HARVARD UNIVERSITY

Library of the

Museum of

Comparative Zoology

u^//

OCCASIONAL PAPERS

of the

MUSEUM OF NATURAL HISTORY
The University of Kansas
Lawrence, Kansas

NUMBER 28, PAGES 1-37 JUNE 17, 1974

A RE-EVALUATION OF THE HYLA BISTINCTA

SPECIES GROUP, WITH DESCRIPTIONS

OF THREE NEW SPECIES

(ANURA: HYLIDAE)

By

Jan CALD^VELL^

Five stream-breeding tree frogs occurring in mountainous areas
of Mexico were assigned to the Hyla bistincta group by Duellman
(1964). Since that time, five species referable to this group have
been described by Adler ( 1965 ) , Duelhnan ( 1968 ) , Straughan and
Wright (1969), and Adler and Dennis (1972). Most of these spe-
cies have relatively limited ranges in mountainous areas, and are
allopatric. While working in Oaxaca, Mexico, in 1969-1970, I dis-
covered three additional species referable to this group.

The five species included in the H. bistincta group by Duellman

(1964) are Hyla bistincta Cope, Hyla charadricola Duellman, Hyla
robertsorum Taylor, Hyla pachyderma Taylor, Hyla crassa (Brocchi).
Duellman used the following combination of characters to separate
this group from all other Middle American species groups: 1) ab-
sence of the quadratojugal, 2) non-projecting prcpollex, 3) long
fingers with little webbing, and 4) stream-inhabiting tadpoles hav-
ing 2/3 tooth rows and two or more rows of labial papillae. Adler

(1965) described Hyla chryses and Hyla pentheter and included
them in the bistincta group; he recognized two subgroups. One
subgroup is composed of medium-sized, slender, thin-skinned frogs

^ Assistant Director, State Biological Survey of Kansas, 2045 x\ venue A,
Campus West, Lawrence, Kansas 66045. Tliis research was carried out when
the author was a graduate student in the Museum of Natural History and the
Department of Systematics and Ecology, University of Kansas.

2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

without nuptial spines in breeding males, and with a distinct axillary
membrane. These frogs occur at intermediate elevations, 2000-2600
m, in oak-pine forests of eastern Hidalgo, northern Puebla, and in
central Guerrero. The second subgroup is composed of frogs which
are medium to large, robust, and thick-skinned and which have
nuptial spines on the first and second fingers of males but have no
axillary membrane. These species are found at intermediate to high
elevations (1600-3050 m) from Jalisco to Oaxaca and north to east-
ern Hidalgo.

Duellman (1968) described Hyla siopela from Cofre de Perote,
Veracruz, Mexico, and placed the seven species then known into
three categories: 1) bistincta and penthefer — long anal sheaths; 2)
charadricola and chryses — small species with axillary membranes
and lacking nuptial excrescences in breeding males; 3) crassa, pachy-
derma, and robertsorum — short heads, round snouts, short anal
sheaths, and nuptial excrescences in breeding males.

Hyla siopela was allocated to the last subgroup although it differs
from the other three species included by the shape of the snout, and
by the presence of a distinct rostral keel. Duellman also suggested
that once the extent of the variation of these four species is known,
they might represent a single species, although crassa has fully
webbed feet and pachy derma has large nuptial spines.

Straughan and Wright (1969) named Hyla bogertae from the
district of Sola de Vega in Oaxaca, and allocated it to Duellman's
(1968) third subgroup. Duellman (1964, 1968) regarded Hyla
robustofemora Taylor as a synonym of Hyla crassa, but Straughan
and Wright considered crassa a nomen dubium until males and fe-
males from an extant population could be associated.

The tenth frog allocated to this group is Hyla tnykter, described
by Adler and Dennis (1972), who suggested that mykter is most
closely related to siopela. Duellman's (1968) third subgroup, con-
sisting of species typically found in moist oak-pine forests at inter-
mediate to high elevations, was further subdivided by Adler and
Dennis into two series: 1) those that are more robust and have
more glandular skin, including bogertae (Oaxaca, 2650 m), crassa
(Oaxaca, 2300 m), and pachydcnna (Veracruz, 1600 m); and 2)
those species that are less roljust and have thinner skin, including
mykter (Guerrero, 1985-2520 m), robertsorum (Puebla and Hi-
dalgo, 2250-3100 m), and siopela (Veracruz, 2500-2650 m).

The purposes of this paper are to: 1) describe the three new
species of tree frogs discovered in Oaxaca, and 2) to discuss phenetic
similarities among these three new species, the other 10 species
allocated to the Hyla bistincta group, and Hyla arborescandens.
The Hyla bistincta group herein is divided into four species groups:
1) the Hyla bistincta group, 2) the Hyla crassa group, 3) the Hyla
charadricola group, and 4) the Hyla arborescandens group (Table 1).

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 3

The plienctic relationships were determined by discriminant
function and discriminant classification analysis. Comparisons of
the different statistical methods are made and the advantages of
each discussed. Results of the statistical analysis provide a different
arrangement of species than the combination of characters used in
the past to define subgroups within the HijJa bistincto group. The
statistical results are considered preliminar\- because of the small
number of specimens available of some species, and the consequent
scarcity of morphological, skeletal, tadpole, and ecological data.
Therefore, the definitions of the four new species groups are based
on combinations of characters previously used by other workers.

Acknowledgments

I thank \\^illiam E. Duellman for man\' profitable discussions
during the course of the study and for reading and improving pre-
liminary drafts of the manuscript. I am grateful to Paul B. Robert-
son for companionship and field assistance during ten months spent
in Oaxaca, Mexico. For advice and assistance while performing the
statistical analysis, and for reading a preliminary draft of this sec-
tion, I thank Jon W. Robinson and Norman A. Slade. Stephen
Edwards also read a preliminar\- draft of the section on statistical
analysis and offered many helpful comments. Linda Basel provided
the illustrations of tadpole's and their moutliparts depicted in figures
1 and 2. I am grateful to Linda Trueb for her advice and encourage-
ment as I prepared the other illustrations.

The following curators kindh' allowed me to examine specimens
in their care: D. F. Hoffmeister and Dorothy M. Smith, University
of Illinois Museum of Natural History (UIMNH); the late James A.
Peters, National Museum of Natural History (USNM); Hymen
Marx, Field Museum of Natural History (FNLNH); and Charles F.
Walker, University of Michigan Museum of Zoology (UMMZ).
William E. Duellman allowed me to examine specimens in the
Museum of Natural History at the Universit>' of Kansas (KU). I
thank Kraig Adler for the loan of specimens from Guerrero, Mexico,
for comparative purposes. I am indebted to Ronald Altig for many
valuable discussions about Mexican hylids and for numerous speci-
mens he generously donated to the Museum of Natural History at
University of Kansas.

Field work in Mexico was made possible by a research grant
from the Organization of Tropical Studies, and by travel funds from
a National Science Foundation grant (GB-4446X, George W. Byers,
principal investigator) awarded through the University of Kansas
Committee on Systematics and Ex'olutionary Biolog)'. Both these
grants were received jointly with Paul B. Robertson. A summer
traineeship was provided by an NSF grant (GB-S785, Robert S.

OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

Ck

p

o

Sh

bo

<3

•K»

u

c

•**

•K>

•2

.S

a

1 — *

»

53

<v

-C

4->

_c

/— **

'V

tn

<D

a

TS

3

?

o

G

&

•Fl

to

>^

A->

<U

C

ft

<D

t«

?:

CA

mh

<a

O

"o

G

(U

O

CO

!U

s

J3

to

4-'

^

O

O

<a

^

■*->

<a

^

-■->

<x>

^

1
^

cr,

TS

to"

4-1

13

*—<

C

a

<3

;ia

JJ

to

_

!»

CS

K

r-

o

O

^5

•J3

>

^

0;

Si

"o!

a:

to

<a

W)

C

«

K

^

CO

<D

■V

13

o

^

<t>

p<

>.

H

i-H

H

h)

c
o

>

5

bc

CS
O
O

T!

C

a

4^

•J' CO

C^

C3 CO

4-»

o m

I u

r^

OJ o

"rt

C"-"

ffi

KcS

oj i

C c

Cl, Pi

O
O
00

to

3

^

'^

O

03

3

O

_o

'T3

1

"o

'o ^

0) «

a

cS

p-5 i

'ij

o

r—

c^ ^

c

U«

O .3

a 0^

4->

03

CO

a

1

S'S

Si

3

a;

03

H g

o

s

O

o

o

o
o

o -

o o

•a ^-

to „ 3

*"' ^ '"'

^ l' rt 3 1:

i2 .« O

^ 2 ^ O -^ C/2 .3

o
o
o

(M

I

o
CO

. Hi

-I Vh

^^ §

a^T3

Q Qj I— '

r- Ci ci

• 3 •-< o

_2 ,<S 03

S f« rt

o

CS

>

OS

p
a

o
S

3

c«

C/2

Z

o

r-«

0)

(— '

a

J2

t>

■y.

CO

■■--I

,^

•y.

•-^

03

73

OO

o
o
CO

CM

I

o

o
o

Ol

T3

3
c«

O
O
CO
CM

lO

0; 3

O '^

Z z

a

3

O

^

bc

'•-i

CO 2i

jj

<u a

'S

<

1^

a M

ta

■w C

c:i -S

S-y

= -S

s

■§-5

:s^

~2

« =*

"ii

1^5

a;

as

a

3

3

O

03

^

CO

3

0/

Q

a

Q

v:

^-2

-2

c

o

•i

"Q

5^

a

^-^

s

5

2

o

"^

-2

.^

^<

"Si

a:

a;

o
_o

a;
a
>.

U

o

CM
O
CM

I

O

0)

a

^ to .3
■2 3 N
kT 0; 1-

to

03

i2 .na

CD «

CO 5

o

3 (D

r, c/2

o

-^ 3 ■=
— , 03 CS

03

g 03 S

p-i s; O

o

60

3 ^

•^ "^ 3

z •• '^

"^ 03 03

;3 0>

to

Q

CO

Si

Si

O

o

Mh

M-l

%■

n3

en

3

i

O

3

u

Oh

o o

in o

O

o

00 CO

o

lO

^ CM

CO

CO

i-H

CM

£8

U

ctf
(D

03

T3

B

•c

CO

« 03

O

rt Z

+J CO JJ

o -^ a
o 1-1 H

3

o

03

o

o^

a
>.

o
o

•a

ID

a.

0^

2

'T^

03

<

u

CO

CO

**

;^

"C

v.

-a

K^

Q

■o

^

O

o

CJ

CO

o

a>

i^

'o

m

ttj

"— '

a

<2

to

^

i

2

^

-2

sj

fc^

•^

~S>

a:

a:

o

r*

SLa
tary

OJ

^

■TJ

•4-'

- 3

r-«

'S

s-a

n

N

-^■£ o

PL,

t^

CO _ 03

H

: 3 >,

o

CS

<4H

o

•5

3
o

ixaca: 1
fradia, i
Rio Ato

^

C/D

& Wright Oa
Co
of

u

3

^

03

'>>

-s,

03

bC

H

3

OS

e

ta

<»

"e

g>

;^

<»

"cj

tjD

a

O

a

-a

^ -2

RE-EVALUATION OF HYLA BISTIXCTA SPECIES GROUP

CO

cS
O
f
<D

c

o

I

o

CO

c
o

■a

O

^ .^

!U

%m

O W5

u

c

-H lU

o

<-M

C3 <j_

<+<

-i^

u

oi

O ,

<^

c

1 "0

C o

i

i

f—

C

O
CO

I— I

Pi o

o
o

Ph

S Q

^ ,P

1 o

Ph S

CO

o

o

O

k:i

lO o

CN

o

lO t-

IC

CO

^6

CM

6

O lO

lO

lO

O CD

00

IC

\n <M

C5

<M

(M

I-^

es

O

13 rt

OJ

■r' 3

-1

c3

a

r-

-a t s

c^

'^

adre
then

o

^

c
-^^

ol3 Si

*5 S

a

Cj

1:2

i*^

0/

pe lo
rroP
Juan

>- c

O

•Z3

^d

C

■4-*

f- U 't:

*-' c (^
*-^ -^ ^^

o o a;

a oj ii.

:- S «
> H S

z

2
.a

C

S 2

C/D

s

Chico
onal

0)

4-1

0)

*-*
c

I— 1

2

La

cq

a

c

C/D

CO

w '1

^H

C/3

i-H

^

1— 1

o

Oh

'v

N

N

60

f3

;. ^

N

ID

o -o' o

Oaxaca
Cerro F

13

c

1

5i Oi

O

u

n3

Hi

u

C5 ^S

O c8

A

3

O

febg

en

W5 ""^

I

'a

'B

^

01 >>

c

CS

Q
=3

U

2

§-^

c*

"^

o

1

Ci

u

i

s
S

D

Q
-2

Si.

o

<

Si

^

-2

^-2

-2

^

-2

"a*

2n

1^"

2>

»>

s>

!l!

!3:

s:

^5

^

6 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

Hoffmann, principal investigator), also awarded through the Univer-
sity of Kansas Committee on Systematics and Evolutionary Biology.
I was the recipient of the 1971-72 Florence Edna Rowe Dissertation
Fellowship awarded by the American Association of University
Women. Without the above support, this study could not have been
completed.

Materials and Methods

A total of 283 frogs, 103 adult females and ISO adult males (Table
2) was analyzed using several multivariate statistical methods.
Thirty-two characters were recorded for each specimen. Fifteen
morphometric characters (continuous variables) were measured to
the nearest 0.1 mm using either dial calipers or an ocular micrometer
on a dissecting microscope. The other 17 characters were coded.
The 32 characters are:

1. Snout-vent length — tip of snout to vent.

2. Tibia length.

3. Radial-ulnar length — elbow to tip of disc of fourth finger.

4. Foot length — proximal edge of metatarsal tubercle to tip of disc of
fourth toe.

5. Head width — width of head at widest point.

6. Head length — distance from end of skull at jaw to tip of snout.

7. Diameter of eye.

8. Diameter of tympanum.

9. Internarial distance.

10. Interorbital distance.

11. Eye to nostril.

12. Eye to tympanum.

13. Length of metatarsal tubercle.

14. Diameter of disc of third finger.

15. Diameter of disc of fourth toe.

16. Lateral view of snout — coded: truncate (1); round (2); sloping (3);
acuminate (4); protruding (5).

17. Dorsal view of snout — coded: truncate (1); round (2); acuminate (3).

18. Rostral keel — coded: al^sent ( 1 ) ; weak ( 2 ) ; present ( 3 ) .

19. Supratympanic fold — coded: absent (1); weak (2); present (3).

20. Tympanum — coded: concealed (1); partiallv concealed (2); distinct
(3).

21. Skin — coded: smooth ( 1 ); weakly tuberculate ( 2 ) ; tuberculate ( 3 ) .

22. Thoracic fold — coded: absent ( 1 ) ; weak ( 2 ) ; present ( 3 ) .

23. Axillary membrane — coded: absent ( 1 ); present (2).

24. Tubercles on ventrolateral edge of forearm — coded: absent (1); pres-
ent (2).

25. Wrist fold — coded: absent ( 1 ) ; weak ( 2 ) ; present ( 3 ) .

26. Tarsal fold— coded : absent ( 1 ) ; weak ( 2 ) ; present ( 3 ) .

27. Subarticular tubercles — coded: single (1); bifid and large (2); bifid
and separate ( 3 ) .

28. Amount of webbing (foiuth toe) — coded: base of disc (1); middle
of penultimate phalanx (2); base of penultimate phalanx (3); middle
of antepenultimate phalanx (4); base of antepenultimate phalanx (5).

29. Coloration of venter — coded: no pigment (1); small amount of pig-
mentation (2); large amount of pigmentation (3).

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 7

30. Dorsal color pattern — coded: plain (1); weakly mottled (2); mottled
(3); large spots (4).

31. Posterior surfaces of thighs (pigmentation) — coded: plain (1); weakly
mottled ( 2 ) ; mottled ( 3 ) .

32. Color of chin — coded: plain ( 1 ) ; weakh' mottled ( 2 ) ; mottled ( 3 ) .

Many characters traditionally used to define subgroups of the
Hyla bistincta group are used in the statistical analysis. Characters
not used in the statistical analysis generally fell into two categories:
1) subjective characters which are not easily quantifiable, such as
degree of thickness of skin or robustness of limbs, and 2) characters
which show no variation within a species.

Because of the small sample sizes of females in many species,
males and females were analyzed together; therefore, characters
specific to one sex were eliminated from the statistical analysis.

Several computer programs were used to examine the data. One
of these (BMD07M; Dixon, 1970) is a discriminant classification
analysis which calculates canonical axes (Anderson, 1958). Indi-
vidual OTUs are projected on the first two canonical axes. Approxi-
mate ninety-five percent confidence circles can be drawn around

the means by using a radius of 1.96/ \/n canonical axis units, and
around the groups by using a radius of 1.96 canonical axes units
(Seal, 1968:137). One disadvantage of this program is that the
groups are plotted on only two axes. If three or more axes explain
additional variation, programs other than BMD07M can be used to
plot the groups on three axes. For this purpose I used a multiple dis-
criminant function program (MULDIS) which is part of the NT-
SYS ( Numerical Taxonom\- System ) package, a group of multi-
variate programs de\eloped at The Uni\ersit\' of Kansas.

Using MULDIS, discriminant fimctions were first calculated
using only those groups with sample sizes greater than five, thus
eliminating five of the original fourteen species used as groups.
Next, the same program was run using all groups regardless of sam-
ple size. Multiplying the discriminant scores from the first run by
the table of means from the second run gives a matrix containing the
means of all OTUs plotted on discriminant functions which were
calculated using only groups with sample sizes larger than five. This
matrix is then used to obtain a three-dimensional plot by a program
called PROJ3D from the Uni\'ersit\' of Kansas Computation Center
Librar)'. This program plots a perspective \iew of a three-dimen-
sional scatter diagram. Another option of this program allows the
shortest simply connected network to be superimposed on the plot
to link up the points.

Because populations of Hyla arhorescandens and Hyla siopeJa in
Oaxaca are superficially so similar, these two species initially
were examined by populations using the discriminant classification
program, BMD07M. As discussed above, Hyla arhorescandens has
one of the largest ranges of all the species under consideration;

8 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

therefore, five populations from Oaxaca, Veracruz, and Puebla were
used in the analysis (Table 2). In addition, a single specimen from
Cerro Machin, Oaxaca, and the type specimen from Acultzingo,
Veracruz, were entered as separate groups. Using an option of the
program, these two specimens were not used in calculating the ca-
nonical axes, but were subsequently projected onto the axes. Speci-
mens from the two known populations of Hyla siopeJa from Cofre
de Perote, Veracruz, and Cerro Felon, Oaxaca, were used. Localities
and sample sizes of all groups are listed in table 2.

DESCRIPTIONS OF NEW SPECIES
Hyla cyanomma new species

Holotype. — KU 137014, adult female, from a mountain stream
1.2 km (by road) N Cerro Pelon, 2650 m, in cloud forest on a north-
ern slope of the Sierra de Juarez, Distrito de Ixtlan, Oaxaca, Mexico;
obtained by Jan Caldwell on 5 April 1970.

AUotijpe. — KU 137032 from the same stream as the holotype, 0.9
km (by road) N Cerro Pelon, 2670 m; Jan Caldwell, 11 May 1970.

Pomtopotijpes.—KU 136997-137007, 137009-13, 137015-31,
137033-34, 21 females and 14 males, from the same stream as the
holotype and the allotype, from 0.9 to 1.3 km (by road) N Cerro
Pelon, 2650-2670 m; Jan Caldwell and Paul B. Robertson, 5 April-12
July 1970.

Diafi^nosis. — A large, robust frog assigned to the Hijla crassa
group because it lacks a quadratojugal, has a short, round head,
thick glandular skin, long, moderately robust fingers without web-
bing, webbing of feet extending to middle or base of penultimate
phalanx of fourth toe, and partially or completely concealed t\anpa-
num. It can be distinguished from all other members of the group
by its uniformly colored, olive-green dorsum which becomes pale
blue on the flanks and posterior surfaces of the thighs. The eyes are
pale blue, and many frogs in the series have a few tiny bright yellow
spots on the dorsum. The only frog around in sympatry with Hyla
cyanomma is Hyla siopela, which is slightly smaller (maximum
snout-vent length 50.4 mm in females and 47.3 mm in males ) has a
dark and light brown mottled dorsum (in some, dorsum is leaf
green), and a distinct rostral keel.

Description of holotype. — Adult female, nongravid, body robust,
limbs stout. Snout-vent length 64.2 mm; tibia length 32.6 mm (50.8%
of snout-vent length); snout rounded in lateral and dorsal profiles;
nostrils not protruding, located closer to tip of snout than to eye;
internarial distance 5.3 mm (24.1% of head width); loreal region
concave, canthal ridge rounded, not prominent; top of head flat;
lips thick, only slightly flared laterally; pupil of eye horizontally
elliptical, diameter of eye 5.9 mm (27.6% of head width); eye slighdy

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 9

Table 2. Numlaer of males and females of each species of the Htjia histincia

group used in the statistical analysis. If a locality is not given after the name,

specimens throughout the range of the species were used in the analysis.

Species Males Females Total

H. arborescandens . . . W slope Sierra de

Juarez, low ele\ations 10 14 24

H. arborescandens . . . W slope Sierra de

Juarez, high elevations 10 22 32

H. arborescandens . . . Cerro Machin 10 1

H. arborescandens . . . Cerro San Felipe 5 2 7

H. arborescandens . . . Tezuitlun, Puebla 19 0 19

H. arborescandens . . . Acultzingo, Veracruz 16 3 19

H. arborescandens . . . type specimen, Acultzingo .— 10 1

H. siopela . . . Cofre de Perote 6 6 12

H. siopela . . . Cerro Pelon 23 7 30

H. sabrina 6 7 13

H. pentheter 29 2 31

H. robertsoruin 15 12 27

H. chryscs 2 13

H. cliaradricola 4 8 12

H. pacliydcrnia 12 3

H. bistincta 22 3 25

H. bogertae 0 2 2

H. crassa 3 4 7

H. mykter 112

H. cyanomma 5 7 12

H. ccnd)ra 10 1

Totals 180 103 283

closer to tympanum than to nostril; palpebral membrane unpig-
mented; supratympanie fold moderately heavy, extending from cor-
ner of eye to insertion of forearm; tympanum small, visible although
covered with skin; diameter of tympanum 1.4 mm (6.5% of head
width), upper edge hidden by supratympanie fold. Tongue cordi-
form. Skin smooth, thick on dorsum, granular on chin, venter,
posteroventral surfaces of thighs; axillary membrane present; tho-
racic fold weak. Cloacal opening directed posteroventrally at mid-
level of thighs; cloacal sheath short; tubercles present below anal
opening intergrading with those on posterior surfaces of thighs.
Forearm heavy; wrist fold present; no row of tubercles along ventro-
lateral margin of forearms; fingers long, moderately slender; vestigal
webbing present between fingers two and three, and three and four,
but no webbing between thumb and second finger; fingers, in order
of length from shortest to longest: 1-2-4-3; disc on third finger 2.5
times diameter of tympanum; subarticular tubercles round, except
distal tubercles on fourth finger bifid; supernumerary tubercles pres-
ent on proximal phalanges; palmar tubercle single, small, flat. Der-
mal fold present on heel; tarsal fold absent; toes long, moderately

10 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

slender; toes, in order of increasing length: 1-2-5-3-4; webbing ex-
tending from base of disc of first toe to base of penultimate phalange
of second toe, from base of disc on third to base of penultimate
phalange of fourth, from base of penultimate phalange on fourth to
base of disc on fifth toe; outer metatarsal tubercle absent; inner
metatarsal tubercle large, flat, oval; subarticular tubercles round ex-
cept distal tubercle on fifth toe weakly bifid; indistinct supernumer-
ary tubercles present on proximal phalanges.

Color (in alcohol) of dorsum dark metallic gray with few tiny
white spots, paler gray on flanks and limbs; venter, chin, and pos-
teroventral surfaces of thighs gray changing to pale tan on first
three toes and webbing, first two fingers, axillary membrane, and
antero ventral surfaces of thighs.

Color (in life) uniform olive-green on dorsum of body and limbs;
few tiny, bright yellow spots on dorsum; olive-green fading to pale
blue around vent and along outer edge of forearm and tarsus; venter
and chin greenish-yellow; ventral surfaces of limbs to first three toes
and first two fingers bright yellow-orange; outer toes and fingers
greenish-yellow; iris pale bluish-gray (Pi. 1 ) .

Variation. — The allotype is an adult male with snout-vent length
53.1 mm, enlarged non-projecting prepollex, and nuptial excres-
cences lacking; other characters and coloration are like those de-
scribed for holotvpe. Among 22 adult specimens, snout-vent length
is 51.9-64.5 mm (x=57.6 mm, N=15 9 9 ), 51.6-56.0 mm (x=53.8
mm, N=7 $ S ). The smallest female and male collected have
snout-vent lengths of 36.7 and 34.5 mm, respectively. Of 39 juven-
iles and adults, 13 (5 juveniles and 8 adults) have the tympanum
completely concealed, whereas the diameter of the tympanum is
0.7-1.6 mm (x=rl.2 mm) in the other 26. Even in the latter speci-
mens, the tympanum always is covered with skin, concealing the
tympanic ring. The axillary membrane is absent in 13 specimens.
The distal subarticular tubercle on the fourth finger is single in 16
specimens, bifid in 21, and divided in one. All males lack vocal slits
and nuptial excrescences. Variation in color in a series of 11 adults
(KU 136997-137007) is as follows: the dorsum is olive-green with
few tiny yellow spots, changing to pale blue along outer edge of
tarsus and forearm. The venter and chin are dark greenish-yellow,
and the limbs are yellow-ochre with a varying amount of brown.
The yellow is brightest on the webbing of toes and fingers, becoming
yellow-orange in a few. The eyes are pale blue or bluish-gray.

Distribution. — The species is known only from the type locality,
a moderate-sized stream in the Sierra de Juarez of Oaxaca, Mexico,
at elevations between 2650-2670 m.

Natural History. — Most specimens were collected between 5
April and 12 July 1970, from a moderate-sized stream in a pine-oak
cloud forest; one was taken from approximately the same locality on

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 11

5 August 1965. The latter part of the dry season and the early part
of the wet season occur in April, May, and June; thus, the stream
was flowing well but was never torrential. At night the frogs were
lying completely submerged on the bottom of large pools or at the
edges of pools with only their heads above water. When frightened,
they attempted to hide by moving to the deepest part of the pool or
by going under large rocks in the pools. During the months of April,
May, and June, 1-10 Hyla cyanomma were present in each large
pool in the stream. On the night of 10 May, 68 frogs were observed
within 0.1 km. The frogs were found by day near the same pools,
usuall)' sitting on rocks several centimeters above the water. The
only other species of tree frog found at this locality, HyJa siopela,
was found only at night sitting on leaves of terrestrial bromeliads,
on branches above the stream, or on mossy rock walls.

The reproductive condition of 22 female H. cyanomma was exam-
ined. Based on the condition of the oviduct, seven of these speci-
mens are juveniles, 10 subadults, and five adults. All adults are
nongravid. No amplexing pairs or egg clutches were found during
the observ^ation period from April to July. Two tadpoles referable to
this species because of their large size were collected in January.
Apparently reproduction occurs at the beginning of the dr)' season,
possibly in December or January.

Description of tadpole. — Two tadpoles, KU 139849, are pre-
sumed to be Hyhi cyanomma. Both are in developmental stage 26
(Gosner, 1960), and hax'e bod\- lengths 20.4 and 20.2 mm, and total
lengths 55.4 and 52.3 mm, respectively; description of larger speci-
men (Figs. lA and 2A) is as follows: body depressed, flattened
dorsally, 1.6 times as wide as deep, body nearly same width through-
out its length; snout broadh' rounded in dorsal view, round in lateral
view; nostrils small, directed anteriorh-; closer to eyes than tip of
snout; eyes small, directed dorsolaterally, widely separated; spiracle
sinistral, directed posterodorsalh-, located at point below midline
slighth' less than two-thirds length of body; caudal musculature
moderately developed, tapering at midlength of tail and ending an-
terior to fin; dorsal caudal fin arches near end of tail; depth of dorsal
and ventral caudal fins two-thirds of caudal musculature at mid-
length of tail.

Mouth \'entral. small, half as wide as body at greatest width; lips
infolded laterally; edge of lips bordered by single row of papillae;
inner row of larger, irregularly spaced papillae; upper beak slender
\\'ith ver>' fine serrations, tapering slighth' to form lateral processes;
lower beak robust with slightly larger serrations; two upper and
three lower tooth rows, all equal in length and extending nearly to
edge of lips, second upper row interrupted medialh'.

Bodv dark grav dorsalK', lighter on sides and snout; venter trans-
parent; caudal musculature with irregular gray spots, absence of

12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

spots forms tan stripe dorsolaterally; caudal fins opaque with irreg-
ular gray spots.

Etymology. — The specific name is from the Greek kyonos mean-
ing blue and the Greek omnia meaning eye, and is in reference to
the blue eyes of this species.

Hyla sabrina new species

Holotype. — KU 137086, adult female, from a mountain stream
15.<S km (by road) S Vista Hermosa, 1990 m, in cloud forest on a
western slope of the Sierra de Juarez, Distrito de Oaxaca, Mexico;
collected by Jan Caldwell on 12 June 1970.

Allotype. — KU 137067, from a mountain stream 11.9 km (by
road) S Vista Hermosa, 1920 m; Jan Caldwell and Paul B. Robert-
son, 17 May 1970.

Pamtypes.—KV 137044, 137053, 137059-60, 137064, 137066,
137069-72, 137076, 1370.85, 137087, 10 females and 3 males, from
mountain streams, 11.1-16.6 km S Vista Hermosa, 1840-2020 m; Jan
Caldwell and Paul B. Robertson, 23 November 1969, 17 July 1970.

Diagnosis. — Hy]a sabrina is a small, slender-limbed member of
the Hyla charadricoJa species group because it lacks a quadrato-
jugal, has thin skin and an axillary membrane, lacks nuptial spines
in breeding males, has long, slender fingers with no webbing, round
to slightly pointed snout, tan venter with dark brown mottling and
a mottled dorsum. It differs from charadricoJa by its smaller size
and the absence of a tympanum and from chryses by coloration and
absence of a tympanum. Hyla charadricola and H. sabrina are most
alike in coloration, have a mottled green and darker green dorsum,
with dull yellow or yellow-orange on the posterior and anterior sur-
faces of the thighs. Hyla chryses resembles these two in coloration
when cold and sluggish, but when acti\'e is a golden yellow overlaid
with small brown flecks and some indistinct green flecks (Adler,
1965). At night, Hyla sabrina may be a bright leaf green. Hyla
sabrina is sympatric with H. chaneque and H. arborescandens. Hyla
chaneque is much larger than sabrina, has a blotched dorsum and a
tympanum. Hyla arborescandens is also larger, has a mottled dark
and light dorsum, a distinct rostral keel, a tympanum and less web-
bing than .sabrina. Hyla sabrina can be distinguished from Ptycho-
hyla ignicolor, which occurs nearby, by the lack of webbing be-
tween the fingers and the absence of the tympanum.

Plate I. Top.— Hyla ciassa $ (KU 148699; 52.4 mm, SVL) from 1.9
km S El Eshidiante, 1850 m, Oaxaca, Mexico. X 1. Photo by William E.
Duellman. Middle.— Hf//a sabrina 9 (KU 137083; 31.2 mm SVL) from 15.8
km S \'ista Hermosa, 1990 m, Oaxaca, Mexico. X 2. Photo by Jan Caldwell.
Lower.— Hyla cyaiwmma 9 ( KU 137008; 55.3 mm, SVL) from 1.2 km N
Cerro Pelon, 2650 m, Oaxaca, Me.xico. X 1. Photo by Jan Caldwell.

PLATE 1

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 13

B

#i» **■

E

/

Fig. 1. Tadpoles of species in the Hijla crassa and Hyla arhorcscandcns
groups. A. Ihjla cyanomma, KU 139849. B. Hi/la cembra, KU 139859. C. Hyla
crassa, KU 139845. D. Hyla aiborescandens, KU 139832. E. Hyla siopela,
KU 139841. X 2.

14

OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

Fig. 2. Months of tadpoles of the Hijh crassa and Hijla arhorcscandcns
groups. A. Hyla cyanomma, KU 1.39849. B. Hyla ccmhra, KU 139859. C. Hyh
crassa, KU 139845. D. Hyla arhorescandens, KU 139832. E. Hyla siopcia,
KU 139841. X 10.

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 15

Description of holoiijpe. — Adult female, gravid, medium-sized,
slender-bodied. Snout-vent length 41.7 mm; tibia length 21.4 mm
(51.3% of snout- vent length); snout in dorsal and lateral views acu-
minate; nostrils not protruding; internarial distance 2.9 mm (23.6%
of head width), canthal ridge angular; loreal region flat; lips thin,
not flared; pupil of eye horizontally elliptical; palpebral membrane
unpigmented; diameter of eye 4.6 mm ( 37.4% of head width ) ; supra-
tympanic fold present, thin, extending from posterior corner of eye
to forearm; tympanum absent; tongue round, posterior one-fourth
free. Skin on entire body smooth, thin; few tubercles around vent;
row of tiny white tubercles present on ventrolateral edge of fore-
arm. Thoracic fold absent; small axillary membrane present. Anal
opening directed posteroventrally at midlevel of thighs.

Forearm slender; prepollcx only slightly enlarged; fingers, in
order of increasing size: 1-2-4-3; subarticular tubercles round on
distal articulation of three fingers, bifid on other three; few tiny su-
pernumerary tubercles present on proximal phalanges; three palmar
tubercles in a triangle at base of third and fourth fingers. Dermal
fold present on heel; tarsal fold absent, indicated by a tiny row of
white spots from heel to inner metatarsal tubercle; inner metatarsal
tubercle small, flat, oval; outer metatarsal tubercle absent; sub-
articular tubercles small, round; few tiny supernumerarx' tubercles
on proximal phalanges; toes, in order of increasing size: 1-2-3-5-4;
toes approximately three-fourths webbed; wt>bbing extending from
base of first toe to middle of penultimate phalange of second toe,
from base of second to base of penultimate phalange of third, from
base of disc of third to base of penultimate phalange of fourth, from
base of penultimate phalange of fourth to base of disc of fifth.

Color (in alcohol) gray on dorsal surfaces of head, body, and
limbs with trace of mottling; white mottling where gray of dorsum
meets dark brown of flanks and lateral surfaces of limbs; venter and
ventral surfaces of limbs tan with overlying brown pigment.

Color (in life) light chocolate brown on dorsum with mostly
green and some dark brown mottling; dark brown stripe extending
from nostril along canthus to back of arm; green behind eye and
below suprat\'mpanic ridge to upper lip; flanks dark brown, inter-
grading with light chocolate of dorsum, giving appearance of white
mottling on flanks; anterior and posterior surfaces of thighs dull
yellow, densely covered with brown; dull yellow extending on legs
to first three toes, including webbing, and to first two fingers; dark
brown around vent, with few tiny white tubercles; white line above
vent; dark brown on ventrolateral edges of forearm, heel, and tarsus;
rows of white tubercles on ventrolateral edge of forearm and outer
edge of tarsus; venter tan with dark brown on chest, slightly lighter
brown on throat; brown on ventral surfaces of legs; iris dark brown
(Pl.l).

16 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

Variation. — The allotype is an adult male with a snout-vent
length of 29.8 mm, and an enlarged prepollex without nuptial spines.
It has a round snout in contrast to the acuminate snout of the holo-
type; also, it has less brown pigment on the venter and chin than
the holotype. Other characters and coloration in alcohol are as de-
scribed for the holotype. Among the paratypic series of 13 adults,
the snout-vent length for 10 subadult and adult females is 33.7-41.3
mm (x=37.6 mm) and for 3 males is 26.9-29.9 mm (x=28.2 mm).
The smallest juvenile female and male collected have snout-vent
lengths of 23.0 and 22.8 mm, respectively. Of the 10 females, 6 have
venters and chins heavily pigmented, 3 lightly pigmented, and one
has almost no pigment. Of the 3 males, one has venter and chin
lightly pigmented and 2 have almost no pigment. Variation in color:
KU 137051-52, dorsum mottled green, thighs dull yellow, flanks gray
and white; KU 137063, dorsum mottled green, thighs light yellow,
iris golden with black reticulations; KU 137038, dorsum mottled
green and brown (bright green when first caught), flanks mottled
brown and white, canthus dark brown with light gray line above,
venter with dark brown pigment.

Distribution. — The species is known only from a restricted area
in the cloud forest of the Sierra de Juarez. It has been taken from
mountain streams 7.8-16-6 km (by road) S Vista Hermosa, 1650-
2020 m.

Natural History. — All specimens were found on vegetation
around fast-moving mountain streams. Of 33 females and 16 males
collected throughout the year, only seven females and two males are
adults. Two females, taken in June and July, are gravid. Tadpoles
of this species are unknown.

Etijniohgij. — The specific name sahrina is Latin meaning river
nymph and refers to the aquatic habitat of this species.

Hyla cembra new species

Holotype. — KU 137035, adult male, from a mountain stream at
Campamento Rio Molino, 33.8 km (by road) N Candelaria Loxicha,
2160 m, in tropical deciduous forest in the Sierra Madre del Sur,
Distrito de Pochutla, Oaxaca, Mexico; collected by Jan Caldwell
and Paul B. Robertson, 10 October 1969.

Diagnosis. — A moderate-sized, robust-limbed frog placed in the
Hyla crassa species group because it lacks a quadratojugal, has ro-
bust limbs, a short, round head, thick, glandular skin, fingers long
and moderately robust without webbing, webbing on feet extending
to middle or base of penultimate phalange, and tympanum partially
concealed. It difters from other members of the group, except cya-
nomma, by its uniform olive-green coloration and from cyanomma
by its smaller size, lack of blue pigment on Hanks or thighs and blue
eyes. No tree frogs have been found in sympathy with Hyla cembra.

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 17

Description of holotype. — Adult male, body moderate in size,
robust; snout- vent lengtli 37.0 mm; tibia length 18.8 mm (50.8% of
snout-vent length); snout rounded in dorsal and lateral profile;
canthus rounded, indistinct; loreal region barely concave; lips mod-
erately thick, not flared laterally. Pupil horizontally elliptical; pal-
pebral membrane unpigmented. Dermal fold from corner of eye to
insertion of forearm; upper edge of tympanum indistinct; t\'mpanum
small, 1.3 mm (10.9% of head width), tympanic ring barely visible.
Tongue cordiform, barely free behind; vocal slits absent.

Skin smooth on dorsum of body and limbs, weakly granular on
venter, chin, and ventral surfaces of thighs. Thoracic fold absent;
axillary membrane present, but not extending to elbow; faint row of
tubercles present on outer edge of forearm; wrist fold moderateh'
developed; fingers long, not especialh" slender, in order of length
from shortest to longest: 1-4-2-3; no webbing between fingers; discs
moderate in size, that on third finger 1.4 times larger than tympa-
num; subarticular tubercles round, single except those on fourth
finger bifid; supernumerarx' tubercles present, numerous; two mod-
erately large palmar tubercles present; prepollex enlarged, non-
projecting, with numerous small nuptial spines; nuptial spines pres-
ent on first and second fingers.

Length of foot 17.4 mm; discs small, about two-thirds those on
fingers; subarticular tubercles round; supernumerary tubercles small,
numerous. Webbing on foot extending from base of disc of first
finger to base of penultimate phalange on second finger, from base
of disc on second to base of penultimate phalange on third, from
base of disc of third to base of penultimate phalange of fourth, and
from the latter to base of disc on fifth; toes, in order of length from
shortest to longest: 1-2-5-3-4, length of tibia 19.5 mm; heels overlap
about one-third length of tibia; anal opening mid-level of thighs;
transverse dermal fold above anus; few moderate!)' large tubercles
lateral and ventral to anus.

Color (in alcohol) uniform dull graxish brown on dorsum and
upper surfaces of limbs; venter, anterior and posterior surfaces of
thighs, first three toes and first two fingers light tan; moderately
large tubercles lateral to anus dull yellow; nuptial spines dark brown.

Color (in life) bright greenish yellow on dorsum when first
caught, later changing to dull green; light brown and green reticu-
lations along sides; throat and chest white; xentral sides of limbs
and belly pinkish tan; iris brownish yellow.

Natural History. — The single known specimen was found calling
at night from under a piece of bark on a large log in a small stream.
The call is a short "wrack" often followed by a series of low chuckles.
Tadpoles, probably referable to this species, were common in
streams in the area in April and July. No other tree frogs are known
to occur in the area.

18 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

Table 3. Measurements (in mm) of Hyla cembra tadpoles in \'arious stages
(Gosner, 1960) from three localities. Single figure is the mean; range is given

below in parentlieses.

Stage N Body Length Total Length

Campamento Rio Molino, 2160 m

26 2 12.6 30.0

(11.5-13.7) (29.2-30.8)

27 1 13.0 29.9

28 2 14.3 32.6

(13.5-15.0) (30.2-37.0)

29 1 14.8

30 3 15.7 39.7

(14.6-16.6) (38.0-41.3)

31 4 16.3 42.1

(14.9-17.6) (36.0-45.5)

33 2 16.4 42.5

(15.8-17.0) (40.4-44.6)

35 2 17.1 47.2

(16.2-18.0) (43.9-50.5)
,36 1 16.3
11.6 km S Campamento Rio Molino, 1970 m

28 1 16.1 40.1

31 2 17.3 44.1

(16.4-18.1)

37 1 19.2

38 1 18.4
12.8 hn S Campamento Rio Molino, 1930 m

34 1 16.9 46.0

36 1 16.8 42.0

Description of Tadpole. — Twenty-five tadpoles from Campa-
mento Rio Molino, 2160 m, to 12.8 km S Campamento Rio Molino,
1930 m, were collected. These tadpoles are assumed to be Hyla
cemhra because no other frogs haxe been observed or collected in
this area. The body and total lengths are given in Table 3. The
following is a description of a tadpole from Campamento Rio Mo-
lino, KU 139859 (Figs. IB and 2B), that is in developmental stage
30 (Gosner, I960) : i3ody length 15.8 mm, total length 41.3 mm, tail
slightly less than two-thirds length of body; body slightly depressed,
nearly ovoid, depth 88 percent of width; venter flat, dorsal contour
sloping anteriorly; snout broadly rounded in dorsal view, rounded
in lateral view; nostrils small, directed anterolaterally, located half-
way between tip of snout and eye; eyes moderately large, directed
dorsolaterally, widely separated; spiracle sinistral, directed postero-
dorsally, located at a point on midline of body slightly less than
two-thirds length of body; anal tube moderately long, dextral; cau-
dal musculature moderately heavy, tapering gradually, ending an-
terior to caudal fin; dorsal and caudal fins normal, large; dorsal fin
1.1 times and ventral fin 1.2 times as deep as musculature at mid-
length of tail.

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 19

Mouth small, less than two-thirds width of body; lips infolded
laterally; several irregular rows of papillae around edge of lips;
papillae especially numerous on lower lip and lateral corners of lips;
one regular row of slightly larger papillae medial to outer rows;
beak moderate in size with very fine serrations; upper beak forms
arch with long slender processes; two upper and three lower tooth
rows; lower rows slightly shorter than upper, but all extending
nearly to lateral edges of lips; second upper row interrupted
medially.

Color (in alcohol) uniform dark brown on dorsum and sides;
spiracle light gray; venter dark but transparent; caudal musculature
light tan; caudal fins light transparent gray, plain except for few
tiny brown spots on outer edge of fin.

Color (in life, taken from KU 139857) body dark brown covered
with light green, nc^IIow, and copper chromatophores; caudal mus-
culature light yellow-brown; caudal fins same color as musculature
with few light yellow and copper chromatophores; eye yellow.

Etymology. — The specific name cembra is Latin meaning timber
and is used in reference to the pine-oak forest this species inhabits.

SYSTEMATICS

Re-evaluation of the Hyla ])istincta group (sensu lato) has re-
sulted in the allocation of these 13 species, plus Hyla arhorescan-
clens, into four groups, as defined and discussed below.

The Hyla bistincta Group

Content. — As redefined, the Hyla bistincta species group in-
cludes Hyla bistincta and //. pentheter.

Definition. — The species in this group are characterized by: 1)
large body size, 2) truncate snout, 3) no quadratojugal, or small
spine-shaped quadratojugal failing to articulate with maxillary, 4)
distinct t\mpanum, 5) xocal slits present or absent, 6) breeding call
present, 7) bold reticulated pattern on flanks, 8) plain dorsum, 9)
venter plain, 10) thick skin, 11) long anal sheath, 12) nuptial spines
on first and second fingers of breeding males, 13) long, slender fin-
gers without webbing, 14) robust limbs, 15) webbing on feet ex-
tending to base of penultimate phalanx or to base of antepenultimate
phalanx.

Distribution. — Hyla bistincta is the most widespread species of
those included in the four groups (Table 1; see Duellman, 1970, for
distribution map); it has an elevational range of 1400 to 2800 m.
Hyla pentheter is confined to a more restricted area in the deciduous
tropical forest in the mountains of the Sierra Madre del Sur in
Oaxaca and Guerrero and has an elevational range of 1320 to
2000 m.

20 OCCASION.4L PAPERS MUSEUM OF NATURAL HISTORY

The Hyla charadiicola Group

Content. — This group is composed of three species: Hyla charci-
dricoJa, H. chryses, and H. sahrino.

Definition. — The following combination of characters separate
this group from all other species groups: 1) small to medium body
size, 2) snout round to slightly pointed, 3) no quadratojugal, 4)
tympanum distinct, partially concealed, or concealed, 5) vocal slits
absent, 6) breeding call absent, 7) fine mottling or spots on flanks,
8) mottled dorsum, 9) venter mottled dark brown and gray or tan,
10) thin skin, 11) transverse anal sheath, 12) nuptial spines absent
in breeding males, 13) long, slender fingers without webbing, 14)
slender limbs, 15) webbing on feet extending to base of penultimate
phalanx or to base of antepenultimate phalanx.

Distribution. — The three species are allopratic and have fairly
restricted ranges at moderate to high elevations. Hyla charadricoh
occurs in northern Puebla and eastern Hidalgo; H. chryses is known
only from the type locality, 45 km (airline) WNW of Chilpancingo,
Guerrero, and H. sahrina inhabits the Sierra de Juarez in Oaxaca.
All have elevational ranges within 1650-2600 m.

The Hyla crassa Group

Content. — This group is composed of five species: Hyla crassa,
H. pachyderma, H. hogertae, H. cyanomma, and H. cemhra.

Definition. — These frogs possess the following combination of
characters: 1) moderate to large body size, 2) snout round, 3) no
quadratojugal, 4) tympanum partially or completely concealed, 5)
vocal slits absent, 6) breeding call present or absent, 7) flanks plain
or reticulated, 8) dorsum plain, flecked or spotted, 9) venter plain,
10) thick, glandular skin, 11) transverse anal sheath, 12) nuptial
spines present on thumb and first finger (except cyanomma, but
breeding season may have been past when males were collected)
13) long, moderately robust fingers without webbing, 14) robust
limbs, 15) webbing on feet extending to middle or base of penulti-
mate phalanx.

Distribution. — Each of the five species has a restricted distribu-
tion. Hyla cembra and H. bogertae occur in the Sierra Madre del
Sur in Oaxaca, H. cyanomma in the Sierra de Juarez in Oaxaca, H.
pachyderma in the Sierra Madre Oriental near Tezuitlan, and H.
crassa in the central mountains of Oaxaca. The five species are
allopatric and are found at relatively high elevations, ^^'ith the ex-
ception of Hyla pachyderma, all are found in pine, pine-oak, or pine-
fir forests between 1850 and 2650 m. Hyla pachyderma (known only
from four specimens) was taken from cloud forest at 1600 m.

Remarks. — Three of the five species allocated to this group
{crassa, bogertae, and cyanomma) are known to be highly aquatic
in nature, and usually are found along streams sitting in water or

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 21

just above water. When frightened, these frogs hide in the water
beneath rocks or other debris. Of the other two species, the four
known specimens of pachy derma were found on bushes and weeds
beside a small bounding stream (Taylor and Smith, 1945), and the
single specimen of Hyla cemhra was calling beneath the bark of a
log in a stream.

As discussed above, Straughan and Wright (1969) regarded
Hyla crassa a nomen duhium and continued to use Hyla robusto-
femora, although Duellman (1964) had compared the types of
Cauphias crassiis Brocchi and Hyla robustofemora Taylor and con-
cluded that they represented the same species.

Five recently acquired specimens of Hyla crassa from 1.9 km S
El Estudiante, 1(S50 m, Oaxaca, allow assessment of the variation in
this species. I also have one specimen from 9 mi NE Oaxaca avail-
able for comparison. Duellman (1964) listed the following charac-
ters of the female H. crassa (type of Cauphias crassus) as differing
from the male H. crassa (H. robustofemora Taylor): 1) the tym-
panum is completely concealed, 2) vomerine teeth are 8-7, com-
pared to 5-5 in the male, 3) more cream-colored mottling on the
flanks and posterior surfaces of the tliighs, and 4) more distinct
mottling on the throat.

The above four characters and snout- vent length are compared
among the Hyla crassa I have examined in table 4. I have not seen
the type specimen of Cauphias crassus (= Hyla crassa) but have
relied on Duellman's (1964) description for comparative purposes.
The >'ariation in size, distinctness of the tympanum, and degree of
mottling of flanks, posterior surfaces of thighs, and throat, among
the five recent specimens encompass the two specimens originally
discussed by Duellman (1964). Therefore, I concur with Duellman
that Hyla crassa and Hyla robustofemora are conspecific.

Color notes from li\dng specimens of Hyla crassa are: KU
148696, 9 , and KU 148697, $ , have dorsum pale greenish-gray ( in
dark), changing to dark brown (in light); mottling along outer edge
of forearm, sides and outer edge of leg same color as dorsum inter-
spersed with pale green; venter bright yellow; yellow extending onto
legs and arms and becoming darker; yellow on venter becoming
cream-colored to pale green on chin with darker mottling; female
has heavier mottling on chin than male; heavy pale green irregular
stripe above and below vent; eyes copper with black reticulations.
KU 148698, 9 , same as above, but dorsum has \'ery distinct irreg-
ular dark brown spots (becoming lighter if animal is kept in dark)
on lighter greenish-brown to brown background; spots always dis-
tinct whether frog kept in dark or light (PI. 1). Ronald Altig, who
collected the specimens, noted that tlie same frog can change from
spotted to uniform coloration.

Two individuals were on rocks and sticks a few centimeters

22 OCCASION-\L PAPERS MUSEUM OF NATURAL HISTORY

CO

0)

o

a;
a

00

bo

c
o

CS

en
;-(

dj
^^
O

03

c

e
o

a
a
g
o
U

P

1— ■

5f

T3

Ti

-0

*^ .IT?

n3

^0 x-t:

-d

-M

ir.

(— -*-•

2 -^

^

2 "i^

-i 2 t;

IJ

;-

J-' (—

0; C

4->

C
' 1

C/^

r>

^

1

^

s ^

s

t*

O
U

0)

c
>

>

en

o
Z

c
a

c/;

1) ^

in

in

I

oc

-^

OJ

P^

"«

>

'«

q;

o

o

y

r^

o

c

"rt

o

Q

U

U

^

Ot

o
in

o
in

z

D

CO
CO

CO

Z

K
Z

:^

Cl,

I

13

Oh Q^

^ CO CO

CO C-'l CO

CL,

CO

'C Tj 'w '^

Ji ^ -i 13 -^

-*. ij X Oj >> gj — >. oj -w

Oh &h Oh U Ph Q

t-

t-

t

in

m

— ; -r

05

CO
in

CO
m

o

CD

CO o\

CD in

d
in

o

ot

ot <fO

in

C5

CO

CD

in

00

(M

Tf

^
«

I-

c:5

CD

00

D

CD CD

00 00

D D

CH-

o
o

CO

S

05

CO
CO 05 fM
2 ■— < l>

CO „ 05

<3 '^ ^

c -S bc 2-

~ "H. = <

^- 5 u ^

^ ^ ^ ,5

*; '^^ ^ <

E (^ /-»

a ii ii

c" O O

0) <U (b

§1 1

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 23

above the water along a large stream by day. Four were along the
same stream at night. These frogs lack vocal slits, but make long,
low calls.

By the process of elimination, three tadpoles from this area are
presumed to be Hyla crassa. One specimen, KU 139850, from 6.3
km SE Ixtlan, 1910 m, is in developmental stage 36 (Gosner, 1960)
and has a body length of 21.8 mm and a total length of 59.2 mm.
Two other specimens, KU 139854, from 2.3 km E and 11.6 km NE
Oaxaca, 1720 m, are in stage 41 and have body lengths of 19.7 and
18.9 mm and total lengths of 60.1 and 54.4 mm, respectively. The
description of the latter specimen (Figs. IC and 2C) is as follows:
shape of body depressed, depth of body 79 percent of width; tail
equal to two-thirds body length; in lateral profile body flattened
dorsally, snout bluntly rounded; snout round in dorsal profile; nos-
trils directed anterolaterally, located closer to eye than tip of snout;
eyes large, directed laterally, widely separated; spiracle sinistral,
directed posterodorsally, located at a point on midline two-thirds of
body length; left forelimb evident beneatli spiracle; anal tube mod-
erately long and dextral; caudal musculature massive, ending before
tip of tail; dorsal and ventral caudal fins normal, narrow at mid-
length of tail, 59 and 48 percent depth of caudal musculature, re-
spectively; fins becoming deepest on posterior fourth of tail.

Mouth x'entral, small, approximate]}' one-half body width; lateral
folds present; double row of small papillae completely bordering
lips, becoming mnnerous on lateral arenas of disc; inner row of larger
widely spaced papillae present; beak slender, broadly V-shaped,
with very fine serrations, small notch on upper beak; tvvo upper and
three lower tooth rows, same length and extending nearly width of
mouth; second upper row interrupted medially.

Color (in life) of dorsum brown with copper and light yellow
chromatophores; tail musculatine grayish; fin grayish, transparent;
tail with large black spots and numerous copper chromatophores
forming blotches; dorsum of legs pale yellow with brown markings;
eyes copper.

Color ( in alcohol ) of body gray; caudal fin and musculature tan
with irregular dark brown flecks and spots; legs light gray above,
whitish below.

The Hyla arborescandens Group

Content. — Four species comprise this group: Hyla arborescan-
dens, H. rohertsortim, H. siopela, and H. mykter.

Definition. — This species group is distinguished by the following
combination of characters: 1) moderate body size, 2) snout slightly
rounded to truncate, 3) no quadratojugal, 4) tympanum distinct.
5) vocal slits present or absent, 6) breeding call present or absent,
not necessarily correlated with presence or absence of vocal slits,

24 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

7) flanks mottled, 8) dorsum mottled or reticulated, 9) venter plain
to strongly mottled, 10) thin skin, 11) long or short anal sheath, 12)
nuptial spines present on thumb and first finger of breeding males,
13 ) long, moderately slender fingers with traces of webbing between
two outermost fingers, 14) slender limbs, 15) webbing on feet ex-
tending to base of penultimate phalanx or to middle of antepenulti-
mate phalanx.

Distribution. — Hyla arhorescandens has a relatively large geo-
graphic range extending from the Sierra de Juarez in Oaxaca north-
westward to the Sierra Madre Oriental in Puebla. HyJa siopela is
known from two isolated mountain peaks: Cofre de Perote in the
Cordillera Volcanica Transversal in Veracruz, and Cerro Felon in
the Sierra de Juarez of Oaxaca. Hylo rohertsorum has been found
in five localities in eastern Hidalgo and northern Puebla, and HijJa
mykter is known from the vicinity of Cerro Teotepec in the Sierra
Madre del Sur of Guerrero. The elevational range of Hyla arhore-
scandens encompasses that of the other three species (Table 1).

Remarks. — Hyla siopeJa was discovered around a small stream on
the west slope of Cofre de Perote, Veracruz, elevation 2500-2550 m
(Duellman, 1968). In the description of this frog, Duellman stated
that in structure and coloration Hyla arhorescandens resembles sio-
pela, but the former is smaller and males of arhorescandens have
vocal slits, whereas those of siopela do not.

While working in Oaxaca, I found an isolated population of Hyla
siopela on the north slope of Cerro Pelon in the Sierra de Juarez at
elevations between 2650 and 2670 m. Thus, the two populations of
siopela are found in isolated mountain ranges separated by the Rio
Quiotepec.

Specimens of Hyla arhorescandens have been taken from local-
ities surrounding both populations of Hyla siopela. While working
a transect from Tuxtepec, Oaxaca, to Cerro Pelon (0-3000 m), I
commonly found Hyla arhorescandens around mountain sti'eams be-
tween 1610 and 2020 m. No frogs were taken from 2020 to 2650
m on the transect, because of a lack of suitable habitat. I orig-
inally identified the specimens from Cerro Pelon as Hyla arho-
rescandens because of their similarity to the frogs occurring at
lower elevations. Coloration and certain structural features, such as
the presence of a rostral keel, are nearly identical in the two popu-
lations. Closer examination, however, revealed that specimens taken
at Cerro Pelon lacked vocal slits, as did those from Cofre de Perote.
Also, the size differences among the Narious populations were con-
sistent (Table 5). An increase in size due to elexational effects was
ruled out because of the size of the male arhorescandens taken on
Cerro San Felipe at an elevation of 2670 m. These specimens have
vocal slits and are similar in size to the H. arhorescandens occur-
ring at lower elevations in the area of Vista Hermosa. Female ar-

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 25

Table 5. Size differences among populations of Htjia aiboiescaudens and H.
siopela. Single figure represents the mean; the sample size is below the range.

Hyla arborescandens

Vista Hermosa

Cerro San Felipe

( 1610-2020 m)

(2670 m)

Snout-vent length

35.4

35.6

(males)

32.5-40.0

34.6-37.1

(i\=z20)

(N=5)

Snout-vent length

42.7

48.9

(females)

39.0-46.4

48.8-48.9

(iN=34)

(N=2)

Hyla siopela

Cerro Pelon

Cofre de Perote

(2650-2670 m)

( 2500-2550 m )

Snout-vent length

43.4

43.6

( males )

40.4-47.2

40.8-45.5

(N=22)

(N=6)

Sno<it-vent length

49.9

48.6

( females )

47.9-51.9

45.1-52.0

(N=5)

(N=6)

horescandens on Cerro San Felipe are much larger than males, but
this could be a function of the small sample size available (N:=2).
Neither specimen is as large as the largest siopela female from either
Cerro Pelon or Cofre de Perote.

Additionalh', the dorsal color pattern of the two populations of
//. siopela appears to be more variable than that in //. arborescan-
dens. Duellman (1968) reported that the dorsal coloration of H.
siopela from Cofre de Perote varied from pale green to olive-green
with darker green or black flecks or reticulations. Variation in dorsal
pattern of the specimens from Cerro Pelon is as follows: one speci-
men leaf green, another light green with brown mottling, another
light brown with few brown flecks, and several mottled light and
dark brown. In contrast, the H. arborescandens from lower eleva-
tions in the area of Vista Hermosa, Oa.xaca, have much less variation
in coloration, usually having the dorsum mottlc>d with light and
dark brown.

Tadpoles of H. arborescandens (Figs. ID and 2D) from the
Vista Hermosa area and of siopela from Cerro Pelon ( Figs. IE and
2E ) are very similar in body shape, mouthparts. and other structural
features. The primary difference is size of the serrations of the
upper beak; those in arborescandens are peglike and are larger than
the serrations on the lower beak, whereas those in siopela are
smaller and equal in size on the upper and lower beaks. The tadpole
figured by Duellman (1970:376-377) as arborescandens has equal-
sized serrations on the upper and lower beak and an additional

26

OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

Table 6. A comparison of size differences between populations of Ilijhi .siopela
tadpoles from Cofre de Perote and Cerro Pelon. The single figure is the mean;

the range is in parentheses.

Stage

N

Body Length

Total Length

Cofre de Perote

37

1

23.8

63.8

41

3

19.4
(18.3-20.9)

60.9
(60.0-61.8)

Cerro Pelon

26

4

16.1
(14.1-18.9)

39.0
(36.8-46.6)

27

3

17.6
(16.0-18.7)

42.3
(39.0-44.1)

28

3

17.4
(15.4-20.0)

44.8
(41. .5-50.0)

30

1

16.9

43.1

32

1

18.3

47.8

33

1

16.0

42.3

40

1

16.5

45.1

lower broken tootli row. It is possible that this tadpole could be that
of H. sabrina.

A comparison of Hyla siopela tadpoles from Cerro Pelon and
Cofre de Perote was made. Four tadpoles, all in advanced stages of
development, are available from Cofre de Perote, and 15 lots of
tadpoles are available from Cerro Pelon. The mouthparts of the
tadpoles from the two populations are identical, as are other struc-
tural aspects such as depth of caudal musculature and fins, and
coloration (see Duellman, 1970, for description and figures of Hyla
siopela tadpoles from Cofre de Perote). The primary difference
between the two populations of tadpoles is size. All but one of the
tadpoles from Cerro Pelon are in earlier stages of development than
the Cofre specimens. The largest Cerro Pelon specimen, in stage 40
(Gosner, 1960), lias a total length of 45.1 mm, compared with three
specimens from Cofre in stage 41, which have a mean total length
of 60.9 mm. Table 6 presents a comparison of measurements of the
four tadpoles from Cofre de Perote and selected developmental
stages from Cerro Pelon.

Several hypotheses can be made regarding the differences in size
of the tadpoles of the two populations. Perhaps there has been a
reduction in size of the H. siopela tadpoles on Cerro Pelon due to
competition with the larger, sympatric //. cyanomma. Two //.
cyanomma tadpoles from Cerro Pelon are approximately the same
size as the H. siopela specimens from Cofre de Perote. In develop-
mental stage 26, the H. cyanomma tadpoles have a mean total length
of 53.9 mm. There are no frogs sympatric with H. siopela on Cofre
de Perote. Alternatively, differences in water temperature may af-

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 27

feet the size of the tadpole. This hypothesis was considered uiiHkely
because the range in water temperature at Cerro Pelon for 12 col-
lections of tadpoles was 10 C to 17.8°C (mean 13^C), and the
water temperature at Cofre when the specimens were collected was
15.5'C. However, little is known about seasonal or daily variation
of water temperatures, or about other physical or chemical factors
at the two localities which may influence developmental rates in
these tadpoles.

RESULTS AND DISCUSSION

In the initial discriminant classification analysis (BMD07M)
using individuals of 5 populations of Hyla arhorescandens and 2
ptjpulations of HijJa siopela, 90 percent of the total amount of vari-
ation is explained by the first three axes as follows: one, 53 percent;
two, 25 percent; and three, 12 percent. The plot of individuals on
canonical axes I against II (Fig. 3) shows an interesting grouping
by locality rather than species. H. siopela (B on Fig. 3) from Cerro
Pelon is closer to H. arhorescandens from the surrounding localities
of Vista Hermosa (C and D) and Cerro Machin (E) than to the
other population of //. siopela from Cofre de Perote (A). Two
other populations of H. arhorescandens from Tezuitlan and Acult-
zingo (both near the Veracruz-Puebla border, but separated by
about 130 km; G and H on Fig. 3) are closer to each other than to
H. arhorescandens from the Vista Hermosa localities. One deviation
from this pattern is that specimens of //. arhorescandens from Cerro
San Felipe (in central Oaxaea; F on Fig. 3) are plotted closer to
//. arhorescandens from Puebla and Veracruz than to those from
other localities in Oaxaea. This could be due in part to small sample
size, because only a few specimens (Table 2) are available from
Cerro San Felipe. Values of the F matrix indicate that all groups
are significantly different (oc^O.05) from each other except the
populations of arhorescandens from low and high elevations north
of Vista Hermosa (C and D on Fig. 3) and populations of arhores-
candens from Tezuitlan, Puebla, and Acultzingo, Veracruz (differ-
ent at 0.001 level; G and H on Fig. 3).

As discussed prex'iously, the two most ob\'ious characters sepa-
rating siopela and arhorescandens are larger size and absence of
vocal slits (in males) of siopela. Presence or absence of vocal slits
was not used as a character in the multivariate analysis, but the plot
of individuals on canonical axes I and II ( Fig. 3 ) may reflect a size
trend. The four populations of arhorescandens from \^ista Hermosa,
Tezuitlan, and Acultzingo are the smallest frogs and are grouped
near the bottom of the plot. Hyla siopela from Cofre de Perote are
the largest specimens and appear at the top of the plot.

All fourteen species under consideration were analyzed next
using BMD07M, MULDIS, and PROJ3D. The populations of Hyla

28 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

Fig. 3. Populations of Hyla siopcJa (A and B) and Htjia arhoicscandcns
( C-I ) plotted on canonical axes I ( horizontal axis ) and II ( vertical axis ) .
Large circles are confidence circles around the groups; small circles are confi-
dence circles around the means. Localities are as follows: A. Cofre de Perote,
Veracruz. B. Cerro Pelon, Oaxaca. C. N Vista Hermosa, Oaxaca, low elevations.
D. N Vista Hermosa, Oaxaca, high elevations. E. Cerro Machin, Oaxaca, (one
specimen). F. Cerro San Felipe, Oaxaca. G. Tezuiltlan, Puebla. H. Acultzingo,
Veracruz. I. Acultzingo, Veracruz (type specimen).

siopela and arhorescandens discussed above were lumped for tliis
part of the analysis. In addition to showing the interspecific rela-
tionships, results from BMD07M and MULDIS can be compared
to show that additional information can be gained by examining
plots with three axes rather than two.

Figures 4 and 5, from the BMD07M program, show the fourteen
species plotted on canonical axes I and II. These two figures were
made from the same graph and can be superimposed by aligning
the zeros of each coordinate. The groups were separated for the
figures by putting two species groups on each plot. Eight axes were
required to explain the total amount of variation, with (SI percent
explained by the first three axes, as follows: one, 41 percent; two, 30
percent; and three, 10 percent. The values of the final F matrix

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP

29

arborescandens

-0

J L

Fig. 4. Indi\idiuils ot the II[ihi histiiicta group (includes pcnthctcr) and
the Hijla arborescandens group (includes siopela, rohertsorum, and nujkter)
plotted on canonical axes I (horizontal axis) and II (vertical axis). Small cir-
cles are confidence circles around the means; large circles are confidence circles
around the groups. This figure can be superimposed on Figure 5 by aligning
the zeros of the coordinates.

indicate that all groups (species) are significantly different from all
others.

Figure 4 includes the bistincta and arborescandens species
groups, whereas figure 5 includes the crassa and charadricola species
groups. Superimposing the two figures, the bistincta and crassa
groups are most distinct, and the greatest overlap occurs between
the cliaradricola and aborcscandens species groups. In their original
description of mykter, Adler and Dennis ( 1972 ) state that mykter
appears to be related most closely to bogertae, crassa, pachydernia,
robertsorum, and siopela. Superimposing the two figures, mykter,
the first four aforementioned species and sabrina occur together in
the upper left quadrat of the plot. Adler and Dennis further say
that niyktcr is probabh' most closeh' related to siopela, which is not
indicated b\' the graph. The following are four characters which
Adler and Dennis use to indicate similarity between nujkter and

30 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

sabrina

chrvses

pachydernia

-0

bogertae

Fig. 5. Same as Fig. 4, Ijut includes the crassa species group (cijanomma,
bogertae, pachyclerma, and cembra) and the charadricola species group (in-
cludes chrijses and sabrina).

siopela: 1) rostral keel present; 2) traces of webbing between outer
fingers; 3) thin skin; and 4) least robust bodies of the six species
named above. Of these four characters, I used only the first in the
multivariate analysis because the other three are not quantifiable.
Adler and Dennis then list the following nine characters which dif-
ferentiate mykter and siopela: 1) snout truncate in dorsal view in
siopela; 2) snout truncate in lateral view of siopela; 3) canthus more
angular in siopela; 4) nostrils of siopela more anterior; 5) nuptial
excrescences present on first and second fingers of siopela, but on all
fingers in mykter; 6) toes slightly more fully webbed in mykter: 7)
cloacal sheath longer in mykter; 8) weak tarsal fold in mykter; and
9) proportionally longer legs in mykter. Of these nine characters,
I used 6 in the multivariate analysis.

Adler and Dennis ( 1972) state further that of the four remaining
species (hofi,ertae, crassa, pachyderma, and rohertsorum), rohert-
sorum most closely approaches siopela and mykter. According to
my analysis, mykter is closest to rohertsorum. This is shown on both

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP

31

<

CD

■^^

a

Jw

iZ

•^

-5

Q

O

^
w

u

CO

'C

o

g

ai

!X

a

z

Ph

-c

«

M

^

_c

a:

ft

'5

^

-o

w

c

§>

CO

u

CS Si

■S ^ ^

Ch

^

^

'^

vj

13

^

c

^

s>

c

?

<u

^

1^

s

■e

1

-5

J'

CO

'^

qj

.

Co

•£

a:

2

c3

u

^

a:

W2

i.r

^

<U

w

'o

la

OJ

r«»

^

a

■^

a

y;

•Kt

^

^

u

^

a;

"7

tx

^

^

C

■^

■"-

^

rs

ai

-a

U

n

ito

o

"^

c

c

V3

Co

Q

^

O

*

-a

t—

1^

a

u

J

■^^

^

"<.

<

H

CQ

«

o

w

u

aa

1 1 1

1

1

1

1

1

1

1

1

1

O

CO

o
d

1 1 1

1

1

1

1

1

1

1

1

=

d

O
CD

CD

1 1 1

1

1

1

1

1

1

1

o

CO
X

00

in
CO

00

CD

in

00

1 1 1

1

1

1

1

1

1

o

o
o

00

X
00

in

X

CD
00

1 1 1

1

1

1

1

1

o

X
CO

CD

CD

CD

CD

CO
CD

o
oc

CO

t-H
C<1

1 1 1

1

1

1

1

o

m

CO

o

X

in

T,

CD
X

X

t-'

t-^

CD

t^

t-^

1 1 1

1

1

1

^^^

X
X
00

CD

in
00

CD

in

o

in

X

X

in

CD

in

l>

in

t-^

CD

1 1 1

1

1

O

CD

m

CD

in

in

CD
X

m

X
X

m

CM

in

o

05

l>

ai

t-^

d

»— *

CD

»— *

d

1 1 1

1

O

05
CO

in

X
CD

C5

CO
CO

in

CO
CO

o

CO

en

CD

9"'

'CO

o

in

CO

05
CD
X

00

Tp"

CD

t-^

00

t-'

in

oc

d

1 1 1

o

CO

in

in

CO

00

X

05

CO

CD

o

CD

in

d

1— 1

t-^

oc

d

in

00

d

o

d

d

1 1 c

CO

CD

lO

o

m
in

CD

o
o

X

s

X

m

in

X

o

X

CD
G5

05

o>

^H

t-^

oc

sC

CD

d

d

C5

in

G5

in

00
1 o 2

CD
O

CD
CO

05

X

CO

CD

in

C5

;:::;

CO

CD

oc
CO

2
P

CO

ao

t-^

t-"

in

in

00

in

00

d

t-^

d

d

CD CD
t- O

o — ; 05

CO

CD
CD

CD

CO
X

CD
G5

o

CO

CD

05

CO

05

m

I"' 3C

t-^

I>

in

^'

00

in

00

o^

t-^

o

I>^

<muCaDL.CKHH'-.«j

32 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

Fig. 6. A three-dimensional plot of the first three canonical axes from
PROJ3D including all species originally in the Hyla histincta species group and
Hyla arborescandens. A. Hyla bisiincta. B. H. pcntheter. C. H. charadricola.
D. H. chryses. E. H. sabrina. F. H. crassa. G. H. pachijderma. H. H. bogertae.
I. H. cyanomma. J. H. cembra. K. H. arborescandens. L. H. siopela. M. H.
robertsorum. N. H. mykter. Species groups indicated by symbols, as follows:
closed circle, Hyla bisiincta group; hatched line, Hyla charadricola group; open
circle, Hyla crassa group; stippled circle, Hyla arborescandens group.

the two-dimensional and the three-dimensional plots. The three-
dimensional plot (Fig. 6) shows that mykter is 5.062 units from
robertsorum and 5.208 from sabrina. Table 7 presents the distances
of all species from all other species, but figure 6 has lines drawn so
that each species is connected to the one closest to it.

Straughan and Wright ( 1969), in their discussion of the relation-
ships of these species, state that siopela is so superficially similar to
robertsorum that they could be considered to belong to the same
species population. The two characters they list which separate
these two species are 1) the presence of the rostral keel in siopela,
and its absence in robertsorum, and 2) the size of the bony prepollex
element, which is twice as large in siopela as in robertsorum. My
analysis reveals that robertsorum and siopela are not especially
close. The two-dimensional plot (Figs. 4 and 5) shows robertsorum
closest to pachijderma, mykter, and charadricola. The three-dimen-
sional plot, however, shows that robertsorum is actually closest to
charadricola ( 4.539 units ) and is 6.968 units from pachyderma. This
example also shows clearly the value of examining a three- rather
than a two-dimensional plot.

Another interesting fact revealed by the three-dimensional plot
is the relationship between arborescandens and siopela. The two-

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 33

dimensional plot (Fig. 4) shows the two overlapping to a great
extent. However, figure 6 reveals that they are separated by a large
distance (4.176 units) along the third axis. From the MULDIS
program, the following combination of characters are found to con-
tribute most significantly to the separation of groups along the third
axis: 1) thoracic fold; 2) tubercles on lateral margin of forearm;

3) tarsal fold; 4) head width; and 5) snout- vent length. These
characters are also important in separating crassa from pachyderma.

Regarding separation of groups along the first and second axes
in the three-dimensional plots, the following combination of char-
acters contribute most to separation along the first axis : 1 ) tubercles
on lateral margin of forearm; 2) subarticular tubercles; 3) thoracic
fold; 4) foot length; and 5) tarsal fold. The characters contributing
most to separation along the second axis are : 1 ) tubercles on lateral
margin of forearm; 2) color of dorsum; 3) pigmentation of thighs;

4) pigmentation of chin; 5) subarticular tubercles; and 6) thoracic
fold.

The characters wliich I used initially to characterize the species
groups are those which are easily observed in the field, and which
other workers have considered important in the past. Some of these
characters could not be used in the multivariate analysis as ex-
plained previously. However, the analysis has shown that, using 30
quantifiable characters, a somewliat different classification can be
obtained. When more data are available, such as tadpole and skel-
etal data, and when larger samples of the adult frogs are available,
we may find that another classification will more accurately repre-
sent phenetic relationships among these species.

SUMMARY

Five tree frogs were assigned to the Hijla bistincta species group
by Duellman in 1964, and fi\'e more species were named and as-
signed to this group between 1965 and 1972. Three new species of
tree frogs discovered in Oaxaca, Mexico, defined herein, are closely
related to frogs of the H. bistincta group. The three new species are
Hyla cyanomma and Hyla sabrina from the Sierra de Juarez, and
HyJa cembra from the Sierra Madre del Sur.

The Hyla bistincta group is reviewed, and with Hyla arbo-
rescandens, is divided into four species groups, as follows: 1) the
Hyla bistincta group, including Hyla bistincta and Hyla pentheter,
2) the Hyla charadricola group, consisting of Hyla charadricola,
Hyla chryses, and Hyla sabrina, 3) the Hyla crassa group with Hyla
crassa, Hyla pachyderma, Hyla bogertae, Hyla cyanomma, and Hyla
cembra, and 4) the Hyla arborescandens group including Hyla
arborescandens, Hyla roberfsorum, Hyla siopela, and Hyla mykter.
Characters used to define these new groups are those which have

34 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

been used in the past to delimit subgroups of the HyJa bistincta
group.

Two multivariate statistical programs, discriminant classification
analysis (BMD07M) and discriminant function analysis (MULDIS),
are used to determine phenetic relationships among these 14 spe-
cies. Comparisons of these statistical methods are made and the
advantages of each discussed. The results of the statistical analysis
provide a different arrangement of species than the combination of
characters used in the past to define subgroups within the HyJa
bistincta group. The statistical results are considered preliminary
because of the small number of specimens available of some species,
resulting in a deficiency of morphological, skeletal, tadpole, and
ecological data. Therefore, the four new species groups are defined
by characters used by previous workers.

RESUMEN

Cinco ranas fueron asignadas por Duellman en 1964 al grupo de
especies HyJa bistincta, y cinco cspecies mas fueron denominadas
y asignadas a este grupo entre 1965 y 1972. Tres nuevas especies de
ranas descubiertas en Oaxaca, Mexico, definidas aqui, estan es-
trechamente relacionadas con ranas del grupo de H. bistincta. Las
tres nuevas especies son Hyla cyanomma e HyJa saJjrina de la Sierra
de Juarez, e HyJa cembra de la Sierra Madre del Sur.

El grupo HyJa bistincta es revisado, y con HyJa arJjorescandens,
es dividido en cuatro grupos de especies, a continuacion: 1) el
grupo HyJa J)istincta, indugendo HyJa bistincta e HyJa pentJietcr,
2) el giTipo HyJa charadricoJa, consistiendo de HyJa charadricoJa,
HyJa chryses, e HyJa saJjrina, 3) el grupo HyJa crassa con HyJa
crassa, HyJa pacJiycJerma, HyJa Jwiicrtae, HyJa cyanomma, e HyJa
cemJjra, 4) el grupo HyJa arJ)orescancJens incluyendo HyJa arJjo-
rescandens, HyJa robertsorum, HyJa siopeJa, y HyJa myJ<.ter. Las
caracteres usados para definir estos nuevos grupos son aquellos cjue
han sido usados en el pasado para delimitar subgrupos del grupo
HyJa bistincta.

Dos programas de estadisticas multivariadas, analisis de clasi-
ficacion discriminante (BMD07M) y analisis de funciones dis-
criminantes (MULDIS), son usados para determinar relaciones
feneticas entre estas 14 especies. Comparaciones entre estos dos
metodos estadisticos son hechas y las ventajas de cada uno son dis-
cutidas. Los resultados de estos analisis estadisticos produccn una
distribucion de especies diferente a la combinacion de caracteres
usados en el pasado para definir subgrupos dentro del grupo HyJa
J)istincta. Los analisis estadisticos son considerados preliminarcs
debido al pequeiio numero de especimenes disponibles de algunas
especies, resultando en una deficiencia de datos morfologicos, de

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 33

esqueleto, de renaciiajos, y ecologicos. Por lo tanto, los cuatro
nuevos giupos de especies son definidos por caracteres usados por
trabajadorcs anteriores.

SPECIMENS EXAMINED

HyJa arhorcscandens

OAXACA: Ceno San Felipe, FMNH 104575-78, UIMNH 27821-23, 30361,
KU 137120-22, 139851-53 (tadpoles); Ceno Machin, UIMNH 50068-70; 3.5
km W Cerro Machin, 2370 m, KU 137119, 139837 (tadpoles); 15.8 mi SE
Nochixtlan, UIMNH 38548; 32.5 mi NW Oaxaca, UIMNH 38546-47; 3.9 km
S Vista Hermosa, KU 86994; 4.2 km S Vista Hennosa, 1580 m, KU 86999-
7007, 139816 (tadpoles); 6.5 km S Vista Hermosa, 1610 m, KU 58453-3,
71216-20, 87008, 137123; 7.8 km S Vista Hermosa, 1650 m, KU 137124,
139817-19 (tadpoles); 9.1 km S Vista Hermosa, 1750 m, KU 137125-29; 9.3
km S Vista Hermosa, 1700 m, KU 137130-31; 9.4 km S Vista Hemiosa, 1710
m, KU 139820-21 (tadpoles); 11.0 km S Vista Hermosa, 2070 m, KU 87012;
11.1 km S Vista Hermosa, 1840 m, KU 137132-35, 139822 (tadpoles), 139825-
27 (tadpoles); 11.6 km S Vista Hennosa, 1870 m, KU 137136 (C & S), 139828-
32 (tadpoles); 11.9 km S Vista Hermosa, 1920 m, KU 137137-42; 12.2 km
S Vista Hermosa, 1920 m, KU 139834 (tadpoles); 12.3 km S Vista Hermosa,
1920 m, KU 137143-45; 14.3 km S Vista Hermosa, 1970 m, KU 137146; 15.0
km S Vista Hermosa, 1980 m, KU 104126 (tadpoles), 137147-51; 15.8 km S
Vista Hermosa, 1990 m, KU 137152-76, 139835 (tadpoles); 16.0 km S Vista
Hermosa, 2180 m, KU 87013-14; 16.6 km S Vista Hermosa, 2020 m, KU
137177-78, 139836 (tadpoles).

PUEBLA: 14.4 km W Huachinango, 2280 m, KU 58912; Paraja Verde,
UIMNH 49119-20; Puente Colorado, UIMNH 49131; Rio Octapa, 3.7 km NNE
Tezuitlan, 1800 m, KU 53818, 55985, 58467-8, 64316-20, 68379-80 (tadpoles).

VERACRUZ: Acnltzingo, FMNH 99261-2, 103285-6, 110596-8; near
Acnltzingo, UIMNH 27820, 27825-27, 49113-16; 0.5 mi above Acultzingo,
FMNH 123916-19; 3 km SW Acultzingo, UIMNH 25045 (type); near Bar-
ranca, 1.2 mi SW La loya, FMNH 70601-18, 70621-3, 70625-30; Barranca, 3
mi SW La loya, FMNH 70624; La loya, FNLNH 70619-20; Pan de Olla, FMNH
125351-8, 172137-42, UIMNH 49117-18, 10135-49; 2 km N Paraja Nuevo, KU
23955; Pica de Orizaba, UIMNH 57635-84; Teocelo, Coscomatepec, Huatusco,
KU 26846-50.

Hiila histincta

GUERRERO: 4.5 km (by road) E El Union (=6 km SW Chilapa), 1525
m, KU 140420-3.

MEXICO: W. Villa \'ictoria, UIMNH 29162.

MICHOACAN: Uruapan, UIMNH 20457, 28167, KU 68077-8, 69093
(skeleton), UMMZ 112839 (14), 115233 (5); 14 mi from Rt 15 on Valle de
BraN o Road, KU 148653.

MORELOS: 2 mi N Cuernavaca, UIMNH 28167.

OAXACA: Cerro San Felipe, UIMNH 28163, 56593, 59236-43, 60076-84,
62826-27, 73714-15, 73883-85, 73907.

SINALOA: 1 mi E Santa Lucia, 5650 ft, KU 44567.

VERACRUZ: Acultzingo, UIMNH 28164-66, 49133-34, 70188-91, 2 mi W
Acultzingo, UIMXH 70192.

IhjJa hogertae
OAXACA: Tributary of Rio Atoyac, below Vi\'ero El Tapanal, 1.6 km S
La Cofradia, LACM 44400 (type), 44401-13, 44414 (tadpoles).

36 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY

Hyla cembra
OAXACA: Campamento Rio Molino, 2160 m, KU 137035 (type); 139857-
9 (tadpoles); 11.6 km S Campamento Rio Molino, 1970 m, KU 139860 (tad-
poles); 12.8 km S Campamento Rio Molino, 1930 m, KU 139861 (tadpoles).

HtjJa charadricoJa

HIDALGO: Lago de Tejocotal, 11 km E Acaxochitlan, 2250 m, KU 58438;
4 km SW Tianguistengo, 2080 m, KU 53811-12.

PUEBLA: 11.7 km W Huachinango, UMMZ 121567 (5); 14.4 km W
Huachinango, 2280 m, KU 53813-15, 58414 (type), 58415-22, 58424-37, 59886
(C&S), 152366-9, UMMZ 118166 (5).

Hylo cJinjses
GUERRERO: between Puerto Chico and Asoleadero, 45 km airline WNW
Chilpancingo, 2540-2600 m, KU 106306, UMMZ 125373, 125375.

Htjla crassa

OAXACA: Cerro San Felipe, UIMNH 25050 (type); 1.9 km S El Estu-
diante, ca 1850 m, KU 148696-700; 9 mi NE Oaxaca, KU 125354.

Hijla cijanomma

OAXACA: 0.9 km N Cerro Pelon, 2670 m, KU 137031-34; 1.0 km N Cerro
Pelon, 2660 m, KU 137025-30; 1.2 km N Cerro Pelon, 2650 m, KU 137008
(skeleton), 137009-13, 137014 (type), 135015-24, 139849 (tadpoles), 148656-
63; 1.3 km N Cerro Pelon, 2650 m, KU 136997-7007; 31.2 mi N Guelatao,
9600 ft, KU 100507.

Hyla mykfer

GUERRERO: Asoleadero, 2520 m, KU 137552; 11.4 km (by road) SW
Puerto del Gallo, 1985 m, KU 1.37553 (type).

Hyla pachydeima

VERACRUZ: Pan de Olla, south of Tezuitlan, USNM 115026-28, 115029
(type).

Hyla pentheter

OAXACA: 0.1 km N Jalatengo, 1280 m, KU 136863-71, 139782-83 (tad-
poles); 3.9 km N Jalatengo, 1490 m, KU 1.39784 (tadpoles); 0.2 km S Jala-
tengo, 1280 m, KU 139780-81 (tadpoles); 0.4 km S Jalatengo, 1290 m, KU
139779 (tadpoles); 0.8 km S Jalatengo at Rio Jalatengo, 1280 m, KU 139778
(tadpoles); 5.1 km S Jalatengo, 1390 m, KU 136872-77, 139775-77 (tadpoles);
29.0 km SSE Juchatengo, 1980 m, KU 86936; 25.8 km N San Gabriel Mi.x-
tepec, 1230 m, KU 1.39785-88 (tadpoles); 27.8 km N San Gabriel Mi.xtepec,
1320 m, KU 136887-89, 139789-91 (tadpoles); 31.0 km N San Gabriel Mix-
tepec, 1860 m, KU 104142 (tadpoles); .32.9 km N San Gabriel Mixtepec, 1530
m, KU 136879-83, 136884 (C&S), 136885-6, 139799 (tadpoles); 33.1 k-m N
San Gabriel Mixtepec, 1530 m, KU 139793 (tadpoles), 139796 (tadpoles); 36.7
km N San Gabriel Mixtepec, 1690 m, KU 136878; 37.0 km N San Gabriel
Mixtepec, 1860 m, KU 100931-33, 117426 (skeleton).

GUERRERO: 5.6 km (by road) NE Yerbabuena, 2000 m, KU 140424-26.

Hyla robertsorum

HIDALGO: El Chico Parque Nacional, 3050 m, KU 57650-71, 59824-5
(skeletons), 59914-5 (C & S), 71757 (skeleton), 110117 (tadpoles); 3.3 km
N Zacualtipan, 2340 m, KU 53810, 60078 (tadpoles); 8.5 km SE Zacualtipan,
2250 m, KU 60079.

RE-EVALUATION OF HYLA BISTINCTA SPECIES GROUP 37

Hyla sabrina

OAXACA: 7.8 km S Vista Hermosa, 1650 m, KU 137038-40, 11.0 km S
Vista Hennosa, KU 87011; 11.1 km S Vista Hermosa, 1840 m, KU 137041-50;
11.6 km S Vista Hermosa, 1910 m, KU 137051-3; 11.9 km S Vista Hermosa,
1920 m, KU 137054-62, 137063 (C & S), 137064-72; 12.3 km S Vista Her-
mosa, 1920 m, KU 137073-77, 13.5 km S Vista Hemiosa, 1920 m, KU 137078-
9; 15.8 km S Vista Hennosa, 1990 m, 137080-85, 137086 (type); 16.6 km S
Vista Hemiosa, 2020 m, 137087-88.

Hyla siopela

OAXACA: N slope Cerro Pelon, KU 59998 (tadpoles), 104125 (tadpoles);
0.9 km N Cerro Pelon, 2670 m, KU 137089-104, 139812-5 (tadpoles); 1.0 km
N Cerro Pelon, 2660 m, KU 137105-9, 139838-41 (tadpoles); 1.2 km N Cerro
Pelon, 2650 m, 137110, 137111 (C & S), 137112, 139842-7 (tadpoles), 139849
(tadpoles), 148664-9; 1.0-1.2 km x\ Cerro Pelon, 2650-60 m, 137113-6; 1.3
km N Cerro Pelon, 2640 m, 137117-8, 139848 (tadpoles).

VERACRUZ: W slope Cofre de Perote, 2500-2550 m, KU 100976-80,
100981 (tvpe), 100982-5, 100986-1000, 105628-34, 110118 (tadpoles), 117428-
30 (skeletons); Rancho El Capulio, 7 km SE Perote, 2920 m, KU 129163.

LITERATURE CITED

Adler, Kraig

1965. Three new frogs of the genus Hyla from the Sierra Madre del Sur
of Me.xico. Occas. Papers Mus. Zool., Univ. Michigan, 642:1-18.
Adler, Kraig, and David M. Dennis

1972. New tree frogs of the genus Hyla from the cloud forests of western
Guerrero, Me.xico. Occas. Papers Mus. Nat. Hist., Uniw, Kansas,
7:1-19.
Anderson, T. W.

1958. An introduction to multivariate statistical analysis. John Wiley and
Sons, New York, 374 pp.

Dixon, W.J. (ed).

1970. Biomedical computer programs. Uni\-. California Press, x + 600 pp.
Duellman, William E.

1964. A review of the frogs of the Hyla histincta group. Univ. Kansas

Publ., Mus. Nat. Hist., 15:469-491.
1968. Descriptions of new hylid frogs from Mexico and Central America.

Univ. Kansas Publ., Mus. Nat. Hist., 17:559-578.
1970. The hylid frogs of Middle America. Monogr. Mus. Nat. Hist., Univ.
Kansas, l:.xii -f 753 pp.
GOSNER, K. L.

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

1968. Multivariate statistical analysis for biologists. Methuen and Co.
Ltd., London, xii + 209 pp.

Straughan, Ian R., and John W. Wright

1969. A new stream breeding frog from Oaxaca, Mexico (Anura, Hylidae).
Contr. Sci., Los Angeles Co. Mus., 169:1-12.

Taylor, Edward H., and Hobart M. Smith

1945. Summary of the collections of amphibians made in Me.xico under
the Walter Rathbone Bacon traveling scholarship. Proc. U.S. Natl.
Mus., 95:521-613.

University of Kansas Publications
MUSEUM OF NATURAL HISTORY

The University of Kansas Publications, Museum of Natural
History, beginning with volume 1 in 1946, was discontinued
with volume 20 in 1971. Shorter research papers formerly pub-
Ushed in the above series are now published as Occasional
Papers, Museum of Natural History. The Miscellaneous Pub-
lications, Museum of Natural History, began with number 1 in
1946. Longer research papers are published 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 publications committee.

Institutional libraries interested in exchanging publications
may obtain the Occasional Papers and Miscellaneous Publica-
tions by addressing the Exchange Librarian, University of Kan-
sas Library, Lawrence, Kansas 66045. Individuals may pur-
chase separate numbers of all series. Prices may be obtained
upon request addressed to Publications Secretary, Museum of
Natural History, University of Kansas, Lawrence, Kansas 66045.

Editor: Richard F. Johnston-

printed BY

UNIVERSITY OF KANSAS PRINTING SERVICE
LAWRENCE, KANSAS

3 2044 093 361 657