HARVARD UNIVERSITY

Library of the

Museum of

Comparative Zoology

OCCASIONAL PAPERS

of the

MUSEUM OF NATURAL HISTORY

The Universit^^ ^i^WMoi.
Lawrence, Kansa^lBRARY

NUMBER 23, PAGES 1-40 FEB 1 1 fflTJ^^^^^^^ ^^^ ^^'^^

SPECIATION IN FROGS C^J^J^^^^^ PARVICEPS
GROUP IN THE UPPER MIAZON BASIN

By
William E. Duellman^ and Martha L. Crump-
As our knowledge of tlie ampliibians in tlie upper Amazon Basin
in South America increases, it becomes evident tliat tlie richness of
the frog fauna is due to tlie occurrence there of representatives of
many species groups and in some cases to the occurrence of several
sympatric species l)c>longing to one group. The latter phenomenon
is especially noticeable in the //(//« leucoplujUata and Ihjla parvi-
ceps groups with five and three species, respectively. The present
paper deals only with the latter group.

The primary purposes of this paper are to present analyses and
interpretations of the morphological features of adults and tad-
poles, mating calls, breeding behavior, and ecological relationships
of the three species at Santa Cecilia in Amazonian Ecuador. How-
ever, as in most Amazonian frogs, some taxonomic problems must
be dealt with before the biological problems can be discussed ef-
fectively. Thus, we also present here a systematic review of the
entire^ Ilyla parviceps group, as presently understood.

Acknowledgments

We are indebted to the following persons for the loan of speci-
mens or for the provision of working space in their respective insti-
tutions: Werner C. A. Bokermann, Alice G. C. Grandison, Konrad
Klemmer, Jean Lescure, Alan E. Leviton, Hymen Marx, Charles W.
Myers, Giinther Peters, the late James A. Peters, Douglas A. Ross-

* Curator, Division of Herpetology, Museum of Natural History, The Uni-
versity of Kansas, Lawrence, Kansas 66045.

- Research Assistant, Di\'ision of Herpetology, Museum of Natural History,
The University of Kansas, Lawrence, Kansas 66045.

2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

man, Dorothy Smith, and Charles F. Walker. The field studies on
these frogs were carried out at Santa Cecilia, Ecuador, where facili-
ties were provided by Ing. Ildefonso Munoz B., to whom we extend
our gratitude. Many of the specimens from Ecuador were collected
by our field associates, who worked with us for varying periods of
time from 1967 to 1972; we especially thank W. C. A. Bokermann,
Stephen R. Edwards, Thomas H. Fritts, John D. Lynch, John E.
Simmons, Linda Trueb, James W. Waddick, and Charles F. Walker.
We are grateful to Jan Caldwell for analyzing the recordings of the
mating calls, Albert E. Fisher and John E. Simmons for recording
data, and Linda Trueb for executing the illustrations of the tad-
poles.

Field work in Ecuador was partly supported by Watkins Mu-
seum of Natural History Grants and the F. William Saul Fund of
the Museum of Natural History, and a grant from the Graduate
School, The University of Kansas; further support was received
from the National Science Foundation (GB 29557) and the Organi-
zation for Tropical Studies. Duellman's study of specimens in
European museums was made possible by the American Philo-
sophical Society (Pem-ose Fund No. 5063). The final work on this
project is part of a study on the herpetofauna of Santa Cecilia, sup-
ported by the National Science Foundation (GB 35483).

Materials and Methods

The systematic aspects of this study are based on the examina-
tion of 720 presei'ved specimens, 26 cleared and stained specimens,
and 19 lots of tadpoles. Recordings were made on a Uher-4000 Re-
corder and analyzed on a Vibralyzer (Kay Electric Company).
All measurements of morphological characters and calls were taken
in the manner described by Duellman (1970). Webbing formulae
were determined in the manner described by Savage and Heyer
(1967). Tadpoles were staged according to Gosner's (1960)
system.

All specimens are referred to by the following abbreviations:

AMNH American Museum of Natinal History

BMNH British Museum (Natural History)

CAS California Academy of Sciences

CM Carnegie Museum

FMNH Field Musemn of Natural History

KU University of Kansas Museum of Natural History

LG Lescure-Guiana ( Jean Lescure, Paris )

LSU Louisiana State University Museum of Zoology

NHMB Naturhistorisches Museum Basel

SMF Senckenbergische Museum Frankfurt

UIMNH University of Illinois Museum of Natural History

UMMZ University of Michigan Museum of Zoology

FROGS OF THE HYLA PARVICEPS GROUP 3

USNM-GO\^ United States National Museum (Gustavo Oices V. collection)
WCAB Werner C. A. Bokemiann, Sao Paulo, Brasil

ZMB Zoologisches Museum Berlin

SYSTEMATICS
The Hijla parviceps Group

Definition. — 1) Great sexual dimorphism in size; snout-vent
lengths to 25 mm in males, 32 mm in females; 2) snout short, blunt;
3) tympanum visible, but tympanic ring indistinct or absent; 4)
hands and feet moderately webbed; 5) axillary membrane abbrevi-
ated; 6) thoracic glands absent; 7) calcars, ulnar and tarsal folds
absent; 8) males having single, median, subgular vocal sac and no
nuptial excrescences; 9) dorsum tan or brown with dark brown
markings, with or without dorsolateral light stripes; 10) pale,
vertical suborbital bars present; 11) thighs marked with cream,
yellow, or orange spots; 12) anal, ulnar, and tarsal stripes absent;
13) iris pale gray with red ring around pupil; 14) nasals small,
abutting or slightly overlapping sphenethmoid; 15) tectum nasi
and septum nasi perichondralh' ossified and synosteotically united
with sphenethmoid; 16) frontoparietal fontanelle covered, or nearly
so, by frontoparietals; 17) squamosals articulating with prootics;
18) quadratojugal reduced to small spur, or absent, not articulating
with maxillary; 19) prevomerine teeth on posteromedially directed
dentigerous processes between choanae; 20) presacral vertebrae
nonimbricate, lacking neural crests, ha\ ing trans\erse processes de-
creasing in length posteriorly (3,6 perpendicularly transverse; 4,5
inclined posteriorly; 7,8 inclined anttuiorh ) ; 21) sacral diapophyses
expanded 53-77°, with convex edges; 22) tadpoles having ovoid
bodies and xiphicercal tails with moderately deep fins not extend-
ing onto body; 23) larval mouths anteroventral with one row of
large labial papillae laterally and ventrally, robust serrate beaks,
and no more than one row of denticles; 24) mating call consisting
of short, high-pitched notes, followed or not by shorter secondaiy
notes; 25) diploid chromosome number 30.

Content. — Six species: Hijla bokernmnni Goin, 1960; H. brevi-
frons new species; H. luteoceUata Roux, 1927; H. microps Peters,
1872; H. parviceps Boulenger, 1882; H. subocularis Dunn, 1934.

Distribution. — The Amazon Basin, coastal lowlands and slopes
of southeastern Brasil, northern South America from French Guiana
to Colombia, and eastern Panama.

Comment. — The Hijla parviceps group is one of four groups of
small Neotropical frogs that share many of the characters noted
above. The other groups are — the Hijla leucopJnjUata group with
about six species in South America and one in Central America, the
Hyla microcephala group with perhaps a dozen species in South
and Central America, and Hyla minuta, a widespread South Amer-

4 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

ican "species" (see Cochran and Coin, 1970, for taxonomic com-
ments). The Hi/ki leucophi/IUita and microcephala groups were de-
fined by Duelhnan (1970)."

Twelve of the 25 species in these four groups have a diploid
number of 30 chromosomes (Duellman, 1970; Bogart, 1973); the
chromosome numbers in the others are unknown. All known tad-
poles (IS species) have xiphicercal tails and reduced mouth parts.
In addition, the species in the four groups are consistent in char-
acters 4, 7, 8, IS, 19, 20, and 22. Members of the Hyla parviceps
group differ from all of the other three groups by having ( numbers
correspond to statements in preceding definition ) : 1 ) more pro-
nounced sexual dimorphism in size; 2) shorter snout; 3) tympanic
ring indistinct or absent; 5) much less extensive axillary mem-
brane; 9) sexual dimorphism in width of dorsolateral stripes; 10)
suborbital bars; 11) thighs patterned; 13) iris color; 15) more
perichondral ossification in the tectum nasi and solum nasi; 17)
squamosals articulating with prootics.

Although tadpoles of all of the groups have reduced mouth
parts, there are consistent differences among the groups. Members
of the Hyla porviceps and minuta groups have antero ventral
mouths; there is one row of papillae in porviceps and two in
minuta. Tadpoles of the Hyla parviceps group have no, or one,
row of denticles below the beaks; minuta has one row. Tadpoles
of the Hyla leucophyllata and microcephala groups have terminal
mouths lacking rows of denticles; labial papillae are present in the
former, absent in the latter.

The mating calls of members of the Hyla leucophyllata and
microcephala groups, and H. bokermanni in the parviceps group
consist of primary and secondary notes; the calls of Hyla minuta
and other members of the Hyla parviceps group consist of series of
short notes.

Cochran ( 1955 ) and Bokermann ( 1964 ) included Hyla microps
in the Hyla marmorata group, although Bokermann (1964) defined
the group in a more strict sense to include only four species ( Hyla
acreana, nuirmorata, melanargyrea, and senicula). These frogs dif-
fer from the groups discussed above by having more extensive
webbing, well-developed ulnar and tarsal folds, less cranial ossifi-
cation, and different kinds of mating calls. They are like the other
groups in general larval and karyological features.

Other named taxa might be associated with the Hyla parviceps
group or one of the other three groups; however, these species are
too poorly known at the present time to be assigned to any group.
These include:

Hyla schubarti Bokermann^ 1963. — Rondonia, Brasil, and Ama-
zonian lowlands of Peru.

FROGS OF THE HYLA PARVICEPS GROUP 5

Hijla leali Bokermann, 1964. — Rondonia, Brasil, and Amazonian
lowlands of Ecuador and Peru.

Hyla grandisonae Coin, 1966. — Guyana.

Hyla oliveae Cochran and Goin, 1970. — Leticia, Colombia.

Hyla riveroi Cochran and Goin, 1970. — Amazonian Colombia,
Ecuador and Peru.

Duellman (1969) named Hyla carnifex from the Pacific slopes
of the Andes of Ecuador and placed the species in the Hyla parvi-
ceps group. Although there are superficial similarities in adult
structure and coloration, tadpoles, and mating calls, Hyla carnifex
seems to be related to Hyla columhiana and variabilis in Colombia
and is now excluded from the Hyla parviceps group.

ANALYSIS OF CHARACTERS

Three categories of characters were analyzed in adult frogs —
morphometric, structural, and coloration. In addition, characters of
the mating calls and tadpoles were utilized.

Morphometric Characters

Seven measurements were taken on rc^presentative samples of
the six species. From these, five proportions were calculated
(Tables 1 and 2). Analyses of variance showed highly significant
differences (P ^ .001) in both sex(>s among species for the follow-
ing: snout-vent length, tibia IcMigth/snout-vent length, and head
width /snout- vent length. E(|ually significant differences exist for
head length/snout-vent length in females and for t\'mpanum/eye
in males. Significant differences (P -= .01) were found for foot
length /snout- vent length in males. The differences among species
in foot length/snout-vent length and tympanum/eye in females,
and in head length/snout-\'ent length in males were not significant.

Statistical comparisons of the sexes of three species revealed
significant differences in snout-vent length between males and fe-
males of all three and in different proportions in different species
(Table 1). Adult females of all six species are much larger than
adult males; this is especially noticeable in Hyla parviceps, in which
the smallest gravid females are larger than the largest males.

Structural Characters

The species comprising the Hyla parviceps group are alike in
most structural features; those features that are constant, or nearly
so, are listed in the definition of the group. The tongue is cordiform
in all species, deeply notched posteriorly in H. microps, and shal-
lowly notched in the otlu^rs. Ulnar and tarsal tubc^rcles are absent
in all species, except females of //. parviceps and both sexes of H.

6

OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

p
&

S"

"p

5

0)

t->

^

rs

O

O

n3

^

en

o

.2

W

'o

<u

cs"

p

c/:

Tjl^

o

0)

Oj

Si

U

f-«

•4-*

aj

^^

M-l

r^

o

CS

(/2

c«

p2

(—

o

^-

4^

s

5

a

(«

p

-i

a

r^

'i;

5

r^

en

CS

X

<

1)

03
,<5

n

en

fa

C/5

0)

iX

M

C

C3

Q

'X

Ml

3

U

m
o

3

u

o c/) c/5 en c/5 in

P z z Z Z P
o o

o CO t^ in t o

CO c o o o — I

CO o o o o o

odd

o c o

O O CO 00 CO CO

o 01 o-i -H ^ CO

p p p p p p

•-i d d CD d d)

00
p

CO
I-

in

in

CO

1 — (
o
CO

05
I-

CO

1— I

d

d

d

d

d

Ol

1—1
in

in

CO

CO
CO

CO

d

d

d

d

^^

in

00

i~^

c^

-4

d

CO <M CM CO

d d d d

■— I o CO in CO fS] t-

■"l 05 C O O O O

. ^ p c o o p

«o d d d o o o

in o in I- o o-i CD

oq t- -H oi 1— I 1— I CO

^ p p o o o p

c d d d d d d

■J o t
!iJ l^ 00

00
c .—I

1^

d

<;c -t 00 o

■— I Ol C ^H

-f CO CO "^^

d d> d d>

o r- 1 CD in CO o

rf in ^ CO CO in

^' d d d d d

I I I I I I

O CD in 00 05 CD

P 'Tf CO CM O] CO

t-^ d d d d d

-1
>

C/3

£ J J J J

a > > > > ^

— ^ > -Q > \

S c a c •$ 3

> ^ S ^ ^ g

I ^^ ^^ "^ TO

ca ^ TJ -a 5*

O rt rt c

^ OJ 0) "

3 .2 ■"

C ^ O ™ TO ^

g t:; o aj «J >>

lyD H fa ffi ffi H

CD '^ m a> m ^

q z 2 p z z

o o

O in CD CO (M ^

CD O O O O — I

CI o o p o o

d d d d d d

O CO 00 T Ol CD

o oi oi ^ ^ in

CO p p p p p

'-^ d d d d d

-t O Ot) CO CO

00 05 fM O (35 ^

-f -r "i^ CO oj -f

n d d d d d

C<1

CO CD

CD in n'
d

in

(M

"=0

in

o

-^ (M CD

CO CO in

in

(M

o o o o

o

CH- C<l
CM-

in -r C5 t^
-r CO ci oi

d d d

o in t

in

1— I

d

O CD

d d

o
d

in tM oi

l^ O '"^ o o

o

CM

-r m -r
—J d

- o^

O CD 00 00

I- in

ci

o
o

d
oi

o o

7= t*

CO

CO

CO I-

o o

c o

d d

o -t" ci 00 I- in

CO

d

ol M C5 00 ^
O Ol CI 35 O

CO ci CO

Cl

CO CO
d d

CD 00 O CD CO

-r CO CO (M CO

d d d d d

>

C/2

■£ J J J J
tt > > > > ^

£ ^. ^ ^ ^ fa

^ a; % G

^ *" c ^ ^ c
<-• T:i c ^ ^^ >,
c/2 H fa S S E-H

O C/5 CA) O O <Z)

q z z P q z

o o o

c=; in t CI oi 00

C^i o o o o o

■-H o o o o o

d d d d d d

c in ^ (M oi c^

t^ CI CJ 1— I ^ Tf

c:;5 o o o o o

d d d d d d

CD

in o

1; in

CO

-f C75 00 CI

CI 05 00 c;5

t CI CI CO

d d d d d

CD

cq

in CD CI -H d

in -t CO CO in

d d d d d

CD 00 00 CD "*

-^ CO CI ci CO

^ d d d d d

p

Cl

ci ~I^

CO CO CI CO

o o o o

o o o o c o

*o

lO CD CI

S f^i f^i

<^ t- o

g_ H ^ -:.,-; ^:

CJ

5

t- -p O Cl

-H ^H 1-H Tfl

o o o o

d d d d

o "-f in 00 CO 00
-t a> CI ci o ci

I-

CD

-S -2

CO CO CO

d

o in t- in CO in
CO in -t CO CO m;

(Xi d d d d d

o Tp 00 c 00 in

C] ^ CO CO ci ci

v^ d d d d d

>

£ J J jJ J

M >

O

■3 rt

^ S ^

0)

^ ^ '^ ^ z.

cT) c/: c/3 c/2 r^'

"S ~ ~ rS ■-

C M tc M -^ S

^ ?5 c £ '5 5

— — — Cj 1; >,

v5 H fa K K H

— a

0;

FROGS OF THE HYLA PARVICEPS GROUP ■;

Table 2. — Measurements and proportions of males of Non-Ecuadorian

species in the Hyla parviceps group.

( See Table 1 for comparable statistics for Ecuadorian species. )

Character Range X SD SE

Hyla luteoccUata 9 $ 6

Snout-\ent length (SVL) ._ 20.00-23.10 21.720

Tibia length/SVL 0.47- 0.50 0.483

Foot length/SVL 0.39- 0.45 0.417

Head length/SVL 0.31- 0.35 0.326

Head width/SVL 0.31- 0.33 0.319

Tympanum/Eye 0.36- 0.45 0.397

Hyla microps 25 $ $

Snout-vent length (SVL) . 18.30-23.20 21.110

Tibia length/SVL 0.45- 0.55 0.499

Foot length/SVL 0.41- 0.46 0.437

Head length/SVL _..... ._-. 0.30- 0.34 0.320

Head width/SVL 0.29- 0.34 0.318

Tympanum/Eye 0.37- 0.52 0.454

Hijla suhocularis I'S $ $

Snout-vent length (SVL) . 20.50-23.10 21.720

Tibia length/S\'L 0.48- 0.55 0.509

Foot length/SVL 0.38- 0.45 0.426

Head length/SVL 0.29- 0.34 0.320

Head width/SVL 0.32- 0.34 0.328

Tympanum/Eye 0.26- 0.44 0.365

microps. In the latter, small tubercles are present on the eyelid in
most specimens (see Lutz, 1973. for discussion of variation).

The webbing is slightly variable within all species (Table 3);
H. parviceps has slightly more webbing than any of the others.
Females generally have slightly more webbing than do males; this
is especially noticeable in H. parviceps, which has the greatest
sexual dimorphism in size.

Coloration

Five aspects of color pattern were analyzed on 40S preser\^ed
adults, representing both sexes of all species, except H. hiteocellata
and microps, for which only males were used. However, partial
data were utilized for two female H. microps; the pattern of female
H. luteocellata was described b>' Rivero ( 1969 ) .

Dorsal Body Pattern. — This is the most intraspecifically variable
pattern character (Table 4). Hyla hokermanni is the most variable
species.

Dorsolateral Stripes. — These stripes are absent in H. microps
and parviceps, and most males of H. hrevifrons. In males of the
other species a narrow cream or pale yellow stripe extends from

1.260

0.420

0.014

0.005

0.017

0.006

0.013

0.004

0.008

0.003

0.035

0.012

1.340

0.270

0.025

0.005

0.018

0.004

0.014

0.003

0.014

0.003

0.037

0.007

0.730

0.150

0.015

0.003

0.018

0.004

0.011

0.002

0.007

0.001

0.041

0.008

8

OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

a

.

bn

C

a

s.

;.-!

U4

w

(U

«•>

<i)

C3

lU

a

<-lH

<-l-l

vs

n

■55

j_,

as

0)

-*-»

3
1=1

M-l

o

II

C/1

t/1

Qi

^

^^

Q)

s

S

3

(—I

o

M

C8

^

<

^

CD

C/1

4-'

bii

n:3

c

•4-*

OS

ctJ

'^

^

oi

0;

(— '

►^

:3

CO

c

w

C8

>-l

ti

a

n

<

«

o

o

T3

c

a

<M

Ol

I

CM

CM

I
CM

CC|

HH (M

(M

I

CM

CM

CM CM

>

(M

I

CM

(M

I— I CM I— (

CM

I

CM

CM

I

CM

CM

CM

CM

> >

>

1

!-H

i-H

■"^^

1— 1

;i*j

. ^

■ — ^

>

(— 1

1

1

^

VM

i—i

OJ

CM

CM

.— 1

^

1

1

,,-^

C>4

CM

CM

i-i<I

-~—^

-._^

-. .^

1— (

CM

^

>

>

"

1

l-H

HH

r^l

CM

'''^

-■— ^

-— ^

>

>

fM

1

CM

Csl

t~^

H- 1

OI

r<i

-'*"*"

CM

z; 1^ f^i S S

CM

I

CM

(>1
i
CM

CM

CM

CM
1

r

CM

1

+
CM

1

CM

1

<M

1
CM

1

CM

1
CM

CM

CM

CM

1
CM

CM

I— 1

;_-r
1— i

tA

i-H

M CM

CM

CM
I

eg

CO
CM

CM

h— 1
HH

CO
1

OI

CO
1

i,t

CM

CO

I

o-i

-^ — ■

CM

CO

^~ — '

"

CM

1
CM

Sit
CM

1
1

Ol

+
CM

T— (

1

(M

CM

CM

1 — 1

?:t
f— (

I-H

>-H

HH

O

CM

o

CM

lO

CO

o

CM

o

CM

a ^ ? a o

FROGS OF THE HYLA PARVICEPS GROUP 9

the eyelid or supratympanic region to the sacral region. In females
of these species the stripe is broad and well defined; the stripe
angles posterolaterally onto the flank in H. brevifrons but continues
posteriorly in the other species.

Table 4. — \'ariation in dorsal color pattern in the Hyla parviceps group.

( Numbers are percentages. )

c 5 «o

I I ==§ a & ^

Pattern | ? § g I §

ir ■=■ 3 c? a =

N 93 HI 9 25 114 56

Plain 20 .... 36 _... 18

Flecks 19 3

Longitudinal marks 11 .— ....

Median blotcli anteriorly;

paired marks posteriorly 9

Transverse bars 41 97

X anteriorly; paired

spots posteriorly .... 100

A-shaped mark .... .... 76

Median blotch anteriorly;

trans, bar posteriorly .... .... .... 24

X anteriorly .... .... .__. 20

X anteriorly; transverse

bar posteriorly .... .... . 62

Irregular dashes .... .... 64

Head Markinfi^s. — All species have vertical, cream or yellow sub-
orbital bars, although the bars are absent in 5 percent of the H.
parviceps. Normally there are two bars in H. bokermanni, brevi-
frons, and subocularis, and only one bar in the other species. The
bars are much broader in H. microps than in the other species.
Canthal and rostral stripes are absent in H. microps and parviceps.
Stripes are present in varying percentages of the other species
(Table 5). The stripes are cream or pale yellow. A canthal stripe

Table 5. — \'ariation in head stripes in the Hyla parviceps group.
( Numbers are percentages. )

H. bokermanni H. brevifrons H. luteocellata H. subocularis

N 93 in 9 56

Canthal only 41 8 .... 29

Canthal & rostral 51 30 62

Rostral only .... 29 100

None 8 33 .„. 9

10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

extends along the canthal ridge from the anterior comer of the
orbit to the nostril; in some individuals the stripes meet on the
snout or fuse with the upper end of the vertical rostral keel on
the snout.

Hind Limb Markings. — Important interspecific differences are
evident in the coloration of the thighs. All individuals of H. Jtiteo-
cellata and subocularis and 80 percent of the H. bokermanni have
a large black-bordered, deep yellow spot on the anterior surface of
the thigh. The anterior surfaces of the thighs are unmarked in all
H. brevifrons and microps and in 20 percent of the H. bokermanni
and 11 percent of the H. parviceps; in the rest of the H. parviceps
(89%) there are one or two small cream spots on the anterodorsal
surface of the thigh, which othenvise is black. The dorsal surface
of the thigh is dark brown or black with 1-3 small cream or vellow
spots in H. bokermanni (91%), Jjrevifrom (94%), luteocellata
(100%), parviceps (3%), and subocularis (91%); in all H. microps,
97 percent of the //. parviceps and less than 10 percent of the other
species, pale spots are absent. The posterior surface of the thigh is
dark brown in H. brevifrons, black with or without small cream
spots in H. parviceps, and usually yellow or tan with or without
black mottling in //. bokermanni, luteocellata, and subocularis. On
the posterior surface of the thigh in H. microps there is a large
black-bordered, orange spot. The dorsal surface of the shank has
transverse dark markings. These consist of 3-4 narrow bars in the
H. bokermanni; 3 broad bars in H. brevifrons, microps, subocularis.
and parviceps (median bar wider than others), or faint lines in
H. luteocellata.

Ventral Coloration. — In H. parviceps the venter is dark gray
fading to white medially on the throat and anterior belly; the
ventral surface of the shank is gray with a bright orange spot prox-
imally. In the other species the venter is cream; gray flecks are
present on the anterior edge of the chin in H. brevifrons (75%),
luteocellata ( 100%), and microps ( 100%), and gray flecks are present
on the belly in 32 percent of H. microps.

Mating Calls

We have analyzed tape recordings of four species (Table 6),
and Rivero (1969) provided an audiospectrogram of //. luteocel-
lata. The call of H. microps has not been analyzed, but Lutz
(1973) described the call as "the chirping of a cricket." The calls
of all of the species consist of short, high-pitched, insect-like notes.
The call of H. parviceps is much higher pitched than the others,
and the call of //. bokermanni is distinctive in having secondary
notes ( Fig. 1 ) . The release call of H. luteocellata illustrated by
Rivero ( 1969 ) may be a series of secondary notes.

FROGS OF THE HYLA PARVICEPS GROUP

11

Table 6. — Coinparison of mating calls of members of the Htjlu parviceps group.
(N = indi\'iduals/notes; means in parentlieses below ranges.)

Note

Pulses

Dominant

Repetition

Duration

per

Secondary

Frequency

Species (N)

Rate

( sees )

Second

Notes

(Hertz)

H. bokernuinni

5-19

0.23-0.28

100-190

2-5

4000-4652

(4/12)

(12.2)

(0.259)

(132.5)

(3.3)

(4478)

H. brevifrons

26-46

0.43-0.49

30- 40

0

4152-5115

(9/25)

(36.8)

(0.467)

(38.9)

(4625)

H. parviceps

54.5

0.12-0.14

140

0

6072-6341

(1/3)

(0.13)

(6221)

H. subocularis

3-20

0

(2/4)

(0.53)

(43)

2200

Tadpoles

Tlie tadpoles of H. luteocellata are unknown. The tadpole of
//. microps was deseribed by Bokermann (1963b) and that of H.
subocularis by Duellman (1970). The mouthparts of the tadpoles
are alike in having robust, finely serrated beaks, bare upper lips
and one row of labial papillae laterally and ventrally. In H. hoker-
manni and su])0cul(iris the papillae lateral to the mouth are par-
tially fused in the form of a \ertieal fold. The mouth is small and
direeted anterovcntralh'; the body is o\()id with blunt snout in
dorsal view. The eaudal fin does not extend onto the bod}', and the
tail is xiphicercal (Fig. 2). The spiracle is sinistral and lateral, and
the anal tube is dextral. Rows of denticles are absent in all except
H. microps, which has a single row of small denticles below the
lower beak.

Tadpoles of this group are brightly colorc^d and patterned.
Longitudinal cream stripes an^ present on the dorsal surface of the
body of H. ))ucro})s and sul)Ocuhiris, and on the snout of H. bokcr-
manni. Two broad cream transverse bars are present on the body
of H. parviceps, and broad yellowish-tan dorsolateral stripes on the
body and a red streak on the dorsal fin characterize the tadpole of
H. brevifrons.

KEY TO THE SPECIES

1. Dorsolateral light stripe absent 2

Dorsolateral light stripe present, narrow in males, broad in fe-
males 4

2. Venter gray, white medially; small orange spot proximally on
ventral surface of shank Hyla parviceps

\^enter cream or white, with or without small gray flecks; no
orange spot on ventral surface of shank 3

12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

8

0

8
NI

£«

UJ

X

04

_J

A

B

C

rW*

imum^

. k. JL L k t g CTfc. ,. .. k. . , . fc. -.. . k: -...-..... y . .

V i J ^f ^ f-f4-i ^"^tt

nii<.n..) I.i.iiiiii*! m t.,Muii,>m;<mm

r,^m

Q L-.M«M»yi^ -^L rf.^ «> .. -1 t. AS. --! i-.i ^C..^— v^U.:i ■ -g::^ .>^l ^. i..^ .-L. ^ ■^■. ^ ^1.

0.2 0.4 0.6

TIME IN SECONDS

Fig. 1. Audiospectrojjranis of mating calls of members of the Hyla pani-
ceps group at Santa Cecilia, Ecuador. A. Htihi hokcrnunini ( KU Tape 688).
B. H. hrcvifnms ( KU Tape 1176). C. H. paiviccps ( KU Tape 1178).

FROGS OF THE HYLA PARVICEPS GROUP

13

«ie*sw;v- ■!.-.**■

^f^ %y^

c

'TiEc/e*

Fig. 2. Tadpoles of nu'inbeis of the //(//« parviceps group at Santa Cecilia,
Ecuador. A. Hijla bokcnminni ( KU 124193). B. //. hrevifnms (KU 125899).
C. //. panircps ( KU 146791). x5.

3. No spots on dorsal surface^ of tlii^h; laiG;e orani2;e spot on pos-
terior surface of thigh J^yld niicrops

Two small creamy yellow spots on dorsal surface of thigh; no

large orange spot on posterior surface of thigh

Hijla hrevifrons { S S )

4. Anterior surface of thigh dark i^l/ln hrevifrons (99)

Anterior surface of thigh tan, usually with large, black-bordered
spot 5

5. One suborbital white bar HyJa Ititeocellota

Usually two suborbital white bars 6

6. Dorsum plain, flecked, or usually marked with large median
dark blotch or transverse bar anteriorly Htjla bokenminni

Dorsum marked with a dark X-shaped pattern anteriorly

Hyla subocularis

14 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

SPECIES ACCOUNTS

In the following accounts, the major distinguishing features are
enumerated in the diagnoses. Sizes given in diagnoses are maxi-
mum snout-vent lengths. With the exception of one new species,
detailed descriptions are omitted. Variation of characters is treated
in the preceding analysis of characters. Colors are of living frogs
and tadpoles. The statements of distribution are based on the
localities in the list of specimens examined.

Hyla bokermanni Coin
(Figure 3)

Hyhi bokermanni Coin, 1960:721 [Holotype.— WCAB 2881 from Tarauaca,
Territorio do Acre, Brasil; W. C. A. Bokeniiann collector].

Hyla rondoniac Bokermann, 1963a: 247 [Holotype.— WCAB 7845 from Ron-
donia, Territorio do Rondonia, Brasil; A. Machado collector]. New
synonym.

Diagnosis. — 1) Snout- vent length 22.7 mm in males, 25.6 mm in
females; 2) webbing on hand II 2-3" III 2-2 IV, on foot I 2-2 II
IJ4-2 III 1^4-2 IV 2-1^ V; 3) ulnar and tarsal tubercles absent; 4)
canthal and rostral stripes (50%) or canthal stripes only (41%) pres-
ent; 5) two suborbital bars (98%); 6) dorsum plain, flecked, or
marked with three transverse bars or one blotch anteriorly and one
bar posteriorly; 7) dorsolateral light stripe present, broad in fe-
males; 8) thigh usually (80%) having large black-bordered yellow
spot anteriorly and 1-3 small yellow spots dorsally; 9) all ventral
surfaces uniform cream.

Hyla bokermanni, luteocellata, and subocularis all have dorso-
lateral light stripes and large yellow spots on the anterior surfaces
of the thighs. Hyla bokermanni differs from H. luteocellata in
usually having canthal stripes and two, instead of one, suborbital
bars; H. subocularis differs from H. bokertnanni by having an X-
shaped pattern anteriorly on the dorsum, instead of a large blotch,
transverse bar, or no markings.

Coloration. — At night, the dorsum is pale yellowish tan; the
dorsal markings and dorsolateral stripes are barely discernable. By
day, the dorsum is tan, and the dorsal markings are dark brown.
The dorsolateral stripe is creamy yellow in males and creamy white
in females. The large spot on the anterior surface of the thigh is
bright yellow, and the small spots on the dorsal surface of the thigh
are creamy yellow. The iris is silvery gray with a narrow red ring
around the pupil. Males have a yellow vocal sac; otherwise the
venter is creamy white.

Tadpoles. — A tadpole in stage 36 has a body length of 7.2 mm
and a total length of 19.8 mm. The ])0(ly is dark brown with two
short, broad, longitudinal yellowish tan bars on the snout. Prox-

FROGS OF THE HYLA PARVICEPS GROUP

15

Fig. 3. Hijla bokennmwi, female, 24.4 nun S\'L ( KU 126552).

imally the caudal musculature and fins are pale orange, unpig-
mented distalh", except for broad gray vertical marks. The iris is
red medially and sih'cr peripherally.

Distrilmtion. — Lowlands (< 1000 m) of upper Amazon Basin
in Ecuador and extreme western Brasil. and presumably in the
interv^ening area of northeastern Peni.

Remarks. — We have examined the holotype of H. rondoniae
(WCAB 12148) and compared it with the figure of the holotype of
H. bokernianni and series of specimens from Ecuador. As sug-
gested by Rivero (1969), the name H. bokermanni is based on a
female and //. rondoniae on a male of the same species.

Hyla brevifrons new species

(Figures 4 & 5)

Holotype. — KU 126370, an adult male, from Santa Cecilia, Pro-
vincia Napo, Ecuador, obtained on 16 May 1969, by William E.
Duellman and Linda Trueb.

Allotiipe. — KU 126371, an adult female, same data as holotvpe.

ParatopoU/pes.—KV 111786-802, 123049-50, 123059, 123072-88.
126342-61, 126363-69, 126546, 126548-51, 136298, 143189, 146281-
87, 146413, 146415-18.

Diagnosis. — 1) Snout-vent length 21.4 mm in males, 23.2 mm in
females; 2) webbing on hand II 2-23^ III 2)3-2" IV, on foot I 2-2 II
1)3-2 III 1)^-2 IV 2-l)i V; 3) ulnar and tarsal tubercles absent; 4)

16 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

canthal and rostral stripes usually present (one or both in 66%);
5) two suborbital bars (95%); 6) dorsal pattern consisting of broad
transverse bars; 7) dorsolateral light stripe absent in males, broad
in females; 8) thigh dark brown anteriorly, having 1-3 small yellow
spots dorsally; 9) venter uniformly cream, except for grav flecks
on chin (78%).

Hyla brevifrons differs from those other members of the group
having dorsolateral light stripes by lacking a large black-bordered
yellow spot on the anterior surface of the thigh.

Description. — The following description is based on the holo-
type and allotype; where differences occur, the characters of the
allotype are in parentheses. Snout-vent length 19.1 mm (23.0 mm);
snout short, truncate in dorsal view, inclined posteroventrally in
profile; canthus rounded; loreal region slightly concave; lips
rounded; internarial area depressed; nostrils slightly protuberant
anterolaterally; supratympanic fold weak; tympanic ring indistinct.
Axillary membrane extending one-fourth length of upper arm; fore-
arm lacking folds or tubercles; fingers moderatelv short, bearing
small discs; webbing on hand II 2 -2M III 2^-2' IV (11 2 -2M III 2)1-2"
IV); webbing on foot I 2-2- II l,'^-2 III 1)^-2 IV 2--1M V (I 2-2 II IK-
2M III 1)4-2 IV 2-lK V). Anal flap short, opening at upper level of
thighs; skin on belly and proximal posteroventral surfaces of thighs
granular; skin on other surfaces smooth. Prevomerine teeth 2-2
(3-4) on processes posteromedially inclined between posterior mar-
gins of ovoid choanae; tongue cordiform, shallowly notched pos-
teriorly; vocal slit extending from midlateral base of tongue to
angle of jaw (absent).

Color ( in alcohol ) : Dorsum tan ( grayish tan ) with brown
markings — blotch in occipital region, square blotch in scapular re-
gion, and transverse blotch in sacral region extending onto flanks;
dorsolateral stripe absent (distinct, broad, creamy white, extending

.4-
Fio. 4. Ilyla ])rcvifi(ms. holotype, male, 19.1 mm S\'L ( KU 126.370).

i

FROGS OF THE HYLA PARVICEPS GROUP

17

Fig. 5. Htjla brcvifixni',

alio

type, female, 23.0 nun S\'L ( KU 126371).

from eyelid to transverse sacral blotch and thence laterad onto flanks,
bordered below by broad brown mark); one (two) sul)orbital white
bar; forearms and shanks tan with two and three transverse brown
bars; thighs pale brown with two creamy yellow spots on dorsal
surfaces; venter white with brown flecks on chin.

Coloration. — The dorsum is olive-tan with brown markings and,
in females, distinct creamy tan dorsolateral stripes. The thighs are
dark brown; the spots on the thighs and the suborbital bars are
cream}' yellow. The venter is unpigmented (>\cept for the pectoral
region and chin (in females), which is white; males have a pale
yellow vocal sac. The iris is red medially and silvery gray periph-
erallv.

Tadpoles. — A tadpole in stage 39 has a body length of 7.6 mm
and a total length of 22.2 mm. In dorsal view the body is dark
brown medially and palc> brown laterally. The venter and sides of
the body are dark brown with cream flecks. The tail is tan with
dark brown mottling and a bright red streak on the dorsal fin. The
iris is gray-bronze with a red ring around the pupil.

Distribution. — The upper Amazon Basin in Ecuador, Peru, and
probably western Brasil, to elevations of about 1300 m on the east-
ern slopes of the Andes. Additionally, the species is known from
French Guiana and Belem, Brasil; it probably occurs throughout
the Amazon Basin.

Etymology. — The specific name is derived from the Latin brevis

18 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

meaning short, and the Latin frons, meaning brow. The name is
used in alhision to the short head of the species.

Hyhi luteocellata Roux

Ilyla hitco-oceUata Roux, 1927:260 [Molotype.— NUMB 3900 from El Mene,
Estado Falcon, Venezuela; Kugler and X'onderschmidt collectors].

Diagnosis. — 1) Snout- vent length 23.1 mm in males, 28.0 mm in
females (Rivero, 1969:132); 2) webbing on hand II 2-3 III 2^^-2
IV, on foot I 2-2 II 1^-2 III 2^-2^ IV 2^-lK V; 3) ulnar and tarsal
tubercles absent; 4) canthal stripes absent, rostral stripe present;

5) one suborbital bar; 6) dorsal pattern consisting of X-shaped
mark anteriorly and pair of dashes posteriorly; 7) dorsolateral light
stripe present, broad in females; 8) thigh having large black-
bordered yellow spot anteriorly and 1 or 2 small yellow spots dor-
sally; 9) venter uniformly cream, except for gray flecks on chin.

Hyla luteocellata is like H. hokermanni and subociilaris in hav-
ing dorsolateral light stripes and a large yellow spot on the anterior
surface of the thigh. It differs from these species by lacking canthal
stripes and in having only one suborbital bar. Furthermore, it
differs from H. hokermanni by having an X-shaped dorsal marking
anteriorly, instead of a median blotch or transverse bar, and from
H. suhociilaris by having paired longitudinal markings posteriorly
on the dorsum instead of no markings or a transverse bar.

Coloration. — Rivero (1969:128) described the suborbital bar as
cream or yellow, the spot on the anterior surface of the thigh
orange, spots on the dorsal surface of the thigh greenish yellow,
brownish, or dirty white.

Distribution. — Northern Venezuela from Lago Maracaibo to the
Unare Depression (Rivero, 1969:1.30). The species also occurs
near Brownsweg and in the Kayser Mountains in Surinam (M. S.
Hoogmoed, pers. comm. ).

Hyla microps Peters

/////<■; microps Peters, 1872:682 [Holotype. — ZMR 7472 Ironi Novo Friliurt^o,
Estado do Rio de Janeiro, Brasil; Salmin collector].

Uilh hilli Boulenjrer, 1920:123 [Holotype.— BMNH 1914.3.20.9 (RR 1947.
2.24.10) from Teresopolis, Estado Rio de Janeiro, Brasil; J. P. Hill col-
lector]. Synonymy fide Cochran (19.55:177).

Hyla fiieslcri Mertens, 19.50:18.5 [Holotype. — SMF 41217 from Barro Branco,
Estado do Rio de Janeiro, Brasil; P. Giesler collector]. Synonymy fide
Bokermann (1966:51).

Diagnosis. — 1) Snout-vent length 24.6 mm in males, 31.3 mm in
females; 2) webbing on hand II 2K-2:*i III 2?4-2 IV, on foot I 2-2 II
V-2 IV 2-1/^4 V; 3) ulnar and tarsal tubercles low in both sexes;
4) canthal and rostral stripes abscMit; 5) one broad suborbital bar;

6) dorsum plain or marked with irregular dashes; 7) dorsolateral

FROGS OF THE HYLA PARVICEPS GROUP 19

light stripe absent; 8) thighs yellow anteriorly, brown dorsally,
having large black-bordered orange spot posteriorly; 9) venter
cream with gray flecks on chin (100%) and on belly (32%).

This species difi^ers from all other members of the group, except
H. parviceps, in lacking dorsolateral light stripes; it differs from H.
pawiceps by having a cream venter and a large orange spot on the
posterior surface of the thigh and a black mark in the axilla. Hijla
parviceps has a gray venter with an orange spot on the proximal
ventral surface of the shank and cream spots anterodorsally on
the thigh.

Coloration. — Lutz (1973:95) described specimens from Serra
da Bocaina, Brasil, as brown above with darker brown dorsal mark-
ings. The spots on the thighs and the webs between the inner toes
are briglit orange. The axilla is black, and the belly is greenish
yellow.

Tadpoles. — Bokermann ( 1963b ) described tadpoles from
Campo Grande, Sao Paulo, Brasil. The body is dark olive with
dorsolateral and lateral light stripes.

Distribution. — Coastal lowlands and lower slopes of the high-
lands in soutlicastern Brasil from Espirito Santo to Parana.

Reniarks. — Lutz (1973) mcMitioned a specimen 38 mm in length
from Angra dos Reis, Rio de Janeiro, Brasil. Although we have not
seen the specimen, we are duljious of the identification reported b\'
Lutz because of the reported size and coloration.

Duellman examined the holotypes of Hyla hilli and Ht/la fi^iesleri
and concluded as did Cochran (1955) and Bokermann (1966) that
they are representatives of Hyla microps. In addition to the locali-
ties given in the specimens examined, Cochran (1955) reported the
species from Rio Novo, Espirito Santo, and from several localities
in Rio de Janeiro and Santa Catarina.

Hyla parviceps Boulenger
( Figure 6 )

Hyhi parviceps BouleiiKer, 1882:393 [Holotype.— BMXH 80.12.5.214 ( RR
1947.2.13.5) from Sarayacu, Pro\ iiicia Pastaza, Ecuador; Mr. Buckley
collector].

Diagnosis. — 1) Snout- vent length 18.3 mm in males, 26.1 mm in
females"; 2) webbing on hand II l?^-2?^ Ill 2-2 IV, on foot I VA-2
II r-2 III 1)^2 IV 2-lJi V; 3) ulnar and tarsal tubercles low on fe-
males, absent on males: 4) canthal and rostral stripes absent; 5)
usually (92%) one suborbital bar; 6) dorsal markings usually (76%)
aA-shaped mark, or (24%) median blotch anteriorly and transverse
bar posteriorly; 7) dorsolateral stripe absent; 8) thigh usually
(92%) having one or two small cream spots anteriorly and unifomi
dark dorsally; 9) throat and belly gray, white medially; ventral sur-
face of shank gray with orange spot proximally.

20

OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

Fig. 6. Hyla parviceps, female, 24.5 mm SVL ( KU 126476).

HyJa parviceps differs from all other species in the group, ex-
cept H. microps, in lacking dorsolateral light stripes; it differs from
H. microps by having a dark venter with an orange spot on the
pro.ximal ventral surface of the shank and cream spots antero-
dorsally on the thigh. Hyla microps has a cream venter and a large
orange spot on the posterior surface of the thigh.

Coloration. — The dorsum is tan with faintly darker markings at
night. By day the dorsum is orange-tan, reddish brown, or brown
with darker brown markings. The flanks are dark brown or black
and white. The thighs are dark brown or black with creamy-yellow
spots. The suborbital bar is creamy white. The venter is gray or
white with gray or black suffusion most prominent laterally on the
throat and belly. The ventral surfaces of the limbs are gray with
an elongate bright orange spot proximally on the shank. Some in-
dividuals also have a small orange spot in the axilla. The iris is red
medially and silvery gray peripherally.

Tadpoles. — The body is dark brown with two broad, transverse
cream bars dorsally (snout and posterior to spiracle). The tail is
pinkish orange or orange-tan with dark brown mottling.

Distribution. — The upper Amazon Basin in western Brasil,
southern Colombia, Ecuador, and Peru, to c^levations of 1300 m on
the eastern slopes of the Andes.

Remarks. — In addition to the localities given in the specimens
examined, Mel in (1941) reported the species from Taracua, Ama-

FROGS OF THE HYLA PARVICEPS GROUP 21

zonas, Brasil, and Cochran and Coin ( 1970 ) reported it from Rio
Apoporis, Amazonas, and Ceilan, Caqueta, both in Colombia, and
from Acre, Brasil.

Hyla subocularis Dunn

Hyhi subocularis Dunn, 1934:2 [Mnlotype.— AMXH 41117 from Rio Chu-
cunaque at first creek al}o\e Rio Tuquesa, Provincia Darien, Panama;
Charles M. Breder collector].

Diagnosis. — 1) Snout- vent length 23.1 mm in males, 26.1 mm in
females; 2) webbing on hand II 2-3 III 2}i-2 IV, on foot I 2-2 II
1)^2)^ III 1-2 IV 2-VA V; 3) ulnar and tarsal tubercles absent;
4) canthal and rostral stripes (64%) or canthal stripes only (29%)
usually present; 5) one (14%) or two (86%) suborbital bars; 6) dor-
sal markings consisting of X-shaped mark anteriorly (83%) with or
without transverse bar posteriorly; 7) dorsolateral light stripe pres-
ent, broad in females; 8) thigh having one large black-bordered
yellow spot anteriorly, and usually (91%) one small yellow spot
dorsally; 9) ventral surfaces imiform cream.

Hyla suJ)()cnlaris is like H. hokennanni and hitcoccllata in ha\'-
ing dorsolateral light stripes and a large yellow spot on the anterior
surface of the thigh. Hyla su])ocuIaris differs from H. luteocellata
in usually having canthal stripes and tvvo, instead of one, suborbital
bars; //. hokermunni differs from //. su])ocularis in having a large
median dark blotch, transverse bars, or no markings anteriorly on
the dorsum, instead of an X-shaped mark.

Coloration. — Duellman (1970:236) described the color of Hyla
subocularis. The dorsum is yellowish tan with brown markings.
The suborbital bars and dorsolateral stripes are creamy white or
creamy yellow. The spots on the thigh are yellow. The iris is rose-
pink medially and pale siKer peripheralK'.

Tadpoles. — The tadpoles were described by Duellman (1970:
237), who noted that the body and tail are black with a dorso-
lateral tan stripe on the body and irregular tan vertical marks on
the tail.

Distribution. — Lowlands to 800 m in Darien, Panama, and
northwestern Colombia.

ECOLOGY

All the members of the Hyla parviceps group inhabit wet low-
land tropical forest. Information on the habitat of Hyla luteocellata
was provided by Rivero (1969), of H. subocularis by Duellman
( 1970), and of H. microps by Lutz (1973). We are concerned here
with the three sympatric species in the upper Amazon Basin (H.
])okermanni, H. brcvifrons, and //. parviceps). The following ob-
sei-vations resulted from intensive field studies from June 1971

22 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

through July 1972 at Santa CeciHa, Provincia Napo, Ecuador. Santa
Ceciha is situated at 76^58' W, 00°02' N at an elevation of 340 m
on the northern bank of the Rio Aguarico, which flows into the
Rio Napo, a major tributaiy of the Rio Amazonas.

At Santa Cecilia the three sympatric species utilize different
breeding sites. Males of each species occasionally were found call-
ing in swamps in open, disturbed areas; no females were found in
such habitats. Hyla parviceps seems to be a habitat generalist,
breeding in great abundance in both forest-edge and forest swamps.
This species calls and breeds sympatrically and synchronously with
either of the other species. On the other hand, H. hokermanni and
H. hrevifrons only rarely were observed together. Hyla hokermanni
breeds mainly in forest swamps, but occasionally in forest-edge
situations. Hyla hrevifrons breeds mainly in forest-edge swamps
and only occasionally in forest swamps.

The calling sites of the three species are similar. Males call
from vegetation usually 1-2 m above the water. Hyla hokermanni
and H. hrevifrons call from stems, branches, or leaves, whereas H.
parviceps almost always calls from leaves. Hyla hokermanni
usually calls from emergent vegetation, whereas the other two
species also call from vegetation peripheral to the water.

There is distinct segregation in egg deposition sites. Hyla
hokermanni and H. hrevifrons deposit eggs in clumps on leaves
overhanging the water. Egg clutches of the former were found at
heights from 50 cm-160 cm (x = 104.0 cm; N = 14), and those of
the latter at heights from 52 cm- 100 cm (x =r 75.4 cm; N = 5)
above water. On the other hand, H. parviceps deposits eggs in
water.

Breeding Patterns at One Study Site

The breeding patterns and relative abundances of the species
were studied during 10-day intervals at a forest swamp from July
1971 to July 1972. Most observations were made between 1900 and
2400 hrs. The swamp is approximately 15 m by 50 m and is choked
with large-leafed Heliconia plants. Twenty-two species of frogs
were found at the swamp, but no more than 12 were there at any
given time.

All three species of the Hyla parviceps group were found at the
site, although Hyla hrevifrons was observed only rarely. Several
males of that species were present in August, whereas only five
were present in October; none was calling. One gravid female was
there in March. There were many H. hokermanni calling whenever
H. hrevifrons were found.

Throughout the year //. hokermanni and //. parviceps were ob-
served calling in great numbers ( more than 50 and 20 respectively ) .

FROGS OF THE HYLA PARVICEPS GROUP 23

Gravid females and/or amplectant pairs of both species were found
on the same nights in November, March, and June. On only one
occasion was the population of H. parviceps estimated to be greater
than 50 individuals; in July 1972, about 100-150 calling males were
present. Approximately 20 gravid females were found; this is about
twice the number found on any other evening. The same evening
30-50 calling H. hokermanni were found; no females were obser\^ed.
The water depth at this time was about 1 m in the middle of the
swamp. Generally, when the two species were found calling syn-
chronously, 77. hokermanni was more than twice as abundant as
//. parviceps (30-100 \'crsus 10-35). IhjU hokernuinni calls from
Heliconia leaves and stems 1-2 m above the water surface. Hyla
parviceps calls from the same sites, but usually at least one-third of
them call at about the same heights from vegetation at the edge of
the swamp. Hyla hokermanni was rarely found at the edge f)f
the swamp.

Hyla hokermanni had two peaks of breeding activity at the
swamp: October-November and February- April. These periods
coincided with the times of greatest water depth in the swamp. On
several occasions density of calling males was estimated between
300-500. Once in late March, the number of gravid females was
estimated at 30-50; usually no more than 10-15 were found in one
evening. Occasionally, wlien the swamp was dry, numerous egg
clutches of //. hokermanni were hanging from the vegetation, and
males were calling.

In contrast, male //. parviceps were obser\'ed calling only when
there was at least 30 cm of water in the swamp. Generalh', the
more water, the greater the number of calling indixiduals. Breed-
ing acti\ity of H. parviceps reached a peak in October-November
(coinciding with that of //. hokermanni) and again in June-July.

The breeding population of //. hokermanni generally seems to
be larger than that of //. parviceps at any given time at the swamp.
Apparently, there is no species replacement, for both species breed
at the same time. This is probably due to differences in mating
calls and in egg deposition sites.

Reproduction

Two modes of reproduction are represented in the Hyla parvi-
ceps group. Hyla parviceps deposits its eggs directly in the water,
whereas both //. J)okcrmanni and H. hrevifrons deposit their eggs
on leaves over water; upon hatching the tadpoles drop into the
water where they complete development (Table 7).

Tlic eggs of H. parviceps are small (1.0-1.2 mm) and heavily
pigmented black and creamy white. The ovarian complement ( 120-
395) is more than twice that of either H. hokermanni or H. hrevi-

18.0-23.5

20.0-23.5

(20.9, n=15)

(23.3, n=101)

48-114

120-395

(79.1, H=15)

(2.34.1, 71 = 101)

0.9-1.2

(1.1, n=6)

0.1-0.2

(0.1, n=6)

8 3-22 2

(11.2, n=6)

48-91

165-385

(70.0, n=6)

(302.8, n = 15)

1.2-1.6

1.0-1.2

(1..3, n=6)

(1.1, n=15)

6.3

1.3-2.3

(n = l)

(1.9, n = 13)

6.0

4.0-4.5

(" = 1)

(4.1, n=13)

24 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

Table 7. — Summary of Reproductive Data for three species in the

Hyla parviceps group at Santa CeciHa.
( Means and sample sizes are in parentheses below observed ranges. )

H. bokermanni H. brevifrons H. parviceps

Snout-vent length (mm) 22.5-25.5

of gravid females (23.9, J)=29)

Number of mature 65-135

ovarian eggs (98.3, n—29)

Volume (ml) of 0.9-1.1

gravid female (LO, n=5)

Volume (ml) of 0.1

mature ovarian eggs __.. (n=5)
% Egg volume of 9.1-11.1

total volume of 9 (10.5, n=5)

60-196

Clutch size (116.8, »=53)

1.5-1.8

Ovum size (1.6, n=53)

Number of days 4.0-5.3

until hatching -(4.4, n=20)

Total length (mm) 4.0-4.5

of hatchling (4.2, n=20)

frons. A correlation analysis was run on 112 females with mature
ovarian eggs in order to determine the relationship between snout-
vent length and fecundity (both variables were normally dis-
tributed). The correlation coefficient was significant at the .01
level. Therefore, on the basis of this sample, it is concluded that
larger female H. parviceps produce more eggs at a given breeding
time. Hyhi parviceps has a slightly greater percent volume of eggs
relative to its total body volume than does II. liokermanni. The
eggs develop rapidly under laboratory conditions (1.3-2.3 days).
The range of variation of total length of the hatchlings (4.0-4.5
mm) is the same as that of H. bokermanni. Gravid females were
found in every month of the year, indicating that breeding is con-
tinuous. Of 146 females found, 125 ((S5.6%) were gravid.

HijJa bokermanni deposits the largest eggs (1.5-1.8 mm) of the
three species. The eggs are lightly pigmented tan and cream. The
ovarian complement ranges from 65-135. A correlation analysis run
on 29 females of this species (both variables normally distributed)
indicated no significant correlation between snout-vent length of
the female and the number of mature ovarian eggs. Thus, it seems
as though there is no selective advantage for female H. boker-
manni to increase in size in order to increase fecundity. It is sug-
gested that perhaps larger females produce larger eggs. Under
laboratory conditions the eggs took from 4.0-5.3 days to hatch. As
prc-\iously indicated, although the eggs are larger to begin with
and require more than twice as long to hatch, the size range of the

FROGS OF THE HYLA PARVICEPS GROUP 25

hatchlings is the same as that of H. pawiceps. Gravid females were
found October-November 1971 and throughout January-June 1972.
Of 92 females found, 85 (92.4%) were gravid.

Hijla brevifrons, the smallest of the three species, deposits the
intermediate-sized eggs (1.2-1.6 mm). The eggs are lightly pig-
mented tan and cream, similar to those of H. bokermanni. The
species has the lowest ovarian complement (48-114). Only one
clutch of eggs hatched in the laboratoiy; the eggs hatched in 6.3
days, and the hatchlings were 6.0 mm in length. Gravid females
were found in July, October, and December 1971, and January,
March, and June 1972. Of 13 females found, 10 (76.9%) were
gravid.

An ovarian size factor'' was calculated from means of egg
clutches deposited in the laboratory. Hyla pawiceps has the largest
value (13.38), H. brevifrons the smallest (4.42), and H. boker-
manni an intermediate value (7.05). The reproductive strategy of
II. parviceps is to produce many small eggs, deposited in open
water. The species is a habitat generalist, breeding both in forest
and forest-edge situations. The other two species are more re-
stricted; H. bokermanni breeds mainly in the forest, and H. brevi-
frons in forest-edge swamps and ponds. The last two species lay
relatively few, large eggs, and ha\e more specialized egg deposi-
tion sites on vegetation above the water. The nearly identical
utilization of egg deposition sites perhaps explains why these
species generally do not breed at tlie same swamps. Neither //.
bokernmnni nor H. I)revifrons was found to be gravid throughout
the year. Perhaps th(\se factors are indicative of more critical re-
quirements for breeding than those of //. parviceps. Obser\'ations
on the relationship between egg deposition sites and sunlight are
inconclusive. It is suggested, however, that differences in pigmenta-
tion of the eggs may be explained by their exposure to different
amounts of solar radiation. The darkest eggs are those of H.
parviceps (deposited in water); perhaps these are subject to more
solar radiation than those of //. bokcrnumni and //. brevifrons (de-
posited on leaves above water) which are protected by surrounding
vegetation.

Food

Stomach contents of adults of both sexes of the three species at
Santa Cecilia were analyzed. Ten orders of arthropods were found
in 66 stomachs. Treating males and females separately, niche over-

''The ovarian size factor is calculated by the formula CS(OD)/SVL, where
CS = tlie mean clutch size, OD = the mean ovum diameter, and SVL = the
mean snout-\'ent length of the females depositing the clutches. This provides
an index for comparing fccundit>' and egg size relati\e to adult body size.

26 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

lap scores were calculated by the method suggested by Colwell aud
Futuyma (1971). Niche overlap scores varied from 0.13 to 0.76
(1.0 is complete overlap). The lowest scores are those of female
H. brevifwiis in which the sample consisted of only two specimens.
Analysis of variance within groups (males and females) and be-
tween groups (males versus females) revealed no significant differ-
ence. The average niche overlap scores for food is 0.451, indicating
a moderate amount of overlap between sexes in a given species and
between species. The niche overlap values must be considered
maximal to actual overlap because 1) small sample sizes (2-23)
minimize overlap, and 2) only gross identifications (to order) were
made on the food items. Because of the fragmentary nature, and /or
degree of decomposition, of the food, insufficient data were ob-
tained on prey size. However, individuals of all three species, in-
cluding both sexes, contained caterpillars that were nearly as long
as the frogs that had eaten them.

PHYLOGENETIC RELATIONSHIPS

In deducing phylogenetic relationships, it is customaiy to estab-
lish criteria for primitiveness. We have relied on the criteria used
by Wagner (1961) (see Kluge and Farris, 1969, and Trueb, 1973,
for discussions). Utilizing these criteria, primitive character states
of a particular group are: 1) likely to be present in many closely
related groups; 2) more likely to be widespread within a group
than is any one derived state; and 3) likely to be associated with
other primitive characters.

As a basis for phylogenetic analysis, we are assuming that the
small, primarily Amazonian Hyla having a diploid number of 30
chromosomes and tadpoles with reduced mouthparts and xiphicer-
cal tails are more closely related to one another than any is to any
other group. Thus, we recognize the leucophijUata, microcephaJa,
minuta, and parviceps groups of Hijla. Eleven structural and color
pattern characters that are variable within the //(//(/ parviceps
group and present in the other groups were analyzed. Of these 11
characters, the primitive states of eight are present in all three of
the other groups, two in two groups, and one in one group.

In the following list of characters, the primitive state is number
0; states 1 and 2 are derived in succession, except F and G, in
which the advanced states (1 and 2) are probably independently
derived.

A. Sexual dinioiphisin in size:

0. S > 90% 9

1. S - 80-90% 9

2. 5 < 80% 9

B. Ulnar and tarsal tubercles:

0. Absent

1. Present

FROGS OF THE HYLA PARVICEPS GROUP 27

C. Suborbital bars:

0. Absent

1. One

2. Two

D. Canthal stripes:

0. Absent

1. Present

E. Rostral stripes:

0. Absent

1. Present

F. Dorsolateral stripes:

0. Absent in both sexes

1. Absent in males; present in females

2. Present in both sexes

G. Anterior thigh:

0. Unicolor

1. Small yellow spots

2. Large yellow spot
H. Dorsal thigh:

0. Unicolor

1. Small yellow spots
I. Posterior thigh:

0. Unicolor

1. Large orange spot
J. Venter:

0. Plain
L Patterned
K. Denticles (larvae):

0. One lower row

1. Absent

The distribution of character states witliin the Hijla parviceps
group reveals variation from 6 to 11 derived characters (Table 8).
The greatest number of priiuitive states occurs in //. microps, fol-
lowed by H. parviceps and hrevijrons\ II. hokermanni, luteocellata,
and subocularis have the fewest primitive states.

A phylogenetic arrangement (Fig. 7) was constructed by a
method proposed by Camin and Sokal (1965). The branching se-
quence defines one group having few derived characters (H. mi-
crops and parviceps) and another with many derived characters
(H. luteocellata, hokermanni, and subocularis). Hyla brevifrons
is intermediate between these two groups.

Although reproductive data are incomplete, the existing infor-
mation supports the proposed phylogeny. Hyla microps and par-
viceps deposit eggs in water, whereas the other species (unknown
for H. luteocellata) deposit eggs on vegetation above water. Quan-
titative data on reproduction of three species in Ecuador reveal
that in most features H. brevifrons is intermediate between H. boker-
manni and parviceps. If the reproductive mode of H. luteocellata
is tlie same as that for H. bokermanni and subocularis, a phylo-
genetic construct based on reproductive characters would approxi-
mate the arrangement based on morphological featui-es.

28

OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

in

Co

«o

A^BC

CD

to

CO

03

G K DF'

Fk;. 7. Most parsimonious pliylogenctic anangonient of species in Iliila
parviccps group, based on 26 states of 11 characters. Letters indicate sliifts
from primitive to derived character states; superscripts indicate direction of
change or degree of change in those characters represented l:)y more tlian two
states; X ^ change in dorsal pattern not accompanied by shifts in any one of
the 11 characters; numbers in parentheses are numbers of evohitionary steps in
each Hne. See Table 8 and text for character states and explanation.

FROGS OF THE HYLA PARVICEPS GROUP

29

Taule 8. — Character states of 11 characters in species in the

Hyla parviceps group.

(0 = primiti\e state; 1 and 2 = derived states;

see text for hst of characters and explanation.)

c

a

•2

i

00

Si

05

^

Character

^

J5

1

2

8

►a:
o

g

.o

-a

►Q

-55

-5

s

§

a:

s:

a:

EC

a:

s:

A

1

0

1

2

2

1

B

0

0

0

1

0.5*

0

C

2

2

2

1

1

0

D

1

1

0

0

0

1

E

1

1

1

0

0

1

F

2

1

2

0

0

2

G

2

0

2

0

1

2

H

1

1

1

0

0

1

I

0

0

0

1

0

0

J

0

0

0

1

1

0

K

1

1

1

0

1

1

Total

11

7

10

6

6.5

11

" Females only.

The presence of a row of denticles in the tadpoles of Hyla
microps is the onl\' lan'al character that can be determined as
primitive in the group. Otherwise, the known tadpoles differ from
one another prineipalK' in coloration. We have no bases for deduc-
ing evolutionary change in the coloration. We have incomplete
data on mating calls, but the calls of //. pawkeps and hrevifrons
consist of series of uniform notes, whereas //. hokcnnanni produces
primary and secondary notes. Presumabl)' the complex call of //.
hokermanni is derived from the simple call of the others, a con-
clusion compatible \\'ith the phylogenies deduced from morpho-
logical and reproductive data.

Zoogeography

Recent works on distributional patterns in the tropical lowlands
of South America emphasize Quarternary climatic fluctuations and
the existence of forest refugia during times when climatic condi-
tions were drier than present (Miiller, 1968; Haffer, 1969; Van-
zolini and Williams, 1970). Miiller and Schmithusen (1970) and
\\iilleumier ( 1971 ) summarized Pleistocene climatic changes and
interglacial sea transgressions in South America. Haffer (1969)
pnnided e\idence from avian distributions and differentiation for
six forest refugia in Amazonian South America. Vanzolini and WW-
liams ( 1970 ) proposed four core areas for the differentiation of

30 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

populations of AnoUs clin/solepis and hypothesized an evolutionary
model of three expansions and two contractions of forest environ-
ments resulting in isolation, differentiation, and introgression of
Aiwlis chnjsolepis.

Accepting the preceding phylogenetic arrangement as a reason-
able approximation of the actual evolutionary sequences of mor-
phological characters in the Hijla parviceps group, we can super-
impose the cladistics of the frogs on the biogeographical patterns in
the lowland tropics. Our paleogeographic evolutionary model of
the Hyla parviceps group is predicated on cyclic alternation of
wetter and drier climates in the Pleistocene and Holocene, resulting
in expansion and contraction of lowland tropical forests, and the
existence of forest refugia during the drier periods (see Moreau,
1966, for paleogeographic evidence from Africa; Vanzolini and
Williams, 1970, for a survey of the literature documenting changes
in South America; and Van der Hammen and Gonzalez, 1960, for
dating of palynological data). Duellman (1972) and Heyer (1973)
suggested probable evolutionary histories of groups of South Amer-
ican hylid and leptodactylid frogs with respect to vicissitudes of
Quarternary climates.

The evolutionary model is outlined, as follows:

1. The prototype of the Hyla parviceps group was widespread
in Amazonia during a wetter period. This prototype had the primi-
tive states of the characters listed in the preceding section, aciuatic
eggs, and a simple mating call.

2. Climatic desiccation resulted in contraction of forests and
isolation of two populations: 1) the H. microps-parviceps stock in
an Amazonian refugium; 2) the H. ])revifr()ns-])okermanni-luteoceI-
lata-suhocularis stock in an upper Amazonian refugium (Napo or
Peruvian refugia of Haffer). The former stock developed more
sexual dimorphism in size, ulnar and tarsal tubercles in females,
one suborbital bar, and grav flecks on the venter. The second stock
developed two suborbital bars, rostral stripes, yellow spots on the
dorsal surfaces of the thighs; retained moderate sexual dimoi-phism
in size; and lost the larval denticles.

3. During the subsequent pluvial period both stocks spread
through the Amazon Basin; the H. microps-parviceps group ex-
panded into eastern and southeastern Brasil, and the other stock
invaded the coastal lowlands of northern South America and east-
ern Panama.

4. Climatic desiccation resulted in the isolation of two popula-
tions of the H. parviceps-microps group (H. microps in the Serra
do Mar refugium of Miiller in southeastern Brasil and //. parviceps
in an Amazonian refugium). The other stock was separated into
four isolates. Tlu> prototypes of //. hokeruunmi and //. ])revifrons
were isolated in separate refugia (probably Haffer's Napo and

FROGS OF THE HYLA PARVICEPS GROUP

80 ^0 60

50

40

D bokermanni

O brevifrons
▲ luteocellato

J \ >" ( >-^p'

\

• microps
■ porviceps

^■, ■ '^ '-'

/W^

▼ subocularis

0

0

A Jo .,--

O bokermanni, brevifrons
C bokermanni, porviceps
9 brevifrons, porviceps
• bokermanni, brevifrons.

porviceps

10

/

10

/

{

^ 1 '^'~1

/

20

20

0 500 1000 7 ^^^^ ^1

•
•

Kilometers ', ' J ^
1 I , \ ,

80 70 60

50

40

Fig. 8. Distribution of species in tlie H(//« parciceps group.

Peruvian rcfugia); //. liitcocellata was isolated in the Cordillera
de la Costa in northcMii \'enezuela ( a core area proposed b\- \'an-
zolini and Williams, 1970), and //. subocularis was restricted to
one or more of the small refugia in northwestern Colombia (Choco,
Nechi, or Catatumbo refugia of Haffer. 1969).

5. The development of more equable climates resulted in the
expansion of the Amazonian forests and of the ranges of the six
populations of frogs. However, due to the dr\' caatiniia and cerrado
of eastern Brasil, H. microps remained restricted to southeastern
Brasil. Hyki liitcocellata expanded its range eastward into the
Guianas, and H. subocularis moved into Panama. In the Amazon
Basin, the ranges of three species (H. bokermanni, brevifrons, and
parviceps) expanded from their respective refugia and became
sympatric in the upper part of the basin (Fig. 8).

EVOLUTION IN THE SYMPATRIC SPECIES

If we accept the proposed phylogenetic arrangement and the
suggested paleogeographic model, we are forced to hxpothesize
that the three species now occurring in sympatry in the upper
Amazon Basin e\oh'ed through geographic isolation at different
times {H. parviceps versus //. brevifrons and bokermanni) and in

32 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

different places (//. hrevifron.s and hokermanni) . Presumably,
genetic differences that evolved when the populations were isolated
have been reinforced through selection in sympatric populations.
Due to lack of data on any one of the species where it occurs in
the absence of the other two, we are unable to determine if there
is any character displacement.

According to Van der Hammen and Gonzalez (1960), the cli-
mate in northern South America has been becoming progressively
wetter for the past 4000 years, thus, it is possible that the three
species in the upper Amazon Basin have been sympatric for less
than 4000 years, before which time they were isolated in forest
refugia (Haffer, 1969).

We are concerned here with those mechanisms which are
operant in maintaining the specificity of the sympatric populations.
Following Miiller's (1942) classification of isolating mechanisms,
as modified by Fouquette ( 1960 ) , we recognize three major cate-
gories of isolating mechanisms — anti-mating, courtship, and post-
mating. The last category includes gametic incompatibility, hybrid
inviability, and hybrid sterility; we have no information on these
factors and no evidence of hybridization, so they are not considered
here.

Of the five kinds of anti-mating mechanisms, one (geographic
isolation) has already been accounted for by the hypothesized
differentiation through geographic isolation. Data on the others
can be summarized, as follows:

Habitat isolation. — Relative to breeding activities, //. parviceps
is a habitat generalist, usually present at sites where the other
species breed. Hijla hokermanni and brevifrons usually breed at
different sites.

Seasonal isolation. — Synchronous breeding activity is common,
but at least H. bokermanni and 77. parviceps have some different
periods of intense breeding activity.

Temporal isolation. — All three species are nocturnal breeders.

Climatic isolation. — All three species have a positive response to
periods of heavy rainfall.

Fouquette ( 1960 ) recognized only two categories of courtship
isolating mechanisms, whereas we recognize three; oviposition sites
are considered:

Auclitonj isolation. — The mating calls of three species are dis-
tinctly different in note repetition rate and duration of notes; fur-
thermore, the call of H. brevifrons is distinctive in pulse rate, that
of H. parviceps in dominant fre(iucncy, and tliat of H. bokermanni
in the presence of secondary notes ( Table 6 ) .

Mechanical isolation. — Because the females of the three species
have nearly identical snout-vent lengths, difference in size might
be an effective mechanical isolating mechanism only in //. parvi-

FROGS OF THE HYLA PARVICEPS GROUP 33

ceps, the males of which are much smaller than males of the other
species.

Oviposiiional isolation. — Ilyla parviceps deposits its eggs in
water and the other species place their eggs on vegetation over
water. The apparent behavioral and probable physiological differ-
ences of females and eggs, respectively, are potentially important
factors in isolation.

The three species broadly overlap in most categories of poten-
tial isolating mechanisms, but there are distinct differences in two
categories, ethological and ovipositional. Fouquette (1960) found
that the mating call was the primary isolating mechanism operant
in three sympatric species of Hyla in Panama. Duellman (1967)
reached the same conclusion in a study of a breeding community
of ten species of hylids in Costa Rica. Martof and Thompson
(1958) and Littlejohn and Michaud (1959) provided experimental
evidence demonstrating that females respond positively to the calls
of the males and that females can discriminate between calls of
males of their own and other species. Littlejohn (1964, 1965) pre-
sented evidence on evolution of differences in mating calls througli
geographic isolation and reinforcement of these differences in sym-
patric populations. We conclude that at least some of the differ-
ences in mating calls in the three species of the HyJa parviceps
group evolved at times when the populations were geographically
isolated from one another; fmthermore, thes(> differences may have
been reinforced through selection upon subsequent geographical
sympatiy.

The differences in o\'ip()sition sites invoKe a modification in
reproductive strat(>gy from the presumed primitive strategy of
placing eggs in water (H. parviceps) to a derived strategy of plac-
ing eggs on vegetation over water (H. hokcrmanni and hrevifrons) .
Salthe and Duellman (1973) showed that within a given anuran
reproductixe mode there exist positive correlations between ovum
size, clutch size and female snout-vent length, and negative correla-
tions between clutch s\zc and o\'um size and between o\'um size
and rate of de\elopment. Among the three species under con-
sideration, HyJa parviceps with aquatic eggs has the largest clutch
size, smallest ova, and shortest duration of development (Table 7).
Females of the other species are no larger than H. parviceps; thus,
as concluded by Salthe and Duellman (1973), the achievement of
a larger ovum in H. hokerinanni and hrevifrons, without an in-
crease in body size of the females, was b\' selection for a change in
the site of intraoval development, in this case vegetation over water.
Larger eggs produce larger hatchlings better adapted to make the
transition from leaf to water. Because the reproductixe modes are
different in the two lineages in the group, it is most reasonable to
assume that tliis differentiation took place only once and that by

34 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

the time the three species became sympatric in the upper Amazon
Basin the differences in reproductive strategy were ah^eady in
existence. However, the differences are important aspects of re-
source partitioning in synchronously breeding frogs within one
community.

Salthe and Duellman (1973) noted that in small anurans: ". . .
it seems to be more advantageous to maintain a certain minimal
number of eggs in the clutch than it is to maintain the size of the
hatchling. Presumably, clutch sizes are already at or near, a lower
limit allowable for the production of an adequate amount of
genetic variability per unit time. Thus, in general, selection in
small species seems to function by maintaining minimal clutch sizes
and maximizing the sizes for relatively tiny hatchlings." Our re-
productive data indicate that despite differences in mode of re-
production and ovum size, the hatchlings are about the same size
in H. hokermanni and parviceps, but larger in H. brevifrons, which
has the smallest clutches ( Table 7 ) .

The problem of production of an adequate amount of genetic
variability, as related to fecundity, per unit time perhaps is an
important factor in the great amount of sexual dimorphism in size
in the Hyla parviceps group. At Santa Cecilia, the three species in
the group occur in swamps inhabited by many other species, of
which 17 Hyla are known to breed in the same ponds or swamps as
do members of the Hyla parviceps group. All of these are larger
than the species in the Hyla parviceps group. Selection might
favor reduction in size in response to competition for food, calling
sites, and oviposition sites. However, reduction in size of the fe-
males is more restricted than in males due to the limitations im-
posed by egg-carrying capacities. Schoener (1967, 1969, 1970)
documented evidence for sexual selection in size of West Indian
Aiiolis, noting that extremes of sexual dimorphism in size are
reached on islands inhabited by only one species of Anolis; on these
islands the sexes exist ecologically as two species using different
resource states, such as perches and food. However, in the Hyla
parviceps group, the presence of many other species utilizing a
variety of resource states negates the hypothesis of intraspecific se-
lection in size solely for resource partitioning and suggests that the
reproductive necessities of the females balance selection for small
size, which is evident in males.

SUMMARY

Among the small Neotropical Hyla having a diploid number of
30 chromosomes and reduced mouth parts in the tadpoles, one
group of six species is characterized by great sexual diTiioiphism in
size, more cranial ossification, and aspects of coloration (patterned
thighs, suborbital bars, and iris color). Six species are included in

FROGS OF THE HYLA PARVICEPS GROUP 35

the group: H. hokermanni Goin, H. hrevifrorvs new species, H.
luteocellata Roux, H. microps Peters, H. parviceps Boulenger, and
H. suhocularis Dunn. Ihjhi rondoniae Bokermann, 1963, is placed in
the synonymy of Hyla Jwkermanni Goin, 1960. The group is wide-
spread in the Amazon Basin, southeastern Brasil, northern lowlands
of South America and eastern Panama.

Although interspecific differences exist in size, proportions,
amount of webbing, and tubercles, the principal taxonomic char-
acters are in the coloration. These include the number of pale
suborbital bars, presence or absence of canthal and rostral stripes,
presence and width of dorsolateral stripes, and pattern on the dor-
sum, venter, and especially the thighs. All of the tadpoles have one
row of labial papillae. Tadpoles of H. microps have one row of
denticles; denticles are absent in the others. The known tadpoles
are brightK' colored and have different patterns of transverse bars
or longitudinal stripes. The mating calls consist of series of short,
high-pitched notes; H. hokermanni produces primary notes fol-
lowed by shorter secondary notes.

It is suggested that the species arose through isolation in forest
refugia in the tropical lowlands during dr\' periods in the Pleisto-
cene. The proposed phylogeny indicates that \\\c two most primi-
tive species now occupy different regions — H. microps in southeast-
ern Brasil and H. parviceps in the upper Amazon Basin. The three
most advanced species also have allopatric distributions — H. lioker-
manni m the upper Amazon Basin, U. luteocellata in coastal north-
ern South America, and //. suhocularis in eastern Panama and
northwestern Colombia.

Three species (//. hokermanni, hrevifrons, and parviceps) occur
at Santa Cecilia in Amazonian Ecuador. Intensive field studies
there continuously for 13 months revealed slight differences in
habitat and seasonal activity. One species {H. parviceps) lays eggs
in the water, whereas the other two place eggs on vegetation over
the water. Clutch size is larger and ovum size is smaller in U.
parviceps than the others.

It is suggested that differences in mating calls and oviposition
sites are the primary isolating mechanisms operating in the sym-
patric species. Furthermore, it is suggested that natinal selection
for small size in comparison with many sympatric congeners is
balanced in females by the necessity for maintaining an adequate
number of eggs to insure genetic variability in the population. In
the absence of such pressure in the males, they have become much
smaller than the females.

RESUMEN

Entre las pequcfias ranas neotropieales del genero Uijla con un
numero diploide de 30 cromosomas y renacuajos con bocas de

36 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

estructuras pequenas, un grupo de seis especies se caracteriza por el
gran dimorfismo sexual de su tamaiio, mayor osificacion craneal, y al-
gunos aspectos de su coloracion (diseiios en los muslos, barras sub-
orbitales, y color del iris). Las seis especies incluidas en el grupo
son: H. hokermanni Coin, H. brevifrom nueva especie, H. luteo-
celkita Roux, H. microps Peters, H. parviceps Boulenger, y H. suh-
ocularis Dunn. Hijla rondoniae Bokermann, 1963, se pone en el
sinonomia de Hijla hokermanni Goin, 1960. Este grupo se encuen-
tra difundido en la Cuenca Amazonica, el sureste brasileiio, las
tierras bajas del norte de Suramerica, y el este panameiio.

Aun cuando existen diferencias interespecificas de tamafio, pro-
porciones, extension de las membranas interdigitales, y tuberculos,
las principales caracteristicas taxonomicas giran alrededor de la
coloracion. Esto incluye el numero de barras suborbitales claras, la
presencia o ausencia de las rayas cantales y rostrales, la presencia
y ancho de las rayas dorsolaterales, y el diseno dorsal, ventral, y
especialmente de los muslos. Todos los renacuajos tienen una fila
de papilas labiales. Los renacuajos de H. microps tienen una fila
de denticulos que no existe en los otros. Los renacuajos que se
conocen son de coloracion luminosa y tienen distintos disenos de
barras transversales y rayas longitudinales. La llamada de reclamo
consiste de series de notas cortas y altas; Hijla hokermanni produce
notas primarias sequidas de notas secundarias mas cortas.

Se sugiere que estas especies surgieron como resultado de su
aislamiento en refugios en la selva de la tierra tropical baja durante
los periodos secos del Pleistoccno. La filogenia propuesta sugiere
que las dos especies mas primitivas ocupan ahora regiones dis-
tintas — H. microps ocupa el sureste brasileiio y H. parviceps la
parte superior de la Cuenca Amazonica. Las tres especies mas
evolucionadas tambien tienen distribuciones alopatricas — H. hoker-
manni en la parte superior de la Cuenca Amazonica, H. hiteocel-
lata en la costa norte de Suramerica, y H. suhocularis en el este
panameiio y noroeste colonibiano.

Tres especies (H. hokermanni, hrevifro7is, y parviceps) se en-
cucntran en Santa Cecilia en la Amazonia Ecuatoriana. Estudios
continuos e intensivos de campo durante un periodo de 13 nieses
revelan pequeiias diferencias en su "habitat" y en la actividad esta-
cioiial. Una de las especies (H. parviceps) pone sus liuevos en el
agua, mientras que las otras dos los ponen en la vegetacion sobre
el agua. En H. parviceps el numero de huevos es mayor y el
tamaiio de los huevos es meiior que en las otras especies.

Se sugiere que diferencias en las llamadas de reclamo y los
lugares de oviposicion son los mecanismos principales de aislami-
ento (Ml las especies sinipatridas. Se sugiere ademas que la selec-
cion natural a favor de un taniano pequeiio en estas especies en
compariacion con muclios otros congeneres simpatridos esta con-

FROGS OF THE HYLA PARVICEPS GROUP 37

trarestada en las hcmbras por la necessidad de mantener un niimero
adecuado do huevos con lo que asegura la variabilidad genetica de
la poblacion. En la ausencia de esta presion, los machos se ban
mantenido mas pequenos.

SPECIMENS EXAMINED

Hyla bokermanni

BRASIL: Acre: Cruzeiro do Sul, WCAB 12148. Rondonia: Rondonia,
WCAB 7845.

ECUADOR: Napo: Limon Cocha. UIMNH 64799-800, 90069; Santa
Cecilia, KU 104432-.3 (skeletons), 105118-9, 105121-2, 105125-6, 105131,
105189, 109355-7, 109452, 109468, 123051-8, 123060, 123062-71, 123089-91,
124193 (tadpoles), 126.362, 126541-5, 126547, 126552-5, 143131, 143188,
146288-95, 146414, 146419, 146800-1 (tadpoles), 1,50026-51, 152290-1 (tad-
poles), 152419-28, 152535, 1,52750-3 (skeletons), UMMZ 129.325 (4).
Pastaza: 3 km S Puyo, KU 127087.

Hyla hicvifrons

BRASIL: Paid: IPEAN, 5 km E Belem, KU 127846.

COLOMBIA: Putumaijo: Santa Rosa de Sucumhios, AMNH 88068-80.

ECUADOR: Na/w; Lago Agrio, 126556-8; Puerto Libre, KU 123092-4;
Santa Cecilia, KU 10,5063, 105107-11, 105120, 105127-30, 1051.32-4, 105190,
10700.3-4, 109449-51, 1094.5,3-67, 111786-802, 112341 (eggs), 12,3049-.50,
12,3059, 12,3072-88, 125899-900 (tadpoles), 126342-61, 126363-71, 126546,
126,548-51, 1,36298, 143189, 146281-7, 146413, 146415-8, 1,500,52-74, 1,52429-
.32, 1,52536-7 (tadpoles), 152754-8 (skeletons), UMMZ 129,324 (2). Pastaza:
Puyo, CAS 85142.

FRENCH GUIANA: Inini: Cri(iue Eleupocigne, Upper River Ovapak,
LG 1119, 1308, 1311-12.

PERO: Loreto: Pampa Hermosa, Rio Cushabatay, AMNH 42.358; Rio
Utoquinia-Rio Tapiche, AMNH 4.3576; Tacsha Huachivacu, Rio Morona,
AMNH 43045; Yurimaguas, Rio Huallaga, BMNH 84.2.18..53.

Hyla luteocelhta
VENEZUELA: Aranua: Cumhoto, I'MMZ 11.3890 (2); Monte Oscuro,
BMNH 1968.73-4, KU 12.5861-5, UMMZ 1,30085-6.

Hyla microps
BRASIL: Param'i: Curitiba, UMMZ 104118 (2), 104129 (6), 104130,
104160. Rio de Janeiro: Barro Branco, SMF 41217; Novo Friburgo, ZMB
7472; Teresopolis, BMNH 1947.2.14.10. Sao Paulo: Campo Grande, Santo
Andre, KU 74266-8, 74269 (.skeleton), 92073-7, 92078 (skeleton); Campos
de Jordao, KU 112398 (tadpoles); 10 km W Casa Grande, KU 129838. Santa
Caiarina: UMMZ 58515; Humboldt, AMNH 15,573-82.

Hyla particeps

COLOMBIA: Piittimayo: Santa Rosa de Sucumbios, AMNH 88081-2.

ECUADOR: Napo: Bermejo No. 4 (well site), Sierra Umbaqui, KU
122963; Cuyabeno, UIMNH 54134-6, 54172, 59625-8, 90108-9; Dureno, KU
105113-4; Lago Agrio, KU 126480; Limon Cocha, KU 99223-8, 99229 (skele-
ton), UIMNH 64805, 90076; Santa Cecilia, KU 104428-31 (skeletons),
105046-62, 10.5064-106, 105112, 107048-9, 109433-46; 11764-84, 112339-40
(tadpoles), 122925-62, 12.3061, 126462-79, 143167-73, 146307-11, 146790-1
(tadpoles), 150,309-56, 15230.3-4 (tadpoles), 152547 (tadpoles), 152759-63
(skeletons), UMMZ 129279 (9); Tena, UIMNH .59628, 900,59-64; UMMZ
123903. Pastaza: Abitagua, UMMZ 90416-7; Chontoa, KU 120910-13,

38 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

121407; Mera, KU 120909; 3 km S Puyo, KU 127086, 127090; 25 km NNE
Puyo, KU 146792 (tadpoles); Rio Conambo, mouth of Rio Shione, USNM-
GOV 7143; Rio Oglan, upper Rio Curaray, USNM-GOV 9526; Sarayacu,
BMNH 1947.2.13.51. Timgumhua: Llanganate area, FMNH 23522-3; Mirador,
BMNH 99.10.30.49; Rio Negro, KU 120913-14, 121408.

PERtJ: Cuzco: Pilcopata, KU 139228. Loreto: Balta, Rio Curanja, LSU
25719-57; Iquitos, AMNH 42086; headwaters Rio Utoquinia, AMNH 43314.

Hyla subocularis

COLOMBIA: Cordoba: Rio Manso, CM 50447-50.

PANAMA: Darien: Laguna, KU 77348-59, 77620; Rio Chucunaque,
AMNH 51777; Rio Chucunaque at first creek above Rio Tuquesa, AMNH
41117; Rio Ucurganti, 7 km above mouth, KU 116785, 116784 (tadpoles);
Tacarcuna, KU 77360-403, 77621, 77688-90 (skeletons), UMMZ 124780.

LITERATURE CITED

BOKERMANN, W. C. A.

1963a. Duas novas especies de Hyla de Rondonia, Brasil ( Amphibia,

Salientia). Rev. Brasil. Biol., 23:247-250.
1963b. Girinos de anfibios Brasileiros — 1. (Amphibia, — Salientia). An.

Acad. Brasil. Cien., 35:465-474.

1964. Notes on tree frogs of the Hyla marmorata group with description
of a new species (Amphibia, Hylidae) Senckenberg Biol., 45:243-
254.

1966. Lista anotada das localidades tipo de anfibios Brasileiros. Impres.
Serv. Documentacao, Sao Paulo, 183 pp.

Bog ART, J. P.

1973. Evolution of anuran karyotypes, in Vial, J. L. (ed. ) Evolutionary
biology of the anurans: contemporary research on major problems.
Univ. Missouri Press, Columbia: 337-349.

BOULENGER, G. A.

1882. Catalogue of the Batrachia Salientia s. Ecaudata in the collection

of the British Museum, ed. 2, London, xvi-|-503 pp.
1920. Descriptions of two new frogs from Brasil. Ann. Mag. Nat. Hist.,
(9)5:122-124.
Camin, J. H. and R. R. Sokal

1965. A mediod for deducing branching sequences in phylogeny. Evo-
lution, 19:311-326.

Cochran, D. M.

1955. Frogs of southeastern Brazil. Bull. U.S. Natl. Mus., 206:xvi+423
pp.

Cochran, D. M. and C. J. Coin

1970. Frogs of Colombia. Ihid., 288 :xii -1-655 pp.

CoLWELL, R. K. and D. J. Futuyma

1971. On the measurement of niche breadth and overlap. Ecology, 52:
567-586.

DUELLMAN, W. E.

1967. Courtship isolating mechanisms in Costa Rican hylid frogs. Her-
petologica, 23:169-183.

1969. A new species of frog in the /////« jxirviceps group from Ecuador.
Ibid., 25:241-247.

1970. The hylid frogs of Middle America. Monog. Mus. Nat. Hist. Univ.
Kansas, l:xi-|-753 pp.

1972. South American frogs of the Hyla lostiata group (Amphibia,
Anura, Hylidae). Zool. Mededel., 47:177-192.

FROGS OF THE HYLA PARVICEPS GROUP 39

Dunn, E. R.

1934. Two new frogs from Darien. Amer. Mus. Novitates, 747:1-2.
FouQUETTE, M. J., Jr.

1960. Isolating mechanisms in tliree sympatric treefrogs in the canal
zone. Evolution, 14:484-497.
GoiN, C. J.

1960. Description of a new frog of the genus Hijla from northwestern
Brazil. Ann. Mag. Nat. Hist., 13(2) :721-724.

GOSNER, K. L.

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

1969. Speciation in Amazonian forest birds. Science, 165:131-137.
Heyer, W. R.

1973. Systematics of the marmoratus group of the frog genus Leptodac-
tijlus (Amphibia, Leptodactylidae). Contri. Sci. Los Angeles Co.
Mus., 251:1-50.
Kluge, a. G. and J. S. Farris

1969. Quantitative phvletics and the evolution of anurans. Svst. Zool.,
18:1-32.

LiTTLEJOHN, M. J.

1964. Geographic isolation and mating call differentiation in Crinia sig-
nifera. Exolution, 18:262-266.

1965. Premating isolation in the Hi/la ewingi complex (Anura: Hy-
lidae). Ibid., 19:234-243.

LiTTLEJOHN, M. J. and T. C. Michaud

1959. Mating call discrimination by females of Strecker's chorus frog
{Pseudacris streckeri). Texas Jour. Sci., 11:86-92.
LuTZ, B.

1973. Brazilian species of Hyla. Univ. Texas Press, Austin, xix+260 pp.
Martof, B. S. and E. F. Thompson

1958. Reproductive behavior of the chorus frog Psctiduciis nigiita. Be-
haviour, 13:243-258.
Melin, D.

1941. Contribution to the knowledge of Amphibia of South America.
Gciteborgs Kungl. Vetensk.-och Vitterh.-Sam. Handl., Ser. B, 1(4):
1-71.

Mertens, R.

1950. Froschlurche aus Rio de Janeiro und seiner Umgebung. Wochen-
schr. Aquar. Terrarien, 44:173-188.

MiJLLER, H. J.

1942. Isolating mechanisms, evolution, and temperature. Biol. Symposia,
6:71-125.

MiJLLER, P.

1968. Die herpetofauna der Insel von Siio Sebastiao (Brasilien). Saar-
bracker Zeit., 68 pp.

MuLLER, P. and J. Schxhthijsen

1970. Probleme der genese siidamerikanischer Biota, in Deutsche geog-
raphische Forschung in der Welt von Heute. Kiel: 109-122.

Moreau, R. E.

1966. The bird faunas of Africa and its islands. London, 424 pp.
Peters, W.

1872. Uber eine Sammlung von Batrachiern aus Neu-Freiburg in Bra-
zilien. Nhmatsb. Akad. Wiss. Berlin, 1872:680-684.

RiVERO, J. A.

1969. On the identity and relationships of Hylo hiteocellata Roux (Am-
phibia, Salientia). Herpetologica, 25:126-134.

40 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

Roux, J.

1927. Contiil)ution a I'erpetologie du Venezuela. Verh. Naturf. Gesell.
Basel, .38:252-261.
Salthe, S. N. and W. E. Duellman

1973. Quantitative constraints associated with reproductive mode in
anurans. in Vial, J. L. (ed. ). Evolutionary biology of the anurans:
contemporary research on major problems. Univ. Missouri Press,
Columbia: 229-249.
Savage, J. M. and W. R. Heyer

1967. Variation and distribution in the tree-frog genus PlnjUomcdusa, in
Costa Rica, Central America. Beitr. Neotrop. Fauna, 5:111-131.
SCHOENER, T. W.

1967. The ecological significance of sexual dimorphism in size in the
lizard Anolis conspersus. Science, 155:474-477.

1969. Size patterns in West Indian Anolis lizards: I. Size and species
diversity. Syst. ZooL, 18:386-401.

1970. Size patterns in West Indian Anolis lizards. II. Correlations with
the sizes of particular synipatric species — displacement and con-
vergence. Amer. Nat., 104:155-174.

Truer, L.

1973. Bones, frogs, and evolution, in Vial, J. L. (ed.). Evolutionary

biology of the anurans: contemporary research on major problems.

Univ. Missouri Press, Columbia:65-132.
Van de Hammen, T. and E. Gonzalez

1960. Holocene and late glacial climate and vegetation of Paramo de
Palacio (Eastern Cordillera, Colombia, South America). Geol.
en Mijnb., .39:737-746.

Vanzolini, p. E. and E. E. Williams

1970. South American anoles: the geographic differentiation and exolu-
tion of the Anolis cJirysolepis species group ( Sauria, Iguanidae).
Arq. Zool., 19:1-298.

Vuilleumier, B. S.

1971. Pleistocene changes in the fauna and flora of South America. Sci-
ence, 173:771-780.

Wagner, W. H., Jr.

1961. Problems in the classification of ferns, in Recent Advances in
Botany. Univ. Toronto Press, Toronto:841-844.

657