to the branchial food traps. Al- 
though they lack secretory ridges in the 
pharynx they have otherwise well-devel- 
oped branchial food traps (Wassersug 
and Rosenberg, 1979). These larvae have 
functional narial valves. The filters, buc- 
cal pump and valves are all essential 
features of the tadpole suspension-feed- 
ing mechanism (Wassersug, 1975). 

Experimental work has shown that 
Ascaphus tadpoles can extract food par- 
ticles from suspension (Altig and Bro- 
die, 1972), although their efficiency (in 
terms of the particle sizes they capture ) 
is below that of more typical pond 
forms (Wassersug, 1972). As noted by 
Gradwell (1973) and Starrett (1973), 
there is no support for the popularly ac- 
cepted (cf. Morton, 1967:42) suggestion 
by Noble (1927) that these larvae take 
in most of their food through their 
naries. 

Some authors (Eaton, 1959; Schmal- 
hausen, 1968) have considered the 
stream tadpoles of Ascaphus to repre- 


122 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


sent the generalized anuran larva. This 
view is based on an effort to explain the 
hard mouth parts of tadpoles as a gen- 
eralized feature initially evolved as an 
aid for hanging onto rocks in streams. 
Such a view is founded on the belief 
that the most generalized extant frog 
must have the most primitive tadpole. 
Hard mouth parts in most other tad- 
poles are primarily used to generate a 
suspension of food particles, and it is 
equally plausible that they did not ini- 
tially evolve as a response to life in 
streams. As with other organisms, an 
anuran may be a mosaic of primitive 
and derived characters; there is no real 
reason why a frog and its larva must be 
at the same grade of specialization. It 
seems most likely that Ascaphus tad- 
poles are secondarily specialized for 
stream life. 


Developmental series of Leiopelma 
have not been available for study. 
Stephenson (1950a, 1950b), who has ex- 
amined ontogenetic material of this 
genus, believed that the direct-develop- 
ing embryo of Leiopelma represented a 
link between the salamander and frog 
type larva. Stephenson (1955) wrote 
that, “no evidence exists of terrestrial 
specializations as shown by certain anu- 
rans which have secondarily acquired a 
terrestrial mode of development follow- 
ing abandonment of a_ specialized 
aquatic tadpole’s stage.” In direct con- 
trast to that statement, it is generally 
accepted by most workers that Leio- 
pelma secondarily acquired direct de- 
velopment from a free-swimming tad- 
pole stage. 

From my analysis of Stephenson's 
(1950a) reconstruction of the embryonic 
chondrocranium and visceral arches of 
Leiopelma larvae, it is fair to say that 
this genus shows a substantial number 
of vestigial features, ones specifically as- 
sociated with an aquatic, suspension- 
feeding existence. The jaw suspension 
in developing Leiopelma is directed an- 
teroventrally and only secondarily ro- 
tates at metamorphosis into a more 


vertical position. The midportion of the 
ceratohyal is expanded in the frontal 
plane as it is in all aquatic larvae with 
a buccal pump. Long ceratohyals asso- 
ciated with the filter plates of suspen- 
sion-feeding forms line the otherwise 
nonfunctional gill slits. In fact, what 
Stephenson identified as a_ branchial 
pouch in his serial sections (1950a, Plate 
1), may be a vestige of the ciliary groove 
used to transport particles trapped in 
mucus to the esophagus in suspension- 
feeding tadpoles. In contrast, Eleuthero- 
dactylus, a direct-developing frog that 
no one doubts evolved from an aquatic 
ancestor, shows far fewer features asso- 
ciated with feeding in its jaw suspension 
and branchial skeleton than Leiopelma 
(See Lynn, 1942). Evidently, Leiopelma 
arrived at terrestrial development from - 
an ancestor with an aquatic, suspension- 
feeding, larva. 


EVOLUTIONARY TRENDS IN THE 
DISCOGLOSSIDAE 


A case can be made for grouping 
Ascaphus closely with the discoglossids. 
These forms are similar not only in the 
way the ventral velum attaches to the 
filter plates, but also in the gross shape 
of the buccal floor arenas, the position 
of the glottis, and the absence of a well- 
developed ridge pattern for the secre- 
tory tissue of the branchial food traps 
(Wassersug and Rosenberg, 1979). 

The multiple pustulations on the 
tongue anlage in Alytes are similar to 
the multiple papillation of the tongue 
anlage in Ascaphus. In features of the 
pharynx, it is Bombina among the disco- 
glossids which is most similar to As- 
caphus. Typical advanced tadpoles have 
well organized ridges running trans- 
versely across the branchial food traps, 
and pits along the posterior margin. 
Ascaphus and Bombina lack both the 
secretory pits and ridges that character- 
ize the branchial food traps of most 
“advanced” anuran families (e.g., hylid, 
bufonid, ranid). Alytes and Discoglos- 
sus, in contrast, have secretory cells clus- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 123 


tered to form ill-defined pits. The micro- 
anatomy of the food traps suggests a 
morphological sequence going from As- 
caphus to Bombina to Alytes and Disco- 
glossus and then to more advanced 
anuran larvae. Phylogenetically, the se- 
cretory pits evidently appeared before 
secretory ridges (Wassersug and Rosen- 
berg, 1979). 

These ideas neither support nor con- 
flict with the proposal of Lanza, Cei and 
Crespo (1975, 1976) to partition the 
Discoglossidae (sensu lato) into two 
families. Sokol (1977b) recognizes the 
suborder Discoglossoidei, to encompass 
the two families Leiopelmatidae (= As- 
caphus and Leiopelma) and Discoglossi- 
dae; the similarity of Ascaphus tadpoles 
to discoglossid larvae makes this an emi- 
nently reasonable association. 


THE PrIeoIDEA AND THEIR RELATIONSHIPS 


Since Orton (1957) first proposed 
that Rhinophrynus and the Pipidae were 
closely related, virtually all herpetolo- 
gists (except for Griffiths and de Car- 
vallo, 1965) have formally or informally 
recognized the superfamily Pipoidea. It 
is thus surprising how little resemblance 
exists between oral structures in Rhi- 
nophrynus tadpoles and pipid larvae. 
The shape of the prepocket buccal sur- 
face, the presence of a valvular ventral 
velum, the shape of the branchial food 
traps, the shape of the dorsal velum 
and pressure cushions, all distinguish 
Rhinophrynus larvae from any pipid 
genus. This does not necessarily mean 
that Rhinophrynus is not closely related 
to the pipids at the superfamilial level 
or that Orton’s Type 1 larva is invalid. 
Starrett (1968, 1973) and Sokol (1975, 
1977a) have both reaffirmed the validity 
of the Rhinophrynus-pipid association. 
Hymenochirus, Xenopus and Pipa larvae 
all differ so much from each other that, 
provided they are validly grouped at 
the familial level, Rhinophrynus could 
still be reasonably united with them at 
the next higher level. 

Certain features of Rhinophrynus are 


similar to the discoglossids; most con- 
spicuous is the full attachment of the 
ventral velum to filter plates cb. 2 and 
3. Rhinophrynus, however, has ad- 
vanced upon the discoglossid grade in 
having developed secretory ridges in the 
branchial food traps (Wassersug and 
Rosenberg, 1979). In several super- 
ficial features, Rhinophrynus closely re- 
sembles the microhylids. These include: 
buccal floor exposure for the glottis; 
absence of posterior projections from the 
velar margin; branchial food traps re- 
stricted to the anterior portion of the 
filter cavities (crescentic organs); larger 
branchial baskets with a denser filter 
mesh; and the absence of postnarial and 
buccal roof papillae. Except for the 
buccal floor exposure of the glottis, the 
restricted size of the branchial food traps 
and the density of the filter mesh, this 
combination of character states is not 
unique to Rhinophrynus and the micro- 
hylids; it occurs, for instance, in certain 
hylids. Although the glottis is exposed 
in Rhinophrynus and microhylids, its 
position is much farther forward in the 
microhylids than in Rhinophrynus. In 
other features, such as the overall shape 
of the visceral skeleton, the resemblance 
breaks down. Presumably, similarities 
between Rhinophrynus and microhylids 
are due to convergence on a largely sus- 
pension-feeding, nectonic way of life. 
If we accept Rhinophrynus as a 
pipoid, then it follows that Rhinophrynus 
has, overall, the most generalized Orton 
Type 1 larva. This view is based on the 
belief that the absence of a valvular 
velum in the Pipidae is best compre- 
hended as a derived character. There is 
little support for the opposite hypothesis 
offered by Gradwell (1975a), which 
would derive virtually all anurans from 
the genus Xenopus. The unusual posi- 
tion and microanatomy of the secretory 
ridges in Xenopus (Wassersug and Ro- 
senberg, 1979), the papillate pressure 
cushions and fused filter rows in Pipa, 
and the complete loss of filter-feeding 
structures in Hymenochirus, must all be 


124 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


highly derived characters. Sokol (1977a, 
1977b) has reached a basically similar 
conclusion with an independent analysis 
of these and other characters. 

Within the Pipidae, the genera 
Xenopus and Pipa bear the strongest 
resemblance in larval oral surface struc- 
tures (Sokol, 1977a). On the basis of 
other characters, however, Sokol places 
Hymenochirus closer to Xenopus. Ac- 
cording to Sokol the filter apparatuses of 
Xenopus and Pipa are extremely differ- 
ent, although one must concede that 
they are more similar to each other than 
either is to Hymenochirus. Sokol views 
Pipa as evolving away from the extreme 
microphagous condition, as exemplified 
by Xenopus. Pipa appears to be a die- 
tary specialist adapted for feeding on 
medium-sized planktonic organisms. Hy- 
menochirus, of course, is a carnivore, 
specialized for feeding on large plank- 
tonic organisms. 

Except for discoglossoid frogs, all 
anurans including the Pipoidea have a 
fused trigeminal and facial ganglia. So- 
kol views this as a uniquely derived 
character and Lynch (as pers. comm. in 
Sokol, 1977b) has gone so far as to sug- 
gest that the Pipoidea are derived from 
pelobatoid frogs. 


It is my suspicion that the fusion of 
the fifth and seventh ganglia is an in- 
direct result of the orientation of the 
palatoquadrate bar in non-discoglossoid 
larvae and is associated with perfection 
of a suspension-feeding larva with a 
large, efficient buccal pump. Investiga- 
tors as far back as Luther (1914; see 
Edgeworth, 1930; Wassersug, 1975) 
have explained the unusual chondrocra- 
nium of tadpoles as an adaptation asso- 
ciated with their feeding mechanism. 
Elongation «nd anterior extension of the 
palatoquadrate has allowed for expan- 
sion of the ceratohyal in the horizontal 
plane to form the piston of the buccal 
pump (Severtzov, 1969). The anterior 
displacement of the palatoquadrate, 
however, limits the area for the emer- 
gence of cranial nerves from the brain 


case in the region between the optic cup 
and the otic capsule. We may hypothe- 
size that evolution has “squeezed” the 
ganglia of the fifth and seventh nerves 
together incidental to the forward dis- 
placement of the palatoquadrate and ex- 
pansion of the ceratohyal. The palato- 
quadrate of the pipids is unlike that of 
the other anuran larvae, but it neverthe- 
less has a relatively anterior position 
compared to discoglossoid larvae. Con- 
sidering the differences in the form of 
the palatoquadrate in pipoid frogs and 
non-pipoid, non-discoglossoid larvae, it 
seems possible that an efficient cerato- 
hyal pump has evolved along two paths 
above the discoglossoid grade. One path 
would have been through the Pipoidea 
and the other through all remaining 
anurans. It seems plausible that the - 
trigeminal and facial nerves could have 
been independently “squeezed” together 
on both evolutionary courses. 

In conclusion, the Pipoidea are an 
extremely specialized and diverse super- 
familial group. Internal oral features 
corroborate Sokol’s conclusion that the 
Pipoidea are too specialized to be either 
an ancestral anuran group, or near the 
ancestral line (contra Starrett, 1973). 
However, in identifying the pipoid lar- 
vae as highly specialized and derived, I 
consider it premature to propose that 
they evolved from any extant family or 
superfamilial group. 


THe MiIcrRoHYLID PROBLEM 


Perhaps the greatest controversy in 
the higher taxonomy of the Anura cen- 
ters on the position of the frogs with the 
Orton Type 2 larvae, the Microhylidae. 
Many authors have commented on this 
problem, most recently Starrett (1973), 
Savage (1973), Sokol (1975), Duellman 
(1975), Lynch (1973), and Blommers- 
Schlosser (1975). Based on the char- 
acters used by Orton and Starrett, the 
microhylids bear the closest resemblance 
to the pipids and Rhinophrynus. Based 
on characters of the adults and addi- 
tional larval features discussed by Sokol, 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 125 


the microhylids are presumed to be 
ranoid-derived frogs. 

In the structure of their branchial 
food traps and visceral skeleton, micro- 
hylid larvae are neither clearly pipoid 
nor obviously ranoid. The buccal floor 
exposure of the glottis in the microhylids 
is most like that of the pipids, but the 
attachment of the ventral velum is more 
similar to that of advanced Type 4 lar- 
vae. My study suggests that the micro- 
hylid larvae are profoundly different 
from both Orton Type 1 and Orton Type 
4 larvae. 

Of the microhylids examined (see 
also Savage, 1952), Microhyla heymonsi 
looks the most like an Orton Type 4 
larva in internal oral features. Specifi- 
cally, it has a relatively long buccal floor 
anterior to the buccal pockets, small 
branchial baskets (compared to other 
microhylids), and a ventral velum vir- 
tually continuous across the midline. 
There is little question that these re- 
semblances are due to secondary con- 
vergence as the ancestor of M. heymonsi 
abandoned total reliance on microscopic 
food. The presence, however, of a con- 
tinuous or at least nearly continuous 
ventral velum in this species raises the 
possibility that microhylids could have 
evolved from forms with a continuous 
ventral velum. This excludes the Pipidae 
(but not Rhinophrynus). 

The ancestor of the Microhylidae 
need not be from an extant family. If 
microhylid larvae evolved from tadpoles 
with keratinized beaks, the presence of 
secretory ridges in the branchial food 
traps points to an origin above the dis- 
coglossoid/pelobatoid grade. Blommers- 
Schlosser (1975) recently described lar- 
vae of Pseudohemisus granulosus, a 
scaphiophrynine frog from Malagasy; 
these tadpoles exhibit a mosaic of micro- 
hylid/ranid features in their external 
anatomy. As more exotic microhylid lar- 
vae are described, I suspect that they 
will support a microhylid-ranoid rela- 
tionship, and resemblances between mi- 
crohylid larvae and pipoid forms will 
prove to be convergences. 


THE ORIGINS OF THE “ADVANCED” 
ANURANS (TYPE 4 LARVAE) 


Although it is generally recognized 
that pelobatoid frogs (families Pelobati- 
dae and Pelodytidae) represent the least 
specialized families with Orton Type 4 
larva (Lynch, 1973), the relationship of 
these frogs to more advanced families is 
not fully understood. Duellman (1975) 
placed the pelobatoid frogs in his sub- 
order Archaeobatrachia, suggesting that 
they are closer to the “ancestral” frogs 
than to the “advanced” anurans of the 
order Neobatrachia. Pelobatids lack se- 
cretory ridges in the branchial food traps 
and in that feature are most similar to 
the archeobatrachian Ascaphus and dis- 


coglossids (Wassersug and Rosenberg, 
1979). 


The most detailed discussion to date 
on the origin of Type 4 larvae is in 
Heyer (1975:40-43). Savage, as a pers. 
comm. in that paper, suggests that lep- 
todactylids arose directly from a leio- 
pelmatid ancestor. Heyer (who evident- 
ly accepts the Starrett-Savage tenet that 
the pipoid larvae represent the ancestral 
larval form) correctly points out that 
one implication of Savage’s suggestion is 
that the Type 4 larva must have evolved 
twice from a beakless ancestor. Extant 
archeobatrachian species with their re- 
lict distributions and exotic larvae pro- 
vide little information that can help 
evaluate Savage's suggestion. To quote 
Heyer (1975:42), “The crux of the argu- 
ment hinges, then, on whether the pelo- 
batid acosmanuran tadpole is really the 
same as a leptodactyloid (bufonid and 
ranoid) acosmanuran tadpole.” Heyer 
argues that given a limited number of 
ways that a tadpole can be efficient at 
scraping, one would not expect great 
differences in morphology between pelo- 
batid tadpoles and leptodactyloid larvae, 
even if they are diphyletic. 

I am not so convinced that the Pi- 
poidea are the stem anuran group (see 
above), and, if they are not, this alters 
the problem. Nevertheless, assuming 
that the pelobatids and leptodactyloids 


126 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


arose from families with beaked larvae, 
the question of the phylogenetic rela- 
tionship of the leptodactyloids to archeo- 
batrachian frogs remains. 

Lynch (1973) has made a case for 
the evolution of primitive leptodactylids 
directly from megophrynine pelobatids. 
I have not examined internal features 
of any leptodactylid tadpoles in detail 
but these tadpoles seem superficially 
similar to other neobatrachian, Type 4 
larvae. The bow-shaped branchial bas- 
kets and abbreviated attachment of the 
ventral velum to the filter plates of the 
second and third ceratobranchials ap- 
pear to be uniquely derived character- 
istics of the pelobatids and may pre- 
clude deriving the pattern of velum 
attachment in other Type 4 larvae from 
this family. I feel that it is most plausi- 
ble that the pelobatid velar structure 
evolved from the common Type 4 pat- 
tern. If Australian myobatrachine larvae 
have pharyngeal morphology like the 
megophrynine pelobatids I would con- 
sider this strong evidence for the phylog- 
eny proposed by Lynch. Until more is 
known about primitive leptodactyloid 
(sensu lato) larvae, it is best to view 
the pelobatids as a sister group rather 
than the ancestors of more advanced 
frogs. 


EVOLUTION OF SPECIFIC LARVAL TYPES 
IN THE GeNus Hyla 


Because of the large and morphologi- 
cally diverse sample of Hyla larvae exam- 
ined it is both possible and appropriate 
to comment on the adaptive radiation of 
larval types in this genus. The evolution 
of two types are singled out for discus- 
sion, the macrophagous herbivorous 
larva and the arboreal larva. These types 
help demonstrate how knowledge of lar- 
val morphology can give some insight 
into evolutionary history below the fa- 
milial level. 

1) Macrophagous, Herbivorous Hyla 
larvae —Members of the Hyla leucophyl- 


lata species group, Hyla ebraccata and 
Hyla sarayacuensis, exhibit larval spe- 
cializations away from ultra-planktonic 
suspension feeding and toward macroph- 
agy. They have reduced or lost the pa- 
pillae of the buccal floor and roof. They 
have a low filter mesh density and ab- 
breviated branchial food traps. H. sara- 
yacuensis has a lower number of filter 
rows and more reduction of the secretory 
ridges in the branchial food traps than 
H. ebraccata and appears to be the more 
macrophagous of the two forms. 

Of all the hylinine tadpoles studied, 
Hyla phlebodes ( Hyla microcephala spe- 
cies group) larvae have the most ex- 
treme reduction in structures associated 
with the ingestion of fine, suspended 
matter. 

If one looks at the general pattern of © 
reduction, a morphological sequence can 
be drawn from Hyla regilla as a typical 
Hyla larva to the leucophyllata group 
and then to the Hyla microcephala group 
as represented by Hyla phlebodes. This 
morphological sequence represents a 
trend toward a particular type of ma- 
crophagy dealing with coarse plant mat- 
ter and detritus rather than large, live 
prey such as eaten by carnivorous tad- 
poles. It is through a similar morpho- 
logical series that the ancestors of the 
Hyla microcephala group might have 
evolved, although I do not mean to im- 
ply that the Hyla microcephala group 
evolved directly from the Hyla leuco- 
phyllata group. 

2) Arboreal Hyla Tadpoles—The 
internal anatomy of Hyla dendroscarta 
larvae sheds some light on the question 
of how arboreal forms of the Hyla 
bromeliacia group may have evolved. 
Duellman (1970:681) stated that the 
“sroup containing bromeliacia in north- 
ern Central America and dendroscarta 
in southeastern Mexico, apparently di- 
verged from the lowland pond-breeding 
picta-stock by adapting to arboreal 
breeding habits in a successful attempt 


1Zynch (as cited in Sokol, 1977a) suggested that the pipoids could also be derived from 
the pelobatoids, but I consider this unlikely for the reasons stated above. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 127 


to invade the foothills in low mountains 
where ponds are scarce.” Members of 
the H. picta group, however, are quite 
different in larval biology from H. den- 
droscarta and H. bromeliacia. H. picta 
larvae have a tall tail fin, moderately 
short pointed tail, and occur in grassy 
pools. Tadpoles of the H. bromeliacia 
group have the long tails of stream 
adapted larvae. Internally, H. dendro- 
scarta larvae have a mosaic of features 
that are associated with either pond or 
stream existence. The lobe-like infra- 
labial papillae and absence of velar mar- 
ginal projections are characteristic of 
certain pond tadpoles. On the other 
hand, the many tall buccal floor and 
buccal roof papillae and the truncation 
of the third filter cavity are features that 
appear in stream larvae. In overall in- 
ternal and external proportions, H. den- 
droscarta tadpoles are more like stream 
than pond larvae. 


Duellman (1970:429) noted that 
members of the H. bromeliacia group 
have cranial characters similar to H. 
miotympanum but considers any alliance 
between these two species tenuous. In 
fact, members of the H. miotympanum 
group have stream larvae that look very 
much like tadpoles of the H. bromeliacia 
group. Hyla arborescandens, the sister 
species of H. miotympanum, even has 
larvae with the unusual 2/4 denticle pat- 
tern that characterizes the middle Amer- 
ican arboreal Hyla larvae. A species of 
Hyla closely related to the H. miotym- 
panum group has a breeding behavior 
that could represent a link between the 
stream breeding of H. miotympanum 
and the arboreal breeding of H. dendro- 
scarta and H. bromeliacia. This is Hyla 
thorectes, which, according to Duellman 
(1970, p. 391), has stream larvae but “is 
unique among members of the genus in 
northern middle America by depositing 
its eggs on vegetation above the 
streams.” It seems most plausible that 
the ancestor of the present bromeliad 
breeding Hyla went through just such an 
evolutionary stage, first leaving eggs 


above water, that dropped into the wa- 
ter after hatching, then later developing 
tadpoles that could survive on the vege- 
tation. This is a more believable evolu- 
tionary sequence than the quantum jump 
from pond to tree that is required if the 
Hyla bromeliacia group evolved from a 
Hyla picta-like stock. 

Arboreal larvae in other genera and 
families, particularly those with elongate 
larvae, could have followed a similar 
evolutionary history. Centrolenid larvae 
presumably represent an intermediate 
stage in this evolutionary scenario. 


The present geographic range of H. 
miotympanum, H. arborescandens, and 
H. thorectes does not exclude the possi- 
bility that their immediate, common an- 
cestor could also have served as the 
ancestor of the H. bromeliacia group. 
Based on external and internal morphol- 
ogy of the larvae and the breeding be- 
havior of the adult, it is highly probable 
that the ancestor of H. bromeliacia spe- 
cies group separated from the nuclear 
central American hylid stock at the level 
of a common branch leading to H. mio- 
tympanum, H. arborescandens, and H. 
thorectes. 


THE EVOLUTION OF ONTOGENIES AND ITS 
ROLE IN LARVAL DIVERSITY 


While this study demonstrated a 
great deal of morphological diversity in 
the feeding structures of anuran larvae, 
it is worth noting that few truly unique 
structures were seen in any of the tad- 
poles. 


Most differences between species are 
gradational differences in the relative 
size, shape, or number of particularly 
common features. 


Three specific processes can account 
for most of the structural diversity: pro- 
liferation, fusion, and reduction. These 
are not only processes in the phylogeny 
of anuran larvae, but are common proc- 
esses in their ontogeny as well. Recent- 
ly, much attention has been given the 
role of ontogenetic shifts in phylogenetic 
evolution (Gould, 1977); the tadpole 


128 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


oral structures helped to illustrate how 
important this heterochrony may be in 
generating evolutionary diversity. 

The opportunity for heterochrony in 
anuran evolution is accentuated by the 
fact that developmental programs can be 
modified not once, as in most vertebrates’ 
development, but twice. They can be 
modified first when the embryo develops 
into a tadpole, then again when the tad- 
pole metamorphoses into the frog. 

In comparing different forms, it is 
sometimes possible to tell which part of 
the developmental process has been dis- 
torted through evolution. Consider, for 
example, the different patterns for the 
gill filters seen in hylid larvae with re- 
duced branchial baskets. Both Hyla 
phlebodes and Anotheca spinosa have 
small branchial baskets with highly re- 
duced gill filters. In the case of Ano- 
theca the filters are thin, wispy struc- 
tures, whereas in H. phlebodes they are 
blunt, knoblike projections. In the nor- 
mal ontogeny of a Hyla regilla both 
morphologies are seen  (Wassersug, 
1976b and pers. obser.). In the early 
ontogeny, as the embryo develops into a 
larva, Hyla regilla gill filters appear rela- 
tively knobby without having a compli- 
cated, folded pattern. At metamorpho- 
sis, however, as the filters degenerate, 
they shrink and develop a wispy appear- 
ance, much like those of Anotheca. We 
can suggest that the reduction of the 
gill filters in Hyla phlebodes results from 
abnormally early and prolonged meta- 
morphosis of the gill filters relative to 
other oral structures. It then follows 
that the pattern seen in Anotheca is due 
to an arrested embryonic development 
of these same organs. 

Other examples can be seen in the 
papillae of the buccal floor and roof. 
The proliferation, attenuation and_bi- 
furcation of buccal papillae commonly 
seen in stream bottom larvae with suc- 
torial mouths suggests differential 
growth of these structures either by 
early, accelerated growth or by prolon- 
gation of the period over which they 
develop. The fusion of papillae into 


ridges, a common pattern in funnel- 
mouthed forms, suggests a relative sup- 
pression of the tendency for the papillae 
to elongate, without a reduction in the 
tendency for papillae bases to grow and 
develop. 

Some of the most unusual morpho- 
logical patterns described in this paper 
can be comprehended in the framework 
of heterochrony. A choice example is 
the uninterrupted gill filter rows that 
bridge the gill slits in Gastrotheca. A\- 
though it is possible that this pattern is 
due to a truly novel fusion of filter rows, 
there is an ontogenetically simpler hy- 
pothesis. I suspect that when the gill 
pouches develop in the embryo they 
perforate only at the bottom of the filter 
canal and these perforations fail to coa- 
lesce into continuous slits. Arrested de- - 
velopment of the larval gill slits in free- 
living Gastrotheca may be precursory to 
direct development in Gastrotheca lar- 
vae that have abandoned the free-living 
tadpole stage. This hypothesis, one of 
arrested development, could be verified 
or refuted with a growth series of Gas- 
trotheca embryos. 

Major advances in our general under- 
standing of anuran larval diversity will 
require attention to subtle differences in 
ontogenetic processes. 


ECOLOGICAL CONSIDERATIONS 


TADPOLE FEEDING ECOLOGY 


The autecology of anuran larvae is 
poorly known. Much of what is dis- 
cussed here constitutes hypotheses pre- 
sented in relatively broad terms. For in- 
stance, the sizes of food particles are 
referred to simply as “small,” “medium,” 
or “large.” This vague, comparative 
rather than quantitive, terminology is 
necessitated by our ignorance. There is 
little hard information on the actual 
particle size distribution either in the 
environments in which tadpoles occur 
or in their alimentary tracts. 

The typical pond larvae is often de- 
scribed as a filter feeder (Griffith, 1961; 
cf. Savage, 1973); but elsewhere I (Was- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 129 


sersug, 1975) have argued that the 
broader term “suspension-feeder” is more 
appropriate. As demonstrated by Ru- 
benstein and Koehl (1977) filter feed- 
ing, that is, sieving, is only one mechan- 
ism by which aquatic organisms can 
extract particles suspended in water. 
Aerosol engineers have identified at least 
five mechanisms by which particles can 
be removed from a fluid. Depending on 
the physical properties of the suspended 
particles and the entrapping structures 
(e.g., particle diameter; mass; pore size 
of filters; electrostatic charge, etc.), cer- 
tain mechanisms will be more effective 
than others. 


Tadpoles can clearly ingest particles 
smaller than the pore size of their gill 
filters (see Dodd, 1950; Jenssen, 1967; 
Kenny, 1969a; Hendrick, 1973; Wasser- 
sug, 1972), so they must use other mech- 
anisms besides simple sieving. Kenny 
(1969b), Wassersug (1972), Wassersug 
and Rosenberg (1979), have all dis- 
cussed mucous entrapment in tadpoles. 
This process may involve inertial impac- 
tion or electrochemical attraction (elec- 
trostatic entrapment; cf. Rubenstein and 
Koehl, 1977; LaBarbera, 1978). Follow- 
ing Rubenstein and Koehl, it is likely 
that tadpoles capture particles by at least 
three methods: direct interception, in- 
ertial impaction, and electrostatic en- 
trapment. 


Although many herpetologists have 
made a distinction between the feeding 
ecologies of tadpoles with and without 
keratinized mouth parts, there may be 
little or no difference between the way 
in which these animals feed. It is fair 
to say that all free-living tadpoles, ex- 
cept for some extremely specialized mac- 
rophagous forms, can suspension-feed. 
Tadpoles with hard mouth parts may 
also graze on periphyton, macrophytes, 
and large detrital material, but their 
grazing activity primarily serves to pro- 
duce a suspension. This suspension is 
then handled internally by basically the 
same mechanisms that tadpoles lacking 
hard mouth parts use. 


PARTICLE SORTING: THE GENERAL 
MECHANISM 


The complexity of internal oral struc- 
tures in anuran larvae suggests that com- 
plex sorting processes take place be- 
tween the oral orifice and the esophagus. 
The variety of palps, flaps, papillae and 
other projections in the tadpole mouth 
are interpreted here as forming a multi- 
tiered sorting system used to capture 
particles of a variety of sizes with great 
efficiency. 

The sorting starts at the oral orifice. 
Particles too large to enter the mouth 
cannot be ingested unless they are re- 
duced to smaller particles by the beaks 
and denticles. Particles small enough to 
enter the mouth may be too large to pass 
between the infralabial papillae and can 
be ejected immediately. 


Particles small enough to pass the 
infralabial papillae enter the space be- 
tween the buccal floor and buccal roof 
arena. From this region there are three 
possible courses a particle can take. If 
the particle is too large to pass between 
any of the papillae of the buccal floor, 
it can be coughed out of the mouth. 
Presumably, all particles rejected would 
be too large to pass down the esophagus. 
If a particle is too large to go between 
the papillae on the right and left side of 
the arena, but small enough to pass be- 
tween the papillae at the back of the 
arena, the general funnel-like shape of 
the buccal floor arena will shunt it di- 
rectly posterior and into the esophagus. 
Any particle small enough to pass _be- 
tween the papillae laterally can go to 
either the left or right side and over the 
ventral velum into the pharynx. 


Once in the pharynx, a particle may 
be trapped by direct interception and 
inertial impaction on the gill filters. Par- 
ticles so small as to pass through the 
gill filters can still be aggregated in 
mucus on the branchial food traps. The 
larger aggregates formed there can then 
be retained by the gill filters (Wasser- 
sug, 1972): 

How particles trapped in the filter 


130 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


cavities are transported posterolaterally 
to the ciliary groove and ultimately to 
the esophagus is poorly understood. A 
reasonable, but admittedly vague, hy- 
pothesis is that particulate matter com- 
pressed between the pressure cushions 
and the gill filters is shunted later- 
ally by a rhythmic pulsing action of the 
pharynx. It should be possible to test 
this hypothesis by giving tadpoles dyed 
food and sacrificing them individually at 
successively longer intervals after feed- 
ing. Dissection of these larvae should 
reveal the path of food through the 
pharynx and clarify the transport mech- 
anism. In any case it is evident that tad- 
poles have alternative methods of trap- 
ping particles that depend on particle 
size. The ingestive system appears ar- 
ranged so that gross structures which 
can trap large particles protect smaller, 
more delicate surfaces which are de- 
signed for handling the finer particles. 
Specifically, buccal structures protect the 
branchial baskets from clogging by chan- 
neling larger particles, that have already 
passed through the infralabial papillae, 
directly to the esophagus. 


It follows from the above general 
outline that interspecific differences in 
the size, shape, and spacing of the vari- 
ous trapping structures will reflect differ- 
ences in the particle sizes upon which 
different larvae feed most efficiently. 


MICROHABITAT IMPLICATIONS OF 
MORPHOLOGICAL PATTERNS 


Several common morphological pat- 
terns in the overall proportions of the 
various sorting and sieving structures of 
tadpoles provide clues to the feeding 
ecologies of these organisms. Many of 
these patterns were alluded to in the 
section on Functional Considerations 
and are summarized here. 

Several anurans have larvae which 
show a loss or reduction in all of the 
particle sorting and trapping structures. 
This morphological pattern is consistent- 
ly associated with extreme macrophagy, 
as illustrated by the carnivorous Ano- 


theca and Hymenochirus larvae. Not 
only obligate carnivores, but several 
herbivore/detritivore larvae have aban- 
doned indiscriminate suspension-feeding 
in favor of macrophagy. Hyla micro- 
cephala and Ooeidozyga larvae are ex- 
amples. 

Another common pattern is the pro- 
liferation of buccal sieving structures in 
larvae which also have reduced filter 
density (high filter porosity). This 
places emphasis on the gross sieving 
structures rather than finer entrapping 
surfaces and appears to be an adaptation 
for feeding on a relatively coarse sus- 
pension of particles. Benthic, thigmo- 
tactic tadpoles commonly show this pat- 
tern. 


Fusion of buccal floor and roof pa- 
pillae into ridges, which is characteristic ' 
of funnel-mouthed tadpoles, enables the 
buccal cavity to serve as a particle guid- 
ing system, rather than a particle sieving 
system. This morphological arrangement 
appears ideal for handling a relatively 
narrow size range of moderately coarse 
particles. 


Enlargement of the branchial baskets 
and a very high density of the gill filters 
is most often associated with relatively 
few, widely spaced buccal papillae. Tad- 
poles with this pattern appear well 
adapted for feeding on a suspension of 
uniformly small particles. Indeed, larvae 
with this morphology are the obligate 
microphagous suspension-feeders; Mi- 
crohyla ornata and Xenopus laevis are 
two examples. 


Each of these patterns correlates well 
with the distribution of particle size in 
the environments where the larvae ex- 
hibiting them live. For example, micro- 
planktonic algae are likely to be rare in 
the aphotic environment of Anotheca 
spinosa. Food source in this habitat will 
be those large organisms near the top of 
the food chain, on which Anotheca is 
clearly specialized to feed. 

Benthic, thigmotactic tadpoles are 
best represented by stream larvae. 
Stream environments are relatively oli- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 131 


gotrophic; with a rapid exchange of wa- 
ter, and constant nutrient washout, there 
is little chance for a phytoplanktonic 
bloom in a stream. Most primary pro- 
ductivity in this habitat is in the form of 
periphyton, which can be gleaned only 
by grazing upon the substrate. The 
grazing activity of tadpoles with beaks 
and denticles produces a_ relatively 
coarse suspension of particles. 


Tadpoles adapted for feeding on 
these self-generated suspensions are not 
likely to encounter large particles or ex- 
ceptionally small particles with any fre- 
quency. Although it may appear advan- 
tageous for stream tadpoles to have large 
and dense gill filters that could com- 
pletely retain all of the smallest particles 
ever encountered, extremely small food 
particles are too rare in a stream habitat 
to warrant the energetic cost of develop- 
ing and maintaining the tissues neces- 
sary for capturing all such particles. At 
the other extreme, thigmotactic stream 
larvae are not likely to run into very 
large particles such as those regularly 
ingested by macrophagous carnivores 
and herbivores. The large gap at the 
back of the buccal floor and buccal roof 
arena, which allows many tadpoles to 
shunt the largest particles they ingest 
directly from the buccal cavity into the 
esophagus, is absent in these forms. 


The funnel-mouth adaptation appears 
even more specialized for selecting rela- 
tively large particles in a relatively nar- 
row size range. Lacking the multi-tiered 
sieving system, funnel-mouthed tadpoles 
have abandoned the ability to feed in- 
discriminately over a broad range of 
particle sizes in favor of the capacity to 
handle large particles which might float 
in the backwashes of quiet streams or 
adhere loosely to surfaces. 

Tadpoles specialized for extreme mi- 
crophagy seem to occur most commonly 
in midwater in small, stagnant pools or 
ponds which have high insolation, a high 
nutrient load, and an abundance of uni- 
cellular phytoplankton (see Heyer, 
1973, 1974; Wassersug, 1973). These 


microphagous suspension-feeders often 
metamorphose at a relatively small size 
and quite soon after hatching. They are 
specialized for ingesting the earliest, pri- 
mary productivity which develops in 
temporary ponds, namely small phyto- 
plankton. The absence of keratinized 
mouth parts in most of these forms can 
be construed as an energy-conserving 
adaptation. While hard mouth parts 
would allow these larvae to graze on 
macrophytes, periphyton, and other de- 
tritus, such material is often absent in 
their ephemeral environment and these 
morphological structures are energeti- 
cally expensive to develop and main- 
tain. 


A few tadpoles are unusual in ap- 
pearing to be specialized for handling 
food of two different size classes. Sca- 
phiopus and Rhinophrynus typify this 
bimodal size selection potential. They 
are both temporary pond breeders in 
seasonally dry areas. In these environ- 
ments much of the potential food is 
either unicellular algae or the larger 
herbivores, such as conspecifics, that 
graze on this algae. On the one hand, 
these larvae are efficient, ultraplanktonic 
suspension-feeders able to handle the 
initial phytoplanktonic blooms in the 
small bodies of water where they breed. 
On the other hand, the ephemeral nature 
of their ponds requires that they be able 
to grow to a large size extremely fast 
and this means they must be able to 
ingest all other possible food items in 
their ponds. In having few, widely 
spaced, buccal floor papillae and dense 
gill filters, these larvae are adapted for 
handling both the largest and the small- 
est particles in their environment; they 
are not likely, however, to feed efficiently 
on the medium sized particles that most 
tadpoles generate while grazing. 

In contrast to the extreme forms just 
described, the majority of pond larvae 
appear adapted for handling a relatively 
broad spectrum of food sizes and types. 
In the “average” temporary pond (e.g., 
a temporate zone, vernal pond) the size 


132 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


distribution of potential food can fluc- 
tuate so drastically that most larvae 
would never reach metamorphosis if 
they ate food in a narrow size range 
only. Because of extreme fluctuation in 
the resource base it is not surprising that 
there has been little documentation of 
direct (exploitation) competition among 
tadpoles for food. Competition among 
most anurans has resulted in spatial and 
temporal separation of larvae rather than 
overt dietary partitioning (Heyer, 1974, 
1976). Only recently has exploitation 
competition been demonstrated for typi- 
cal pond larvae in either the laboratory 
(Steinwascher, 1978; Seale and Beckvar, 
in press) or the field (Seale, 1973). 


One can use slight variations in the 
morphological patterns delineated above 
to hypothesize about the feeding ecolo- 
gies of specific anuran larvae. Hyla ru- 
fitela and Hyla femoralis, for instance, 
which have denser gill filters than either 
Hyla regilla or Acris crepitans, are prob- 
ably more efficient at ultraplanktonic en- 
trapment than these latter forms. As 
another example, the Bufo larva illus- 
trated by Savage (1952) has many tall 
buccal floor papillae, a feature associated 
with benthic, thigmotactic grazing. It is 
not surprising, then, that Bufo tadpoles 
are often found feeding on the bottom, 
grazing on detritus. They appear spe- 
cialized for handling medium to small 
fragments of material and may not be 
particularly efficient at feeding on either 
extremely large or extremely small parti- 
cles. 

From all that has been said so far, 
we can conclude that tadpole oral fea- 
tures reflect the breadth of the larval 
diet, at least in terms of particle size, 
and, as indicators of niche breadth, oral 
structures can provide some insight into 
species packing problems for tadpoles. 
To illustrate this latter point we can 
compare two distinct larval ecotypes. 
Larvae of moderately large size from 
cool-temperate climates (e.g., many ran- 
ids) require a relatively long develop- 
mental time, over which they are likely 


to be exposed to a very broad range of 
particle sizes. They may have few po- 
tential competitors for food resources 
and are understandably dietary general- 
ists. In contrast, in wet, lowland, tropi- 
cal ponds developmental time can be 
rapid but the anuran fauna is very large, 
and competition for breeding sites severe 
(Crump, 1974, and others cited therein). 
This is a situation where many anuran 
species use the same breeding pond at 
the same time and the larvae may not be 
able to avoid competition; here one is 
most likely to find pond tadpoles which 
are dietary specialists. 

Consider Hyla phlebodes; the ex- 
tremely specialized oral morphology in 
this larvae and its obligatory, narrow 
dietary range suggest that it occurs in 
ponds with larvae of many other species 
that use other parts of the resource base. 
Contrast this with Gastrotheca riobam- 
bae—based on its generalized morphol- 
ogy this large larvae would appear to 
have a very broad diet and to feed effi- 
ciently on a variety of food types. One 
can predict that it would occur with 
larvae of relatively few other species. 
Although I have personally collected 
neither species, Duellman (pers. comm. ) 
has indeed confirmed that H. phlebodes 
larvae are always found inhabiting ponds 
with tadpoles of several other species, 
while G. riobambae (which occurs at 
high elevation) is found alone. 


The feeding currents of a tadpole are 
also its respiratory currents, so it is ap- 
propriate to add here some comments 
on ecological patterns in the respiratory 
systems of anuran larvae. Tadpoles that 
are specialized for feeding on large par- 
ticles often have a large, exposed glottis 
on the floor of the mouth and evidently 
depend extensively on aerial respiration. 
It is a reasonable hypothesis that, in the 
absence of buccal papillae to act as 
coarse sieves, these larvae should reduce 
their aquatic respiratory pumping activ- 
ity in order to prevent accidental clog- 
ging of the filter cavities or glottis by 
particles that would otherwise be too 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 133 


small to be ingested. Macrophagous lar- 
vae, such as Anotheca, probably do not 
exhibit regular aquatic pumping. It is 
known that Hymenochirus larvae do not 
exhibit regular respiratory pumping and 
only open their mouths to take in either 
air or prey (Sokol, 1962). Many of the 
macrophagous, herbivorous Hyla larvae 
(H. microcephala and H. leucophyllata 
species group) are small and may rely 
extensively on cutaneous respiration. 


Obligate midwater, microphagous 
anuran larvae often occur in extremely 
eutrophic and, potentially hypoxic, habi- 
tats; they also exhibit large glotta and 
may depend heavily on aerial respira- 
tion. 


In contrast, stream larvae, which live 
in well aerated habitats, can depend on 
gill and cutaneous respiration to meet 
their metabolic needs. In these forms, 
aerial respiration could lead to positive 
buoyancy, which may be maladaptive 
(Wilder et al., 1920; Wake, 1966; Was- 
sersug and Seibert, 1975). 

Differences in both feeding and res- 
piration are often reflected in the rela- 
tive size of the buccal floor since buccal 
floor area is one parameter affecting the 
volume displaced by the buccal pump 
(Wassersug and Hoff, 1979). This, 
in turn is intimately related to both 
feeding and respiratory patterns. 

All macrophagous larvae exhibit a 
general expansion of the buccal floor in 
the prepocket region relative to the rest 
of the floor and pharynx that allows 
them to suck in large particles. This 


expansion is well exemplified by the 
differences between the funnel-mouthed 
Microhyla and other Microhyla. Most 
Microhyla have a small buccal displace- 
ment compared to the large volume of 
their branchial baskets. These non-fun- 
nel-mouthed forms cannot completely 
clear the oral cavity with a single stroke 
of the buccal pump, but can maintain a 
continuous, gentle flow past the gills 
throughout a buccal cycle. Thus, they 
are equipped to clear constantly a sus- 
pension of rather fine matter, as well as 
to irrigate continuously their respiratory 
surfaces. In contrast, M. heymonsi, with 
its comparatively large buccal pump and 
small branchial baskets, can suck in a 
large volume of water with each stroke 
of the buccal pump. H. heymonsi has 
the ability to use gape-and-suck feeding 
to selectively take in rather large parti- 
cles but not to maintain a gentle con- 
tinuous respiratory flow while it is feed- 
ing. Its relatively large buccal pump 
displacement is well adapted for spo- 
radic feeding, but could cause the ani- 
mal to lurch conspicuously forward and 
backward in the water column if the 
tadpole pumped maximally during nor- 
mal respiratory activities. 

More work on the feeding and respi- 
ratory ecology of anuran larvae will be 
necessary before we will completely un- 
derstand the ecological and evolutionary 
implications of larval morphology. It is, 
however, already clear that knowledge 
of tadpole oral morphology can provide 
major insight into the way of life of 
these organisms. 


SUMMARY AND CONCLUSIONS 


This study has been limited to gross 
surface features in the mouths of tad- 
poles—surfaces where particles are ex- 
tracted from water currents and gas 
exchange takes place. Internal oral sur- 
faces are described for larvae of eight 
anuran families. When combined with 
descriptions elsewhere in the literature, 
the anatomy of this region is now known 
for representatives of a dozen families. 


In the following pages the morpho- 
logical variation seen in the oral struc- 
tures of the larvae examined is first sum- 
marized, and morphological patterns 
reviewed. Then major conclusions con- 
cerning anuran systematics and tadpole 
evolution and ecology are enumerated. 


MORPHOLOGICAL VARIATION 


An effort was made to interpret the 


134 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


function of all major morphological 
structures: 

(1) Keratinized spurs and _projec- 
tions—Two keratinized structures are 
seen inside the mouths of certain tad- 
poles. The first, which occurs in many 
species and may be used in shredding 
food is a bilateral, medially directed 
keratinized spur at the edge of the oral 
orifice between the upper and lower 
beaks. The second, a medial cornified 
projection from the prenarial arena in 
the macrophagous Scaphiopus bombi- 
frons, is presumably used in holding and 
shearing prey. 

(2) Infralabial papillae —In thigmo- 
tactic, benthic larvae these are typically 
large, elongated and dendritic, whereas, 
in larvae adapted for macrophagy, they 
may be short and globose or lacking 
altogether. These structures have a me- 
chanical and possible chemoreceptive 
function; they appear particularly im- 
portant in the positioning of food as it 
comes into the mouth. 

(3) Lingual papillae—Most species 
examined have two premetamorphic pa- 
pillae, but some have one or four. As- 
caphus and the discoglossids have multi- 
ple lingual papillae. The pipids, lacking 
a tongue, also lack lingual papillae. Fol- 
lowing Hammerman (1969), I have as- 
sumed these structures serve a chemo- 
receptive function. 

(4) Buccal floor arena (BFA) and 
Buccal roof arena (BRA) papillae.—Pa- 
pillae of the buccal roof and floor are 
usually arranged to form a V- or U- 
shaped area. Species may have no buc- 
cal papillae (e.g., Pipidae, macrophagous 
Hylidae) or as many as eighty or more 
papillae on the buccal floor alone. Pa- 
pillae seem to be important in the pri- 
mary sorting and sieving of particles 
from the water; they may actually be 
the major food sorting structures in the 
mouth of stream-adapted species. 

In funnel-mouthed larvae, papillae 
often are either absent or fused into 
interlocking dorsal and ventral ridges. 
I interpret these structures as chutes for 
separating out larger food particles. 


(5) Ventral velum.—tThe ventral vel- 
um is an important valve which regu- 
lates one-way flow from the buccal to 
the pharyngeal cavity in all anuran lar- 
vae except the Pipidae. There are six 
major configurations of the ventral 
velum which can be associated with spe- 
cific families or superfamilial groups. 

(6) Gill filters——Interspecific varia- 
tion, in overall size and in relative den- 
sity of the gill filters reflects differences 
in tadpole feeding ecology. Tadpoles 
generally recognized as extremely spe- 
cialized for microphagy have very large 
branchial baskets and large filter plates 
with very dense gill filters; these serve 
to extract oxygen and food from the 
water. Obligate macrophagous larvae 
have reduced branchial baskets and gill 
filters; in the extreme macrophagous © 
forms, such as Hymenochirus, gill filters 
are completely absent. Funnel-mouthed 
larvae typically show some reduction in 
the gill filters compared to close rela- 
tives with more typical oral features. 

Two unusual patterns involving fu- 
sion of filter rows are documented. In 
Pipa, neighboring filter rows on each 
filter plate are fused so that the filter 
canal between the rows is canopied. In 
Gastrotheca, the filter rows are fused 
ventrally with opposing filter plates over 
the ventral gill slits, which means that 
they have fenestrated gill pouches rather 
than continuous gill slits. 

(7) Branchial food traps——tThere are 
two basic patterns for these organs; 
one characterizes the Archaeobatrachia 
(sensu Duellman, 1975), and the other 
the Neobatrachia. Overall size and 
shape of the branchial food traps does 
not seem to correlate strongly with any 
aspect of tadpoles’ diets. The texture of 
the secretory surface itself, however, 
shows some correlation with larval feed- 
ing ecology. Forms specialized for mi- 
crophagy tend to have distinct, widely 
spaced ridges, while extreme macropha- 
gous larvae tend to lose the secretory 
tissues in general, and the secretory 
ridges specifically. Exceptions to the 2 
basic forms are the unique secretory 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 135 


ridge patterns seen in Rhinophrynus, 
Xenopus, and the Microhylidae. As- 
caphus and the discoglossids have an 
extensive field of secretory epithilia in 
the branchial food traps but lack secre- 
tory ridges (see Wassersug and Rosen- 
berg, 1979, for details of the micro- 
anatomy of these surfaces). 

(8) Glottis—The size of the glottis 
and the extent of its exposure on the 
floor of the mouth appears to correlate 
both with aerial respiration and the tad- 
pole’s buoyancy. Bottom larvae tend to 
have a small glottis, while obligate mid- 
water larvae, and those that depend 
heavily on aerial respiration (e.g. Ano- 
theca), invariably have a large glottis. 

(9) Esophageal funnel—tThe size 
and shape of the anterior portion of the 
esophagus reflect the size of the parti- 
cles upon which tadpoles commonly 
feed; macrophagous larvae tend to have 
a large esophagus, whereas obligate mi- 
crophagous forms have a small, narrow 
esophagus. 

(10) Prenarial arena.—Diverse spe- 
cies-specific patterns in ridges, pustula- 
tions, projections, etc. are found in the 
prenarial arena of anuran larvae. Their 
exact function is not well understood 
but it is hypothesized that they assist the 
lower beak in holding and positioning 
food as it enters the mouth. 

The most distinctive pattern for these 
structures is seen in funnel-mouthed 
forms, regardless of family; here there 
is invariably a distinct, posteriorly di- 
rected, V-shaped ridge in the prenarial 
arena. 

(11) Internal nares.——The internal 
nares are perforated structures in all but 
microhylid larvae. Their orientation 
seems to vary, in part, with the shape of 
the tadpole head. The nares are always 
associated with valves. In the Dis- 
coglossidae, large flaps extend under 
the nares and presumably protect them 
from particles flowing into the mouth. 
Several species have pockets of pre- 
sumed sensory epithelia (Jacobson’s or- 
gans) arising from the margins of the 
internal nares. Sensory epithelium is evi- 


dent in the unperforated nares of the 
Microhylidae. Tadpoles of this family 
have a stiff flap that extends down from 
the posterior margin of the internal nares 
and may help direct currents toward 
these sensory regions. 

(12) Postnarial arena. — Tadpoles 
typically have a symmetrical arrange- 
ment of papillae or ridges that are tight- 
ly grouped posterior to the internal 
nares. Projections are large and numer- 
ous in this region in forms which have 
many large and numerous buccal papil- 
lae elsewhere in the mouth. Postnarial 
structures, however, are reduced in 
forms which typically lack papillae or 
show reduction in papillae elsewhere on 
the buccal roof or floor. Blunt, shallow 
papillae typically replace rows of post- 
narial papillae in the postnarial arena of 
funnel-mouthed larvae. 

Postnarial papillae are aligned to 
form a cap over the lingual anlage when 
the mouth is closed. Postnarial struc- 
tures may serve chemoreceptive, me- 
chanical-receptive, and particle-sorting 
roles. They may be important in direct- 
ing water currents to regions of sensory 
epithelia and particulate matter to other 
regions of the buccal cavity for sorting. 

(13) Glandular zone.—The ventral 
velum of tadpoles meets the buccal roof 
during closure along a region of secre- 
tory epithelia, the glandular zone, which 
has been implicated in particle capture. 
The secretory tissue of the zone may 
assist the branchial food traps in aggre- 
gating particulate matter in mucus. 
There appears to be a slight inverse 
correlation between the size of the secre- 
tory pits in the glandular zone and the 
particle size that a tadpole typically 
feeds upon. It is proposed here that the 
glandular zone may also assist the velum 
in its valvular function. The zone varies 
extensively in shape, but in part reflects 
the shape of the ventral velum beneath 
it. Secretory tissue of the glandular zone 
often continues posteriorly onto the 
pressure cushions of the dorsal pharynx. 
Macrophagous larvae often lack a glan- 
dular zone. 


136 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


(14) Dorsal velum.—A continuation 
of the ventral velum onto the roof of 
the mouth is characteristic of all non- 
pipid tadpoles but is reduced in extreme 
macrophagous forms. The dorsal velum 
is invariably positioned slightly posterior 
to the ventral velum in such a way that 
it must deflect water from the buccal 
cavity into the anterior part of each 
filter cavity when the buccal floor is ele- 
vated. It is proposed here that, in de- 
flecting water currents ventrally and an- 
teriorly, the dorsal velum helps prevent 
the resuspension of particulate matter 
aggregated and precipitated in the pos- 
terior parts of the pharynx. 


(15) Dorsal pharynx.—The dorsal re- 
gion of the pharynx is small in all tad- 
poles, and the only gross structures of 
this region are pressure cushions. These 
are folds of epithelium that extend ven- 
trally and complement the shape of the 
filter cavities below. The larger and 
deeper the filter cavities are, the larger 
and deeper are the pressure cushions. 
The secretory tissue on these structures 
suggests that they have a role in food 
processing. They may help collect food 
in the filter cavities or in some way 
assist in the lateral movement of food 
from the filter epithelium to the ciliary 
groove. 


SYSTEMATIC CONCLUSIONS 


Major points from the discussion on 
Systematic Considerations are summar- 
ized below in the framework of the 
Archaeobatrachia-Neobatrachia  distinc- 
tion. Patterns in oral structures gener- 
ally support but in some cases refute, 
prevailing ideas on the taxonomic rela- 
tionships within and among anuran fam- 
ilies. 

(1) Internal oral features of As- 
caphus larvae and Leiopelma embryos 
provide no evidence for the inclusion of 
these genera in the same family. 

(2) When considering the structure 
of their internal nares and the shapes of 
their pharynxes, the current members 
of the Discoglossidae appear to form a 


natural family or superfamily. Bran- 
chial food trap anatomy suggests a close 
relationship between Ascaphus and the 
Discoglossidae. Among the Discoglossi- 
dae, Bombina appears to have the most 
generalized larva. 


(3) Rhinophrynus larvae are exter- 
nally similar to Xenopus and Pipa, but 
internally they are as different from 
pipid larvae as they are from the larvae 
of any other family. While the super- 
familial association of the Rhinophryni- 
dae and the Pipidae may be valid, Rhi- 
nophrynus must be accepted as the most 
generalized of the Pipoidea. 


Pipoid frogs are so specialized that 
the superfamily could not serve as the 
immediate ancestors of any extant anu- 
ran families. The recent suggestion that . 
the Pipoidea is more closely related to 
the Neobatrachia than to the Archaeo- 
batrachia (cf. Sokol, 1976b) finds no 


support in my observations. 


(4) Many features of the ventral 
velum, branchial food traps, internal 
nares, and glottal position distinguish 
the Microhylidae from all other families. 
Internally, microhylid larvae do not look 
like pipid tadpoles, although their bran- 
chial food traps are grossly similar to 
those of Rhinophrynus. Ranoid rather 
than pipoid associations of the Mirco- 
hylidae are suggested by larval features 
(specifically, the shape of the ventral 
velum in Microhyla heymonsi). 


(5) The Pelobatidae have larvae 
which are unique in the shape of their 
branchial baskets and the attachment of 
their ventral vela. The absence of se- 
cretory ridges in this family is consistent 
with an Archaeobatrachian assignment 
(cf. Duellman, 1975). I interpret the 
shape of the branchial baskets in the 
Pelobatidae as a unique, derived char- 
acter state, which makes it unlikely that 
other anuran families evolved directly 
from this family. Primitive leptodacty- 
loid (e.g. myobatrachian) larvae, how- 
ever, have not been examined and if 
their branchial baskets are of the pelo- 
batid form, this will support the idea 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 137 


that leptodactyloid frogs have their ori- 
gin in Asiatic (megophrynine) pelo- 
batids. 

(6) Interspecific variation in oral 
structures in the Hylidae is enormous. 
Most patterns can be interpreted along 
ecological lines. Externally similar lar- 
vae from different genera are often very 
similar internally, suggesting convergen- 
cies or parallelism. On the other hand, 
tadpoles of the same genus which differ 
only slightly in external morphology, 
may differ greatly in internal morphol- 
ogy. The following are systematic con- 
clusions at the subfamilial level: 

(a) Anotheca and Gastrotheca are 
extremely different in internal larval 
morphology, supporting Maxson’s (1977) 
removal of Anotheca from the Amphi- 
gnathodontinae. However, Anotheca 
larvae are so bizarre, they provide little 
evidence for inclusion of this genus in 
the Hylinae. 

(b) Tadpole oral structures support 
the idea that Phyllomedusa and Agalych- 
nis are closely related but do not settle 
the issue of the status of the family 
Phyllomedusidae. 

(c) Hylinine larvae with similar 
ecologies are internally very similar, re- 
gardless of genus or species. 

(d) Internal oral features support 
certain phylogenetic relationships pre- 
viously suggested by biochemical data 
only, and which otherwise conflict with 
traditional taxonomy based on adult os- 
teology. An example is the similarity 
between Acris crepitans and Hyla regilla 
larvae compared to certain other Hyla 
tadpoles. The basis for this “tadpole: 
biochemical” taxonomic congruence is 
not well understood, and more tadpoles 
will have to be studied to know whether 
it is a general pattern. 

(e) The genus Hyla is artificial and 
should be partitioned into several genera. 
Certain taxa such as the Hyla mixomac- 
ulata and Hyla microcephala species 
groups, probably form valid genera. 


EVOLUTIONARY CONCLUSIONS 
In addition to the specific taxonomic 


considerations just reviewed, some gen- 
eral evolutionary conclusions may also 
be drawn from this work. 

(1) The unusual chondrocranium of 
tadpoles is associated with the develop- 
ment of an expanded ceratohyal and an 
efficient buccal pump. The development 
of an enlarged buccal pump, the pres- 
ence of an elongated ceratohyal, and the 
presence of a buccal valve in anuran 
larvae are all features interpreted as 
associated with a generalized suspension- 
feeding way of life. This is considered 
the primitive and generalized way of life 
for anuran larvae. 

(2) A unique way of life for anuran 
larvae, not specifically discussed by her- 
petologists in the past, is macrophagous 
herbivory. Extreme reduction in the 
branchial food traps, loss of buccal papil- 
lation, development of large, pad-like 
infralabial papillae, and the presence of 
a massive buccal floor area characterize 
these tadpoles. This morphology, best 
exemplified in Hyla phlebodes, is in- 
terpreted as an adaptation for feeding 
on coarse plant matter and detritus. 
Larvae of the Hyla leucophyllata species 
group show moderate development of 
this type of larva. Ooeidozyga in the 
Ranidae has independently converged 
on this morphology. 

(3) The arboreal larvae of the Hyla 
bromeliacia group evolved from an an- 
cestor with a stream-adapted larva. It 
is suggested that the elongated arboreal 
larvae, in general, had stream-adapted 
ancestors, and that the Hyla bromeliacia 
group specifically evolved from a lineage 
that was ancestral to Hyla miotympan- 
um, H. arborescandens, and H. thorectes. 

(4) Much of the anatomical differ- 
ence between larvae of different species 
can be explained by heterochronic 
changes in the early (embryonic) and 
late (metamorphic) ontogeny of the 
tadpoles. Certain morphological pat- 
terns can be identified as due to evolu- 
tionary modifications of specific early or 
late ontogenetic events when compared 
with more generalized larval develop- 
mental patterns. Further understanding 


138 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


of the evolution of anuran larvae diver- 
sity will require greater knowledge of 
differences in larval ontogenies than is 
presently available. 


ECOLOGICAL CONCLUSIONS 


The majority of internal oral features 
in tadpoles are interpreted as part of a 
multi-tiered particle entrapping system 
which sorts ingested particles by size. 
Direct interception and inertial impac- 
tion are used to different extents on dif- 
ferent surfaces. The mucous surfaces of 
the branchial food traps and the gill 
filters of the pharynx together can effi- 
ciently trap the smallest particles in- 
gested by a typical pond larva. Buccal 
papillae strain larger particles from the 
water and funnel them directly to the 
esophagus, bypassing most of the 
pharynx. Delicate pharyngeal surfaces 
that could be clogged or damaged by 
large particles are, thus, protected by 
this size-sorting mechanism. 

The size of the particle upon which 
a species feeds most efficiently may be 
inferred from the size, shape, number 
and spacing of buccal and pharyngeal 
structures. Interspecific differences in 
these features may reflect differences in 
the size distribution of particles in the 
microhabitats in which the tadpoles live. 

In comparison with generalized tad- 
poles, extreme macrophagous larvae 
show reduction in all pharyngeal struc- 
tures associated with planktonic entrap- 
ment. On the other hand, extreme 
microphagous larvae, which live in mid- 
water, have large branchial baskets and 
dense gill filters well designed for cap- 
turing small phytoplankton that many 
abound in their habitat. Benthic, thig- 
motactic larvae typically inhabit streams 
which are probably devoid of phyto- 
plankton but rich in periphyton. These 


larvae can generate a coarse suspension 
with their keratinized mouth parts; they 
have closely spaced, supernumerary buc- 
cal papillae for straining coarse particles, 
but highly porous gill filters not well 
adapted for ultraplanktonic entrapment. 
Funnel-mouthed tadpoles feed selective- 
ly on large particles floating on the water 
surface; they have buccal ridges in place 
of papillae that seem well adapted for 
sorting moderately coarse particles from 
a relatively narrow size range. 


Most pond larvae are adapted for 
handling the broad spectrum of food 
types and sizes that may occur in un- 
predictable environments. Pond larvae 
specialized for ingesting a narrow size 
range of particles (e.g., Hyla phlebodes) 
are most likely to occur in species rich 
environments where competition may 
have been intense during their evolu- 
tion. 


Respiratory structures vary with the 
availability of dissolved oxygen. Since 
the amount of phytoplankton in any 
aquatic body affects the amount of dis- 
solved oxygen, there appears to be a 
relationship between a tadpole’s diet 
and its respiratory structures. Oligo- 
trophic streams are likely to be well- 
aerated, and the benthic tadpoles from 
these environments tend to have few of 
the morphological features associated 
with aerial respiration. On the other 
hand, obligate midwater microphagous 
larvae may live in extremely eutrophic 
ponds, and typically have a large glottis, 
a feature associated with a strong de- 
pendence on aerial respiration. 

Internal oral structures of anuran 
larvae can be used to make reasonably 
sound predictions about the feeding and 
respiratory ecology of anuran larvae 
when field data are not available. 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 139 


LITERATURE CITED 


ALCALA, A. C. 

1962. Breeding behavior and early develop- 
ment of frogs of Negros, Philippine 
Islands. Copeia, 1962:679-726. 

AuTic, R., and E. D. Bronte, Jr. 

1972. Laboratory behavior of Ascaphus truei 

tadpoles. Jour. Herp., 6:21-24. 
BLOMMERS-SCHLOSSER, R. 

1975. Observations on the larval develop- 
ment of some Malagasy frogs, with 
notes on their ecology and _ biology 
(Anura: Dyscophinae, Scaphiophry- 
ninae and Cophylinae). Beaufortia, 24 
(309 ) :7-26. 

BOKERMANN, W. C. A., and I. SazIMa 

1978. Anfibios da Serra do Cipd, Minas 
Gerais, Brasil. 4: Descricao de Phyl- 
lomedusa jandaia sp. n. (Anura, Hy- 
lidae). Rev. Brasil. Biol., 38:927-930. 

BOULENGER, GEORGE A. 

1891. A synopsis of the tadpoles of the 
European Batrachians. Proc. Zool. 
Soc. London, 1891:593-627. 

Brace, A. N. 

1965. Gnomes of the night, the spadefoot 
toads. Univ. Pennsylvania Press, 
Philadelphia, 127 pp. 

BRESLER, J., and A. N. Bracc 3 

1954. Variation in the rows of labial teeth 

in tadpoles. Copeia, 1954:255-257. 
CEI, J. M. 

1968. Notes on the tadpoles and breeding 
ecology of Lepidobatrachus (Am- 
phibia: Ceratophryidae). Herpeto- 
logica, 24:141-146. 

Crump, M. L. 

x 1974. Reproductive strategies in a tropical 
anuran community. Univ. Kansas 
Mus. Nat. Hist. Misc. Publ. 61:1-68. 

DeEJoncu, H. S. 

1968. Functional morphology of the jaw 

«apparatus of larval and metamorphos- 
ing Rana temporaria L. Neth. Jour. 
Zool., 18(1):1-103. 

Dopp, J.M. 

1950. Ciliary feeding mechanism in anuran 

larvae. Nature, London, 165:283. 
DUELLMAN, W. E. 

1970. The hylid frogs of Middle America. 
Monog. Mus. Nat. Hist., Univ. Kan- 
sas, 1:1-753. 

1974. A systematic review of the marsupial 
frogs (Hylidae: Gastrotheca) of the 
Andes of Ecuador. Occas. Papers 
Mus. Nat. Hist., Univ. Kansas, 9:1-37. 

1975. On the classification of frogs. Occas. 
Papers Mus. Nat. Hist., Univ. Kansas, 
42:1-14. 

DuELLMAN, W. E., and T. H. Fritts 
1972. A taxonomic review of the Southern 


Andean marsupial frogs (Hylidae: 
Gastrotheca). Occas. Papers Mus. 
Nat. Hist., Univ. Kansas, 9:1-37. 
Dunn, E. R. 

1924. Some Panamanian frogs. Occas. Pa- 
pers Mus. Zool., Univ. Michigan, 151: 
1-17. 

EATON, T. H., Jr. 

1959. The ancestry of modern Amphibia: a 
review of the evidence. Univ. Kansas 
Publ. Mus. Nat. Hist., 12(2):155-180. 

EpGEworrtH, F. H. 

1930. On the masticatory and hyoid muscles 
of larvae of Xenopus laevis. Jour. 
Anat., 64:184-188. 

E1pL-EIBESFELDT, I. 

1953. Die Bestimmung von Kaulquappen 
nach ihrem  Verhalten. Deutsch. 
Aquar. Terr., 6:16-18. 

Gans, Car. 

1975. Tetrapod limblessness: evolution and 
functional corollaries. Amer. Zool., 
15:455-467. 

Gaupin, A. J. 

1974. An osteological analysis of holarctic 
tree frogs, family Hylidae. Jour. 
Herp., 8:141-152. 

GoeETTE, A. 

1874. Atlas zur Entwicklungsgeschicte der 
Unke. Verlag von Leopold Voss, 
Leipzig. 

Gorn, C. J., O. B. Goin, and G. R. Zuc 

1978. Introduction to Herpetology, 3rd ed. 
W. H. Freeman and Co., San Fran- 
cisco, 378 pp. 

Gosner, K. L. 

1960. A simplified table for staging anuran 
embryos and larvae with notes on 
identification. Herpetologica, 16:183- 
190. 

Gou Lp, S. J. 

1977. Ontogeny and Phylogeny. Belknap 
Press of Harvard Univ. Press, Cam- 
bridge, Mass., 501 pp. 

GRADWELL, N., and V. M. PasztTor 

1968. Hydrostatic pressures during normal 
ventilation in the bullfrog tadpole. 
Can. Jour. Zool., 46:1169-1174. 

GRADWELL, N. 

1970. The function of the ventral velum 
during irrigation in Rana catesbeiana. 
Can. Jour. Zool., 48:1179-1186. 

1971a. Ascaphus tadpole: experiments on the 
suction and gill irrigation mechanisms. 
Can. Jour. Zool., 49:307-332. 

1971b. Xenopus tadpole: on the water pump- 
ing mechanism. MHerpetologica, 27: 
107-123. 

1972a. Gill irrigation in Rana catesbeiana. 
Part I. On the anatomical basis. Can. 
Jour. Zool., 50:481-499. 


140 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


1972b. Gill irrigation in Rana catesbeiana. 
Part II. On the musculoskeletal mech- 
anism. Can. Jour. Zool., 50:501-521. 

1972c. Comments on gill irrigation in Rana 
fuscigula. Herpetologica, 28:122-125. 

1973. On the functional morphology of suc- 
tion and gill irrigation in the tadpole 
of Ascaphus, and notes on hibernation. 
Herpetologica, 29:84-93. 

1974. Description of the tadpole of Phry- 
nomerus annectans, and comments on 
its gill irrigation mechanism. Herpe- 
tologica, 30:53-62. 

1975a. The bearing of filter feeding on the 
water pumping mechanisms of Xeno- 
pus tadpoles (Anura: Pipidae). Acta. 
Zool., 56:119-128. 

1975b. Experiments on oral suction and gill 
breathing in five species of Australian 
tadpole (Anura: Hylidae and Lepto- 
dactylidae). Jour. Zool. Lond., 177: 
81-98. 

GriFFITHS, I. 

1961. The form and function of the fore-gut 
in anuran larvae (Amphibia, Salien- 
tia) with particular reference to the 
manicotto glandulare. Proc. Zool. Soc. 
London, 137:249-283. 

1963. The phylogeny of the Salientia. Biol. 
Rev., 38:241-292. 

GrirFitTus, I., and A. L. DE CarvALHO 

1965. On the validity of employing larval 
characters as major phyletic indices 
in Amphibia, Salientia. Rev. Brasil 
Biol., 25:113-121. 

HAMMERMAN, D. L. 

1964. Occurrence of premetamorphic papil- 
lae in ranid tadpoles. Amer. Zool., 
4:319. 

1967. Lingual premetamorphic papillae as 
larval taste structures in frogs. Na- 
ture, London, 215:98-99. 

1969. The frog tongue: I. General develop- 
ment and histogenesis of filiform pa- 
pillae and mucous glands in Rana 
catesbeiana. Acta. Zool., 50:11-23. 

HE.LFF, O. M., and M. C. MELLICKER 

1941. Studies on amphibian metamorphosis. 
XIX. Development of the tongue in 
Rana sylvatica, including histogenesis 
of “premetamorphic” and filiform pa- 
pillae and the mucous glands. Amer. 
Jour. Anat., 68:339-366. 

HENpricks, F. S. 

1973. Intestinal contents of Rana pipiens, 
Schreber (Ranidae) larvae. South- 
west. Nat., 18:93-114. 

Heyer, R. W. 

1973. Ecological interactions of frog larvae 
at a seasonal tropical location in Thai- 
land. Jour. Herp., 7:337-361. 

1974. Niche measurements of frog larvae 


form a seasonal tropical location in 
Thailand. Ecology, 55:651-656. 

1975. A preliminary analysis of the inter- 
generic relationships of the frog fam- 
ily Leptodactylidae. Smithsonian Con- 
trib. Zool., 199:1-55. 

1976. Studies in larval amphibian habitat 
partitioning. Smithsonian Contrib. 
Zool., 2421-27. 

HoocGMoep, M. S. 

1967. Mating and early development of 
Gastrotheca marsupiata (Dumeril and 
Bibron) in captivity (Hylidae, Anura, 
Amphibia). Brit. Jour. Herpetology, 
4:1-7. 

Hora, S. L. 

1930. Ecology, bionomics and evolution of 
the torrential fauna, with special ref- 
erence to the origins of attachment. 
Phil. Trans. Roy. Soc., London, Series 
B., 218:171-288. 

INGER, R. F. 

1966. The systematics and zoogeography of 
the Amphibia of Bomeo. Fieldiana, 
Zool., 52:1-402. 

JENSSEN, T. A. 

1967. Food habits of the green frog, Rana 
clamitans, before and during meta- 
morphosis. Copeia, 1967:214-218. 

Kenny, J. S. 

1969a. Feeding mechanisms in anuran lar- 
vae. Jour. Zool., London, 157:225- 
246. 

1969b. Pharyngeal mucous secreting epithe- 
lia of anuran larvae. Acta Zool., 50: 
143-153. 

KRATOCHWILL, K. 

1933. Zur Morphologie und Physiologie der 
Nahrungsaufnahme der Froschlarven. 
Zeit. wiss. Zool., 144:421-468. 

LaBarsera, M. C. 

1978. Particle capture by a Pacific brittle 
star: an experimental test of the 
aerosol suspension feeding model. 
Science, 201:1147-1149. 

LANZA, B., J. M. Cei, and E. G. Crespo 

1975. Immunological evidence for the spe- 
cific status of Discoglossus pictus, 
Otth 1837 and D. sardus, Tschudi 
1837, with notes on the families Dis- 
coglossidae, Giinther 1858 and Bom- 
binidae, Fitzinger 1826 (Amphibia 
Salientia). | Monitore Zool.  Ital., 
(N.S.), 9:153-162. 

1976. Further immunological evidence for 
the validity of the family Bombinidae 
(Amphibia Salientia). Monitore Zool. 
Ttale (NES) pelO: Silesia 

LATASTE, F. 

1879. Etuda sur la discoglosse. Actes Soc. 
Linn. Bordeaux., 33:275-341. 

bingy, (Ca (Ge 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 141 


1950. Amphibians of western China. Fieldi- 

ana Zool. Memoirs, 2:1-400. 
LUTHER, A. 

1914. Uber die vom N. trigeminus versorgte 
Muskulatur der Amphibien, mit einem 
vergleichenden Ausblick iiber den Ad- 
ductor mandibulae der Gnathostomen, 
und einem Beitrag zum Verstandnis 
der Organisation der Anurenlarven. 
Acta Soc. Sci. Fennicae., 44(7):1-151. 

Lyncu, J. D. 

1971. Evolutionary relationships, osteology, 
and zoogeography of Leptodactyloid 
frogs. Univ. Kansas Publ. Mus. Nat. 
Hist., 53: 1-238. 

1973. The transition from archaic to ad- 
vanced frogs. pp. 133-182. In: Evo- 
lutionary Biology of the Anurans. 
J. L. Vial, ed., Univ. Missouri Press, 
Columbia, 470 pp. 

Lynn, W. G. 

1942. The embryology of Eleutherodactylus 
nubicola, an anuran which has no 
tadpole stage. Contributions to Em- 
bryology, 30( 190) :29-62. 

1961. Types of amphibian metamorphosis. 
Amer. Zool., 1:151-161. 

Magni, E. 

1959. Anatomie du tetard d’ Alytes obste- 
tricans. Actes Soc. Linn. Bordeaux, 
98: 1-60. 

Maxson, L. R. 

1976. The phylogenetic status of phyllome- 
dusine frogs (Hylidae) as evidenced 
from immunological studies of their 
serum albumins. Experientia, 32: 
1149-1150. 

1977. Immunological detection of conver- 
gent evolution in the frog Anotheca 
spinosa. Syst. Zool., 26:72-76. 

Maxson, L. R., and A. C. WiLson 

1975. Albumin evolution and _ organisimal 
evolution in tree frogs (Hylidae). 
Syst. Zool., 24:1-15. 

MENZIES, J. I. 

1967. An ecological note on the frog Pseud- 
hymenochirus merlini Chabanaud in 
Sierra Leone. Jour. West Afr. Sci. 
Association, 12:23-28. 

MILLARD, N. 

1943. The development of the arterial sys- 
tem of Xenopus laevis, including ex- 
periments on the destruction of the 
larval aortic arches. Trans. R. Soc. 
S. Afr., 30:217-239. 

Morton, J. 

1967. Guts. Studies in Biology no. 7. St. 

Martin’s Press, New York, 59 pp. 
NELSON, C. E., and H. S. CuELLAR 

1968. Anatomical comparison of tadpoles of 
the genera Hypopachus and Gastro- 
phryne (Microhylidae). Copeia, 1968: 
423-424, 


Nos ie, G. K. 

1927. The value of life history data in the 
study of the evolution of the Am- 
phibia. Ann. New York Acad. Sci., 
30:31-128. 

Oxapa, Y. 

1931. Tailless Batrachians of Japanese Em- 
pire. Imp. Agric. Exp. Stan., pp. 1- 
215. 

Orton, G. L. 

1943. The tadpole of Rhinophrynus dorsalis. 
Occas. Papers Mus. Zool., Univ. Mich- 
igan, 472: 1-7. 

1944. Studies on the systematic and phylo- 
genetic significance of certain larval 
characters in the Amphibia Salientia. 
Unpublished Ph.D. thesis. Univ. 
Michigan, Ann Arbor, Michigan, 254 
pp. 

1953. The systematics of vertebrate larvae. 
Syst. Zool., 2:63-75. 

1957. The bearing of larval evolution on 
some problems in frog classification. 
Syst. Zool., 6:79-86. 

ParkER, H. W. 

1931. Reports on an expedition to Brazil 
and Paraguay in 1926-7, supported by 
the trustees of the Percy Sladen Me- 
morial Fund and the executive com- 
mittee of the Carnegie Trust for Scot- 
land—Amphibia and Reptiles. Jour. 
Linn. Soc., London, Zool., 37:285- 
289. 

1934. A monograph of the frogs of the fam- 
ily Microhylidae. British Mus., Lon- 
don, 208 pp. 

ParKER, W. K. 

1881. The structure and development of the 
skull in the Batrachia. Phil. Trans. 
Roy. Soc., 172:1-266. 

Pope, C. H. 

1931. Notes on amphibians from Fukien, 
Hainan and other parts of China. 
Bull. Amer. Mus. Nat. Hist., 61(8): 
397-611. 

RICHMOND, N. D. 

1947. Life history of Scaphiopus holbrookii 
holbrookii (Harlan). Part I: Larval 
development and behavior. Ecology, 
28 :53-67. 

RUBENSTEIN, D. I., and M. A. R. KoEHL 

1977. The mechanism of filter feeding: some 
theoretical considerations. Amer. Nat- 
uralist, 111:981-994. 

SAVAGE, R. M. 

1952. Ecological, physiological and ana- 
tomical observations on some species 
of anuran tadpoles. Proc. Zool. Soc., 
London, 122:467-514. 

1955. The ingestive, digestive and respira- 
tory systems of the microhylid tad- 
pole, Hypopachus aguae. Copeia, 
1955:120-127. 


142 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


1961. The ecology and life history of the 
common frog. Sir Isaac Pitman & 
Sons, Ltd., London, 228 pp. 

SAVAGE, J. M. 

1968. The dendrobatid frogs of Central 
America. Copeia, 1968:745-776. 

1973. The geographic distributions of frogs: 
patterns and predictions, pp. 325-445. 
In: Evolutionary Biology of the Anu- 
rans. J. L. Vial, ed., Univ. Missouri 
Press, Columbia. 470 pp. 

SCHMALHAUSEN, I. I. 

1968. The origin of terrestrial vertebrates 
(Translated from Russian). Academic 
Press, New York, 314 pp. 

SCHULZE, F. E. 

1870. Die Geschmocksorgane der Frosch- 
larven. Arch. mikr. Anat., 6:407-419. 

1892. Uber die inneren Kiemen der Ba- 
trachierlarven. II. Mittheilung. Ske- 
let, Musculatur, Blutgefasse, Filterap- 
parat, respiratorische Anhange und 
Athmungsbewegungen — erwachsener 
larven von Pelobates fuscus. Abhdl. 
der Konig]. Preuss., Akad. Wiss., Ber- 
lin 1892(3) :1-66. 

1889. Uber die inveren Kiemen der Ba- 
trachier Jarven. I. Mittheilung. Uber 
das epithel der Lippen, der Mund-, 
Rachen, und Kiemenhdhle erwach- 
sener Larven von Pelobates fuscus. 
Abhdl. der Konigl. Preuss., Akad. 
Wiss., Berlin, 1889(3):715-768. 

SEALE, D. B. 

1973. Impact of amphibian larval popula- 
tions on an aquatic community. Un- 
published Ph.D. thesis. Washington 
Univ., St. Louis, Mo., 167 pp. 

SEALE, D. B., and N. BECKvAR 

1980. The comparative ability of anuran 
larvae (Genera: Hyla, Bufo and 
Rana) to ingest suspended blue-green 
algae. Copeia (in press). 

SEALE, D., and R. WASSERSUG 

1979. Suspension feeding dynamics of anu- 
ran larvae related to their functional 
morphology. Oecologia, 39:259-272. 

SEVERTZOV, A. S. 

1969. Food seizing mechanism of Anura 
larvae. Dokl. Akad. Nauk. SSSR, 187: 
211-214 (Trans. ). 

Sir, M. A. 

1916. Description of five tadpoles from 
Siam. Jour. Nat. His. Soc. Siam, 2: 
37-43. 

1926. The function of the “funnel” mouth 
of the tadpoles of Megalophrys, with 
a note on M. aceras Boulenger. Proc. 
Zool. Soc., London, 96:983-988. 

SoxoL, O. M. 

1962. The tadpole of Hymenochirus boett- 

geri. Copeia, 1962:272-284. 


1975. The phylogeny of Anura larvae: a 
new look. Copeia, 1975:1-23. 

1977a. The free swimming Pipa larvae, with 
a review of pipid larvae and_pipid 
phylogeny (Anura: Pipidae). Jour. 
Morph., 154:357-426. 

1977b. A subordinal classification of frogs 
(Amphibia: Anura). Jour. Zool., 182: 
505-508. 

STARRETT, P. H. 

1960. Descriptions of tadpoles of Middle 
American frogs. Misc. Publ. Mus. 
Zool. Univ. Michigan, 110:1-38. 

1968. The phylogenetic significance of the 
jaw musculature in anuran amphib- 
ians. Ph.D. dissertation, Univ. Michi- 
gan. 

1973. Evolutionary patterns in larval mor- 
phology, pp. 251-271. In: Evolution- 
ary Biology of the Anurans. J. L. Vial, 
ed., Univ. Missouri Press, Columbia. 
470 pp. 

STEBBINS, R. C. 

1951. Amphibians of western North Amer- . 
ica. Univ. California Press, Los An- 
geles, 539 pp. 

STEBBINS, R. C., and J. R. HENDRICKSON 

1959. Field studies of amphibians in Co- 
lombia, South America. Univ. Cali- 
fornia, Pub. Zool., 5:497-540. 

STEINWASCHER, K. 

1978. Interference and exploitation compe- 
tition among tadpoles of Rana utric- 
ularia. Ecology, 59:1039-1046. 

STEPHENSON, N. G. 

1950a. On the development of the chondro- 
cranium and visceral arches of Leio- 
pelma archeyi. Trans. Zool. Soc., Lon- 
don, 27:203-253. 

1950b. Observations on the development of 
the amphicoelous frogs, Leiopelma 
and Ascaphus. Jour. Linn. Soc. Zool., 
42::18-28. 

1955. On the development of the frog Leio- 
pelma hochstetteri Fitzinger. Proc. 
Zool. Soc., London, 124:785-801. 

STERBA, G. 

1950. Uber die Morphologischen und Histo- 
genetischen Thymusprobleme bei 
Xenopus laevis Daudin nebst einigen 
Bermerkungen tber die Morphologie 
der Kaulquappen. Abh. Sachsischen 
Akad. Wiss. Leipzig Math.-naturwiss. 
Klasse, 44(1):1-54. 

Stuart, L. C. 

1961. Some observations on the natural his- 
tory of tadpoles of Rhinophrynus dor- 
salis Dumeril and Bibron. Herpeto- 
logica, 17:73-79. 

UEck, M. 

1967. Der Manicotto glandulare (“Driisen- 
magen”) der Anurenlarve in Bau, 
Funktion und Beziehung zur Gesam- 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 143 


tranges des Darmes. Eine mikroskop- 
isch-anatomische, histochemische und 
elektronenoptische Studie an der om- 
nivoren und mikrophagen Larve von 
Xenopus laevis und der carnivoren 
und makrophagen Larve von Hymen- 
ochirus boettgeri (Anura, Pipidae). 
Zeit. wiss. Zool., 176:173-270. 
VAN EEDEN, J. A. 

1951. The development of the chondro- 
cranium of Ascaphus truei Stejneger 
with special reference to the relations 
of the palatoquadrate to the neuro- 
cranium. Acta Zool., 32:41-176. 

VILLA, J. 

1971. Anfibios de Nicaragua. Instituto Geo- 
grafico Nacional and Banco Central 
de Nicaragua, Managua, 217 pp. 

WAKE, D. B. 

1966. Comparative osteology and evolution 
of the lungless salamanders, family 
Plethedontidae. Mem. S. Cal. Acad. 
Sci. 4:1-111. 

WASSERSUG, R. J. 

1972. The mechanism of ultraplanktonic en- 
trapment in anuran larvae. Jour. 
Morph., 137:279-288. 

1973. Aspects of social behavior in anuran 
larvae, pp. 272-297. In: J. L. Vial, 
ed. Evolutionary Biology of the Anu- 
rans. Univ. Missouri Press, Columbia. 
470 pp. 

1975. The adaptive significance of the tad- 
pole stage with comments on the 
maintenance of complex life cycles in 
anurans. Amer. Zool., 15:405-417. 

1976a. Oral morphology of anuran larvae: 
terminology and general description. 
Occ. Papers Mus. Nat. Hist. Univ. 
Kansas, 48:1-23. 

1976b. Internal oral features in Hyla regilla 
(Anura: Hylidae) larvae: an onto- 
genetic study. Occ. Papers Mus. Nat. 
Hist. Univ. Kansas, 49:1-24. 

Wassersuc, R. J., and K. Horr 

1979. A comparative study of the buccal 
pumping mechanism of tadpoles. Sub- 
mitted to Biol. Jour. Linn. Soc., 12: 
225-259. 


Wassersuc, R. J., and K. ROSENBERG 

1979. Surface anatomy of branchial food 
traps of tadpoles: a comparative 
study. Jour. Morph., 159:393-426. 

WassersuG, R. J., and E. SEIBERT 

1975. Behavioral responses of amphibian 
larvae to variation in dissolved oxy- 
gen. Copeia, 1975:86-103. 

WEiIsz, P. B. 

1945. The development and morphology of 
the larva of the South African clawed 
toad, Xenopus laevis. Jour. Morph., 
77:163-217. 


Wivper, IJ., L. WHIPPLE and E. R. DunN 
1920. The correlation of lunglessness in sal- 
amanders with a mountain brook hab- 
itat. Copeia, 1920:63-68. 
Wricut, A. H. 
1932. Life-histories of the frogs of Okefi- 
nokee swamp, Georgia. Macmillian 
Co., New York, 497 pp. 


APPENDIX: GLOSSARY OF TERMS 


Most terms used in this monograph 
are explained briefly here; fuller descrip- 
tions and definitions are provided for the 
majority of terms in Wassersug (1976a). 
Morphological variation for internal oral 
features listed here is reviewed under 
“Functional Considerations” in the Dis- 
cussion section of this paper. 


Advanced tadpoles—Orton’s Type 4 
larvae. 

Anterior filter valve—see ventral 
velum. 

Anterior narial papillae—papillae 
arising from the anteromedial corner of 
the internal naries and projecting pos- 
teroventrally over the narial passage. 

Arboreal tadpoles—any larvae that 
live in small pools of water which form 
in leaves of vegetation above ground. 

Atrial chamber—the chamber sur- 
rounding the gill filaments between the 
gill slits and the spiracle. 

Beaks—the keratinized structures 
sheathing the supralabial and infralabial 
cartilage in most tadpoles, except Orton 
Type 1 and 2. 

BFA—see buccal floor arena. 

BRA—see buccal roof arena. 

Branchial food traps—regions of se- 
cretory mucosa covering the roof and 
the anterior wall of the filter cavities in 
most tadpoles. The traps usually cover 
the whole ventral surface of the ventral 
velum. 

Buccal cavity—the internal portion 
of the mouth above the ceratohyal and 
the hypobranchial plate, anterior to the 
dorsal and ventral vela. 

Buccal pockets—deep depressions in 
the buccal floor between the posterior 
margin of the ceratohyal and the anterior 
margin of the first ceratobranchial (= 
first gill cleft). 


144 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Buccal slits—perforations of the buc- 
cal pocket (first gill cleft) which con- 
nect the buccal cavity to the atrial cham- 
ber. 

Buccal floor arena (BFA)—a large, 
circumscribed area in the center of the 
buccal floor. 

Buccal floor arena papillae (BFA pa- 
pillae )—rows of papillae that define the 
BFA. 

Buccal roof arena (BRA)—defined 
area on the buccal roof directly above 
the buccal floor arena. 

Buccal roof arena papillae (BRA 
papillae)—papillae that circumscribe 
and thus define the buccal roof arena. 

cb.—see ceratobranchials. 

Ceratobranchials (cb.)—the cartila- 
ginous, skeletal elements of the branchial 
baskets; the gill bars. 

Ceratohyal—the major cartilaginous 
skeletal elements of the buccal floor in 
tadpoles. 

Choanae—see internal naries. 

Ciliary groove—a horizontal groove 
that runs in the posterior and _ lateral 
margin of the pharynx from the anterior, 
lateral corner of the ventral velum to the 
esophagus. It is covered with cilia that 
transport mucus and food particles to 
the esophagus. 

Collecting organs—see crescentic or- 
gans. 

Crescentic organs—small, isolated 
branchial food traps of crescentic shape 
that lie at the anterior limit of the filter 
cavities in microhylid and rhinophrynid 
larvae. 

Denticle formulae—numerical expres- 
sion of the number of rows of denticles 
above and below the oral orifice (e.g., 
two rows above, three rows below = a 
2/3 denticle formula). 

Denticles—keratinized structures ar- 
ranged in transverse rows that surround 
the beaks of most tadpoles. 

Dorsal velum—a_ transverse flap 
across the posterior part of the mouth 
that separates the roof of the buccal 
cavity from the roof of the pharynx; the 
dorsal continuation of the ventral velum 
(= posterior filter valve). 


Esophageal funnel—the region where 
the pharynx narrows into the esophagus. 

Filter canals—partially or fully coy- 
ered passages between neighboring filter 
folds on the filter plates. 

Filter chambers—sections of the 
branchial baskets bounded laterally and 
medially by the filter plates, superiorly 
by the ventral velum, and ventrally by 
the gill slits. There are three filter cavi- 
ties on each side. 

Filter crevices—small passageways 
between two neighboring side folds on 
a single filter row. 

Filter folds—narrow rows of gill fil- 
ters that run from dorsal to ventral down 
the filter plates. 

Filter niches — open, multi-sided 
spaces between secondary and tertiary 
filter folds on filter canals. 

Filter plates — vertically oriented 
plates of connective tissue that arise 
from the ceratobranchials, which sup- 
port the gill filter folds. 

Filter rows—see filter folds. 

Filter ruffles—see filter folds. 

Funnel-mouth tadpoles—any larva 
with an exceptionally large expanded 
oral disc that is directed anterior or 
anterodorsally. The oral disc in these 
forms is usually free of denticles. 

Gill filters—ruffled, epithelial organs 
associated with the posterior surfaces of 
cb. 1, the anterior surfaces of cb. 4, and 
both the anterior and posterior surfaces 
of cb. 2 and 3. 

Glandular zone—a transverse band 
of secretory membrane that covers the 
posterior portion of the buccal roof to 
the dorsal velum. 

Glottis—longitudinal slit-like open- 
ing to the bronchi that lies on the mid- 
line between the anterior portions of the 
4th ceratobranchials. 

Hypobranchial plate—the cartilagi- 
nous skeletal element between the cera- 
tobranchials and the ceratohyal. It sup- 
ports the posterior portion of the buccal 
floor. 

Infralabial papillae—the most ante- 
rior projections in the buccal cavity, 
which lie over the infralabial cartilage 


ORAL FEATURES OF LARVAE FROM EIGHT ANURAN FAMILIES 145 


and its articulation with Meckel’s carti- 
lage. 

Internal naries—slit-like openings of 
the anterior buccal roof. 

Laryngeal disc—the circular, tra- 
cheal outline surrounding the glottis. 

Lateral ridge papillae—subsidiary 
projections that lie lateral to the median 
ridge on the buccal roof. 

Lateral roof papillae—any small pro- 
jections that lie in the midportion of the 
buccal roof lateral to the BRA papillae. 

Lingual papillae — premetamorphic 
projections that occur on the tongue an- 
lage. 

Main filter fold—the large, central 
ruffle of each filter row. 

Median notch—a notch in the mid- 
dle of the ventral velum. 

Median ridge—a transversely orient- 
ed ridge or epithelial flap that lies in 
the midline of the buccal roof, approxi- 
mately halfway between the upper beak 
and the dorsal velum. 

Narial valve—extensive, free flaps of 
the posterior walls of the internal naries. 

Narial valve projection—single, tall 
projections of the narial valve. 

Opercular chamber—see atrial cham- 
ber. 

Operculum—flap of skin arising from 
the hyoidean arch which covers the gill 
filaments. 

Oral cavity—the region of the ali- 
mentary and respiratory system between 
the oral orifice anteriorly, the gill slits 
ventrally, and the esophagus posteriorly. 

Oral disc—an expanded flap of skin 
surrounding the oral orifice. Denticle 
rows occur on the oral disc in most 
Orton Type 3 and 4 larvae. 

Oral orifice—the entrance into the 
mouth or oral cavity. 

Orton Type 1 Larvae—tadpoles that 
have paired spiracles, and lack kerati- 
nized mouth parts (Pipidae and RBhi- 
nophrynidae ). 

Orton Type 2 Larvae—tadpoles lack- 
ing keratinized mouth parts and having 
a single, medial spiracle ( Microhylidae ). 

Orton Type 3 Larvae—tadpoles with 
keratinized mouth parts and a medial 


spiracle (Ascaphidae and Discoglossi- 
dae). 

Orton Type 4 Larvae—tadpoles hav- 
ing keratinized mouth parts and a single, 
sinistral spiracle (all families except 
those listed under Orton Types 1-3). 

Pharyngeal by-pass—see buccal slits. 

Pharyngeal cavity—the region of the 
oral cavity that lies between the dorsal 
and ventral vela anteriorly and the 
esophagus posteriorly. It is bounded 
ventrally by the pharyngeal gill slits. 

Pond larvae, typical—tadpoles such 
as Hyla regilla and Acris crepitans which 
have a 2/3 denticle formula, live in small 
or medium-size ponds, and feed on sus- 
pended particles in the water column 
and graze on the substrate. 

Posterior filter valve—see dorsal vel- 
um. 

Postnarial arena—the region of the 
buccal roof between the nares anteriorly 
and the transverse median ridge poste- 
riorly. 

Postnarial papillae—papillae between 
the internal naries and the median ridge. 
These are usually in distinct, obliquely 
oriented rows from anteromedial to pos- 
terolateral behind the narial valve. 

Postnarial ridges—epithelial folds in 
the postnarial arena. 

Prenarial arena—the region of the 
buccal roof between the supralabial car- 
tilages anteriorly and the internal naries 
posteriorly. 

Prenarial papillae—see anterior na- 
rial papillae. 

Prepocket papillae—papillae of the 
buccal floor that lie over the body of 
the ceratohyal anterior to the buccal 
pocket. 

Pressure cushions—loose folds of epi- 
thelium descending from the roof of the 
pharynx posterior to the dorsal velum. 
There are usually two pressure cushions 
on each side of the pharynx. 

Secretory pits—clusters of secretory 
cells that open in distinct pits along the 
posterior margin of the ventral velum 
and the glandular zone of the buccal 
roof. 


146 MISCELLANEOUS PUBLICATION MUSEUM OF NATURAL HISTORY 


Secretory ridges — mucus-secreting 
ridges that cover the branchial food trap 
surfaces on the ventral side of the ven- 
tral velum in most tadpoles (see Wasser- 
sug and Rosenberg, 1979). 

Secretory zone—the exposed regions 
of mucus cell apices in the bottom of 
secretory pits and at the top of secretory 
ridges. 

Side folds—secondary and _ tertiary 
branching patterns of the gill filters of 
each filter row. 

Spicules—cartilaginous supports for 
the free, posterior surface of the ventral 
velum. 


Spiracle—the opening in the opercu- 
lum where water is expelled from the 
atrial chamber. 

Suctorial tadpoles—larvae with an 
enlarged, oral disc used to adhere to 
substrate often in lotic environments. 

Tongue anlage — the embryonic 
tongue pad that develops in the an- 
terior portion of the buccal floor. 

Ventral velum—a distinctive, mova- 
ble (but nonmuscular) flap arising from 
the floor of the buccal cavity; it is at- 
tached anteriorly to the branchial bas- 
kets and to the superior margins of the 
filter plates (= anterior filter valve). 


ERNST MAYR LIBRARY 


TU 


4 110 326 212 _ 


RECENT MISCELLANEOUS PUBLICATIONS 


32 oo. cycles in lizards and snakes. By Henry S. Fitch. Pp. 1-247, 16 fig- 
ures in text. June 19, 1970. Paper bound, $5.00. 


ne " i olutionary relationships, osteology, and zoogeography of leptodactyloid frogs. 
By John D. Lynch. Pp. 1-238, 131 figures in text. June 80, 1971. Paper bound, 
‘87. 00. 


54. The dentition of glossophagine bats: development, morphological characteristics, 
__-variation, pathology, and evolution. By Carleton J. Phillips. Pp. 1-138, 49 figures 
ius) am text. September 24, 1971. Paper bound, $5.00. 


ss Middle American lizards of the genus Ameiva (Teiidae) with emphasis on geo- 
graphic variation. By Arthur C. Echternacht. Pp. 1-86, 28 figures in text. De- 
_ cember 14, 1971. Paper bound, $3.00. 


Fis 56. Systematics of the chiropteran family Mormoopidae. By James Dale Smith. Pp. 
'___—- 1-182, 40 figures in text. March 10, 1972. Paper bound, $5.00. 


: 87, A systematic review of the Teiid lizards, genus Bachia, with remarks on Heter- 
_ odactylus and Anotosaura. By James R. Dixon. Pp. 1- 47, 15 figures in text. 
; Ree uaty 2, 1973. Paper bound, $1. 50. 


(58. Systematics of three species of woodrats (genus Neotoma) in central North Amer- 
ica. By Elmer C. Birney. Pp. 1-178, 44 figures in text. April 13, 1973. Paper 
___ bound, $7.50. 


_ 59. Systematics and evolution of ihe Andean lizard genus Pholidobolus (Sauria: 
_ Teiidae). By Richard R. Montanucci. Pp. 1-52, 8 figures in text. May 14, 1978. 
Paper bound, $1.75. 


- 60. Mammals of the Black Hills of South Dakota and Wyoming. By Ronald W. 
Tumer. Pp. 1-178, 14 figures in text. April 3, 1974. Paper bound, $7.50. 


61. Reproductive strategies in a tropical anuran community. By Martha L. Crump. 
Pp. 1-68, 13 figures in text. November 15, 1974. Paper bound, $4.50. 


a7 62. A demographic study of the ringneck snake (Diadophis punctatus) in Kansas. By 
Henry S. Fitch. Pp. 1-53, 19 figures in text. April 3, 1975. Paper bound, $3.00. 


63. Morphology of the bony stapes (columella) in the Passeriformes and related 
groups: evolutionary implications. By Alan Feduccia. Pp. 1-34, 16 plates, 7 
figures in text. May 30, 1975. Paper bound, $3.00. 


64. The phylogency and biogeography of fossil and Recent gars (Actinopterygii: 
Lepisosteidae). By E. O. Wiley, Pp. 1-112, if figures in text. November 12, 1976. 
Paper bound, $2.75. 


_ 65. The biology of an equatorial herpetofauna in Amazonian Ecuador. By William 
: E. Duellman. Pp. 1-352, 198 figures in text. August 30, 1978. Paper bound, $15.00. 


66. Leptodactylid Frogs of the genus Eleutherodactylus from the Andes of southern 
Ecuador. By John D. Lynch. Pp. 1-62, 23 figures in text. February 28, 1979. 
Paper bound, $2.75. 


? Postage and handling 5% of total order or a minimum of $0. 50. Publications of this and other 
series are available from the Publications Secretary. 


OS