HARVARD UNIVERSITY

Library of the

Museum of

Comparative Zoology

1

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE

Vol. XCI, No. 1

THE LIZARDS OF BRITISH SOMALILAND

By H. W. Parker, M.A.

Department of Zoology, British Museum

(Natural History)

With an appendix on Topography and Climate
By Capt. R. H. R. Taylor, O.B.E.

CAMBRIDGE, MASS., U.S.A.

PRINTED FOR THE MUSEUM

August, 1942

PUBLICATIONS
OF THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

The Bulletin and Memoirs are devoted to the publication of
investigations by the Staff of the Museum or of reports by spec-
ialists upon the Museum collections or explorations.

Of the Bulletin, Vols. I to LXXXIX have appeared and of
the Memoirs, Vol. I to LVI.

These publications are issued in numbers at irregular intervals.
Each number of the Bulletin and of the Memoirs is sold separately.
A price list of the publications of the Museum will be sent upon
application to the Director of the Museum of Comparative
Zoology, Cambridge, Massachusetts.

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE

Vol. XCI, No. 1

THE LIZARDS OF BRITISH SOMALILAND

By H. W. Parker, M.A.

Department of Zoology, British Museum

(Natural History)

With an appendix on Topography and Climate
By Capt. R, H. R. Taylor, O.B.E.

CAMBRIDGE, MASS., U. S. A.

PRINTED FOR THE MUSEUM

August, 1942

No. 1 — The Lizards of British Somaliland

By H. W. Parker, M.A.

AYith an appendix on Topography and Climate
By Capt. R. H. R. Taylor, O. B. E.

CONTENTS

Page

Introduction 1

Taxonomic Discussion of the Species 21

Appendix on Topography and Climate 92

Bibliography 97

INTRODUCTION

The herpetology of British Somaliland offers a peculiarly interesting
field for research, and a glance at the bibliography appended to the
present paper indicates that it has already received a considerable
amount of attention. The absence of roads and railways has hindered
the complete exploration of the country so that, until recently, the
collections obtained have been meagre and restricted to the more
accessible districts. But the British Museum was singularly fortunate
in the fact that Capt. Taylor, who served on the two boundary com-
missions which have surveyed the whole of the boundaries of the pro-
tectorate, took a keen interest in herpetology and has amassed what is
undoubtedly the finest collection which has yet been obtained in the
region. The author has already reported upon a small part of this col-
lection (1932), but the larger amount of material now available, makes
it possible to present a tentative account of the complete lizard fauna.
The author has already drawn attention to the interesting zoo-geo-
graphical affinities of the herpetological fauna of Somaliland and its
outlier in the Lake Rudolf basin (1932, 1936) and the conclusions
reached have been fully confirmed by the additional material. But
now, though the fauna can by no means be considered thoroughly
known, it is possible to make a preliminary estimate of the distribution
of the species within the territory; for a better understanding of this
matter a knowledge of the topography and climate is essential, so
that brief sketches of these subjects by Capt. Taylor are given as
appendices herewith.

4 bulletin: museum of comparative zoology

Even a cursory glance at the taxonomic section of the present paper
will reveal two important facts ; first, the very large number of species
to be found in such a relatively small area and, secondly, the very high
proportion of endemic species and races. Excluding from consideration
the two introduced oriental geckos, Hemidactylus flavwiridis Riippell
and H. frenatus Dum. & Bib., no less than 72 species of lizards are
found in an area comparable in size with Great Britain and one which
is, on the whole, not remarkable for diversity of climatic and ecological
conditions. Forty-five of these species (62 per cent) are endemics, some
of which show definite indications of differentiation into geographical
races within the country, and, of the non-endemic species, an appreci-
able proportion also shows differentiation in the Somaliland area. Such
a very high degree of endemism on part of a continental land-mass is
remarkable in the extreme and suggests specialisation or isolation.
Probably both of these factors and others, have operated in the pro-
duction of the present fauna. There is manifestly quite a high degree
of geographical isolation, the triangular peninsula being bounded on
two sides by the sea and on part of the third side by the Abyssinian
mountains and Rift valley. This purely geographical isolation is rein-
forced by the climatic conditions. The rainfall of the north of the pen-
insula is for the most part less than 20 inches annually, but in the coun-
tries to the west and south the annual precipitation rises to an average
of at least 20-30 inches except in the Lake Rudolf basin which, as has
been pointed out, is an outlier of the Somali arid zone and bears a
strong faunal resemblance to it. (fig. 1). The geographical isolation
has, apparently, existed since the beginning, for the Abyssinian Plateau
is regarded by Gregory (1921) as dating from the Jurassic whilst Soma-
liland itself is of later date. From a geological standpoint it is a rela-
tively new land; Gregory believes that the main elevation of the land-
mass which includes the peninsula, dates from the Cretaceous, but the
presence of Miocene limestone (Dubar series) in the north of British
Somaliland and large masses of Middle Eocene limestone in the Sol
Haud and Eastern Haud (Macfadyen 1933) indicates that some of the
country, at least, is of more recent origin. Separation from Southern
Arabia appears to date from the Middle Pliocene (Macfadyen op. cit.
p. 15).

The climatic history is not so well understood, but it is certain that
there have been considerable changes since the land came into exist-
ence. A great deal of attention has been devoted to climatic changes
during the Pleistocene in the countries surrounding Somaliland, but
the absence of lakes renders this study difficult in the peninsula itself.

PARKER: LIZARDS OF BRITISH SOMALILAND

There is, however, no reason to suppose that a succession of climatic
changes which extended from Palestine to Tanganyika Territory did
not affect Somaliland also, and it is to be presumed that the succession

■ * * +i

O" to 10".

*' * ' *\ 20" to 30'!
40"to 60'!

o o o

o o
o o o

:•:•:

10" to 20'!

30" to 4-0'!
60" to 80".

Fig. 1. Mean Annual Rainfall of N. E. Africa. (After Knox).

6 bulletin: museum of comparative zoology

was the same there as elsewhere. Parkinson (1932) and Macfadyen
(1932) have accumulated definite evidence that there were indeed wet
periods in the area, whilst Fuchs (1939) has demonstrated a complex
set of such changes in the Lake Rudolf basin (fig. 2).

Wayland (in Hale Carpenter, 1935, p. 437), basing his conclusions
on the evidence obtained in Uganda and Kenya, believes that Abys-
sinia underwent two major pluvial periods during the Pleistocene,
with a prolonged Interpluvial, and "two wet and two dry post-Pluvial
(or epi-Pluvial) phases of lesser intensity". Fuchs finds evidence of 5
maximal lake-levels in Lake Rudolf and the suggestion of a very much
less marked sixth (fig. 2). The first of these, which occurred at the end
of the Lower Pleistocene, he correlates with Wayland's Pluvial I; the
next three, placed as Chellean, Acheulian and Gamblian are correlated
with Wayland's Pluvial II and the two most recent and least marked
would accordingly correspond with the post-Pluvial Makalian and
Nakuran wet phases. The evidence so far obtained in British Somali-
land is not extensive, but the succession found by Macfadyen in the
Bihendula-Dagah Shabell-Daban area is the most complete. The
earliest fresh-water deposits are in the Daban Conglomerates and these
are tentatively referred (Macfadyen 1933) to the late Eocene. This
was followed by a phase of Posthumous faulting in the Gulf of Aden
trend along the Dagah Shabell fault, which may, perhaps, be con-
nected with the major movements which took place in Mid-Pliocene
times and resulted in the separation of the peninsula from Arabia.
There followed in succession (1) a period of boulder deposition, (2) a
period of erosion "during what must have been a long-continued wet
period", (3) the infilling of some of the valleys with the Younger
Gravels and finally (4) a second period of erosion which Macfadyen
believes may be the period existing to the present day. The significance
of these various phases from a climatological point of view is uncertain;
but it can safely be said that there must have been at least one and
probably more wet periods of considerable duration during the
Pleistocene.

Thus so far as British Somaliland is concerned, it appears from the
geological evidence that, after the emergence of the land in the Middle
Eocene, there was opportunity for faunal invasion both from the north
and south; to the westward the Abyssinian plateau probably acted as a
barrier to the free movement of many animals. This state of affairs
continued until the middle of the Pliocene when the formation of the
Aden Gulf inhibited or seriously restricted the possibilities of colonisa-
tion from the north, and a degree of geographical isolation was imposed

PARKER: LIZARDS OF BRITISH SOMALILAXD 7

which has persisted until the present time. Subsequently, alternations
of wet and dry climatic conditions must have profoundly affected the

Hvsnsvtf

IN3D3H

Fig. 2. Pleistocene Fluctuations in Lake Rudolf. (After Fuchs).

flora and fauna. It seems highly improbable that any of the original
migrants into the country would have persisted unchanged through

8 bulletin: museum of comparative zoology

these environmental changes and it is reasonable to suppose that the
present lizard-fauna, adapted to the semi-desert conditions now pre-
vailing has to a considerable extent been evolved in situ since the last
major wet phase. It might be argued that the bulk of the changes
have taken place since the latest, Nakuran, wet phase which is esti-
mated by Brooks (1922) at about 850 B.C. If this were so it would
involve a rate of evolutionary change far in excess of that usually
deemed probable, but it is a possibility which cannot be entirely
neglected. Moreau (1933), however, calculates that during the Na-
kuran wet phase the average annual rainfall of the Lake Nakuru basin
was probably not more than, at most, 5}/2 inches greater than it is to-
day and so can have been of little consequence. Even the Makalian
period, dated between 10-20,000 B.C. by Brooks, is not believed to
have had a rainfall "sufficiently greater than the present to permit of
general forest growth" (Moreau op. cit. p. 431), but the Gamblian
(ending about 20,000 B.C.; Brooks) seems to have been an "epoch of
continuous forest" in Kenya. Of the present known lizard fauna in
Somaliland not a single species can be regarded as a forest form, and
if, as seems likely, conditions were similar in Kenya and Somaliland,
it seems legitimate to conclude that the whole fauna has been com-
pletely changed since that time.

Analysis of the elements composing the fauna lends support to many
of these views. The species, excluding from consideration the intro-
duced species mentioned above (p. 4), can be divided into three
groups as follows:

I. Northern species with an Eremian or Irano-Turanian distri-
bution.

II. Ethiopian species.

III. Endemics.

These are listed in the accompanying tables and the approximate
distribution of each species in Somaliland and elsewhere is given. The
topographical divisions used by Capt. Taylor (Appendix I) are slightly
modified for convenience in indicating distribution, and Ethiopia is
divided into two parts, the highland plateau indicated by Ethiopia I
and the Ogaden which is, of course, continuous with the Haud.

The following divisions are used in indicating distribution in British
Somaliland. (1) The Gubati (" a " of Capt. Taylor's topographical
divisions) ; (2) The Mountains of the Ogo. Captain Taylor has pointed
out that the northern mountain chain is interrupted eastward of the
Golis ranges by the Huguf Plain and this interruption appears to coin-
cide with the distributional limits of several species. Consequently

PARKER: LIZARDS OF BRITISH SOMALILAND 9

the mountains to the westward of this plain are included with the high-
land slopes of the Ogo (Taylor's "c") as a single unit. (3) The Eastern
Mountains, comprising the highlands east of the Huguf Plain. (4) The
Sol Hand, including the Daror Valley, (d & g). (5) The Xogal J' alley (e)
(6) The Haud (f). In addition, Italian Somaliland is considered under
three divisions: (7) Migiurtina being the sultanate of that name in
the extreme north; (8) Obbia, the central area; (9) Shebeli & Juba,
being the southern zone comprising the provinces of Shebeli, Lower
Shebeli, Upper and Lower Juba.

I. THE EREMIAN SPECIES

(Table I)

The presence of these species indicates that the present fauna is
not wholly a development of the peninsula itself. They are obviously
migrants from the north-west, from the true desert region where
desiccation has progressed even further than in Somaliland. But from
the fact that they form only about 13 per cent of the total lizard fauna
it seems probable that the wet Abyssinian plateau does indeed form
an effective barrier to faunal movement. It should also be noted (a)
that many of them are limited to the extreme north-west, (b) that none
have penetrated as far as the Lake Rudolf region and (c) that none of
them show any tendency towards the formation of local races in So-
maliland. This latter fact offers a very marked contrast with both the
other faunal groups and the three above-mentioned facts together
rather suggest the possibility that these Eremian species are recent
migrants into the country and have had insufficient time to spread
widely and become differentiated. But it must also be remembered
that the northern coastal plain is also by far the hottest and driest part
of the country (appendix p. 96) and so is essentially similar to the
Saharan region; species adapted for life in the latter area will be par-
ticularly well adapted for the conditions prevailing in the Guban.

II. ETHIOPIAN SAVANNAH SPECIES
(Table II)

The second group contains 18 species (25 per cent) which have, for
the most part, a "Sudanese" distribution (using the term in reference to
Engler's botanical sub-province of the Savannah province) but also
includes some ubiquitous species such as Hemidactylus mabouia de

10 bulletin: museum of comparative zoology

Jonnes and Agama agama (Linn.). It will be noticed that (a) Many of
these species with a wide range form distinct geographical subspecies
and in four instances (Platypholis fasciata, Latastia longicaudata, Rham-
pholeon kersteni and Riopa modesta) one or more of these races is con-
fined to the Somali peninsula, (b) In two instances the species has an
"Eremian" subspecies {Latastia longicaudata, Agama agama) which
enters Somaliland only in the dry, hot, northern zone, exactly like the
species of group I (fig. 3) ; otherwise very few of these species are re-
corded from the Guban, and only two have been recorded from South-
ern Arabia.

III. ENDEMIC SPECIES
(Table III)

The 45 endemic species account for 62 per cent of the lizard fauna.
It will be noticed that very few of them are uniformly distributed
throughout the whole area and that a few appear to have, in the light
of our present knowledge, a very limited distribution. Further
exploration will no doubt prove that in many instances this is an illu-
sion, but there are evidently faunal differences within the country,
and a few ill-defined faunal provinces can be recognised. Best defined
of all is the Guban, which, as has already been pointed out, has such
marked Eremian affinities; of the 11 Somali forms found there, two are
not known to occur elsewhere and at least two others (Latastia boscai
boscai and Philochortus spinalis) are essentially north-western forms of
Eritrea and Ethiopia. The mountains across the north of the country
might be expected to form a distinctive faunal zone, and there are,
indeed, a number of montane species which extend across the whole
area and into the mountains of Ethiopia and Eritrea. But the Huguf
plain, which breaks the mountain chain, appears to form a slight
barrier, the mountains of the Ogo, to the westward of this dividing
line having a richer fauna than those to the eastward. To the south-
ward of the mountain chain, where the land slopes away gradually to
the south-east, there are no obvious barriers and there is no abrupt
faunal change. The species of the eastern mountains pass into the Sol
Haud (and Migiurtina), whilst those of the west encroach upon the
western and central Haud. But the Haud itself, together with the
Nogal Valley, the Sol Haud and the Ogaden, forms what appears to
be a distinctive faunal zone with a very rich fauna, of which 9 species
are not found elsewhere. A number of the other species which do occur
elsewhere show a tendency towards race-formation in this area though

PARKER: LIZARDS OF BRITISH SOMALILAND

11

►J

<

c
o

a

u

.a

S p

2 o>

■R *

o *

2i2

a oj
ft

(Records badly confused with H.

turcicus)

Borders of Mediterranean from
Spain to N.W. India

33

—

X

o

-t->

-5
q

s.

o3

►— 1

o

03

'C

<

CO

CD

'8

g

X

o

+3

93

>.
XI

93

fl
M

O

C
c,

—
o
■ — i

0)

=
*3

DC

X

93

-

c

93
-u
ft
>.
hC

w

2

X

o

03

CB

he

Morocco and Sierra Leone to N.W.
India

fjopny ayvq

vqnf op vpq^iS

>

>l

>

wiqqo

>

>

U9pv6Q

>

pnvj]

ft9W°A V°B°H

pnvH log

>

vut}j.ni5ij\[

>

>l

nw a

>

> ,

oSq fa '8}J\[

>

">

>

>

">|>

uvqnf)

>

>

>

>|>

>

>

>

>

viqvjLy mg

>

>

>

>

>

i>

I widotifjfl

>

>

>

>

>

>

V9J,JUfl

>

>

>

>

>

>

>

>

uvpng -fl ■ y

|>

>

>

>

>

>

>

CO

G
'55
to

3

1— H

03

a

(II

to

3
Cj

'o

3

W

to

DC
—

B

3

go

'-3

CO

'3

"53

CO
CO

o3
_G

CO

3
">>
o

03

-a
o

CO

j3

3
§

a

93

-

c
d

CD
i-
93

H

"C

CD

'>

0

03
—
o3

3

+3

a
5

CO

-

CD

—

c3

93

a

C

~

o

=

a

CO

-

o
/.

'33

03

=
O
09

U
93

a

CO

J3

"33
o

o

CO
CD

'G

"3

o

12

bulletin: museum of comparative zoology

Approximate Distribution
elsewhere

c
o

'So

aj
u

s

03

a

■~-

H
-u
a

CD

X

V

a

03

'o

a.
o

S

CO

O

o

Coastal belt of Kenya and Tang.
Territory

03
t.
c3
,0
c3

o

• +3

• 3

• O

• to

: c3

• 'C

• <!

'3

X

3
o3

M

Coastal plain and upland savannahs
s. to Cape Province

Littoral rocks, coastal plains and
upland savannahs in Kenya and
Tanganyika Territory

Coastal plains and upland savan-
nahs in Kenya and Tangan. Terr.

Coastal plains and upland savan-
nahs of Kenya

fl

X

3
+a

p3

OJ
b£

—

S3

~~
hi

o
3

X

Upland savannahs of Kenya and
Tang. Territory

Coastal plain and upland savannahs
Kenya and N. Tang. Terr.

/ppny 33/07

>l

>

>

>

cqn/ y 7j9q9i(g

">

*

>

>

>

>

viqqQ

>l

o~*

>

>

UBpvtiQ |

>l

>

">

>

L

1

>

>

>

>

>l

>

>

Pn»# Z0^ )

>

>

>

vui}uni6i}\[

>

siW -a

>

>

oBo fo s)j\[

>

>

>

>

>

> |>

uvqnf)

1

>

>

>

viqojy 'g

1 1

|

>

j vidoiiiifl

>

|

■>l>

>

>

V9J,}Ufl

>

> |>

?

>

uvpng fl ■ y

>

|>

>

>

II

Hemidactylus mabouia

03

.5
'3

CO

o3

£ "i

O (

S3 -
«« j

CO

O

a

a

h a

-, CO

0 -Q
3 co
5 03

3 S

1 bC
c^ 03

n a

33 s

H S3
. hC

L. <

33

00

o
'a

X

-

a

03
b£

03

a

£

<

o

CO

03

be
o
a

S3
U
s3

s

03
be

cn

3

"3
o
o

CO

3
3
03

t-C

S3
>

a

CO

.3

a

CO
CO

o3

a

03
S3

5

c3
+j
X
0)
CO

M

CD

a

CO

C3
+3

03

—
3
7i
■_

'3

=

—

/.

03

C3

-;

'o
>

cu

s->

o3

4

3
03
O

'So

3

c
~

b

c

H

73

7i
*J
u

a

3

z

3
if

5
-5

'5
o

CO

cu

-a
c

S3
bC

a
_o

u

o

'o?

a

o

•<— •

03

a

co

3
<-

3
s3

CO

O
-3
t-.

0)

O

PARKER: LIZARDS OF BRITISH SOMALILAND

13

S

*<S>

e

PQ

<

c

3
jQ

"C

■*3

to

T3

oj

■R g

o OJ

C OJ

c

Savannah from Gold Coast to
Uganda and N. Kenya

Zululand, Transvaal, Mozambique
and Tanganyika Territory

03
:-

:;

—

O

43

•—

s
o

OQ

o3

—

<
—

tc
O

<

Uganda, Kenya and Tanganyika
Territory

[East and south of the Rain Forest
to Angola and Cape Province

Coastal plains and upland savannah
southwards to Tang. Territory

c

'3
-^

3
3

'3
be e

OJ —
C 3
CJ-G

X£

•S3 O

oj ^
cc

- t
>*&

( 'oastal plain and upland savannahs
of Kenya, Uganda and Tang. Terr.

East and south of the Rain Forest
to Angola and Natal.

Coastal plain and upland savannahs
of Kenya and Tang. Territory

Jjopny dywj

|

>

>

~>

vqnf %> ipq&qg

e>-

>

>

>

>

>

mqqO

>

>

>

uapvt)()

>

>|>

>

~>

~^

pnvfj

>

>

>

>

>

>

>

fVd\\V\ \b6os[

>

pnvH PS

>

>

~>

■DUi]j.ni6ij\[

>

>

>

'W ff

>

>

>

O0Q /w -s^

I
> 1

>

>

>

>

>

>

>

>

>

uvqrif)

c-

~S»

>

!

C*"»

Diqouy -g

|

>

>

I

j Dzdotyjfl

>

|

~v

>

>

■>

""?■

>

>

e-

VdJLIllQ

>

~>

>

>

uvpng sv

>

■>

•

II — Continued.
Gerrhosaurus m. bottegoi

2

*3

c

c3

bC

G

—

3
tc

tC

—

5
6

"5
3

Ei

_tc

_^
§

5

tO

1)
M

to
-.

y

o
*

"o

c3
X!

c

CJ

c

K

50

"3

x

tc

_r

IT
<

ci

tc

cS
>>

3

-

>

DC
2

s3

s

OJ

3

cr

03

CJ

o>
Xi

"3
>

OJ

"2
3

tc

a

Q

'■$

a

tc
OJ
"3
O

S

S3
■J)

"3
o

s

•J

o3

CJ

"3

S
o

tc

03
*j

tc
_u

o

a

14

bulletin: museum of comparative zoology

PQ

<

'■♦3

3

.O
• **

-3

ts —
s «
S —

'S *

£J

3. -

<

c3

s

r3

o

—

Z

03
>>

Z

fjopny difvj

>

">

vqnf y }pqayg

">

>

c~

>■

>

>

tnqqo

^—

~>

>

U9pDBQ

>

>

>

>

>

pnvfj |

>

>

>

>

>

?

>

>

>

%Z»4 ;o/5oAr 1

>

>

>

>

>

">

">

>

>

pno^/o^j

>

>

>

>

>

>

>

vuipniB-ij,^

>

>

">

>

'w #

>

>

o^o /° 'nw

>

>

J»

>

>

>

uvqnf)

>

mqoj.y -g

[ vidoiyig

-

>

>

V9J.1Ufl

uvpng -g • y

III

Hemidactylus taylori

33

3

■I.

c

•r.

O

—

a;

IE
—

w

CO

—

X

p

3
O

=

X

bC

03
/.
U

o

—

—

c;

JT

Eh

a:

n

o

3-

c
Eh

'C

s

>.

+3

0

:-
i.

c

CO

c

H

m
3
C
o3
_o
'C

-—
S3

X

PARKER: LIZARDS OF BRITISH SOMALILAND

15

•3

£

8

©

Approximate distribution
elsewhere

CO

—

co
0

fiopny dyvq

>

vqnf vp n3<PHS

>

>

>

>

>

>

viqqo

>

>

>

">

>

>

UdpDBQ

p-

c^-

>

>

>

e-

pnvfj |

>l>

>

>

>

~^

>

>

>

1

>

>

>

>

pnvff }og 1

>

>

>

>

>

>

>

vuipniHijx

>

>

>

>

s^-

~^

>

>

■SIR' '3

>

>

>

>

>

>

>

>

060 fo -sift

>

>

>

>

>

>

>

uvq.ifj

>

>

>

>

viqvjy -g

>

>

■ I mdoinig

>

>

>

>

D9.l)Ug

">

>

>

>

uvpng g Y

">

III t~ Continued.
Lygodactylus somalicus

'7.
g.

CO

3
3

CO

3
—

2-i

■r.

3

—

—

u

3

'>

c3

~

CO

3

CO
CO

—

—

c

CJ

"5
aa

5
<

'53

<

co

—

<

3
•—

"2

E7
3

<

CO

—

s
0

3

CO

— 1

'C

—

co

CO

"53

CD

3

<

CO

"53

cp

<

S3

—
<

_5

o3
43

<:

■r.

0"

-s.

'■+3

CO

c3
f=>

z

—

1=5

3

=

0

_3

O

~.
U

<

16

bulletin: museum of comparative zoology

s

o

o

<

a

_o

'■&

3
£i

to

11

■a £

g JO

a, <»
a

o;

//opn# a^»7

>

vqnf sp }pq^i{S

>

>

>

>

>

>

viqqo

>

>

>

>

>

uapvdg

>

>

?—

>

>

*

pnvfj

>

>

>

>

^,

>

^

>

>

>

ItoapA 1dSoN i

i

>

>

>

>

prwH 20s |

>

>

>

VM)J7ll6l}\[

>

'W '#

>

>

>

ofiO /o -s;/^

>

>

>

>

>

>

>

uvqnQ

>

>

>

>

>

>

viqvj,y -g

j vidoiytg

>

>

>

Z)9JfU^

c^-

~^

>

uvpng h ■ V

III — Continued
Eremias smithi

c3

an

Z

a

CO

C
93

+5

.2

-
.3

X

a

oa

'3

CO

Z
Xi

-

CO

c
Xi

'E
'3

Cj
CO

O

*i
(— i

-~

'S

CD

s-
-

3

C3

X

c
—

h4

'E
">>

—

CO

CO

V-

3

■♦3

t-

3

«

-5

-

CO

3

-3

03

£
e

03

G
to

03

E
OJ

c

a

X

3

CO
X

g

—

S

Ih

03

-

.22

CO

C

0/
'—

"o

-a

3

*-<

CO

3

E
03

-*j

X

ft

03

M

if.

o
—

SB

Pm

X

—

03

'_

'>

03

_£!

oa
a

03

03
3
U

"C
'>

PARKER: LIZARDS OF BRITISH SOMALILAND 17

often it is not sufficiently marked for the race to receive taxonomic
recognition.

There can thus be recognised four faunal divisions and the following
is a summary of the composition of their lizard-faunae as far as at
present known.

A. The Guban with 26 species and subspecies of which 2 are confined
to this zone. Of the remainder, 12 are Eremian forms and 11 Somali
endemics; 19 of the forms are also found in the Ogo, 11 penetrate to
the Haud zone and 6 to the Eastern Mountains.

B. The Western Mountains and Ogo which, though no greater in
area than the Guban, has 44 species and subspecies, 4 of which are
not known to occur elsewhere; only 8 are Eremian forms, and 21 are
Somali endemics; 19 of the forms penetrate into the Guban, 15 into
the Eastern Mountains and 20 into the Haud.

C. The Eastern Mountains, the smallest and probably least collected
area, has only 19 recorded species and races, of which none are confined
to the area; three are Eremian forms, and 12 Somali endemics.
Fifteen of the forms are common to the western Mountains, 6 to the
Guban and 15 to the Haud.

D. The Haud, including the Sol Haud, Daror and Nogal Valleys and
the Ogaden, is the largest area and has also the largest fauna. There
are 52 recorded species and subspecies, 8 of which are not known to
occur elsewhere; only 2 of the Eremian species penetrate into this
zone and 36 of the forms are Somali endemics; 26 of the forms occur
also in the Western Mountains and Ogo, 15 in the Eastern Mountains,
but only 12 in the Guban. The fact that only 23 per cent of the lizards
of the Haud are found in the Guban, 60 miles or less distant, indicate
how very effective the mountain and climatic barriers are to the dis-
persal of animals of this group.

The almost complete absence of endemic genera in an area which
otherwise shows such very pronounced endemism supports the geo-
logical evidence that this is a relatively new country; the endemism
itself emphasises the effectiveness of the geographical and climatic
barriers in imposing isolation. The distinctiveness of the fauna of
the Guban points clearly to a very close linkage between the lizard
fauna and the ecological conditions. The fact that 77 per cent of the
species of the Haud, with a mean annual rainfall less than 20 inches,
have been unable to colonise the dry Guban, and, conversely, that
more than half of the forms found in the latter area have failed to
penetrate into the Haud, makes it seem highly improbable that any
of these animals could have tolerated the vastly different, wet condi-

18

bulletin: museum of comparative zoology

tions of Gamblian times. As we have seen, migration into the area from
both the north-west and the south has taken place, but the number

Fig. 3. Distribution of the Subspecies of Latastia longicaudata. Areas over
1500 metres stippled; over 2000 metres hatched.
° Latastia longicaudata longicaudata an Eremian subspecies rang-
ing from Nigeria to Sinai.
• L.l. revoili ranging southwards into Tanganyika Territory;

also reported from Eritrea.
+ L.l. caeruleopunctata
* L.l. doriai
f L.l. andersoni

of colonists which still exist in the countries to the north-west, where
desiccation has progressed even further than in Somaliland, is small.

PARKER: LIZARDS OF BRITISH SOMALILAND

19

If, as seems probable, the process of desiccation has spread from the
north, immigrants from the dry territories will also have spread south-

Fig. 4. Distribution of the Subspecies of Latastia boscai and L. taylori.
Highlands as in Fig. 3.
° L. boscai boscai, a subspecies of the mountains of Ethiopia and
Eritrea.
• L.b. burii
+ L.b. arenicola
* L. taylori

wards, to be replaced in the north by a succession of new immigrants
better able to live under the progressively more arid conditions develop-

20 bulletin: museum of comparative zoology

ing there. This would account for the large numbers of Lacertidae, a
predominantly Palaearctic family, and for such forms as the Euble-
pharid geckoes (Hemitheconyx) and the genus Teratolepis, both of which
have discontinuous Eremian distributions. But there is also a strong
Ethiopian element in the fauna and, once again, the amount of sub-
specific and specific differentiation which exists between the Somali
forms and their relatives in the damper countries to the south empha-
sises the close connection between climatic conditions and faunal
change.

In the following systematic section will be found synoptic keys to
all the species which have been recorded from British Somaliland.
Many of these are modified from tentative keys which Capt. Taylor
utilised in the field, and so have been subjected to some testing. There
are also notes on the taxonomy of some of the species and references to
the original description, and to other names which have been used in
reference to the same species in the Somaliland Peninsula. Notes on
habits and habitats which have been incorporated are all based upon
the excellent field notes provided by Capt. Taylor, and the specimens
listed are those obtained by him during his journey along the southern
and western boundaries. Occasionally, when revisionary work has
embraced a wider field and led to a new and different opinion con-
cerning the status of the forms involved, it has been found necessary
to include keys to, and descriptions of, species which do not occur in
north-east Africa. The bibliography given is probably far from com-
plete, but it is hoped that it contains references to all the more impor-
tant papers dealing with the lizards of British Somaliland.

Needless to say this paper could never have been written without
the generous co-operation of a number of people, especially of Cap-
tain Taylor; acknowledgement of the author's gratitude to him is
tendered herewith; also to Dr Thomas Barbour for having it published
by the Museum of Comparative Zoology, as well as to Mr. Arthur
Loveridge and Dr. A. ,F. Carr, Jr. for seeing it through the press,
Dr. Guiseppe Scortecci, Dr. W. A. Macfadyen, Dr. V. E. Fuchs and
M. F. Angel, who have all rendered help by correspondence and the
loan of valuable material, and to Col. M. Simon for the figure of
a vertebra of Brookcsia superciliaris.

PARKER: LIZARDS OF BRITISH SOMALILAND 21

GECKONIDAE

Hemidactylus*

I. Unregenerated tail depressed, root-shaped, marked off by a basal
constriction, and usually shorter than the head and body.

A. Back with rows of very large, trihedral, strongly-keeled tubercles;
regenerated tail leaf-shaped.

1. 8-9 lamellae beneath the inner, and 10-12 beneath the fourth
toe; 8 preanal, but no femoral pores H. taylori Parker

2. 4-6 lamellae beneath the inner, and 6-7 beneath the fourth

toe; a long series (28-36) of femoral and preanal pores

H. ruspolii Boul.

B. Back with rows of moderately large, smooth or slightly keeled

tubercles ; regenerated tail tapering.

4 lamellae beneath the inner, and 6-7 beneath the fourth toe;

12-18 femoral and preanal pores

H. laticaudatus Andersson

C. Back with uniform or nearly uniform dorsal scales.

1. Dorsal scales granular; 7-9 lamellae beneath the inner, and
12-13 beneath the fourth toe; 5-6 femoral pores on each side;
no preanal pores H. flaviviridis Riippel

2. Dorsal scales sub-imbricate; 5 la\ ellae beneath the inner, and
8-9 beneath the fourth toe; no f ■ loral pores; 4 preanal pores . .
H. curlei sp. nov.

II. Unregenerated tail conical, longer than the head and body, and
without any basal constriction.

A. Back with rows of strongly enlarged and usually more or less
strongly keeled tubercles, often trihedral.
1. Free distal joints of the digits moderately long, the claw ex-
tending well beyond the basal dilatation.

a. Tail strongly depressed, the outermost row of tubercles

forming a sharp ventro-lateral edge proximally

H. barodanus Boul.

b. Tail but slightly depressed, without ventro-lateral crest.

i. A long series of femoro-preanal pores (20-40); 4-6
lamellae beneath the inner, and 7-8 under the fourth
toe H. brookii Gray

ii. A series of preanal (2-12), but no femoral pores.

♦Written prior to the trinomial treatment of certain of these species in Copeia, 1941, pp. 245-
248, which the author, being engaged in war work, has not had the opportunity of seeing.

**

22 bulletin: museum of comparative zoology

Digits strongly dilated basally; 6-9 lamellae under
the inner, and 9-12 under the fourth toe; preanal

pores 6-12; adult size 54-80 mm

H. macropholis Boul.

Digits strongly dilated basally (fig. 6); 5-8 lamel-
lae under the inner, and 8-11 under the fourth toe;

preanal pores 6-8; adult size 40-59 mm

H. turcicus Linn.

Digits very feebly dilated basally (fig. 5); 5-8
lamellae under the inner, and 8-1 1 under the fourth

toe; preanal pores 2-6; adult size 40-59 mm

H. sinaitus Boul.

***

2. Free distal joints of the digits short, the claw scarcely extend-
ing beyond the basal dilatation ; 4-5 lamellae under the inner
and 6-7 under the outer toe; preanal pores 4-7; adult size 26-
36 mm H. citernii Boul.

B. Back with uniform granules, or with small conical or scarcely
keeled tubercles.

1. Back and base of tail with definite enlarged scales or tubercles.

a. Basal ^4 of toes granular like the sole of the foot; tubercles
extending forward on to the occiput and neck ; 5-7 lamellae
under the outer, and 6-8 under the fourth toe; femoro-
preanal pores 30-60 H. mabouia Mor. de Jonn.

b. Subdigital lamellae extending on to the sole of the foot.

i. Dorsal tubercles extending forwards on to the occiput
and neck; 8-9 lamellae beneath the outer, and 9-11
under the fourth toe.

* Femoro-preanal pores 28-32; adult males up to 53
mm. from snout to vent; dorsal tubercles conical. . .
H. smithi Boul.

** Preanal pores 8; no femoral pores; adult males circa

70 mm. ; dorsal tubercles flattened

H. jubensis Boul.

ii. Dorsal tubercles not extending on to the head and neck;
femoral pores 24-28 H. fremitus D. & B.

2. Back with uniformly small granular scales, or with a few very
indistinct, slightly enlarged ones.

PARKER: LIZARDS OF BRITISH SOMALILAND 23

a. Distance from the tip of the snout to the anterior border
of the bony orbit a little longer than the distance between
eye and ear; eye distinctly shorter than its distance from
the nostril; claw extending well beyond the lamellar por-
tion of the digits.

i. Ventrals large, about 7 longitudinal series at mid-body
corresponding to a distance equal to the diameter of
the eye; subdigital lamellae abruptly differentiated from
the granules of the sole H. laevis Boul.

ii. Ventrals small, 12-16 longitudinal series at mid-body,
corresponding to a distance equal to the diameter of
the eye; subdigital lamellae very small proximally,
scarcely larger than the granules on the sole, 5-6 be-
neath the inner, and 9-11 beneath the fourth toe; no
femoral or preanal pores H. somaHcus Parker

b. Distance from the tip of the snout to the anterior border of
the bony orbit as long as the distance between eye and ear;
eye as long as its distance from the nostril; claw only just
extending beyond the basal lamellae, of which there are 5
under the inner, and 8-9 beneath the fourth toe; preanal
pores 2 H. megalops Parker

1. Hemidactylus taylori Parker

H. taylori Parker, 1932, Proc. Zool. Soc. London, p. 342.

This species has only been found in the Sol Haud, 9°10/N x 49°E
and 9°35'Nx49°E; it is, apparently terrestrial, and probably noc-
turnal.

2. Hemidactylus ruspolii Boulenger

H. ruspolii Boulenger, 1896, Ann. Mus. Civ. Genova, (2), 4, p. 3, pi. i, fig. 1.

4 c? cf, 4 9 9 , juv. Haud, 2100 ft. 46°20'E x 8°15'N

2 9 9 Haud, 3050 ft. 45°09'E x 8°37'N

c? Haud, 3100 ft. 45°04'E x 8°39'N

4 & d" , 4 9 9 Ado, 2100 ft., Ogaden.

This very conspicuously marked black and yellow gecko appears to
be confined to the eastern parts of the Haud in British Somaliland
(east of the 45th meridian), but extends southwards through the
Ogaden and Italian Somaliland to northern Kenya Colony and the
Lake Rudolf basin. It is chiefly arboreal, being usually found beneath

24 bulletin: museum of comparative zoology

the bark of dead trees, though Scortecci (1931, p. 127) records one as
having been collected in the walls of an old building.

3. Hemidactylus laticaudatus Andersson

H. laticaudatus Andersson, 1910, Jahrb. Nassau. Ver. Nat., 63, p. 200.

Known from a single specimen collected at Harrar, this gecko may
perhaps be found in the mountains of the Ogo.

4. Hemidactylus flaviviridis Ruppel

H . flaviviridis Ruppel, 1835, N. Wirbelthiere Fauna Abess., p. 18, PL vi, fig. 2.

Though originally described from Massaua, this is in reality an
oriental gecko, which appears to have been introduced along the Red
Sea littoral in the neighbourhood of the ports.

5. Hemidactylus curlei spec. nov.

Holotype a 9 , number 1937.12.5.295 in the British Museum, from
the Borama District (43°10'E. x 9°55'N), 5000 feet, collected among
stones and rocks on Dec. 2nd, 1932.

Habitus very depressed; head broad and flat, its maximum width
much greater than the distance between the tip of the snout and the
posterior border of the eye. Nostril pierced between the rostral, first
upper labial and 3 nasals, of which the upper is separated from its
fellow by a single scale. Rostral much broader than deep, with a me-
dian cleft. Snout flat, covered with rounded juxtaposed granules,
which are much larger than those on the very flat occiput ; eight upper,
and six or seven lower labials; mental triangular, nearly as long as
broad; median chin-shields very large, forming a long median suture;
a much smaller second pair of chin shields is followed by some enlarged
scales bordering the lower labials. Ear opening small, oblique, its dis-
tance from the eye equal to the distance between nostril and eye.
Body depressed, with a slight, median dorsal furrow, covered above
with somewhat irregular, flat, subcircular, sub-imbricate scales and
below with cycloid imbricate scales which are nearly twice as large as
the dorsals ; the latter are not absolutely uniform in size, but there is no
trace of any definitely enlarged scales or tubercles; about 87 dorsals
and ventrals in a series round the middle of the body. Tail strongly
depressed, with a median furrow, root-shaped and with a distinct basal
constriction; it is covered above with imbricating scales arranged in

PARKER: LIZARDS OF BRITISH SOMALILAND 25

transverse rows and is indefinitely annulate, 5 rows of dorsal scales
forming an annulus; beneath there is a series of transversely enlarged
subcaudals of which two occur on each annulus; the tip of the tail is
regenerated, and this portion is not annulate, but covered above with
quincuncially arranged imbricate scales and has transversely enlarged
subcaudals. Limbs short, the adpressed hindlimb reaching the wrist.
Digits well dilated basally and with moderately long terminal pha-
langes which extend well beyond the lamellar portion ; lamellae extend-
ing on to the palms and soles, 5 under the inner, 8 under the fourth, and
7 or 8 beneath the fifth toe.

Yellowish above, heavily blotched and spotted with purplish black.
A dark streak from the nostril through the eye and above the ear to
the sides of the neck from whence it is continued as a row of spots to
the base of the tail; flanks with an indefinite row of spots, and mid-
dorsal area with series of rather irregular transverse blotches ; tail with
alternately darker and lighter cross-bars; limbs and flanks with scat-
tered, circular, white spots. Lower surface white.

Length from snout to vent 43 mm.

Tail (posterior third regenerated) 42 mm.

Fore-limb 13 mm.

Hind-limb 14 mm.
Paratypes a male and 3 females from the type locality, collected
26.XI.32, a male collected between the type locality and Hargeisa
(4100 ft., 44°E x 9°35'N) in October 1932, and a male and female
from 43°E x 10°05'N, 16.IX.33.

These specimens were all collected amongst stones and rocks and
agree in essentials with the holotype. The number of upper labials
varies from 8 to 10, of lower labials from 6 to 8, of scales round the
body from 81 to 93, and of subdigital lamellae beneath the fourth and
fifth toes from 8 to 9 and 7 to 9 respectively. Males have 4 preanal
pores. The colour is similar to that of the type, but the markings are
usually quite irregular. No specimen has a completely unregenerated
tail, but the fully regenerated appendage is similar in shape to the un-
regenerated, but without annuli and without transverse scale-rows.

The species resembles the Sokotran H. homoeolepis Boulenger in its
flat, subimbricate scales, a character in which it approaches the genus
Teratolepis. But it differs from both homoeolepis and its ally laevis
in its much more depressed habitus and the root-shaped tail. No
species with this type of caudal appendage has such uniform flat scales
or so few subdigital lamellae; the species which it approaches most
nearly is H. zolii Scortecci.

26 bulletin: museum of comparative zoology

6. Hemidactylus barodanus Boulenger

H. barodanus Boulenger, 1901, Proc. Zool. Soc. London, p. 48, pi. vii, fig. 2.

4cfd\ 7 9 9 &2 juveniles from the Borama district (9°55'-10° 20' N x
42°25'-43°15'E) at altitudes of from 4000-6000 feet.

Previously this gecko was known from but a single specimen col-
lected at Gan Lebar ( = Gaan Libah) in the mountains S.S.W. of
Berbera; in addition to the foregoing series from the mountains of the
Ogo it has now also been found in the mountains of western Ethiopia.

The Ogo series shows a range of variation in the size and strength of
keel of the dorsal tubercles which is quite comparable with that noticed
in H. turcicus (q.v.). In some individuals the tubercles are large and
trihedral, as in H. macropholis but in others they are less than half this
size and only feebly keeled. These latter agree with the unique type of
barodanus. Hitherto the shape of the tail has not been definitely known,
but owing to the fact that the basal portion of it, which is the only part
existing in the type, is strongly depressed, the author (1932, p. 346)
placed the species amongst those in which this organ is root-shaped
and constricted basally. This is not the case, however, for the com-
plete tail has no such constriction and is slightly longer than the head
and body. It is strongly depressed, its depth being about half its
breadth, and the outermost row of tubercles forms a sharp ventro-
lateral edge in the proximal part. This character distinguishes the
species readily from both turcicus and macropholis which it resembles
in the number of subdigital lamellae (6-8 under the inner, 9-11 under
the fourth and 10-12 under the fifth toe), femoral pores (6-11) and
tubercles; in both the tail is nearly circular in section and the outer-
most row of tubercles is not, or scarcely, ventral to the lateral mid-
line.

The largest male and female each measures 70 mm. from snout to
vent.

All the specimens were collected in, or under, stones and rocks.

7. Hemidactylus brookii Gray

H. brookii Gray, 1845, Cat. Lizards Brit. Mus., p. 153.

It seems peculiar that this common house gecko, wide-spread
through the Sudanese and northern Savannah provinces of Africa and
in the Indo-Malayan region should not have been discovered in British
Somaliland. It has been reported from the Harrar region of Ethiopia
(Jaldessa) and from Italian Somaliland (Lugh).

PARKER: LIZARDS OF BRITISH SOMALILAND 27

8. Hemidactylus macropholis Boulenger

H. macropholis Boulenger, 1896, Ann. Mus. Genova, (2), 17, p. 3, pi. i, fig. 2.

9 cf d\ 6 9 9,5 juvs. Haud, 2100-4000 ft., between 44°15' — 46°20'E

x 8°15' — 8°55'N.
juv. cT Ado, 2100 ft., Ogaden.

Originally described from Dolo, Italian Somaliland, this species
has since been recorded from other localities in that colony, from
northern Kenya (Merelle River) and from the eastern districts of
British Somaliland, east of the 44th meridian. It may ultimately
prove to be a geographic race of turcicus which, in British Somaliland
appears to be confined to the coastal zone of the north-west and the
adjacent mountains of the Ogo. H. turcicus has, however, been re-
corded from Italian Somaliland (Scortecci, 1929, 1931; Calabresi,
1927) within the known range of macropholis and further collecting
alone can decide whether or not the two intergrade. All the specimens
were taken in dead trees or in termites' nests.

9. Hemidactylus species

9 , juv. Guban, 1500 ft. 43°E x 11°N.

These two specimens represent a species similar to macropholis, but
differing in the possession of an even larger number of subdigital lame-
lae (10 under the inner, and 14 beneath the fourth toe), in having
larger, more numerous dorsal tubercles, smaller ventrals, and a very
distinctive colour pattern of broad, dark, transverse bars. In many
characters, and especially in their colour, they resemble the Indian
H. triedrus and H. subtriedrus, but do not appear to be referable to
any described species. In the absence of a male, however, it is not
possible to give an adequate description or to assess fully the relation-
ships of the animal; so that, until further specimens are obtained, it
must remain nameless. It appears to frequent dry, stony ground in a
region completely devoid of vegetation.

10. Hemidactylus turcicus (Linn.)

Lacerta turcica Linn., 1758, Syst. Nat. (10), p. 202.

t Hemidactylus puccionii Calabresi, 1927, Atti Soc. Ital. Sci. Nat., 76, p. 23,
pi. i, fig. 3.

cT, Berbera

11 d" d\ 8 9 9 42°50' — 43°15'E, 9°55' — 11°25'N, 150-4500 ft.

2 cf cT, 4 9 9 Island off the coast at Zeilah.

28 bulletin: museum of comparative zoology

The species was not found along the southern and eastern boundaries
of British Somaliland, but appears to be confined to the north-western
corner of the territory and to the coastal zone. With the exception of
the specimen taken at Berbera, which was found in a bungalow, all the
specimens were found among rocks or stones or hiding under dead logs.

The series shows a great deal of variation in the size of the dorsal
tubercles and the extent to which they are keeled. In a series of 8
specimens from 43°15' x 11°25'N, there is every variation between
examples with relatively large, regularly arranged, strongly keeled
tubercles, and one specimen in which the tubercles are only about half
as large and are but feebly keeled; among 7 others from 42°50'E x
10°10'N most show the latter condition and two have the tubercles so
small as to be scarcely perceptible and with barely a trace of a keel to
be discerned. These recall, very strongly, the described condition of
Hemidaetylus puccionii Calabresi, from Italian Somaliland, and it is
probable that this name is based on a small, smooth-tuberculate speci-
men of H. turcicus. No such individuals occur in the northern (Medi-
terranean) part of the range of turcicus, but in Somaliland they appear
to form a considerable percentage of the total population and their
occurrence may be taken to indicate the beginnings of racial differen-
tiation, though this does not appear to be sufficiently well-marked to
justify the use of trinomials.

11. Hemidactylus sinaitus Boulenger

H. sinaitus Boulenger, 1885, Cat. Lizards Brit. Mus., 1, p. 126.

c? 43°15'E x 10°30'X, 3000 ft. 18.iv.33.
juv. c? 43°15'E x 11°25'X, 150 ft. 9.i.34.

Anderson (1898, p. 83) showed that the characters originally used to
distinguish H. sinaitus from H. turcicus were untenable when a large
series was examined. He accordingly considered H. sinaitus to be a
"variety" of turcicus and more recent authors have used the name
sinaitus in a subspecific sense. But this practice does not seem to
reflect the true relationships between the two, for the geographical
range of sinaitus falls within that of turcicus. The above mentioned two
specimens differ considerably from a series of turcicus (q.v.) found in
the same region at the same time, notably in the degree of dilatation
of the basal portion of the digits, and the length of the claw. They
agree well with the type of sinaitus and, unless it can be shown that
the digits of turcicus may vary enormously, it seems more reasonable
to regard the two as distinct species. The accompanying figures, (5

PARKER: LIZARDS OF BRITISH SOMALILAND

29

and 6) one of the type of sinaitus and the other of a specimen of turcicus
of comparable size, but selected on account of the unusually small

Fig. 5. Hemidactylus sinaitus. Left foot of Holotype.

Fig. 6.

Hemidactylus turcicus.
with the above.

Left foot of a female of a size comparable

number of its lamellae, reveal the differences better than any descrip-
tion. There are also certain other minor differences, thus:

H. turcicus

a. Digits strongly dilated basally

b. Claw short and stout

c. Lamellae on the hallux 6-8,
usually 7.

d. Supranasals usually separated

e. Scales on the snout small

f. Preanal pores 6-8

a.

b.
c.

e.

f.

H. sinaitus
Digital dilatations small
Claw long and slender
Lamellae on the hallux 5-8,
usually 5.

Supranasals in contact
Scales on the snout larger
Preanal pores 2-6

Both the present examples were found among stones.

30 bulletin: museum of comparative zoology

12. Hemidactylus citernii Boulenger

H. citernii Boulenger 1912, Ann. Mus. Genova, (3), 6, p. 329.

d\ 2 9 9 43°15'E x 11°25'N, 150 ft.

The discovery of this species in the extreme north-west of British
Somaliland is surprising. Hitherto it has only been reported from the
Upper Juba province of Italian Somaliland (Rahanuin Country) and
the eastern boundary of the British territory. The north-western
specimens are, however, quite typical, and it may be that the species
has a continuous range around the coastal zone ; it must be either very
rare in, or completely absent from, the Haud and the mountains of
the Ogo.

13. Hemidactylus mabouia (Mor. de Jonnes)

Gecko mabouia Moreau de Jonnes, 1818, Bull. Soc. Philom., p. 138.

This cosmopolitan gecko, common throughout the Savannah pro-
vinces of Africa, is apparently rare in British Somaliland. It has been
recorded at "Lasgore" ( = Las Khoreh) by Vaillant, and Scortecci
(1931) mentions two individuals collected close to the Anglo-Italian
boundary at El Donfar and Gardo; strangely enough it was not found
by Capt. Taylor in this region. It seems very probable that humidity
may be the limiting factor in the distribution of the species, its place
being taken in the dry areas of the Eremian region and Somaliland by
H. turcicus.

14. Hemidactylus smithi Boulenger

H. smithi Boulenger, 1895, Proc. Zool. Soc. London, p. 532, pi. xxix, fig. 2.

4 d1 d1, 6 9 9 Haud, 2100-2800 ft., 45°24'-46°25'E x 8°15'-8°32'N.

An arboreal gecko of the eastern and central Haud, Nogal Valley
and Buran districts extending southwards to the Webi Shebeli (type
locality). The largest specimen of the present series, a female, measures
57 mm. from snout to vent.

15. Hemidactylus jubensis Boulenger

H.jubensis Boulenger, 1895, Ann. Mus. Genova, (2), 15, p. 10, pi. iii, fig. 1.

The record of this species from the Golis Mountains (Boulenger
1895b) appears to be based on a small-tubercled H. turcicus similar to
those mentioned above. True jubensis is known only from the Ganale
Doria, Milmil and Biornal, but since so many species with a similar
distribution range into the Haud, its presence there may be expected.

PARKER: LIZARDS OF BRITISH SOMALILAND 31

16. Hemidactylus frenatus Dum. & Bib.

H. frenatus Dumcril & Bibron, 1836, Erpet. Gen., 3, p. 366.

An oriental species which appears to have been introduced into the
south of Italian Somaliland (Lower Shebeli province); it may also be
found in British territory.

17. Hemidactylus laevis Boulenger

H. laevis Boulenger, 1901, Proc. Zool. London, p. 48, pi. vii, fig. 1.

Originally found at "Ganlebar" (=Gaan Libah) in the Golis moun-
tains this species has not been rediscovered in the British territories
but has been reported from the south of Italian Somaliland (Dolo and
the Rahanuin Country, Boulenger, 1912). If these latter two records
are correct the species must have a considerable range and may be
expected in very different types of country. But a mis-identification
is to be suspected, and it seems highly probable that there has been a
confusion with H. fragilis Calabresi, a closely similar species described
from Bur Melde in the same general region.

18. Hemidactylus somalicus Parker

H. somalicus Parker, 1932, Proc. Zool. Soc. London, p. 344.

A species closely allied to the preceding but known only from the
Sol Haud, Nogal Valley and eastern Haud. Nocturnal and terrestrial.

19. Hemidactylus megalops Parker
H. megalops Parker, 1932, Proc. Zool. Soc. London, p. 345.

Another little known species of the Sol Haud and Nogal Valley.

Teratolepis Gunther 1869

Teratolepis Gunther, 1869, Proc. Zool. Soc. London, p. 504 (type species

Homonota fasciata Blyth).
Bunocnemis Gunther, 1894, Proc. Zool. Soc. London, p. 85 (type species

Bunocnemis modestus Gunther).
Lophopholis Smith & Deraniyagala, 1934, Ceylon Journ. Sci., B, 18, p. 235

(type species Teratolepis scabriceps Annandale).

In north-east Africa there are known a few geckoes usually referred
to the genus Hemidactylus, in which the dorsal scales are strongly im-

32 bulletin: museum of comparative zoology

bricate. In this character they differ from all other members of the
genus, and, in addition, they all have rather feebly dilated digits,
though in this respect there is complete intergradation with the con-
dition of normal Hemidactylus, through species such as fragilis Cala-
bresi, somalicus Parker and me galops Parker. But there are also in the
same zoo-geographical region two other geckoes in which the scales
are imbricate and in which the digital lamellae are smaller still, and are,
for the most part, undivided mesially; these have been referred to a
distinct genus, Bunocnemis. But a new species in the present collec-
tion, described below as taylori, completely bridges the gap between
Bunocnemis and the imbricate-scaled species of Hemidactylus. It is
very closely allied to the type species of Bunocnemis, even to possess-
ing enlarged scales on the back of the legs, but the digital lamellae are
almost all divided and thus exhibit the same condition as is found in
species like ophiolepis and isolepis. It might accordingly be considered
advisable to unite Bunocnemis with Hemidactylus, or alternatively, to
refer the imbricate-scaled species of Hemidactylus to Bunocnemis.
But reference to the geckoes in other .areas reveal the fact that similar
species have received different generic names. Thus the Oriental
Teratolcpis differs from Hemidactylus in its imbricate scales, and un-
divided subdigital lamellae, and so is exactly comparable, and cannot
be distinguished from a Bunocnemis whilst Lophopholis is simply an
Oriental imbricate-scaled Hemidactylus. It thus becomes necessary
to reconsider the status of these different genera. At first sight it
might seem advisable to merge them all into Hemidactylus, for there
is nowhere any hard and fast line whereby any one or all of them can
be easily cut off. But the combination of imbricate scales and some-
what less highly developed digital pads suggests that we are, in fact,
dealing with a natural group of primitive species, and it may perhaps
be significant that the distribution of these primitive forms — West
Africa, Somaliland, India — is somewhat similar to that of the primi-
tive "Eublepharid" geckoes of the Old World.

The oldest group-name available for these imbricate-scaled species
is Teratolepis and the various species may be distinguished thus:

I. Tail root-shaped; dorsal scales obtusely keeled; digital lamellae
undivided; preanal pores 6 T. fasciata Blyth

II. Tail conical, tapering.

A. Dorsal scales quite smooth.

1. Hinder side of thighs with enlarged tubercles; scales about the
middle of the body 76-83 in d* cf , 96-102 in 9 9 .

PARKER: LIZARDS OF BRITISH SOMALILAND 33

a. Occiput and nape with numerous enlarged, flat scales, two
or three times as large as the granules between them l ; sub-
digital lamellae undivided T. matschei Tornier

b. Occipital scales small, homogeneous.

f Colouration uniform brown; subdigital lamellae for the
most part undivided T. modesta Giinther

ft A marked colour pattern of vertebral light spots and a
labial stripe; subdigital lamellae almost all completely

divided T. taylori sp. nov.

2. Hinder side of thighs without enlarged tubercles.

a. Scales about the middle of the body 50-58; preanal pores
8-22 T. ophiolepis Boulenger

b. Scales about the middle of the body 62-65 in cf cf, 67-77
in 9 9 ; preanal pores 6-8 T. isolcpis Boulenger

B. Dorsal scales keeled or striated; subdigital lamellae divided.

1. Dorsal scales uniform, striated; scales at midbody 55 in cT,
55-62 in 9 9 ; preanal pores 6 T. scabriceps Annandale

2. Dorsal scales heterogeneous, strongly keeled.

a. Enlarged scales feebly keeled, widely separated from one
another; preanal pores 13-19 T. squamulatus Tornier

b. Enlarged scales strongly keeled, larger than the inter-
spaces between themselves; preanal pores 6-10

T. tropidolcpis Mocquard

The only species of the genus found in British Somaliland are as
follows:

1. Teratolepis taylori spec. nov.

Holotype a cf , number 1937.12.5.305, in the British Museum, col-
lected in the Haud (2100 ft., 46°20'E x 8°15'N) by Capt. R. H. R.
Taylor on Jan. 26th, 1932.

Head once and a half as long as broad, somewhat depressed; snout
longer than the distance from the eye to the ear, which is small and
subcircular. Scales of the snout juxtaposed, large and polygonal;
rostral quadrangular, with a median cleft; nostril pierced between
the rostral, first upper labial and three nasals, of which the uppermost
forms a suture with its fellow; seven or eight upper and six lower labials;
mental large, pentagonal, followed by two large pairs of chin shields,
of which the anterior pair form a long median suture; remainder of the

1 No typical material has been examined, but the character is found in a specimen from
Zaria Province, N. Nigeria.

34 bulletin: museum of comparative zoology

lower labials bordered by some enlarged scales. Scales of occiput and
nape homogeneous, granular; dorsals smooth, a little larger than the
ventrals; about 78 scales around the middle of the body. Limbs short,
overlapping when adpressed. Digits short, very slightly dilated proxi-
mally and with divided lamellae, which grade insensibly into the granu-
lar scales of the palms and soles. The numbers under the digits ' are :
fingers 5, 6, 7, 7, 7, and toes 6, 7, 8, 8, 6, counting from the first to the
fifth respectively. Distal phalanges of the fingers very short, but
those of the toes very much longer and armed with very long, slender
claws. Posterior surfaces of femora and tibiae with some enlarged,
irregularly arranged tubercles. Tail conical, with smooth scales similar
to those of the body. An angular series of 15 preanal pores and a single
conical tubercle on each side of the base of the tail; post-anal bones
and sacs present.

Brownish grey above; a light line along the upper lip continuing to
the region above the arm. This light stripe is bordered above by a
blackish stripe from the nostril, through the eye and above the ear,
and beneath by an interrupted black line along the lower labials and
backward as a series of small, widely spaced dots to the arm. A short
whitish stripe from the eye along the temple; a V-shaped transverse
bar on the occiput, followed by a series of six distinct, round or trans-
versely oval, white spots along the vertebral line; tail with a median
series of transversely oval spots anteriorly which give place to two al-
ternating dorso-lateral series behind. Hinder side of the hind-limbs
with circular white spots. Lower surfaces uniform white.
Length from snout to vent 46 mm.
Tail (reproduced terminally) 42 mm.
Fore-limb 11 mm.

Hind-limb 14 mm.

Para types :

A male with the same data as the holotype.

A female collected in the Haud (2700 ft., 45°29'E x 8°30'N) 4. VI.32.

A male collected in the Haud (3300 ft., 44°54'E x 8°42'N) 23. VII.32.

A female collected in the Haud (3800 ft., 44°24'E x 8°52'N) 1 1.IX.32
This short series shows some variation though the general features of
pholidosis and proportions are similar. The number of scales about
the body varies from 76 to 80 in males and from 96 to 102 in females.
The relative sizes of the chin-shields and their shape varies, but the
anterior pair always form a long median suture; upper labials vary from

•These figures can only be given as approximations, since the proximal lamellae cannot be
determined.

PARKER: LIZARDS OF BRITISH SOMALILAND 35

7 to 8 and lower labials from 6 to 7; femoral pores may be from 12 to
19; subdigital lamellae on the fingers and toes: Fingers I 4-5, II 6-7,
III 7-8, IV 7, V 6-7; toes 1 5-6, II 6-7, III 6-8, IV 6-8, V 5-6. The colour
pattern is always similar to that of the type, the white labial stripe
and its dark upper border being very constant. The temporal white
mark is often indistinct and the vertebral white spots are usually trans-
versely dilated and vary in number from 5 to 7. The dorsum may also
have indefinite darker marks which may form a vertebral line between
the white spots.

The species inhabits sandy country with scrub and patches of grass.
All, except one of the specimens, were found on the ground in grass,
the exception being discovered beneath a stone.

The nearest relative of this species is undoubtedly T. modestus from
which it may be distinguished by the smaller size of the post-femoral
tubercles, the divided sub-digital lamellae and the very distinctive
colour-pattern. T. modestus is classed by Loveridge (1937, p. 492)
as an inhabitant of the "wet and humid coastal plain", where it is
found in piles of rubbish and rotting vegetation.

2. Teratolepis ophiolepis Boulenger

Hemidactylus ophiolepis Boulenger, 1903, Ann. Mag. Nat. Hist. (7), 11, p. 55.

<?, 9 Haud, 2900 ft. 45°15'E x 8°34'N

juv. cf Borama District, 4500 ft. 43°10'E x 9°55'N

These three specimens agree closely with the type and hitherto only
known example of this species, except in the number of femoral pores.
In the type there are 8 only, but in both the above-mentioned males
there are 22. This is a greater range of variation than would have been
expected, but, in the absence of any other distinguishing feature cannot
be taken to indicate specific difference.

It appears to be a montane form, ranging from the Hawash River
to the Mountains of the Ogo and the western Haud ; the two examples
from this latter region were found in a termite nest.

3. Teratolepis isolepis Boulenger

? Hemidactylus homoeolepis (non Boulenger) Boettger, 1893, Zool. Anz. 426-417'

p. 114.
Hemidactylus isolepis Boulenger, 1895, Proc. Zool. Soc. London, p. 531, pi. xxix

fig. 1.

cT1 Borama District, 5000 ft., 43°10'E x 9°55'N
<?, 9 Borama District, 5300 ft., 43°15'E x 9°50'N

36 bulletin: museum of comparative zoology

These examples show little variation beyond that recorded for
the species in the Lake Rudolf region (Parker, 1932a, p. 223). In the
two males the number of scales about the middle of the body is 59
or 60 and in the female about 81, rather irregularly disposed; the fe-
moral pores are six and the number of subdigital lamellae slightly
smaller.

In British Somaliland the species appears to be a montane form,
only having been reported at Gaan Libah (5900 ft.) in addition to the
foregoing records from the Borama District. It has also been recorded
from the mountainous Harar region, but in the south it extends into
the lowlands at Lugh, around Lake Rudolf and at Archer's Post
(Kenya Colony).

4. Teratolepis tropidolepis Mocquard

Hemidactylus tropidolepis Mocquard, 1888, Mem. Cent. Soc. Philom. Paris,
p. 113.

6 cf d\3 9 9 Haud, 2200-2500 ft., 45 °50'-46°9'E x 8o00'-8°21'N
2 & a* Ado, 2100 ft., 45°15'E x 7°20'N

These 11 specimens agree with the series previously collected by
Capt. Taylor in the Nogal Valley (Parker, 1932, p. 342) except that the
differences in pholidosis noted there are found not to be correlated with
sex, as was then supposed, and in a slightly greater range of femoral
pores, 6 to 10. Probably the value of femoral pores in distinguishing
the species of this genus and of Hemidactylus has been over-rated
(c.f. Teratolepis ophiolcpis, supra) so that the view previously expressed,
that squamulatus Tornier may be distinguished from tropidolepis by
the greater number of pores, has less to support it. The very strong
keeling of the dorsal scales, their size and linear arrangement are, how-
ever, quite characteristic of tropidolepis, though the geographical dis-
tribution of the two suggests that Calabresi's (1924) interpretation of
them as subspecies may prove to be a correct one. T. tropidolepis is
known from the Haud and Nogal Valley in British Somaliland, ranging
southwards through the Ogaden to Bardera, Dolo, the lower Webi
Shebeli and northern Kenya; it appears to frequent sandy country
with patches of grass and thorn scrub and is probably nocturnal, being
found by day concealed under stones or logs. T. squamulatus has been
recorded from the Anglo-Egyptian Sudan, Uganda, Kenya and Tan-
ganyika Territory; Loveridge (1937) classes it as a form of the coastal
plains and upland savannahs.

PARKER: LIZARDS OF BRITISH SOMALILAND 37

Hemitheconyx Stejneger

This genus contains but two known species, one confined to the
Sudanese districts west of Nigeria, and the other to the Somali Pen-
insula.

Hemitheconyx taylori Parker

H. taylori Parker, 1930, Ann. Mag. Nat. Hist. (10), 6, p. 603; Scortecci, 1931>
Atti Soc. Ital. Sci. Nat., 70, pi. iii.

2 tf> <?, 9 Oadwenia, 3500 ft., 45°05'E x 9C24'N
9 Haud, 2100 ft., 46°19'E x 8°14'N

<? Hargeisa, 4100 ft., 44°E x 9°35'N

This species seems to be almost confined to British Somaliland. It
has been recorded in addition to the above-mentioned specimens from
the Nogal Valley, Buran district and Sol Haud, in the British Protec-
torate and from Gardo in Italian Somaliland close to the border. It
is probably extensively distributed in the northern parts of the Ogaden
also, but owing to its nocturnal habits has not been collected. During
the day time it is usually found concealed beneath stones and when
disturbed emits a coughing noise; the three examples taken at Oadwe-
nia in January were found together 3 feet underground where they were
presumably aestivating through the dry season.

Holodactylus

I. Upper surface of the head flat; 13-18 upper labials

H. cor nii Scortecci

II. Upper surface of the head convex; 9-11 upper labials

H. africanus Boettger

1. Holodactylus corxii Scortecci
H. cornii Scortecci, 1931, Atti Soc. Ital. Sci. Nat., 70, p. 137. fig.

The status of this species is somewhat doubtful (c. f. H. africanus
below) ; it is known from the vicinity of Obbia and near Gardo in
Italian Somaliland and from the Xogal Valley in the British Pro-
tectorate.

2. Holodactylus africanus Boettger

H. africanus Boettger, 1893, Zool. Anz., 16, p. 114.

H. aculeatus Calabresi, 1915, Mon. Zool. Ital., 26, p. 238, fig.

4 cf c?,8 9 9 Haud, 2100-3800 ft., 44°24'-46°20'E x 8°15'-8°52'N
& Guban, 3000 ft., 42°35'E x 10°40'N

2 juvs. Burao, 3500 ft., 45°33'E x 9°31'N

38 bulletin: museum of comparative zoology

Scortecci (1931, p. 134) has drawn attention to the fact that, in a
series of Holodactylus africanus from Italian Somaliland, one example
from the Lower Shebeli Province (Villagio Duca Abruzzi), in the ex-
treme south, has the dorsal granules perfectly smooth, whereas in
others from near Obbia and Gardo (i.e. further north) the granules
are covered with confluent rugosities. In the series available to the
present author it is also noticed that all the specimens from the Haud
and Nogal Valley have the dorsal granules beset with rugosities which
often culminate in a distinct median keel, whereas in another example
from the Rahanuin country (approximately 3°30'N x 43°30'E) the
dorsal granules are quite smooth. But, in addition, eight examples
from the extreme north (Guban and mountains of the Ogo) resemble
the southern form in having the granules quite smooth. Thus, once
again, there is apparently the beginning of racial differentiation, with
the mountain chain south of the Guban forming a boundary line.

The series also shows variation in other respects, though these differ-
ences cannot be correlated with geography. The supranasal shields
are by no means constant. In the present collection those from the
western Haud show every condition between large, well marked supra-
nasals and their complete absence, and among those from the Guban
only one specimen has them developed. The length of the tail, too,
shows considerable variation and this cannot, as has been suggested
(Parker, 1932, p. 350) be correlated with seasonal change, but appears
rather to be an age character. In two juveniles, measuring 42 and 47
mm. from snout to vent, it is 38 per cent and 36 per cent of the length of
the head and body, and there is an almost regular decrease with in-
creasing size to 29 per cent and 30 per cent in the two largest individ-
uals which measure 83 and 85 mm. respectively from snout to vent.
The length of the tail relative to its breadth varies from 2.8 to 4.0 in
the nine specimens collected in the Haud during April, and this range
of variation is sufficiently great to include all variations at other times
of the year, except in very young specimens where, on account of the
proportionally greater length of the organ, the length-breadth ratio
may be as high as 5.0 to 5.3.

The fact that the supranasals are so inconstant in their degree of
development and the tail so variable in its size indicates the possibility
of H. comii no longer being tenable. But no examples with the flat
head ascribed to comii have been seen which can be definitely linked
with africanus.

The species ranges from the sea in the north and east to the Juba
River and westwards to about 43°30'E (type locality) in the central

PARKER: LIZARDS OF BRITISH SOMALILAND 39

Ogaden. It is nocturnal and was found by Capt. Taylor usually in the
vicinity of termites' nests feeding on the winged insects as they emerge.

Lygodactylus

I. A single series of enlarged plates beneath the unregenerated tail . .
L. picturatus

II. A double series of enlarged plates beneath the tail .... X. somalicus

Lygodactylus picturatus

Hemidactylus picturatus Peters, 1870, Mon. Ak. Berlin, p. 115.

Lygodactylus picturatus, or one of its races, has been recorded on
several occasions from Somaliland (Scortecci, 1931, idem, 1930; Bou-
lenger, 1891; idem, 1895; idem, 1896; Mocquard, 1888; Calabresi, 1915;
idem, 1927) but the majority of these records are from the south of the
Italian colony and none are definitely within the British area. It is
almost certain that the most northerly record (Boulenger, 1891) is in
reality based on an example of the following species.

Lygodactylus somalicus Loveridge

L. somalicus Loveridge, 1935, Proc. Biol. Soc. Washington, 48, p. 195.
L. somalicus annectens Loveridge, loc. cit., p. 197.

7 tf d, 5 9 9 Haud, 2100-3900 ft., 44°20'E-46°25'E x 8°15'-8°52'N

Until 1935 this species had been recorded from Somaliland as
L. capensis, L. scheffleri, L. conradti or L. picturatus (Boulenger, 1891).
But Loveridge then showed that the Lygodactylus of Somaliland is
distinct. At the same time he recognised two races, one, the typical,
ranging from the Nogal Valley to Bera in Italian Somaliland and the
other, annectens, from the Buran District to the Juba River, i.e. from
localities to the north, east and south of the typical form. The two
were said to be distinguished by the condition of the mental, fissured
only in annectens, but completely divided in the other. The present
series, from localities to the west of the type locality, has the mental
deeply fissured, and re-examination of five paratypes of the typical
race shows that in three of these also the suture is incomplete, though
a fold, probably due to the manner of preservation, connects the ends
of the fissures and simulates a complete suture. Accordingly the valid-
ity of the two races must be considered very doubtful indeed; in some
individuals the two fissures may unite to sever the mental completely,
but normally this is not the case.

40 bulletin: museum of comparative zoology

This species does not appear to extend westwards or northwards
over the mountain ranges which border the Haud, and ranges south-
wards to the Juba River. In British Somaliland it has been recorded
across the Haud from 44° 20' E to the Nogal Valley northwards to the
Buran District and Burao. It is essentially diurnal and arboreal.

Platypholis fasciata erlangeri (Steindachner)

Homopholis erlangeri Steindachner, 1906, Ann. Hofmus. Wien, 21, p. 149,
pi. vii.

Until recently (1932) erlangeri, originally described from specimens
taken near "Umudu" (=?Afmadu) close to the Juba River, was con-
sidered a synonym of fasciata and all Somali records of the former
appear under the latter name. Typical fasciata is a southern form,
found in Kenya and Tanganyika Territory, whereas the specimens
from British Somaliland examined by the author agree in colour with
erlangeri. Where the two intergrade is uncertain.

The species is arboreal and known in British territory only from the
Sol Haud and Buran.

Pristurus

I. Rostral shield entering the anterior border of the nostril.

A. Tail strongly compressed with a crest of lanceolate scales longer
in males than in females; adult males usually have the crest car-
ried forwards on the dorsal line of the body and females may have
a row of enlarged scales P. flampunctatus.

B. Tail scarcely compressed, with a feeble crest or row of enlarged
scales in males only, never extending on to the body P. rupestris.

II. Rostral shield excluded from the nostril.

A. Claws much longer than the ungual scales; digits long and slender;
/ 18-23 lamellae beneath the fourth toe P. crucifer.

B. Claws shorter than the ungual scales; digits short; 14-19 lamellae
beneath the fourth toe P. phillipsi.

PARKER: LIZARDS OF BRITISH SOMALILAND 41

1. Pristurus flavipunctatus Riippel

Pristurus flavipunctatus Riippel, 1835, N. Wirbelthiere Fauna Abess., Amph.,

p. 17, pi. vi, fig. 3.
Pristurus percristatus Boulenger, 1896, Ann. Mus. Genova, (2), 16, p. 547.
Pristurus percristatus pseudoflavipunctatus Scortecci, 1935, Atti Soc. Ital. Sci.

Nat., 74, p. 123.

cf, 9 Burao, 3500 ft.

ci\ 9 Bohodle, 2100 ft.

8 & o", 2 9 9 Haud, 2200-3900 ft. 46°25'-44°20'E 8°15'-8°55'N

4 d" cf, 2 9 9 Borama District, 5000 ft. 43°20'E 9°50'N
2(j,d' Ado, 2100 ft. 45°15'E 7°20'N

d\ 9 NearZei'lah, 150 ft.

In a previous paper (1932, p. 347) the author referred certain speci-
mens from British Somaliland to P. percristatus Boul. with an element
of doubt, for they had much less developed vertebral and caudal crests
than typical percristatus and so appeared to approach the condition of
flavipunctatus. Scortecci discovered a similar condition in specimens
collected at various localities in Italian Somaliland and considered the
differences to be constant and to indicate the existence of a distinct
southern subspecies which he called pseudoflavi punctatus (Scortecci,
1935, p. 123) and which differed from the typical Eritrean form in its
smaller size, less developed caudal and dorsal crests and the abdominal
scales being granular instead of flat and subimbricate.

The new material now available makes it evident that the height
of the dorsal and caudal crests is variable, probably with age. Thus
of the four males collected together in Lat. 8 35' N., Long. 46° 25' E
two have long spines in the dorsal crest (measuring 0.04-0.045 mm. at
midbody) whereas in the other two the spines are scarcely more than
enlarged granules and are only half as long (0.015-0.025 mm.) ; further,
the two with the shorter crests are smaller animals (28-30 mm. as
compared with 31 and 32 mm. from snout to vent). Similarly of the
four males collected in the Borama District, two examples measuring
35 mm. from snout to vent have dorsal spines of 0.030 to 0.045 mm.,
whereas the smallest of the series measuring 30 mm. has a very much
shorter crest, 0.015 mm. high, and of the two from Ado the larger,
measuring 32 mm. has a crest of 0.04 mm. and the smaller, 30 mm., has
a crest of only 0.02 mm. Of the whole series examined, no male of
less than 30 mm. has a crest of more than 0.025 mm., whereas in exam-
ples of more than 30 mm., the length of the crest varies from 0.03 to 0.045
mm. The only remaining character utilised by Scortecci to distinguish
"pseudoflavipunctatus" is the condition of the ventral scales. Boulenger

42 bulletin: museum of comparative zoology

(1896, p. 549) has considered the same character to be valueless in the
allied species P. collaris, and the present series has examples which by
the length of the dorsal crest are typical percristatus but have the ven-
tral condition of pseudo flavipunctatus. Accordingly it seems probable
that the length of the dorsal crest increases with age and the range of
variation with growth is large enough to include the described condi-
tions of percristatus, percristatus pseudo flavipunctatus and flavipunctatus.
Thus, the 5 cotype males of pcrcristatus available to the author measure
34-35 mm. from snout to vent and have spines in the dorsal crest of
0.034 to 0.045 mm., and so resemble exactly the largest males of the
present series. The cotypes of pcrcristatus flavipunctatus are smaller,
ranging in size from 24 to 34 mm. and have a shorter crest of the order
of 0.01 to 0.02 mm., and the largest type of flavipunctatus, also with a
short crest, is 29.1 mm. from snout to vent. There can be little doubt
that these three names all refer to the same species of the genus in
northeast Africa. It is probable that some records of flavipunctatus
refer to the terrestrial rupestris.

The species is, as mentioned above, arboreal and, like all members
of the genus, diurnal and particularly agile and swift in its movements.
It ranges from the Anglo-Egyptian Sudan to Italian Somaliland as far
south as Dolo, and south-eastern Arabia. In the British Protectorate
it appears to be widely distributed in all areas where acacias are to be
found from sea-level to 5000 feet.

2. Pristurus rupestris Blanford

P. rupestris Blanford, 1874, Ann. Mag. Nat. Hist. (4), 13, p. 454.

P. flavipunctatvs (non Rtippel) Parker, 1932, Proc. Zool. Soc. London, p. 347.

f P. migiurtinicus Scortecci, 1935, Atti Soc. Ital. Sci. Nat. 74, p. 121.

6 & o\4 9 9 Borama District, 5000 ft., 43°10'-43°25'E x 9°50'-9°55'N
2 9 9 Ogo, 5000 ft. 43°E x 10°05'N

9 Guban, 150 ft., 43°15'E x 11°25'N

This species has only once previously been recorded from Somaliland
under the name rupestris, by Scortecci (1935, p. 119) who gave a full
description of his single male in the belief that it might represent a
distinct geographical race. It seems probable that he was led to this
conclusion by comparison with the so-called rupestris of Sokotra,
from which Somaliland examples of rupestris do indeed differ in the
characters he mentions, i.e. a less depressed head, the length of the
digits and a lower number of subdigital lamellae. But as the author
has recently shown (1938) the gecko of Sokotra is really referable to a

PARKER: LIZARDS OF BRITISH SOMALILAND 43

distinct species, and comparison of the Somali material with a long
series from Arabia, from Aden to the Persian Gulf, fails to reveal any
differences which might indicate racial differentiation. The number of
subdigital lamellae in the present series varies from 18-21 beneath the
fourth toe and in the 30 Arabian examples from 18-24.

These small, diurnal geckoes are terrestrial, and though widely dis-
tributed from Karachi through southern Iran and the islands of the
Persian Gulf to south-eastern Arabia, they appear to be confined to
the coastal plain and mountain zone of Somaliland; no specimens have
been recorded from the districts to the south of the mountains.

3. Pristurus crucifer (Val.)

Gymnodactylus crucifer Valenciennes, 1861, Compt. Rend. Ac. Paris, 52, p. 433.
Pristurus stefaninii Calabresi, 1927, Atti Soc. Ital. Sci. Nat., 66, p. 21.

17 cf c? 9 9 9 from various localities in the Haud from Burao to Ado
and in the mountains of the Borama District at altitudes of from 2100 to
5300 ft.

Scortecci (1935, p. 133) has drawn attention to the fact that, al-
though this species is generally characterised by a claw much longer
than the penultimate phalanx there are certain individuals which
have a shorter claw, produced by wear. As might be expected, this is
largely correlated with the substratum on which the animal lives, and
in the present series very long, pointed, scarcely curved claws are
found on those specimens collected in sandy country, shorter, blunter
claws on individuals from localities of stony sand and the shortest
claws of all on those from rocky localities.

A wide-spread species occurring throughout the Somali Peninsula
and extending north-westwards to Eritrea and into south-eastern
Arabia.

4. Pristurus phillipsi Boulenger

P. phillipsi Boulenger, 1895, Ann. Mag. Nat. Hist. (6), 16, p. 165, pl.vii,fig. 1.

P. crucifer (part) idem, 1891, Ann. Mus. Genova, (2), 12, p. 6.

P. somalicus Parker, 1932, Proc. Zool. Soc. London, p. 349.

P. phillipsi somalicus Scortecci, 1935, Atti Soc. Ital. Sci. Nat., 74, p. 152.

9 Haud, 2100 ft., 46°20'E x 8°15'N

Scortecci (loc. cit. supra) has given reasons for considering somalicus
and phillipsi to be conspecific, though he prefers to retain the two
names, using them subspecifically. But both forms occur together in
the same localities so that they can hardly be true subspecies. They

44 bulletin: museum of comparative zoology

may well be ecological races confined to different habitats, but whether
trinomials can properly be used for such modifications seems doubtful.
There is no generally recognised method of designating such races and
to use the same notation as is used for subspecies, must inevitably lead
to confusion. For the present, at least, it seems advisable not to
attempt to label these little understood variants.

P. phillipsi is known only from the Somali Peninsula, ranging from
Dorianle in the south (Calabresi, 1915) to the Golis Mountains near
Berbera and thence eastwards to the coast.

Ptyodactylus hasselquisti (Donndorff)

Lacerta hasselquisti Donndorff, 1789, Zool. Beytr. Leipzig, 3, p. 133.

2 cf cf, 4 juvs. Guban, 150-3500 ft., 42°40'E-43°15'E x 10°30'-11°25'N

This gecko has only once previously been collected in Somali land,
by Neumann in the area between Zeila and Jaldessa (Tornier, 1905;
Neumann, 1905). Though widely distributed over Northern Africa,
Palestine, Syria and Arabia, it does not appear to have extended its
range eastwards beyond the Danakil depression. Tornier and Neu-
mann referred their specimens to the "var. ragazzii", but Flower (1933,
p. 763) has recently pointed out the impossibility of connoting the
numerous local races of this species in Egypt by trinomials.

Tarextola annularis (Geoffroy)

Gecko annularis Geoffroy, 1827, in Savigny, Descrip. d'Egypte, 1, Rept., p. 130,

pi. v, figs. 6, 7.
? Platydactylus delalandii Yaillant, 1882, Miss. Revoil Pays Comalis, p. 14.
Tarentola ephippiata (non O'Shaughnessy) Boulenger, 1895, Ann. Mag. Nat.

Hist., (6), 16, p. 106; idem, 1901, Proc. Zool. Soc. London, (i), p. 49.

9 c* d , 8 9 9,3 juvs. from the Borama District and various localities
along the boundary between this region and Zeilah at altitudes of from
sea level to 5000 ft.

Vaillant's record of Tarentola delalandii from Lasgore ( = Las Koreh)
cannot be accepted; as Boulenger (1921, p. 60) has pointed out there is
no doubt that many of the specimens recorded by Yaillant did not
originate in Somaliland. But if the specimens really were taken at Las
Koreh, they ought probably to be referred to T. annularis. In addition
to this doubtful record two other species of Tarentola have been re-
ported from Somaliland; but reexamination of the material in the
British Museum fails to reveal more than one, for the records of
T. ephippiata (indicated above) are based on specimens of annularis.

PARKER: LIZARDS OF BRITISH SOMALILAXD 45

There seems to be no doubt that the West African T. cphippiata ranges
into the Anglo-Egyptian Sudan, whence it has been recorded many
times, but it seems probable that it ought to be deleted from the list of
species inhabiting the Somali Peninsula.

The genus Tarcntola is remarkable among the geckoes for the degree
of development of the osteoderms. Hitherto these have not been
recorded in the family except as irregularly scattered developments.
In T. annularis, however, they form a complete armour over the body
and limbs. In other families of lizards where a similar armour is devel-
oped the bony scutes are large and correspond in size and position with
the external scales. But here again Tarentola is unique, for only on the
upper eyelid is this correspondence to be found. Elsewhere the osteo-
derms are minute, and bear no relation to the scales overlying them.
In a female specimen measuring 78 mm. from snout to vent the ventral
osteoderms are roughly square, with rounded corners and measure
approximately 0.1 to 0.13 mm. along each side; from 30 to 40 corre-
spond with each scale at the middle of the belly. Dorsally the osteo-
derms are similar in size but are less regular in shape, being subhexa-
gonal or rounded, and the number underlying the external scales
varies, naturally, with the size of the latter.

The species has only been reported from the coastal plain and the
adjacent mountains in British Somaliland; its centre of distribution is
the Egyptian region and the Ogo and Guban are the limits of its dis-
tribution to the east, where it has apparently failed to penetrate into
the Haud and Ogaden.

'&■-

TROPIOCOLOTES

Since Boulenger recorded a rough-scaled Tropiocolotes from Biji,
Somaliland, in 1901, the species has not been re-discovered there until
the present time, and no further records have appeared which bridge
the distributional gap between this area and northern Egypt. These
rough-scaled geckoes have a wide distribution over northern Africa,
but there is some evidence that there may be several species or geo-
graphical races. The material available can be subdivided as follows :
I. Anterior chin-shields large, reaching and forming a long suture
with the second lower labial; posterior chin-shield completely
separated from the first lower labial. Scales about the middle of
the body 44^48. Distribution: Tunisia, Libya, X. Egypt
(8 specs.) T. tripolitanus Peters1

■Vinciguerra (1931, p. 251) regards T. nattereri Steindachner as a synonym of tripolitanus;
but it is described and figured as having smooth scales so that this disposition of the name
cannot be maintained.

46 bulletin: museum of comparative zoology

II. Anterior chin-shields usually not reaching the 2nd lower labial;
posterior chin shields, when present, usually in contact with both
the first and second lower labials. Scales about the middle of the
body 35-41.

A. Posterior chin-shields obvious; supraorbital scales smaller than
those on the interorbital region, juxtaposed. Distribution:
Western British Somaliland T. somalicus spec. nov.

B. Posterior chin-shields hardly distinguishable from the gular
scales. Supraorbital scales larger than those of the interorbital

region, imbricate. Distribution: Rio de Oro

T. occidentalis spec. nov.

It may ultimately prove that these two new "species" are only
geographical races of tripolitanus, or that there is a single race ranging
from the extreme east to the west of the continent to the south of the
typical form. Certainly the two bear a greater resemblance one to
another than either does to tripolitanus, but the differences appear
sufficient to justify the present tentative arrangement.

Tropiocolotes somalicus spec. nov.

Tropiocolotes tripolitanus (non Peters) Boulenger, 1901, Proc. Zool. Soc.
London, p. 48.

The holotype of T. somalicus is a male, number 1937.12.5.693 in the
British Museum from 42° 50'E, 10° 20'N at an altitude of 3000 ft.;
collected by Capt. R. H. R. Taylor 27. vii.33.

Snout a little longer than the distance between the eye and the ear.
Head covered with keeled granules; rostral broader than high, with a
median cleft ; nostril pierced between the rostral, first upper labial and
2 small nasals; seven upper and six lower labials; mental large, triangu-
lar, followed by a pair of large chin-shields which form a median
suture; posterior chin-shields much smaller and widely separated,
making contact with the first and second lower labials. Scales of the
body, both dorsal and ventral, large and strongly keeled, about 37
round the middle of the body, 4 of the largest dorsals equal in length
to the distance between the nostril and the anterior border of the eye.
Limbs covered with strongly keeled scales; hind-limb reaching the
elbow; subdigital lamellae tricarinate.

Grey-brown above, with a few darker flecks; a dark brown band
from the nostril through the eye, above the shoulder; upper lip white,

PARKER: LIZARDS OF BRITISH SOMALILAND 47

brown spotted; anterior border of eye white; tail with alternating
light and dark transverse bars; lower surfaces white, with a few, very
small dark dots.

Length from snout to vent 28 mm.
Tail (regenerated) 35 mm.

Fore-limb 9 mm.

Hind-limb 14 mm.

The paratypes are all very similar in general proportions; their locali-
ties, sexes and scales about the middle of the body are:
43°E x 10°45'N, 3000 ft. 2 9 9 , & 37, 39, 41.
43°15' x 11°25'N 150 ft. 2 cT d\ 9 39, 35, 35.
Biji 9 36.

The species appears to have a very limited distribution and to be
confined to the very arid country of the coastal zone; all the specimens
were taken in the ground in sandy country with patches of stones and
almost devoid of vegetation.

Tropiocolotes occidentalis spec. nov.

Tropiocolotes tripolitanus (non Peters) Gunther, 1903, Novit. Zool., 10, p. 298.

Holotype a male, number 1908.6.13.15 from the Rio de Oro, col-
lected by Herr Riggenbach.

Snout a little longer than the distance between the eye and the ear.
Head covered with feebly keeled, imbricate scales ; rostral broader than
high, with a median cleft; nostril pierced between the rostral, first
upper labial and two small nasals; seven upper and six lower labials;
mental large, triangular, followed by a pair of chin shields which form
a median suture and do not reach the second lower labial; posterior
chin-shields absent. Scales of the body, both ventral and dorsal, large
and strongly keeled, about 41 around the middle of the body, 4^ of the
largest dorsals equal in length to the distance between the nostril and
the eye. Limbs covered with strongly keeled scales ; hind limb reaching
between the elbow and the wrist; subdigital lamellae tricarinate.

Pale straw-colour above; a brown band from the eye, above the
shoulder; upper lip and anterior border of the eye whitish; dorsum
with a series of about seven indefinite, wavy, transverse brown bars;
tail with regular brown cross-bars; lower surfaces of head and body
white, of the tail pale brown.

Length from snout to vent 30 mm.

Tail 36 mm.

Fore-limb 9 mm.

Hind-limb 13 mm.

48 bulletin: museum of comparative zoology

The paratype is a juvenile male from the same locality as the holotype.
It agrees with the latter except in having 40 scales round the body, a
slightly longer hind-limb (the third toe reaches the elbow) and in
having the dorsal cross-bars less distinct and broken up into blotches.

AGAMIDAE

Agama

I. Tail longer than the head and body, not depressed.

A. A distinct crest of enlarged, lanceolate scales on the nape

agama spinosa

B. No nuchal crest, or only a few spinose, not lanceolate scales.

1. A dorso-lateral fold from the sides of the neck to the groin;

dorsal scales heterogeneous.

a. Whole of the dorsum beset with large spines each of which
has a ring of smaller spines at its base A. robecchii

b. Spines, if present, confined to the dorso-lateral fold.

i. Dorsum, between the dorso-lateral folds and flanks,

with scattered, slightly enlarged scales

A. cyanogaster

ii. Mid-dorsal area without scattered enlarged scales.
t Flanks with a few scattered, enlarged scales; a
golden mid-dorsal stripe. Adult males circa 70-90

mm A. phillipsi

ft Flanks covered with uniform, small smooth scales.
Adult cf circa 110-120 mm A. annectans

2. No dorso-lateral fold; dorsal scales large and homogeneous.

a. Spines about the ear partially concealing it, and as long as
the diameter of the eye opening; adult cfc? 58-88 mm.;
9 9 76-88 mm A. rueppelli rueppelli

b. Spines about the ear only half as long as the eye opening
and not encroaching upon the tympanic area; tympanum
fully exposed; adult cf cf 43-54 mm., 9 9 56-64 mm.. . .
A. persimUis

II. Tail much shorter than the head and body, strongly depressed,
with a disc-like basal portion.

A. Basal portion of the tail longer than broad, merging gradually into
the terminal filament, with marginal spines not markedly greater
than those on its upper surface A. batillifera

PARKER: LIZARDS OF BRITISH SOMALILAND 49

B. Basal portion of the tail broader than long, abruptly differentiated
from the terminal filament and with very large marginal spines . .

A. taylori

In addition to these species Vaillant (1882) reports Agama agilis
and A. ruder at a from Las Gore (= Las Khoreh). But these records
seem very improbable and the species have been omitted from con-
sideration.

The following species may ultimately be found in the Protectorate
also:

Agama bottcgi Boulenger Type locality: Lugh
Agama cornii Scortecci Om-Ager, Eritrea

Agama lionotus Boulenger S. E. of Lake Rudolf

Agama zonura Boulenger "Wardergubbernor"

(This may be the "Wal da Gubora's Capitol" in 7°10'N x 40°41'E,
approximately, shown on Donaldson Smith's map in the Geographical
Journal 1895, 5, p. 124, otherwise known as Gineh, Ginea, or Ginir.
The species has also been recorded from Harrar (Tornier 1905; Erlanger
1905).

1. Agama agama spinosa Gray

Agama spinosa Gray, 1831, in Griffith, Cuvier's Animal Kingdom, 9, Syn.,

p. 57, pi.
Agama colonorum Boulenger, 1895b; idem, 1898; Tornier, 1905; Neumann,

1905; Scortecci, 1929.
Agama hartmanni (nee Werner, nee doriae Boul.) Boulenger, 1901.

15 d* d\ 6 9 9, 11 juvs. from various localities along the boundary
northwards and westwards from the Borama District (43°10'E x 10°N)
to 43°15'E x 11°15'N, at altitudes of from 500 to 5000 feet.

Flower (1933, p. 772) has expressed the opinion that Agama spinosa
Gray may be one of the numerous local races of the wide spread Agama
agama. This certainly seems very probable, and explains how it is that
the three closely allied forms A. spinosa, A. colonorum (= A. agama)
and A. hartmanni have all been recorded from British Somaliland.
These three forms have all been recorded from the same general area,
the costal plain and adjacent mountains, and it seems highly probable
that only a single species is involved. The correct name to apply to
this would appear to be spinosa Gray, though it must be admitted that
many specimens from Somaliland differ widely from typical spinosa in
having a much shorter nuchal crest, shorter spines about the ear and
smaller, less strongly keeled dorsal scales. But the above-mentioned
large series shows an enormous amount of variation in these characters

50 bulletin: museum of comparative zoology

and yet appears to represent but a single species ; the variations cannot
be correlated with diversity of habitat nor with altitudinal distribu-
tion. Loveridge (1936, p. 54) has stated that in his experience the
variation of the number of scales around the middle of the body of any
one race of Agama agama is 10, rarely as many as 13, in a given area.
Yet the present series varies from 59 to 80 in this respect and other
examples from near Berbera and the Golis Mountains extend this
range up to 90; I am unable to detect any discontinuity which would
indicate that two species are involved. In typical spinosa from the
Egyptian region the number of scales around the middle of the body
varies from 62 to 74 (18 specs, examined).

In a previous publication (1932) the author placed agama smithi in
the synonymy of A. spinosa, but it is doubtful whether this disposition
is justified.

Agama agama spinosa is strictly terrestrial, and an inhabitant of
rocky and stony districts. As pointed out above, it is confined to the
coastal plain and maritime mountains, as far east as the Wagar range,
a distribution characteristic of so many other Eremian species which
have failed to extend their range into the Haud and south-eastern
drainage.

2. Agama robecchii Boulenger

Agama robecchii Boulenger, 1891, Ann. Mus. Genova, (2), 12, p. 6, pi. i, fig. 1.

This species appears to have a very limited distribution in the sul-
tanates of Obbia and Migiurtina and has only once been recorded in
British Somaliland, from Gumbi Hill (10°16'N x 47°12'E).

3. Agama ctanogaster (Ruppell)

Stellio cyanogaster Ruppell, 1835, N. Wirbelthiere Fauna Abess., Amph., p. 10,

pi. v.
Agama atricollis A. Smith, 1849, 111. Zool. S. Africa, Rept., App., p. /144.

9 43°E x 10°10'X.

Boulenger as long ago as 1896c, pointed out that cyanogaster appeared
to pass completely into atricollis and suggested that the two might have
to be united. Specimens have been recorded from the Ethiopia-
Somali region under both names and Loveridge (1936, p. 57) maintains
that the two are distinct species, cyanogaster being a much smaller
form, measuring not more than 70 mm. from snout to vent. It seems
highly probable, however, that Loveridge's "cyanogaster" is really
phillipsi, for the type of cyanogaster measures 113 mm.

PARKER: LIZARDS OF BRITISH SOMALILAND 51

A. cyanogaster, including atricollis, is widely distributed over eastern
Africa from the Cape Province to the mountains of Ethiopia and
Eritrea. It has only previously been recorded (as atricollis) from
British Somaliland in the extreme east, at Buran and Taleh by Cala-
bresi (1927, p. 27); one of the same specimens was apparently referred
by Scortecci (1931, p. 142) to cyanogaster. It is possible, however, that
this identification is incorrect, for Capt. Taylor failed to collect the
species in that region though the allied A. annectans was plentiful, but
was not reported by either Scortecci or Calabresi.

4. Agama phillipsi Boulenger

Agama phillipsi Boulenger, 1895, Ann. Mag. Nat. Hist., (6), 16, p. 567, pi. vii,

fig. 3.
A. cyanogaster (non Ruppell) Loveridge, 1936, p. 57.

7 cf cf, 12 9 9 , 10 juvs. from various localities in the Borama district
(43°-43c25'E x 9°50'-10°10'N) at altitudes of from 4500 to 6000 ft.

This series is very constant in its general characters and colour.
There is, however, an appreciable amount of variation in the number
and size of the enlarged, keeled scales on the flanks and femora; but
they are constantly present and serve to distinguish this species from
annectans which has uniform small scales in both these regions. The
absence of enlarged scales between the dorso-lateral folds distinguishes
it from cyanogaster.

A. phillipsi appears to be a montane species, ranging from Eritrea
through the mountains of north-eastern Ethiopia to the Ogo as far
east as the Golis Range; no specimens have ever been reported from
the Haud or the lower lying areas to the south and east. Its depressed
habitus reflects its habit of hiding in rock crannies; the diet appears
to consist principally of ants.

5. Agama annectans Blanford

Agama annectans Blanford, 1870, Zool. Abyssinia, p. 446, fig.

? Agama atricollis Calabresi, 1927, Atti Soc. Ital. Sci. Nat., 66, p. 27.

? Agama cyanogaster Scortecci, 1931, Atti Soc. Ital. Sci. Nat., 70, p. 142.

juv. cf Borama District, 4000 ft., 42°25'E x 10°20'N.

A. annectans was originally described from the Suru (or Sooroo)
Pass in Eritrea {circa 39°32'E x 15°05'N), but has not since been re-
corded from that region, though it has been reported from Harar,
1900 meters, and the Fulla Valley, 500 meters (Tornier, 1905; Neu-

52 bulletin: museum of comparative zoology

mann, 1905), various localities in southeastern Ethiopia and in the
eastern parts of British Somaliland (Sol Haud, Buran District, Al
Mado Mts. and Las Elan; Parker, 1932). It does not appear to range
into the Haud and has not been reported from the Ogaden or coun-
tries to the south and east.

6. Agama rueppelli rueppelli Vaillant

Agama rueppelli Vaillant, 1882, in Revoil, Mission Pays Comalis, Rept. p. 6,

pl.i.
Agama vaillanti Boulenger, 1895, Ann. Mus. Genova, (2), 15, p. 12, pi. iii, fig. 1.
Agama rueppelli rueppelli Parker, 1932, Journ. Linn. Soc. London, Zool. 38,

p. 224.

4 cf cf,5 9 9 , 2 juvs. from various localities along the boundary west-
wards and northwards from 43°15'E to 42°40'E and 9°55'N to 10°45'N at
altitudes of 2000 to 5000 ft.

This series is homogeneous and typical. The species appears to be
widely distributed along the maritime plain and adjacent mountains
through British Somaliland and to extend southwards through the
eastern districts of the Sol Haud, Nogal Valley and eastern Haud to
the Ogaden, Italian Somaliland and Kenya Colony. Its place appears
to be taken in the central Haud by a smaller, allied species which has
not yet received a name.

7. Agama persimilis spec. nov.

Holotype a female, number 1937. 12. 5. 64 in the British Museum,
from 45°50'Ex8°N in the Haud; collected by Capt. R.H.R. Taylor
on November 25th, 1934.

Head convex, the upper surface of the snout flat; nostril tubular,
directed upwards and backwards in the posterior part of the nasal
above the canthus rostralis. Upper head scales moderately large, those
On the supraocular region longitudinally elongate, with a blunt keel;
occipital scale enlarged; ear superficial, completely exposed, its diam-
eter equal to that of the eye-opening; lower and posterior borders of
the ear and sides of the neck with tufts of spines, the longest of which
is about half the length of the eye-opening; these tufts of spines do not
encroach upon the ear-opening. Two transverse gular folds, but no
pouch. Body strongly depressed, with large, homogeneous, imbricate,
keeled and mucronate scales, convergent towards the middle line except
on the nuchal and occipital regions where they are reversed and point
forwards; about 30 scales on the vertebral line between the anterior

PARKER: LIZARDS OF BRITISH SOMALILAND 53

limits of the insertions of the fore and hind limbs; 57 scales round the
middle of the body; 9 scales of an oblique series corresponding to a
length equal to that from the tip of the snout to the anterior border of
the ear. Tip of the fourth toe reaching the posterior corner of the eye;
third finger a little longer than the fourth, and the fourth toe slightly
longer than the third; tibia as long as the skull. Tail 2.4 times as long
as the distance between the posterior gular fold and the vent, de-
pressed and tapering very abruptly in its proximal sixth, slender and
cylindrical posteriorly. Ventral scales smooth, much smaller than the
dorsals. General colour rufus brown, with a strongly-marked geo-
metrical colour-pattern . Upper surface of the snout to the level of the
anterior borders of the eyes pink; a broad brown bar, the full width
of the supra-ocular region, across the head with an oblique extension
from its posterior corners towards the ears; an oblique brown bar from
beneath the eye to the anterior border of the ear; occiput pink mesially
and a longitudinal stripe of the same colour extends down the midline
of the back onto the tail ; an oval dark brown spot on each side of the
nape; three transversely elongate brown diamond-shaped blotches on
the back, decreasing in size caudally, the last on the sacral region,
each edged by a narrow light line and bisected by the mid-dorsal light
line. An elongate, triangular, dorso-lateral brown blotch on each side
runs from the shoulder to the middle of the flank where its apex is
narrowly separated from the apex of a similar blotch which extends
backwards to the groin; flanks beneath these two markings occupied
by a single large, triangular, brown blotch, the 3 lateral and dorso-
lateral markings being all more or less distinctly edged with lighter.
Tail with transverse dark bars, bisected mid-dorsally. Limbs with
dark brown, light-edged markings. Lower surfaces uniform white.

Length from snout to vent 56 mm. Tail 95 mm. The following
specimens, all collected by Capt. Taylor are the paratypes :

4d"d\ 5 9 9 , 1 juv. from near Bohodle (46°20'E x 8°15'N, 2100 ft.)
collected between Jan. 21st and April 1st 1932.

1 9 collected at the same time and place as the holotype.

1 gravid female from 46°09'E x 8°17'N, 2200 ft., taken on Aug.
21st, 1932.

1 juvenile from 45°09'E x 8°37'N, 3050 ft., taken on July 9th, 1932.

This series of specimens, collected in the central Haud, is very
homogeneous. There are some variations in scale size, the number of
scales on the mid-dorsal line varying from 26-31 and the number in an
oblique series corresponding to a length equal to that between the tip
of the snout and the ear, from 8-11. Adult males vary in size from

54 bulletin: museum of comparative zoology

43 to 54 mm. and females from 56 to 64 mm. ; in a series of twenty-
three rueppelli rueppelli males measure from 58 to 88 mm. and females
from 76 to 88 mm. The species is undoubtedly very closely allied to
Agama rueppelli rueppelli Vaillant, which was not found in the central
Haud, though common in the coastal mountains and in the countries
to the east and south of the area occupied by persimilis. The two may
be distinguished readily by the characters mentioned in the accom-
panying key.

8. Agama (Xenagama) batillifera (Vaillant)

Uromastix batillijerus Vaillant, 1882, in Revoil, Mission Pays £omalis, Rept.,
p. 10, pi. ii.

12 c? 6% 11 9 9,4 juvs. from the mountains of the Ogo between
43°10' and 43°30'E and 9°30' and 9°50'N at altitudes of 4500 to 5300 ft.

This species is chiefly nocturnal in its habits, being found by day in
burrows which it constructs in soft earth. The tunnels are just wide
enough to permit the animal's passage, are 2 to 3 feet long and reach
a depth of about a foot; there does not appear to be any definite ter-
minal chamber or nest. The species seems to be confined to the moun-
tains bordering the Haud on the north from 43°10' to 50°27/E (Vail-
lant, 1882) and to extend into the Ogaden on its western side as far as
Sasabanch (Boulenger 1895c). Within this area there may prove to be
some geographic variation for the present series from the extreme west
of the range of the species have a slightly narrower, less distinctly dis-
coid tail than others from the east (10°25'N, 47°12'E). To the south,
in the Haud, A. batillifera is replaced by

9. Agama (Xenagama) taylori Parker

Aporoscelis batillijerus (non Vaillant) Peel, 1900, Somaliland (London), pp. 175,

176, 334, fig.
Agama (Xenagama) taylori Parker, 1935, Ann. Mag. Nat. Hist. (10), 16, p. 525.

d\ 9 from the Haud at 3400-3500 ft., near 44 °49'Ex8°43'N.
These two examples are the only ones known to exist in Museums,
but the specimen described and figured by Peel almost certainly belongs
to this form, which differs from the typical montane A. batillifera in
its broader, more discoid tail. At first it was thought to represent a
distinct species, but, as pointed out above, A. batillifera shows some
geographical variation in this respect which suggests the possibility
that A. taylori may ultimately prove to be only racially distinct. At

PARKER: LIZARDS OF BRITISH SOMALILAXD 00

the present moment, however, no intermediates have been found and
the degree of difference between it and any specimen of A. batillifera
is far greater than has been recorded in the latter species over a much
wider geographical area. Peel's specimens were seen in the Haud at
Bally Maroli, a short distance south-east of Odaweina.

Uromastiz

I. Tail only slightly shorter than the head and body; a series of

preanal and femoral pores U. macfadyeni

II. Tail only half the length of the head and body; no femoral or
preanal pores U. princeps.

1. Uromastix macfadyeni Parker

Uromastix macfadyeni Parker, 1932, Proc. Zool. Soc. London, p. 353.

U. ocellatus (non Licht.) Tornier, 1905, Zool. Jahrb., Syst., 22, 4, p. 372.

4 <? d\ 3 9 9,7 juvs. from the Guban between 10° and 11°15'X and
42°40' and 43°15'E, at altitudes of from 1000-3800 ft.

The relationships of this species to various other members of the
genus Uromastix have already been discussed in the original descrip-
tion, but its nearest relative appears to be the southern Arabian
Uromastix {Aporoscelis) bcnti. It appears to be so closely allied to this
species that the retention of Aporoscelis, even subgenerically, does not
seem to be justified.

It is a herbivorous lizard usually found hiding in burrows by day
and seems to be confined to the foot hills of the Guban as far east as
Dagah Shabell, 24 miles S.E. of Berbera.

2. Uromastix prixceps O'Shaughnessy

Uromastix princeps O'Shaughnessy, 1880, Proc. Zool. Soc. London, p. 445,

pi. xliii.
Aporoscelis princeps Boulenger, 1885, Cat. Lizards Brit. Mus., 1, p. 410.

A diurnal, terrestrial species, partly herbivorous, partly insectivor-
ous, which frequents the rocky and stony districts of the extreme north-
east corner of the Somali Peninsula; in the British Protectorate it has
only been recorded from the eastern boundary, from the coastal plain
to the Nogal Valley (Parker, 1932, p. 351). The doubtful records of
the species from Zanzibar (O'Shaughnessy loc. cit. supra), Aden
(Vaillant, p. 10) and Asmara (Scortecci, 1933, p. 1) are probably to be
ascribed to specimens either wrongly labelled or transported in coastal
vessels.

56 bulletin: museum of comparative zoology

VARANIDAE

Varanus
I. Nostril three times as far distant from the tip of the snout as
from the eye; dorsal scales very large, especially on the neck. . . .

V. ocellatus

II. Nostril a little nearer the eye than the end of the snout; dorsal
scales small, not appreciably enlarged on the nape . . . . V. niloticus

1. Varanus ocellatus Ruppell

Varanus ocellatus Ruppell, 1827, Reise N. Afr., Rept. p. 21, pi. vi.
Varanus albigularis Boettger, 1893, Zool. Anz., p. 115.

3 skins from the Haud 42°40'-44°24'E x 10o30'-8°52'N, 3500-4800 ft.

Schmidt (Bull. Amer. Mus., 1919, 39, p. 483) has called attention
to the fact that Meek's record (1897, p. 181) of Varanus albigularis is
really based on a specimen of V. ocellatus and it seems very possible
that other Somali records of the white-throated monitor are due to
similar misidentifications. The two species are very similar and at the
present time it seems almost impossible to delimit accurately the
ranges of the two; ocellatus certainly ranges over the whole of the
Somali Peninsula and the Sudan, southwards into Kenya Colony and
Tanganyika Territory.

2. Varanus niloticus (Linn.)

Lacerta nilotica Linnaeus, 1766, Syst. Naturae, (12), p. 369.

There is no record of the Nile monitor from British Somaliland,
though it has been reported from the Ogaden. Since it is a species
seldom found far from a permanent water-supply, it is more than
likely that it does not occur in the protectorate.

AMPHISBAENIDAE

Agamodon

I. Annuli on the body 133-137; on the tail 15-19; 0-6 preanal pores
A. anguliceps

II. Annuli on the body 143-160; on the tail 18-23; preanal pores 0-2
A. compressum

Neither of these species has yet been reported from the British
Protectorate though both are not uncommon in southern Italian Soma-

PARKER: LIZARDS OF BRITISH SOMALILAND 57

liland. Since burrowing species, such as these, are liable to be over-
looked very easily, it is conceivable that they may yet be discovered;
but the difference in the humidity of the soil in the Shebeli and Juba
regions, as compared with British Somaliland, suggests that if the
genus really does occur in the latter area it will be represented by a dis-
tinct species.

Ancylocranitjm genus nov.

Type species Anops somalicus Scortecci, 1930, Boll. Mus. Tornio, 41,
3, p. 6., figs. 1-5.

Stejneger (1916, p. 85) has drawn attention to the fact that Anops
Bell 1833 (type species A. kingi) is preoccupied by Anops Oken 1815,
and proposed Anopsibaena as a substitute generic name for the South
American lizard A. kingi. The only other species which at that time
was commonly referred to Anops Bell, was a West African lizard
originally described by Gray in 1865 as Baikia africana. On purely
geographical grounds Stejneger was convinced that africana and kingi
could not be congeneric so that the name Baikia was available only for
the African species. With the discovery of somaliea there are now three
known species of the "Anops" type and examination of their skulls
confirms Stejneger's contention and also suggests that the third species
represents yet another evolutionary line. These three may be distin-
guished thus:

I. The whole skull laterally compressed with a median crest extend-
ing from the premaxilla to the occiput; premaxilla extending back-
wards, separating the nasals completely and the frontals anter-
iorly; extra-columella not dilated anteriorly (fig. 7b) .Anopsibaena.
II. The skull compressed anteriorly only, the median ridge extending
from the premaxilla to the parietal only, the upper surface of the
latter being quite flat; premaxilla small, the nasals forming a
median suture; extra-columella not dilated anteriorly (fig. 7c)
Baikia

III. The skull compressed anteriorly only the parietal being flat above;
nasals forming a median suture; compressed anterior elements of
the skull forming a high crest which extends backwards over the
parietal region as a pronounced hook; extra-columella dilated an-
teriorly to form a large, thickened pad attached to the lower labial
(fig. 7a) Ancylocranium

These differences may be appreciated by a comparison of the ac-
companying figures, but some explanation is necessary with regard to

58

bulletin: museum of comparative zoology

the extra-eolumella. A feature of all three genera, and probably of the
majority of Amphisbaenids is the peculiar condition of the ear (Vers-
luys, 1898, p. 82). There is no tympanum, the merest rudiment of a

Fig. 7. Skulls of the three species formerly referred to the Amphisbaenid
genus Anops Bell.

a. Ancylocranium somalicum (Scortecci). Somaliland. Extra-
eolumella indicated by dotted outline.

b. Anoysibaena kingi. South America.

c. Baikia africana (Gray) . West Africa.

Type and only known specimen; outline from X-ray photo-
graph; sutures anteriorly determined by dissection.

cavum tympanum and no eustachian tube; the columella auris is
short, and owing to the forward rotation of the lower end of the
quadrate the fenestra ovalis lies ventral to, rather than behind, this

PARKER: LIZARDS OF BRITISH SOMALILAXD 59

bone. Attached to the outer end of the columella is a long cartilagin-
ous rod, the extra-columella; this rod runs forwards over the outer sur-
face of the quadrate and along a groove on the ventro-lateral surface
of the mandible and is attached anteriorly by fibrous ligaments to the
corium of the lower labials. In Ancylocranium the anterior end of this
extra-columellar rod is expanded to form a large thickened pad under-
lying and attached to the whole length of the enlarged second lower
labial. It seems probable that this modification permits the ear to
function more effectively in detecting vibrations in the soil with which
the chin is normally in contact.

Ancylocranium somalicum (Scortecci)

Anops somalicus Scortecci, 1930, Boll. Mus. Torino, 41, p. 6, figs. 1-5.
Baikia somalica Loveridge, 1941, Bull. Mus. Comp. Zool., 87, p. 370, fig. 6.

Ad. Haud, 2100 ft. 46°E x 8°20'N, 29. iv. 32.
Ad. Haud, 2300 ft. 45°58'E x 8°24'N, 5. v. 32.
Ad. Haud, 2350 ft. 45°43'E x 8°26'X, 18. v. 32.

These specimens show some slight variation beyond that already
described for the species. The segments in an annulus vary from 30/21
to 32/23 (vice 27/22, 28/24), there are from 186 to 193 annuli on the
body (197-199) and from 7 to 8 on the tail (vice 6-7). Thus these
three examples from the Haud differ appreciably from the only other
known examples of the species, the types from Caitoi on the YYebi
Shebeli. It is not improbable that the differences are really indicative
of racial differentiation but far too little is at present known of the
stability or variability of these lizards for it to be possible to form any
definite opinion on the point.

LACERTIDAE

Eremias

I. Upper head-scales smooth or only slightly rugose; dorsal scales
smooth or feebly keeled, never tricarinate.

A. No granules between the supra-oculars and the frontals and
fronto-parietals. Usually three nasals E. guttulata olirieri

B. A ring of granules bordering the supra-oculars. Usually four
nasals.

1 . Dorsal scales smooth. Dorsal pattern of 5 stripes of which the
two laterals on each side may be broken up.

60 bulletin: museum of comparative zoology

a. Posterior subcaudals smooth; ventrals in 6 longitudinal
series with traces of an additional row on each side; sub-
ocular usually entering the lip E. mucronata

b. Posterior subcaudals keeled; ventrals in 8 longitudinal
series with traces of an additional row on each side; sub-
ocular excluded from the upper lip E. smithi

2. Dorsal scales smooth or keeled; posterior subcaudals smooth
or keeled; dorsal pattern of 7 stripes, none broken up. Sub-
ocular entering the lip E. septemstriata sp. n.

3. Dorsal scales keeled; posterior subcaudals keeled; dorsal
pattern of irregular spots; subocular excluded from the lip. . .

E. crythrosticta

II. Upper head scales rugose or strongly striated ; dorsal scales usually
tricarinate.

A. Snout long; toes unicarinate or with additional keels much smaller
than the median.

1. Fronto-nasal distinctly longer than broad; femoral pores 17-
25; three or four nasals; head shields finely or strongly striated;
subocular excluded from the lip; a pattern of 5 dark longi-
tudinal stripes of which at least the outermost encloses rows
of circular white spots E. brenneri

2. Fronto-nasal a little broader than long; femoral pores 13-17;.
four, rarely three, nasals; head shields rugose; subocular
usually entering the lip; 5 dark longitudinal stripes without
any trace of light spots E. striata

B. Snout short, the fronto-nasal much broader than long; three
nasals ; toes bicarinate ; five or seven dark stripes bearing rows of

black blotches ; subocular excluded from the lip

E. spekii sextaeniata

Eremias guttulata olivieri (Audouin)

Lacerta olivieri (part) Audouin, 1829, Descr. d'Egypte, Rept. Supp., p. 175,

, pi. ii, fig. 2.
Eremias guttulata Tornier, 1905, Zool. Jahrb. Syst., 23, p. 379.
Eremias guttulata martini Neumann, 1905, Zool. Jahrb. Syst., 23, p. 396.

14 d" cf , 6 9 9,1 juv. from various localities on the boundary between
42°40' and 43°15'E x 10°20' and 11°25'N, at altitudes of from 0-3500 ft.

Criticism has already been levelled at the varieties of Eremias
guttulata recognized by Boulenger (1921), many of which occur to-

PARKER: LIZARDS OF BRITISH SOMALILAND 61

gether in the same localities and so do not appear to be true sub-
species. On geographical grounds the form occurring in Somaliland
should be martini, but in the foregoing series, whilst some individuals
have the striped colour of martini, others are not distinguishable from
the typical form of olivieri, and the morphological characters are those
of the latter; there are 4 to 7 enlarged plates in the lower eyelid, 37 to
46 scales across the middle of the back and 9 to 14 femoral pores.
Until the whole species can be satisfactorily revised, it is difficult to
know how to treat the various "races." Flower has pointed out (1933,
p. 798) that the forma typica is the normal form in Egypt but that
specimens referable to martini and olivieri occur as individual varia-
tions, and it seems possible that the apparent overlapping of the races
is due to the prevalence of similar individual variants in populations
composed for the most part of a different form; martini is probably
not to be distinguished from olivieri.

Like other northern species it only enters Somaliland in the coastal
zone north of the mountains whose northern slopes it ascends to an
altitude of 3500 ft.

Eremias mucronata (Blanford)

Acanthodactylus mucronatus Blanford, 1870, Obs. Geol. and Zool. Abyssinia,

p. 453, fig.
Eremias lugubris (non Smith) Vaillant, 1882, Mission Revoil Pays Qomalis,

Rept. p. 23.
Eremias brenneri (non Peters) Meek, 1897, Field Columbian Mus., Zool.

Series, 1, 8, Publ. 22, p. 181.

26 cf cf , 14 9 9 Borama Distr., 5000 ft., 42°45'E x 10°20'X, 19-24.2.33.
cf , 150 ft., 43°15'E x 11°25'X, 14.1.34.

This series is very constant in morphological characters and in
colour. Only one specimen, a male, is in any way abnormal, and there
the small post-nasal is fused with the supra-nasal so that there are
only three shields surrounding the nostril. The scales across the back
vary from 61 to 78, ventrals are in 6 or 8 longitudinal series, the outer-
most always being small, and the femoral pores vary from 17 to 24.
The colour is of the normal type with an unbroken median dorsal
stripe, flanked with white and primitively with two broad, dark lateral
stripes enclosing numerous small, circular white spots; this pattern
occurs in juveniles and persists to a greater or less extent in adult
females, but in males the white spots are so numerous that almost all
trace of the lateral stripes is lost.

62 bulletin: museum of comparative zoology

The species is a northern form which extends along the mountain
chain of the north of the country and in the coastal zone. It has not
yet been recorded from the region south of the mountains in British
Somaliland but occurs in the Ogaden (Boettger, 1893; Boulenger, 1895,
1896) and the Sultanate of Obbia (Calabresi, 1927).

Eremias smithi Boulenger

Eremias smithi Boulenger, 1895, Proc. Zool. Soc. London, p. 534, pi. xxix, fig. 4.

3 <? cf , 5 9 9 Burao, 3500 ft.

8 c? c\ 11 9 9 from the mountains around Borama, 4500-5300 ft.

This series shows little variation beyond that already described for
the species (Boulenger, 1921, p. 247) except that there may be as few
as 61 scales across the body and only 16 femoral pores; there are never
less than 8 rows of ventrals and frequently indications of an additional
series on each side are present.

The species appears to range from the coastal plain through the
mountains and the Ogaden southwards to Kenya Colony.

Eremias septemstriata spec. nov.
Eremias striata (non Peters) Parker, 1932, Proc. Zool. Soc. London, p. 354.

Holotype a male, number 1931. 7. 20. 300 from the Halin district,
Somaliland (9°7'N x 48°38'E) at 2000 ft., collected 14. III. 30 by
Capt. R. H. R. Taylor.

The specimens mentioned in the above reference were noted to
deviate from the normal of striata, and the additional comparative
material of that species now available makes it appear more probable
that they represent a distinct form which is in many ways intermediate
between brenneri and striata, but differs from either in its smooth or
only feebly striated head shields and dorsal scales and in the distinc-
tive colour pattern.

Head shields smooth or very feebly rugose, disposed and propor-
tioned as in E. brenneri, except that the subocular, though narrowed
inferiorly, broadly enters the lip; snout long, with a fronto-nasal
slightly longer than broad; constantly 4 nasals. Dorsal and lateral
scales juxtaposed, smooth or feebly unicarinate, in 54 to 68 series
across the middle of the body; ventrals in 8 longitudinal series; en-
larged brachials and subcaudals smooth or keeled in conformity with
the dorsals. Femoral pores 15-20.

PARKER: LIZARDS OF BRITISH SOMALILAXD 63

Colour white or pale brown with 7 dark brown longitudinal stripes
arranged as follows : A broad lateral band from the subocular through
the upper part of the ear and above the limbs, extending on to the tail,
and usually enclosing a single row of small circular, white spots; a
dorso-lateral stripe, similar in width, extends from the posterior corner
of the eye, but does not extend on to the tail and may, or may not,
have a row of white spots included in it; three narrower dorsal stripes,
of which the median alone extends on to the tail, the two outer varying
in length, being as a rule shortest in juveniles where they may not
extend backwards beyond the nape; there may also be traces of an
indistinct dark, ventro-lateral stripe between the fore- and hind-limbs;
lower surfaces uniform white; limbs brown, with numerous large, cir-
cular white spots.

Length from snout to vent 42 mm.

Tail (partly regenerated) 107 mm.

Fore-limb 16 mm.

Hind-limb 35 mm.

The paratypes are:

1931.7.20.301 d1 3300 ft. Buran Distr. (10°13'N x 48°46'E) 7.i.30.

1931.7.20.298 9 2000 ft. Haud (8°Nx48°E) 2.V.30.

1931.7.20.299 9 1900 ft. (8°54'N x 48°54/E) 17.iii.30.
1931.7.20:306 9 2375 ft. Buran Valley (10°21'N x 49°E) 2.x.29.
1931.7.20.302-305 4 juvs. 2600 ft. Buran Valley (10°16'N x 48°55'E)

29.xi.29.

The largest female measures 50 mm. from snout to vent, with an
unregenerated tail of 125 mm., a fore-limb of 17 mm. and a hind-limb
of 37 mm.

As indicated above, this species is allied to E. striata but also shows
affinities with the northern, smooth-scaled species such as E. mucro-
nata and E. smithi.

Eremias erythrosticta Boulenger

Eremias erythrosticta Boulenger, 1891; Ann. Mus. Genova, (2), 12, p. 10, pi. i,
fig. 2.

This species was not collected by Capt. Taylor, but is probably an
inhabitant of the eastern part of the country south of the maritime
mountains. It was originally described from specimens collected
"between Obbia and Berbera" and has since (Calabresi, 1927) been
recorded from Obbia, the dunes between Obbia and Sissib and the
Xogal Valley.

64 bulletin: museum of comparative zoology

Eremias brexneri Peters

Eremias brenneri Peters, 1869, Mon. Ak. Berlin, p. 432.

Eremias edwardsi Mocquard, 1888, Mem. Cent. Soc. Philom. Paris, p. 115,
pi. xi, fig. 1.

2 d> & Burao, 3500 ft.

rf", 9 Ado (45°15'E x 7°20'X) 2100 ft.

c? Milmil (43°40'E x 8°10'X) 3000 ft,

10 cf &, 12 9 9 Haud (46°20'E x8°15'X and 44°54'x8°42'X) 2100 and
3300 ft.
This series shows some slight variations beyond those already re-
corded. The dorsal scales which are always strongly keeled and often
tricarinate, vary from 58 to 72 and the femoral pores from 17 to 25.
It is most surprising, however, to find that in 11 of the 27 species
there are only three nasals, the lower being undivided, and in two others
the condition is asymmetrical and intermediate with three on one side,
and four on the other. A similar variation in the number of nasals,
which completely links the subgenera Lamperemias and Psenderemias
also occurs in E. striata and has been recorded in Eremias nitida quad-
rinasalis Chabanaud, by Loveridge (1937, p. 283). Another interesting
feature of the series is that yet again, there is a marked tendency to-
wards erythrism in specimens from the Haud, though it is by no
means as clearly marked as, for instance, in the genus Latastia. In the
specimen showing the condition most markedly the whole of the dor-
sum is brick-red, with white stripes and longitudinal rows of round
white spots, but without any trace of the usual darker stripes.

The species appears to have a range from the mountains south of
Berbera through the Ogaden and Italian Somaliland as far south as
Brava; the record from the Tana River (Stejneger, 1893) has been
shown by Loveridge (1929, p. 65) to be based on an example of E.
smithi.

Eremias striata Peters

Eremias brenneri var. striatus Peters, 1874, Mon. Ak. Berlin, p. 370.
3 cT cf,2 9 9 Haud, 46°20'Ex8°15'X, 2100 ft,
cf, 8 9 9 Haud, 44°54'E x 8°42'X, 3300 ft,

This species which has a geographical range not unlike that of the
preceding, is closely allied to it and not easily distinguishable. The
foregoing series makes it evident that the condition of the subocular,
whether entering or excluded from the labial margin, is by no means a
constant or satisfactory character, since it is excluded in 4 specimens

PARKER: LIZARDS OF BRITISH SOMALILAND 65

though entering in the remainder. The nasal condition, too, resembles
that of brenneri for though there are usually four shields surrounding
the nostril, one specimen has only three and another is asymmetrical
with 4 on one side and three on the other. These facts would seem to
indicate that striata cannot be maintained as a distinct species, but
there are certain differences utilised in the accompanying "key"
(p. 59) which have not yet been shown to be invalid and a conservative
attitude has been adopted until more definite evidence is forthcoming.

Eremias spekii sextaeniata Stejneger

Eremias sextaeniata Stejneger, 1894, Proc. U.S. Nat. Mus., 16, p. 718.
Eremias spekii sextaeniata Tornier, 1905, Zool. Jahrb., Syst., 22, p. 377.
Eremias spekii Scortecci, 1931, Atti Soc. Ital. Nat. Sci., 70, p. 145.

3 cf <?, 4 9 9 Haud at various localities between 46° and 46°25'E x
8°15' and 8°30'N at altitudes of from 2100-2400 ft.

2 9 9 Borama 43°05'E x 9°55'N, 4500 ft.

3 & d%3 9 9 Dakhato River 42°25'E x 9°10'N, 4700 ft.
c? Daghabur 43°30'E x 8°13'N, 3500 ft.

This series collected at very different altitudes and in different types
of country shows some local variations, of which the most pronounced
is found in the Haud. Here, in sandy country, at altitudes of less than
2500 feet, the scales are relatively larger, in 55 to 64 series at the middle
of the body, and in other examples from the eastern Haud and Nogal
Valley (Parker, 1932) the same condition is found. But in those from
the rocky uplands, inland of Berbera, around Borama, Daghabur and
Harrar, the scales are very much smaller and arranged in from 63 to 73
series. This difference is insufficient to warrant the nomenclatorical
recognition of a distinct race since examples of sextaeniata from north-
ern Kenya show a range of variation in the same character from 59 to
75 ; but it serves to emphasise, once again, the tendency towards differ-
entiation which is to be found in the animals of the Haud.

E. spekii sextaeniata is considered by Loveridge (1937, pp. 493 and
495) to be a form of the coastal plain and upland Savannahs in Kenya,
but its geographical range from the Gulf of Aden to northern Kenya
westwards into the Abyssinian highlands is typical of his Zone 4, the
Northern Desert.

66 bulletin: museum of comparative zoology

Latastia

The various subspecies, which grade into one another, cannot con-
veniently be treated in key-form so that they have not been considered
in this synopsis but are dealt with under specific headings.

I. Usually a group of small irregular scales in the middle of the pec-
toral region interrupting the regular linear arrangement of the
ventral plates.

A. Dorsal scales not larger than the laterals, smooth, tectiform or
keeled L. longicaudata

B. Dorsal scales larger than the laterals, strongly keeled. L. carinata

II. No group of small scales in the centre of the chest which is covered
by regularly arranged plates similar to those on the belly.

A. Frontal with a median groove; dorsal scales keeled or tectiform. .
L. boscai

B. Frontal without median groove; dorsal scales smooth and flat. . .
L. taylori sp. nov.

Latastia longicaudata Reuss

(fig- 3)

Lacerta longicaudata Reuss, 1834, Ber. Mus. Senck., 1, p. 29.

Lacerta samharica Blanford, 1870, Geology and Zoology Abyss., p. 449, fig.

Eremias revoili Vaillant, 1882, Miss. Revoil, Pays Comalis, Rept., p. 20 pi. iii,

fig. 2.
Latastia doriai Bedriaga, 1884, Ann. Mus. Genova, 20, p. 313.
Latastia caeruleopunctata Parker, 1935, Ann. Mag. Nat. Hist., (10), 16, p. 527.

This species is notoriously very variable, but a survey of the ma-
terial brought back by Capt. Taylor from all the boundaries of British
Somaliland does suggest that the variations are, to a large extent,
correlated with geographical distribution, and that many races can be
recognised. But whether or not all these can be named satisfactorily
is another matter, and, for the moment, it is only proposed to use
names which are already in existence.

In the coastal zone both in the east and west, at altitudes up to
3500 feet, there exists a race characterised by smooth dorsal granules,
11 to 14 femoral pores and a colour pattern based on the following
scheme :

Dorsum brown with 5 longitudinal series of small black or dark
brown spots which tend to fuse longitudinally, more especially the

PARKER! LIZARDS OF BRITISH SOMALILAND 67

mid-dorsal series; a broad dark brown dorso-lateral stripe or series of
spots from the ear to the tail, above the limbs, which may include a
series of blue ocelli ; flanks below this with one or two more longitudinal
series of spots; the outermost dorsal, the dorso-lateral and lateral
series show a strong tendency to form transverse (vertical) bars which
in juveniles may be very pronounced and intensified by black lines.
This is generally regarded as the typical form of the species (type
locality Tor, Sinai) though there are some differences between speci-
mens from the west (Guban) and the east. The latter have 8 longitu-
dinal series of ventrals, although the outermost series on each side is
small and rounded, and there are very numerous small scales in the
centre of the pectoral region whereas in those from the Guban the
pectorals are scarcely differentiated and there are only 6 rows of ven-
trals with, sometimes, traces of another row on each side.

On the eastern boundary of the country immediately to the south
of the Al Mado Mountains in the "Buran District" (Taylor in Parker,
1932, p. 335) is a form essentially similar to the preceding in its dorsal
pholidosis and colouring, but with 8 rows of ventrals, the outermost
large and rectangular, and 6-9 femoral pores. The area from which
these specimens came is in the general region of the type-locality of
Latastia rewili (Vaillant, 1882) and they agree very closely with the
description of that species. Immediately to the south of this zone, in
the region of the Sol Haud, Nogal Valley and the eastern Haud, there
is a tendency for the outermost ventral series to be reduced and the
dorsals to become tectiform.

An analogous, but slightly different, form occurs to the south of the
typical subspecies in the western boundary in the Ogo at altitudes of
from 3500 to 5000 feet. The dorsal scales and colouring are essentially
similar, though the latter is much more intense and defined, but the
femoral pores are fewer in number, varying from 4 to 7, 5 and 6 being
the commonest. If this montane race can be distinguished from revoili
the name doriai Bedriaga (cotype examined) may be available. It
should be pointed out that at altitudes of about 3500 ft. where the
mountains drop down to the coastal plain both doriai and the typical
form were taken side by side, but that, among a series of 12 of the
former and 7 of the latter, there were no intermediates, there being a
gap in the number of femoral pores between 7 and 10.

Between revoili on the east and doriai on the west, there occurs in the
Haud at altitudes of from 2100 to 3500 ft. a very differently coloured
race which also has some slight morphological differences. The dorsal
scales are more or less distinctly keeled, there are only six ventral

68 bulletin: museum of comparative zoology

series, and femoral pores vary from 6 to 10. In these characters there
seems to be complete intergradation with revoili but the most char-
acteristic difference is in the colour. The dorsum is bright, brick red
with 4 longitudinal series of small pale blue or whitish spots which may
also form transverse rows or be arranged quincuncially; rarely in
juveniles, there may be transverse black lines between the blue spots.
On the flanks are two or three longitudinal series of larger, darker blue
ocelli which may fuse to form bands ; more usually, however, there are
transverse black lines between the ocelli which fuse to form vertical
dark bars. This is the recently described caeruleopunctatus Parker,
and it appears to be confined to the northern and western Haud;
Neumann's (1905, p. 393) record of a red specimen from Modjo
(Western Ogaden) might be based on this form, and not, as Boulenger
suggests (1921, p. 28) on an erythristic individual. In the original
description of caeruleopunctata it was stated that it occurred side by
side with longicaudata and so the two could not be regarded as sub-
species. This statement was incorrect and it is nov found that all
specimens collected in the area indicated are of the red form but there
is somewhat more variability in colour pattern than was originally
described.

The distribution of the various races along the boundaries and adja-
cent territories of British Somaliland and the material examined is as
follows :

1. Latastia longicaudata longicaudata (Reuss)

d" ll°9'Nx49°E 200 ft. 26.XI.23

9 ll°5/Nx49°E 600 ft. 23.XI.29

juv. Between Hargeisa & Borama X.32

& 10°Nx43°E 1300 ft. 15.VI.34

9 10°30'N x 42°40'E 3500 ft. 22.111.33

2 & d\ 9 10°30'N x 42°45'E 3500 ft. 28.V-6.VII.33

juv., 9 ll°Nx43°E 1500 ft. 19.XII.33

d" ll°15'Nx43°15'E 500 ft. 28.1.34

2. Latastia longicaudata revoili (Vaillant)

2 c? c?\ 5 9 9 10°5'-10°50'N x 49°E 2375-3300 ft.

d" 9°14'N x 48°30'E 2200 ft.

6 c? d\ 5 9 9 8°20'-9°N x 48°5'-48°43'E 1400-1800 ft.

3c?d%39 9 8°-8°10 N x 48°-48°10'E 2000-2500 ft.

PARKER: LIZARDS OF BRITISH SOMALILAND 69

3. Latastia longicaudata doriai (Bedriaga)

9 9°50'N x 43°20'E 5000 ft. 22.XII.32

( 24X11.32
2 & cf, 9, juv. & 9°50'N x 43°10,E 5000 ft. 8-12.V.33

( 5.VI.33
cf, 9 9°55'N x 43°10'E 4500 ft. 27.VIII.-2.IX.34

<?, 9 9°55'N x 43°05'E 4500 ft. 26.1.33

9 10°Nx43°E 4500 ft. 2.X.33

9 10°20'N x 42°50'E 3500 ft. 29.VII.33

9 10°30'N x 42°40'E 3500 ft. 25..III.33

4. Latastia longicaudata caeruleopunctata (Parker)
<? 8°30'N x 46°25'E 2400 ft. 18.11.32

5 cf cf , 10 9 9 8°15/Nx46°20'E 2100 ft. {302]II32

9 8°Nx45°50'E 2500 ft. 26.XI.34

2juvs. 8°34'N x 45°18,E 2900 ft. 17.VI.32

6 cT d\ 9 7°20'N x 45°15'E 2100ft. 23.XI.34
tf, 9 8°37'N x 45°09'E 3050 ft. 26.VI.32
3(^^,29 9 8°42'N x 44°54'E 3300 ft. 12.XII.34

9 8°24'Nx44°E 3500 ft. 20.XII.34

2c?c?,2 9 9 8°10'N x 43°40'E 3000 ft. 20.XII.34

9 8°13'N x 43°30'E 3500 ft. 24.XII.34

2 9 9,2 juvs. Bohodle 2100 ft. 12.111.32

Latastia carinata (Peters)

Lacerta carinata Peters, 1874, Mon. Ak. Berlin, p. 368, pi. fig. 1.

1 Latastia ventralis Werner, 1917, Mitt. Zool. Mus. Hamburg, 34, p. 32.

This species, originally described from a single specimen from
Brava, Italian Somaliland, has since been recorded from the type-
locality and from Berbera (Boulenger, 1921, p. 32), from the Danakil
depression near Zeila, and from near Jaldessa (Tornier, 1905 and Neu-
mann, 1905). Latastia ventralis based on a single specimen obtained
by Hildebrandt, who also collected the type of carinata, is possibly a
synonym.

Latastia boscai Bedriaga
(fig. 4)

Latastia boscai Bedriaga, 1884, Ann. Mus. Genova, 20, p. 322.
Latastia burii Boulenger, 1907, Ann. Mag. Nat. Hist. (7), 19, p. 393.
Latastia wachei Werner, 1913, Jahrb. Hamb. Wiss. Anst., 30, p. 16.

70 bulletin: museum of comparative zoology

Boulenger, (1921, p. 22) and the author (1932, p. 355) in discussing
this species have referred to two very different colour-phases; but
there appears to have been confusion of at least two distinct forms,
one of which probably represents a distinct species (see Latastia taylori
sp.n.). The larger series now available throws considerable light on the
problem and also on the status of Latastia wachei Werner and Latastia
burii Boulenger. The whole of the material can be subdivided into the
following groups :

1.

Latastia boscai boscai Bedriaga

3 d1 <?,

2

9 9

43°25'E x 9°50'N

5000 ft.

5.xi.32-16.i.33

<?, ?

43°20'E x 9°50'N

5000 ft.

20.-21. xii.33

2 <? d",

9

43°10'Ex9°55'N

5500 ft.

8.1.-5.V.33

2 9 9

43°05'Ex9°55'N

4500 ft.

31.-1.2.ii.33

d\2 9

,9

43°E x 10°N

1500 ft.

19.xii.33

<?

43°10'E x 10°10'N

5000 ft.

6.U.33

3 c? <?,

9

42°50'E x 10°10'N

4500 ft.

6.13.-viii.33

2 <? d\

9

(preg.)

42°40'E x 10°30'X

3500 ft.

20.-24.iii.33

<?, 9

43°15'Exll°15'X

1000 ft.

31.i-5.ii.33

This series from the Ogo and the Guban, not far from the type-
localities of L. wachei Werner, has keeled or tectiform dorsal scales
arranged in from 36 to 48 rows across the body and from 8 to 11
femoral pores. The dorsum usually has 5 dark brown longitudinal
lines of which the outermost are the broadest, and these and the
median dorsal almost always extend onto the tail; the other two
dorsals are frequently indistinct or broken up posteriorly. The
flanks are black with three longitudinal rows of white ocellar
spots, the uppermost series commencing behind the eye and the
middle series below the eye on the lip. The lowermost, which is often
indistinct, runs from axilla to groin. These lateral spots tend to form
longitudinal lines but there is frequently a tendency for them to form
vertical bars also. The pattern varies considerably in intensity and in
three examples is so reduced that the dorsum is uniform grey and the
flanks are only relieved by faint traces of white spots.

This series certainly represents the form to which the name wachei
(from Hara and Diredawa) was given, but the range of morphological
variation is sufficiently great to include both boscai (from Eritrea) and
burii from near Berbera. The differences in the colour pattern between
boscai and wachei do not seem to be sufficiently marked to warrant
their retention as separate races, but in burii the ocellar lateral spots
are completely fused to form two white longitudinal lines and the

PARKER: LIZARDS OF BRITISH SOMALILAND 71

outermost two dorsal dark lines are reduced to short streaks just
behind the parietals. A short series of specimens from the north-east
of British Somaliland resembles burii except that the degree of develop-
ment of the outer dorsal stripes is variable and they may be quite as
extensive as in typical boscai. Accordingly burii cannot be regarded as
more than a race of boscai and the following have been examined:

2. Latastia boscai burii Boulenger

2 cf c? 400 ft. near Berbera Types.

o\ 9 1300 ft. Beretableh, Xogal Valley 7.iv.30.

cf 1400 ft. Xogal Valley, 8°43'N x 48°43'E 6.iv.30.

& 1000 ft. Al Mado Mts., 11°3'N x 40°47'E 20.xii.29.

Dorsal scales keeled or tectiform, in 40 to 45 series across the middle
of the body; femoral pores 10-14.

The last four .specimens are those referred by the author (1932,
p. 355) to L. boscai as series A. In the same paper a number of other
uniformly coloured specimens were referred to the same species as
series B. Of these 8 were considered to be immature, but it seems
probable that this was incorrect and that they represent a distinct,
smaller species, L. taylori. The two originally considered adult (specs,
n & o) appear to represent a southern race of boscai which may be
known as:

3. Latastia boscai arenicola subspec. nov.

Latastia boscae Boulenger, 1912, Ann. Mus. Genova, (3), 5, p. 330; idem, 1921,

Mon. Lacertidae, 2, p. 22 (part).
L. boscai Parker, 1932, Proc. Zool. Soc. London, p. 355 (part).

<?, 9 Haud, 1900 ft., 7°55'Nx47°50'E ll.v.30.

& Haud, 2100 ft., 8°15'N x 46°20'E 4.3.-9.iv.32.

<? Ado, 2100 ft., 7°20'X x 45°15'E 26.xi.34.

9 Dolo, Italian Somaliland, Coll. Citerni.

These specimens are the co-types. The race has the head shields
and proportions of the typical form, but has somewhat smaller dorsal
scales, which are in 45 to 48 series at the middle of the body and are
tectiform or keeled; femoral pores vary from 10 to 12. The colour
pattern is very different from that of normal individuals of either the
typical subspecies, or of burii. The dorsum is reddish brown, with, at
most, the merest traces of longitudinal stripes, which instead of being
darker than the background are a very pale blue, and the flanks and
sides of the head are marked with regular narrow vertical bars of

72 bulletin: museum of comparative zoology

black and white; these markings vary in their intensity and may be
almost completely absent or indicated by a dusky, dorso-lateral
smudge. Specimens such as these can hardly be distinguished from
occasional examples of the typical form such as have been mentioned
above, but where the colour pattern is fully developed the two are
markedly different.

The name indicates that this southern subspecies inhabits a zone of
sandy country with thick scrub, whereas both the northern races are
found in stony or rocky localities. This difference of habitat probably
accounts for another morphological difference; in arenieola the claws
are very long and acutely pointed, whereas in the other races from the
rocky or stony areas the claws are only about half as long and are
comparatively blunt. The colour differences are closely paralleled by
the conditions found in L. longicaudata. In both species the race
found in the Haud and Ogaden is distinctly reddish and dark dorsal
markings are absent or replaced by light ones. The specimens taken
in May were found in copula.

Latastia taylori spec. nov.

(fig. 4)

Latastia boscai (part) Parker, 1932, Proc. Zool. Soc. London, p. 355 (specs,
e-m).

Holotype a male, number 1931.7.20.337, in the British Museum,
from the Buran Valley, 2500 ft. (10°20'N x 49°E); collected by Capt.
R. H. R. Taylor, 17.X.1929.

Head flat, depressed, once and three quarters as long as broad, its
depth a little less than the distance between the tip of the snout and
the anterior corner of the eye, and its length contained 4.25 times in
the length from snout to vent. Nostril pierced between four shields;
upper nasals forming a suture half the length of the fronto-nasal which
is a little broader than long and broader than the internarial space;
prefrontals forming a median suture shorter than that between the
nasals; frontal not grooved, a little longer than its distance from the
rostral, once and two thirds as long as broad; interparietal not quite
twice as long as broad, in contact with an occipital half its length; 4
supraoculars, the first divided into two, second and third large and
subequal, fourth very small; a row of granules separating the supra-
oculars from the 5 supraciliaries. Lower eyelid scaly, translucent.
Rostral not entering the nostril ; two superposed post-nasals, the lower
in contact with the first and second labials ; anterior loreal half as long

PARKER: LIZARDS OF BRITISH SOMALILAND 73

as the second; five or six labials anterior to the subocular, which is
much narrowed on the lip and separated by two scales from the pos-
terior loreal ; lateral edge of the parietal bordered by 3 elongate, narrow
scales, of which the anterior is much the longest; anterior margin of the
ear bordered by 3 or 4 scales of which the uppermost is the largest.
Four pairs of chin-shields, the first 3 in contact and the last the largest;
31 gular scales between the chin-shields and the collar which has about
nine scales on its edge, the median very large and the laterals grading
into the granules of the neck. Dorsal scales oval, or subhexagonal, flat
and smooth, in 39 series across the middle of the body and in 105
series between the occipital and the base of the tail (vertically over the
vent) ; twenty-two in a transverse series between the hind limbs. Ven-
trals in six longitudinal series, with straight posterior borders, the two
median series much narrower than the others; no group of small pec-
torals; twenty-five transverse series of ventrals; one very large preanal
bordering the vent, preceded by another, but much smaller, shield. A
series of enlarged plates beneath the fore-arm; upper tibial scales
small, imbricate, keeled; ten or eleven femoral pores on each side; sub-
digital lamellae strongly bicarinate 26 beneath the fourth toe. Caudal
scales in equal whorls, oblique and strongly keeled above, smooth be-
neath. Tip of the fourth toe reaching to midway between the arm and
the ear.

Pale reddish brown above and on the tail, faintly marbled with grey
anteriorly and on the head; flanks anteriorly and side of the neck with
very irregular brown and greyish-white vertical marblings. Lower sur-
faces uniform white.

Length from snout to vent 43 mm.

Fore-limb 13 mm.

Hind-limb 26 mm.

Tail (regenerated in part) 87 mm.

The following specimens are paratypes of this species:
1931.7.20.339 9 2000 ft. 10°15'N x 49°E 13.1.1930

1931.7.20.338 9 20 ft. 11°14'N x 49°E 3.XII.1929

1931.7.20.340-342 3 & & 2000 ft. 9°40'N x 49°E

[Ex. Field Mus.j 8.II.1930
19317.20.335-336 d\ 9 3100-3300 ft. 10°13'N x 48°46'E 8.1.1930

This series shows the following variations from the holotype: The
head may be once and two thirds as long as broad, and its depth equal
to the distance from the snout to the eye; supraciliaries 5 to 7; one or
two scales between the posterior loreal and the subocular; 5 or 6

74 bulletin: museum of comparative zoology

labials anterior to the subocular; dorsals in 36 to 41 series across the
middle of the body; ventrals in 23 to 26 transverse series; gular scales
28 to 32; plates in the collar 5 to 7; femoral pores 9 to 12; subdigital
lamellae beneath the fourth toes 24 to 27. The fourth toe extends to
some point between the shoulder and the middle of the neck. The
colour is usually olive, almost uniform, but with traces of lighter
marblings anteriorly and on the sides of the neck and anterior part of
the flanks, the latter having a tendency towards the formation of verti-
cal bars. The subcaudal scales are smooth proximally, but keeled dis-
tally and an unregenerated tail is a little more than twice as long as the
distance from snout to vent.

These specimens were originally believed to be all immature, but a
female of 42 mm. from snout to vent is pregnant and the species ap-
pears to be consistently smaller than boscai. It is closely allied to the
latter but may be distinguished by its broader, natter dorsal scales,
the absence of a frontal groove and different colour; it appears to be
restricted to the north-eastern districts of Somaliland from the Sol
Haud to the coast, an area close to that in which the strongly striped
L. boscai burii also occurs (fig. 4).

Philochortus

I. Usually no granules between the supraoculars and the frontal;
dorsal and lateral scales in 28-46 rows at the middle of the body;
10-16 enlarged scales between the hind-limbs.

A. Parietals in contact, the interparietal small or absent, not reach-
ing the occipital. Scales usually smooth. Colour pattern of six
white lines, the median two bifurcating on the nape ... P. spinalis

B. Parietals separated by the interparietal and occipital which are in
contact. Scales keeled. Colour pattern of six white lines, the
median two bifurcating on the nape .... P. intermedins intermedins

C. Parietals usually separated by the interparietal and occipital,
rarely in contact; scales smooth or obtusely keeled. Colour pat-
tern of five white lines, the median bifurcating on the nape

P. phillipsi

II. Usually a more or less complete ring of granules between the
supraoculars and the frontal; dorsal and lateral scales in 22-28
(?30) rows at the middle of the body; keeled; 6-8 enlarged scales
between the hind-limbs. Colour pattern of five white lines, the
median bifurcating on the nape P. hardeggeri

PARKER: LIZARDS OF BRITISH SOMALILAND 75

1. Philochortus spinalis (Peters)

Lacerta spinalis Peters, 1874, Mon. Ak. Berlin, p. 369, pi. fig. 2.
Latastia spinalis Tornier, 1905, Zool. Jahrb., Syst., 23, p. 375.

2 cf tf1 Guban 42°40'E x 10°30'N, 3500 ft.

Previously this species has only once been recorded from British
Somaliland (Tornier, loc. cit.). also from the Guban, close to Zeilah;
otherwise it is Eritrean and Ethiopian in distribution1 only.

2. Philochortus intermedius intermedius Boulenger

Philochortus intermedius Boulenger, 1917, Proc. Zool. Soc. London, p. 152,

pi. ii, figs. 2 & 3.
Latastia hardeggefi (non Steindachner) Boulenger, 1898, Ann. Mag. Nat. Hist.,

(7), 2, p. 130; Anderson, 1901, Proc. Zool. Soc. London, 2, p. 145 (part).

2 d" d" Borama District 43°10'E x 9°55'N, 4500-5500 ft.

3 & d" , 2 9 9 Borama District 43°E x 10°05'-10°10'N, 5000 ft.

This series agrees closely with the types of the species from the
mountains inland from Berbera. The species is an inhabitant of stony
mountainous territory in the north and appears to range along the
coastal mountains as far as Berbera where it also enters the Guban;
it may also enter the Eastern Haud (see below) and extends south-
wards to the lower Juba (Calabresi, 1927), but is replaced in the Lake
Rudolf area by a closely allied subspecies.

Philochortus intermedius subspec?

cf Haud, 45°38'E x 8°28'N, 2500 ft.

This single specimen (Parker, 1932) is referred to the species inter-
medius on account of the number of keeled scales at mid-body, and
between the hind-limbs, the absence of granules between frontal and
supraoculars and the presence of an interparietal in contact with the
occipital granule. But there is no colour pattern of white lines to clinch
the specific status and there are notable differences from intermedins.
It seems probable, however, in view of the large numbers of other
lizards which are racially differentiated in the Haud, that this also is a
local race of intermedins. It has a typical Hand colour of rufous
brown, dotted with black and is remarkable among the species of
Philochortus for its very strongly keeled scales.

>Boulenger, 1909, Ann. Mus. Geneva, (3), 4, p. 310 records a "Latastia spinalis Peters" from
Bardera, Ital. Somaliland. This record is not repeated or quoted elsewhere in the Monograph
of the Lacertidae and is probably an error; Philochortus intermedius is the only species of the
genus known from Italian Somaliland. .

76 bulletin: museum of comparative zoology

3. Philochortus phillipsi (Boulenger)

Latastia phillipsi Boulenger, 1898, Ann. Nag. Nat. Hist., (7), 2, p. 131.
Philochortus hardeggeri taylori Parker, 1932, Proc. Zool. Soc. London, p. 354.
7 Philochortus intermedius Scortecci, 1931, Atti Soc. Ital. Sci. Xat., 70, p. 145.

Comparison of the large series of Philochortus hardeggeri, mentioned
below, with the co-type series of P. hardeggeri taylori Parker, and the
two cotypes of P. phillipsi Boulenger seems to indicate that the two
latter cannot be distinguished. P. phillipsi occurs together with
hardeggeri in the region of Berbera so that the two cannot be regarded
as subspecies. In morphological characters phillipsi is scarcely dis-
tinguishable from intermedius, but it has the 5-lined pattern of hardeg-
geri; the specimen determined as intermedius by Scortecci (loc. cit.
supra) came from El Donfar (49°04'E x 10°30'X) which is considerably
beyond the known north-eastern limits of typical intermedins, but very
close to the area from which P. hardeggeri was described and is prob-
ably therefore referable to phillipsi.

The species ranges from Berbera eastwards to Migiurtina and
southwards into the Sol Haud, Xogal Valley and eastern Haud. In
the first and last of these localities it overlaps the range of intermedins
and the two may ultimately be found to intergrade.

4. Philochortus hardeggeri (Steindachner)

Latastia hardeggeri Steindachner, 1891, Ann. Mus. Wien, 6, p. 371, pi. xi.

Eremias heterolepis Boettger, 1893, Zool. Anz., pp. 115, 193.

Latastia carinata (non Peters) Meek, 1897, Field Columb. Mus., Zool. (1), 8,

p. 181.
Latastia degeni Boulenger, 1903, Ann. Mag. Nat. Hist. (7), 11, p. 55.

juv. 9 Burao, 3500 ft,

7(?(? Haud, 2100-2700 ft,, 45°29'-46°20'E x 8°30'-8o15'N
9 Ogo, 3500 ft., 42°50'E x 10°20'N.

The last-mentioned female is almost topotypical for the species and
agrees well with the adult co-type especially in having the series of
granules between the supraoculars and the frontal incomplete. In
the series from the Haud and Burao, however, these granules always
form a complete series and like many other lizards from that region,
differ in a much lighter, rufous-brown colouration. If it should prove
possible and desirable to recognise a distinct race the name heterolepis,
type locality Lafarug, might be available; the type certainly has a com-
plete circle of granules.

The species, which seems to frequent sandy country, ranges along

PARKER: LIZARDS OF BRITISH SOMALILAND 77

the coastal mountains from the Danakil depression to Berbera and
into the western Haud. Tornier (1905) and Neumann (1905) also
record it from the coastal plain near Zeilah.

GERRHOS AURID AE *

Gerrhosaurus

I. Ventral scales in 8 longitudinal series; dorsum between the dorso-
lateral light lines, uniformly coloured G. flavigularis

II. Ventral scales in 10 longitudinal series; dorsum dark brown or

black with regular longitudinal series of yellow flocks

G. major bottegoi

Gerrhosaurus flavigularis Wiegmann

Gerrhosaurus flavigularis Wiegmann, 1828, Isis, p. 379.

This species has not actually been recorded from British Somaliland.
It is, however, a wide-spread species and has been recorded from Har-
rar (Tornier, 1905; Neumann, 1905) so that its presence in the country
and especially in the western mountains is to be expected. Loveridge
(1937) classes it as a creature of the coastal plains and upland sa-
vannahs.

Gerrhosaurus major bottegoi del Prato

Gerrhosaurus bottegoi del Prato, 1895, Atti Soc. Ital. Sci. Nat., 35, p. 19, fig.
G. major zechi Tornier, 1901, Arch. Naturg., 67, p. 74.

3 Ads. Borama District 42°45'-43°E x 10°10'-10°20'N 5000 ft.

These specimens appear to be indistinguishable in morphological
character from G. major, but have the coloration of bottegoi, of which
zechi Tornier appears, as suggested by Schmidt (1919, p. 519) to be a
synonym.2 According to the material, admittedly rather scanty, in the
British Museum, there are 3 recognisable races of G. major, as follows :

I. Uniform brownish above, or with only irregular dark markings;
tail with alternating darker and lighter annuli. This form is confined to
Kenya Colony, Zanzibar and northeastern Tanganyika Territory and
is the typical form. G. major major Dumeril.

Written prior to the revision of this family in Bull. Mus. Comp. Zool. ,89, pp 483-543, which
Mr. Parker, being engaged in war work, has not had the opportunity of seeing. A. L.

2Schmidt's (1919, p. 519) distinction between bottegoi and grandis of the yellow dorsal spots
being between the scales in the former but on the scales in the latter, is not tenable. They are
between the scales even in the type of grandis.

78 bulletin: museum of comparative zoology

II. Black above, with longitudinal series of yellow spots between
the dorsal scales; a more or less distinct yellow dorsolateral stripe;
flanks brown with longitudinal light flecks forming regular series.
Head black above with small yellow spots. This is a northern Sudan-
ese subspecies, ranging from the Gold Coast to Eritrea and Somaliland.
It enters the Savannahs of the Congo around Garamba (Schmidt,
1919, loc. cit), Uganda (Kyagwe and Kaiso) and probably north-
eastern Kenya (? U.S.N.M. 42216 recorded by Loveridge, 1929, p. 66).
In British Somaliland it appears to be confined to the mountains from
the Borama district as far east as the Golis Range: G. major bottegoi
del Prato.

III. Similar in colour posteriorly to the preceding, but anteriorly
the light markings are more extensive, obliterating the darker colour
and the whole of the upper surface of the head is pale brown, uniform
or with small black or chocolate-brown spots. This race occurs in
Zululand, Transvaal, Mozambique and Tanganyika Territory and
should apparently be known as G. major grandis Boul.

It seems very probable that all records of zechi from the latter area
(e.g. Cott, 1934, p. 165; Loveridge, 1933, p. 311) really refer to grandis
whilst records of zechi and major from the northern area are based on
specimens of bottegoi. The specimen from southern Italian Somali-
land (Villagio Duca degli Abruzzi) recorded by Scortecci (1931, p. 146)
may well, to judge from the description, be intermediate between the
northern bottegoi and typical major.

CHAMAELEONTIDAE

Chamaeleon

I. Casque strongly raised posteriorly, with a strong, median, parietal
crest; no occipital dermal lobes C. basiliscus

II. Casque scarcely raised posteriorly, almost flat above, with but a
feeble parietal crest. Well developed, movable, occipital lobes
present C. dilepis ruspolii

These are the only two true chamaeleons which have been recorded
from British Somaliland. But a number of other species, such as C.
gracilis, C. senegalensis, C. affinis and C. bitaeniatus are known from
Ethiopia and Italian Somaliland and some of these may ultimately

PARKER: LIZARDS OF BRITISH SOMALILAND 79

prove to occur there. C. affinis has been recorded from the Harrar
region (Tornier, 1905; Neumann, 1905) and may be expected in the
mountains of the Ogo.

1. Chamaeleon basiliscus Cope

Chamaelcon basiliscus Cope, 1868, Proc. Acad. Philad., p. 316.

<?, 3 9 9 along the boundary between 43°10'E x 9°55'N and 42°40'E x
10°35'N at altitudes of from 3000-5500 ft.

This is a species with an "Eremian" distribution which enters the
protectorate in the coastal zone and mountains as far east as Dagah
Shabell (SS.E. of Berbera).

2. Chamaeleon dilepis ruspolii Boettger

Chamaeleon ruspolii Boettger, 1893, Zool. Anz., p. 116.

3 cf <?, 4 9 9 , 5 juvs. from the Haud, between 45°43'E x 8°26'N and
44°44'E x 8°45'N, at altitudes of from 2350-3500 ft.

The status and distribution of the various races of Chamaeleon
dilepis still present an unsolved problem. Loveridge (1936, p. 330)
recognised only three races in eastern Africa, and finds considerable
difficulty in accounting for the distribution of some of these; this
difficulty may be, in part, due to the fact that some of his races are
really composites. Certainly the series of chamaeleons listed above
appears to represent a race distinct enough from the typical form to
deserve subspecific recognition. As Boettger (1893, p. 116) pointed
out in describing ruspolii from the Ogaden, the most characteristic
feature is the very large size of the scales on the head and particularly
on the occipital lobes. This character also appears in certain specimens
found in the highlands of Nyasaland {isabellinus Giinther) and this
subsequently led Boettger to place his ruspolii as a synonym of the
latter. But the two are apparently quite distinct, for the northern
specimens differ from isabellinus and from any other examples of
dilepis examined by the author, in their very feebly developed gular-
ventral crest. This is composed of scales a little larger than those
found elsewhere on the lower surfaces, but does not form the usual
pendant crest and is quite indistinct posteriorly.

This race is apparently confined to the northern part of the Ogaden
and Haud, and has not been recorded from the mountains or the
coastal plain. Specimens collected to the south of the Haud, in the
Jubaland-Tanaland area have small head scales and a prominent
gular-ventral crest.

80 bulletin: museum of comparative zoology

Rhampholeon kersteni robecchii Boulenger

Chamaeleo kerstenii Peters, 1868, Mem. Ak. Berlin, p. 449.

Rhampholeon robecchii. Boulenger, 1891, Ann. Mus. Genova, (2), 12, p. 13,

pi. i, fig. 3.
Rhampholeon mandera Meek, 1897, Field Columbian Mus., Zool., 1, 8, p. 183.

2 cf cf, 10 9 9 from various localities along the boundary in the
Haud and Ogo between 46°20'E x8°15'N and 43°E x 10°N at altitudes
of from 2100 to 4500 ft.

Loveridge (1933, p. 329) has shown that bicuspid or simple claws
and spinose or smooth scales on the soles of the feet can no longer be
used to distinguish the genera Brookcsia and Rhampholeon. But his
suppression of the latter genus in consequence may, perhaps, be pre-
mature. Werner, in his account of these two genera in "Das Tier-
reich" mentions a number of osteological characters which also may be
diagnostic but which Loveridge has not taken into account. Examina-
tion of the skeletons of the type species of Brookesia (superciliaris)
and of Rhampholeon (spectrum) shows some profound osteological
differences. (1) In B. superciliaris the nasal is large and unpaired and
the vertebrae are modified in a unique manner. The zygapophyses, are
carried on processes projecting laterally from the centrum, these
processes being unconnected only on the first twro cervicals. From the
last (third) cervical backwards almost to the end of the tail (the last ,
four excepted) a longitudinal arcade connects the two, enclosing a
foramen between itself and the centrum; this arcade from the third
dorsal backwards is apposed to the skin and its lateral face is heavily
tuberculated. In addition dorsals 3 to 12 inclusive have a process as-
cending and curving backwards and inwards from the prezygapophy-
sis, meeting and fusing with its fellow of the opposite side and con-
nected by a median longitudinal bar with the tip of the neural spine,
so that the whole centrum is enclosed within a basket-like framework
(fig. 8). The first 7 of these peculiar dorsals (i.e. 3 to 9 inclusive) have
a strong spine projecting laterally from the prezygapophysis and pene-
trating the skin, the spines becoming weaker caudad and in the most
posterior sometimes being absent from one side or the other; a similar
spine is developed on the sacrum which consists of two vertebrae more
or less fused. On the caudal vertebrae the lateral arcade is broadened
ventrally so as to come almost or quite into contact with the tip of the
transverse process. (2) In Rhampholeon spectrum on the other hand
the nasals are very small and paired whilst the vertebrae are of the
normal saurian type without any of the excrescences of Brookesia
superciliaris.

PARKER: LIZARDS OF BRITISH SOMALILAXD

81

Fig. 8. Dorsal vertebrae of Urookesia super ciliaris, seen from the side and
front.

Prezy. = prezygapophysis; po.zy. = postzygapophysis.

82 bulletin: museum of comparative zoology

Sufficient osteological material is not available to the author to de-
termine whether the various species of Brookesia and Rhampholeon fall
into two categories corresponding to the foregoing, but there seems to
be a reasonable probability that they do. Thus all the continental
species (spectrum, temporalis, platyceys, kersteni, marshalli, brevicauda-
tvs and boulengeri) have the nasals small and paired, or absent, whilst
there is no trace of the vertebral basket-like work. On the other hand
the Malagasy species super ciliar is, stumpffi, and ebenaui have large un-
paired nasals and vertebrae similar to those described above; B. den-
tata which has vertebral spines projecting through the skin may be
presumed to belong to this group also.

But the remaining Madagascar species {nasus, minima, tuberculata
and betsileana) have no such spines and have, moreover, smooth scales
on the soles of the feet, characters in which they agree rather with the
type of Rhampholeon.

In British Somaliland this small, grass-frequenting chamaeleon ap-
pears to be confined to the districts south of the coastal mountains
which it does not cross. To the south of the country it is wide-spread
through the Ogaden and is replaced in Kenya Colony by the typical
form; where the transition takes place has not been determined.

A peculiar and unexplained feature of this species is the preponder-
ance of females over males in collections; this may indicate that the
sexes have different habits or habitats.

SCINCIDAE

Ablepharus wahlbergii (Smith)

Cryptoblepharus wahlbergii A. Smith, 1849, 111. Zool. S. Afr., Rept., App. p. 10.

1 Ad. 42°50'E x 10°20'N 3500 ft.

Curiously enough this appears to be the first record of this common,
wide-spread species from British Somaliland, though it has been re-
ported from the Harrar region (Tornier, 1905; Neumann, 1905), from
the southern districts of Italian Somaliland and from the latter coun-
try close to the eastern border of the British protectorate, near
49°04/E x 10°30'N (Scortecci, 1931). Loveridge (1937) classes it as a
species found in the coastal plains and upland savannahs in Kenya,
but since it has been reported up to altitudes of 2900 metres in Abys-
sinia (Tornier, 1905; Neumann, 1905) as well as from the barren, stony
Darror Valley (El Donfar) it must obviously be an animal of very wide
tolerances.

PARKER: LIZARDS OF BRITISH SOMALILAND 83

Chalcides ocellatus (Forskal)

Lacerta ocellata Forskal, 1775, Desc. Anim., p. 13.

Chalcides ocellatus var. ragazzii Boulenger, 1890, Ann. Mag. Nat. Hist., (6),

5, p. 444.
Lygosoma akeleyi Meek, 1897, Field Columbian Mus., Zool. Ser., 1, 8, p. 181.
Chalicides bottegi Boulenger, 1898, Ann. Mus. Genova, (2), 18, p. 719, pi. x,

fig. 1.

30 specimens from various localities along the boundary northwards
and westwards from 43°20'E x 9°50'N, to the sea and around Zeilah,
at altitudes of from 0 to 5500 ft.

Scortecci (1928, p. 324) in discussing the variation of this species in
Eritrea arrives at the conclusion that until sufficient material is
available for a thorough and complete re-examination of the whole
species it is impossible to recognise any distinct sub-species in the
eastern part of its range. His remarks apply with equal force to the
present series, for the species appears to have a vertical range of 7000
feet from sea level and the various "varieties" bear no apparent rela-
tion to distribution. A curious fact of its general distribution in
Somaliland is its abundance in the coastal mountain and littoral
zones and its apparent absence from the interior and especially from
the Haud. This is probably to be correlated with the presence or ab-
sence of water for Flower (1933, p. 789) notes that in Egypt the lizard
is widely distributed wherever there is water, though this need not
necessarily be fresh or perennial.

Mabuya

I. Subocular either excluded from the edge of the lip or so much
narrowed that its lower margin is less than half the upper.

A. Two or three long, lanceolate scales on the anterior margin of the
ear; first supraocular in contact with or narrowly separated from
the frontal. A mid-dorsal and a pair of dorso-lateral light lines;
three large dark spots in a longitudinal row behind the ear.
Scales tricarinate in 30 to 34 series at the middle of the body
M.hildebrandtii

B. Ear-lobules, if present, short; first supraocular well separated from
the frontal; a pair of broad, dorso-lateral stripes but no median
dorsal; a dark lateral band from the eye to the groin. Scales
usually tricarinate, sometimes 5-7 carinate, in 32 to 40 series at
the middle of the body M. striata

84 bulletin: museum of comparative zoology

II. Subocular bordering the lip, not narrowed appreciably inferiorly,
so that its lower margin is always more than half as long as the
upper.

A. Scales on the soles smooth, but with one to three terminal spines;
subdigital lamellae with a strong median keel and sometimes with
traces of additional lateral keels ; dorsal scales bicarinate, laterals

tricarinate, in 30 to 34 rows at the middle of the body

M.brevicollis

B. Scales on the soles without keels or terminal spines; subdigital
lamellae smooth or obtusely keeled. Dorsal scales tri- or quinque-
carinate.

a. Scales in 25-32 rows at mid-body. A broad, light dorso-
lateral stripe, but no mid-dorsal stripe M.planifrons

b. Scales in 32 to 46 rows at mid-body. Juveniles with 5 light
stripes, one being median; in adults these stripes are lost
and the whole dorsum is almost uniform

M .quinquetaeniaia

In addition to the above species Mocquard (1888, p. Ill) has re-
corded Euprepcs (= Mabuya) comorensis from Somaliland, but this
may be erroneous. Mocquard was reporting upon collections from
Zanzibar and Somaliland and there is internal evidence in the paper
that confusion of the localities did occur; for instance on p. Ill Coron-
ella olivacea is recorded as from Somaliland, but on p. 28 it is definitely
stated that the single specimen of this species collected came from
Zanzibar.

1. Mabuya hildebrandtii (Peters)

Euprepes hildebrandtii Peters, 1874, Mem. Ak. Berlin, p. 372, pi. fig. 4. —
Mabuya varia hildebrandtii Scortecci, 1929, Atti Soc. Ital. Sci. Nat., 68,
p. 257; Calabresi, 1927, Atti Soc. Ital. Sci. Nat., 66, p. 30.

The exact status of this form and its presence in British Somaliland
are both uncertain. Loveridge (1923, Proc. Zool. Soc. London, p. 859
and 1929, p. 73) has expressed the opinion that hildebrandtii is a
synonym of varia, but Scortecci (loc. cit supra) prefers to use the name
subspecifically. The two are certainly very similar and Scortecci's
view may be the correct one; all records of hildebrandtii are from be-
tween the route from Obbia to Berbera (Boulenger, 1891) and the
Jubaland region, — Lugh (Boulenger, 1896c) Brava (idem, 1896b) and
Goscia (Scortecci, 1929); M. varia has not been recorded within this
area except by Calabresi (1927, p. 30) who was following Loveridge

PARKER: LIZARDS OF BRITISH SOMALILAND 85

in nomenclature, but whose specimens agreed with hildebrandtii.
Boettger (1893), however, records both species from the Webi (Shebeli)
Valley, and so, for the time being, it seems advisable to continue to
recognise hildebrandtii as a distinct species. It may be regarded as a
derivative of M. raria with particularly long auricular lobules, a modi-
fication which appears to have taken place independently in the super-
ficially very similar M. damarana (Peters) of Angola and S. W. Africa.
Boulenger's (1S91) record indicates the possibility of its occurrence in
eastern British Somaliland.

2. Mabuya striata (Peters)

Tropidolepisma striatum Peters, 1844, Mem. Ak. Berlin, p. 36.
Mabuya varia (non Peters) Meek, 1897, Field Columbian Mus., Zool., 1, 8,
p. 181.

14 specimens along the boundary in the Haud between 46°25'E and
43°10'E, at altitudes of 2100-4250 ft.
3 specimens from Burao, 3500 ft.
5 specimens from the Borama District, 4500-5000 ft.

This common, widely-distributed skink appears to be arboreal in
most districts and this habit may explain its comparative rarity in the
coastal plain from which it has only been recorded once and that on
rather unreliable authority (Meek, loc. cit.).

3. Mabuya varia (Peters)
Euprepes varius Peters, 1867, Mem. Ak. Berlin, p. 20.

1 specimen from the Haud between 44°24'-44°44'E and 8°45'-
8°52'X, 3500 ft .-3800 ft.

16 specimens from the Borama district between 42°50'-43°15'E and
9°50'-10°15'N, 4000-5400 ft.

In the north of its range this wide-spread species appears to be
rather a montane form; it has not been recorded from the coastal
plain; it is rare in the Haud and is replaced in the low-lying districts
of the eastern Ogaden and Italian Somaliland by M. hildebrandtii
(q.v.). This apparent preference does not hold, however, in territories
further to the south for Loveridge (1937) records it as a species of the
coastal zone, upland savannahs, the grassy uplands and alpine
meadows.

The single specimen from the Haud is aberrant in the possession of
only 28 scale-rows, a character only once previously recorded by

86 bulletin: museum of comparative zoology

Loveridge (1936a, p. 317) in a specimen from Gongoni, Kenya Colony.
Females collected in the Borama district in September were pregnant,
but as Loveridge (loc. cit) records females in this condition in Kenya
Colony in December, March and April, the breeding season appears
to be extended.

4. Mabuta brevicollis (Wiegmann)

Euprepes brevicollis Wiegmann, 1837, Arch. Naturg., p. 133.
Mabuya somalica Calabresi, 1915, Mon. Zool. Ital., 26, p. 242.

10 specimens from the boundary in the Haud between 46°25' and
45°09'E, 2100-3050 ft.

1 juv. Ado, Ogaden, 45°15'E x 7°20'N, 2100 ft.

<?, 9, juv. along the boundary between 42°50'E x 10°20'N and
42°45' E x 10°45'N, 2000-3500 ft.

Loveridge's suggestion that somalica is a northern race of planifrons
appears highly improbable, for somalica with 32 scalerows, bi- or
tricarinate dorsal scales, and a dorsal pattern of transverse dark zones
beset with white spots is quite unlike planifrons which seldom has
more than 30 scale rows, has tri- or quinque-carinate scales, and well-
marked dorso-lateral light stripes as its salient characters. Instead
somalica has so many features in common with brevicollis that it is
impossible to avoid the conclusion that they are synonyms. The only
morphological differences between the two, to judge from existing
descriptions, appear to lie in the number of supraoculars in contact
with the frontal and the condition of the subdigital lamellae. In the
present series the first of these characters is found to be very variable ;
in six specimens the third supraocular touches the frontal, in four
these two scales are separated and in one both conditions occur. The
subdigital lamellae of brevicollis are described as smooth, those of
somalica as unicarinate; actually they are unicarinate in brevicollis too,
as pointed out by Boulenger (1896b, p. 19) although in older indi-
viduals they may become smooth through wear, so that this character
also falls to the ground. There remains only colour, which is suffi-
ciently variable in brevicollis to include the described condition of
somalica. The white spots, so characteristic of the very young tend
to persist in transverse bands, more particularly in males than in
females, but are almost completely lost in the largest individuals
(over 150 mm. from snout to vent) of both sexes.

The juvenile taken in July measures only 36 mm. from snout to
vent and was, apparently, but newly hatched. The species was found

PARKER: LIZARDS OF BRITISH SOMALILAND 87

in dead trees and under piles of stones, and is widely distributed in all
parts of the protectorate. It ranges from the coastal plain and upland
savannahs of Kenya Colony and Tanganyika Territory (Loveridge,
1937) northwards to Eritrea and southeastern Arabia.

5. Mabuya planifrons (Peters)
Euprepes planifrons Peters, 1878, Mem. Ak. Berlin, p. 203, pi. ii, fig. 3.

1 Bohodle, in the Haud, 46°20'E x 8°15'N, 2100 ft.

2 Borama district, 5000 ft.

1 Ado, 45°15'E x 7°20'N, 2100 ft.

This species appears to be absent from the coastal plain but to be
distributed throughout the rest of the protectorate from which it
ranges southwards in the coastal plains and upland savannahs to
Tanganyika Territory and westwards through the mountains of
Abyssinia.

6. Mabuya quinquetaeniata quinqtjetaeniata (Lichtenstein)

Scincus quinquetaeniatus Lichtenstein, 1823, Verz. Double. Mus. Berlin, p. 103.
Mabuya semicollaris Werner, 1917, Mitt. Zool. Mus. Hamburg, 34, p. 33.

8 cf d" , 4 9 9,2 juvs. Borama Distr., 4000-5000 ft.
4 9 9 42°40'E x 10°30'N, 3500 ft.

2 cf cf 43°Exll'N, 1500 ft.

This species appears to be confined to the mountains as far east as
Berbera, but has not been recorded from the Haud or any part of the
area south of the mountains. These skinks from British Somaliland
agree closely with the typical quinquetaeniata from Egypt in possessing
32-38 scales at mid-body, the pre-frontals almost always (88 percent)
in contact to form a distinct suture and the throat of the male heavily
mottled with black; this is in marked contrast to the immaculate-
throated race of obsti of the countries south of Uganda and Kenya.
As is often the case, the mountains seem to have been subject to in-
vasion from the north and west and the present species provides further
evidence of this. M. quinquetaeniata has almost a complete circum-
rainforest distribution and extends northwards along the Nile Valley
to Lower Egypt; but although it is common in northern Uganda,
Kenya and Abyssinia, it appears to be completely absent from the
Ogaden and Italian Somaliland. Unless the species formerly had a
continuous distribution and has only recently become exterminated in
this region, the population of the montane region in British Somaliland

88 bulletin: museum of comparative zoology

must have been derived from the north-west and it is with examples
from this area that the Somali population shows the closest affinity.
Flower (1933, p. 758) concludes that the evidence of its distribution
in Egypt "points to its being a tropical African form which has entered
Egypt by the Nile and is still extending its distribution when oppo-
tunity offers".

Riopa

I. Snout not depressed; fronto-nasal in contact with the rostral;
nostril between the supra-nasal and two small nasals. Scales in
24-26 rows R. producta Boul.

II. Snout depressed, wedge-shaped; rostral separated from the
fronto-nasal.

A. Nostril between three shields. Two superposed preoculars

jR. simdevallii Smith

B. Nostril between two shields (i.e. the anterior and supra-nasals
fused). Two superposed preoculars.

1. Four labials anterior to the subocular; frontal very narrow

posteriorly; dorsal scales of juveniles tricarinate

R. laeviceps Peters

2. Three labials anterior to the subocular; frontal broader; dorsal
scales smooth R. modesta somalica

C. Nostril in a single nasal; a single preocular . R. vinciguerrae Parker

Riopa sundevalii (Smith)

Eumeces (Riopa) sundevallii Smith, 1849, 111. Zool. S. Africa, Rept., Appendix,

p. 11.
Lygosoma ferrandii Boulenger, 1898, Ann. Mus. Genova, (2), 18, p. 718 pi. ix,

%■ 2.

Ad. Haud, 4000 ft., 44°15'E x 8°55'N 28.ix.32.

Ad. Borama Distr., 5000 ft., 43°10'E x 9°55'N 9.V.33.

These two specimens have 24 scale-rows which appears to be the
normal condition in British Somaliland and the surrounding territo-
ries (Parker, 1932, p. 359); but it seems doubtful whether the use of
trinomials is justifiable to distinguish a subspecies in this area.

PARKER: LIZARDS OF BRITISH SOMALILAND

89

Riopa laeviceps (Peters)
(%. 9)
Euprepes (Tiliqua) laeviceps Peters, 1874, Mem. Ak. Berlin, p. 371, pi. fig. 3.
9 specs. Haud 2100 ft. 46°20'E x 8°15'N 10.ii.-30.iii.32.

2 '

' " 2200 ft.

46°10'Ex8°15'N

8.iv.32.

1 '

' " 3100 ft.

45°04'E x 8°39'N

14.vii.32.

1 '

' " 2300 ft.

45°58'Ex8°21'N

ll.viii.32.

2 '

' " 2250 ft.

46°04'E x 8°19'N

12.viii.32.

3 '

' " 2200 ft.

46°09'E x 8°17'N

12.viii.32.

2 '

' Ado 2100 ft.

45°15'E x 7°20'N

23-26.xi.34.

These 20 specimens, compared with other specimens collected in the
mountains of northern Somaliland, show conclusively that Boulenger
(1896, p. 20) was wrong in uniting laeviceps and modesta and the author
(1932, p. 360) was equally at fault in regarding the former as a northern

Fig. 9. Head scales of Riopa laeviceps Peters.
Adult measuring 80 mm. from snout to vent.

race of modesta; the above series, which agrees admirably with Peters's
description of laeviceps, represents a distinct species which is recogniz-
able by its longer, less depressed snout, larger fronto-nasal, much
longer and narrower frontal, the presence of 4, instead of 3, labials an-
terior to the sub-ocular, and the tricarinate scales in the young. The
head shields are very constant in all the specimens examined, and the
differences in proportions are readily appreciable by a comparison of
the accompanying figures (figs. 9 & 10).

90

bulletin: museum of comparative zoology

The series shows little variation and the original description is ade-
quate in most respects. There may, however, be 26 or 28 scales at the
middle of the body, and the dorsals are tricarinate only in juveniles.

Fig. 10. Head scales of Rio-pa modesta somalica Parker.
Cotype 1937.12.5.747, measuring 65 mm. from snout to vent,
drawn to the same scale as Fig. 9.

The largest specimen measures 83 mm. from snout to vent and 3
females collected at the end of March have each two embryos in an
advanced state of development. The species is cryptozoic, being found
under logs and stones or in termites nests by day.

Riopa modesta somalica subspec. nov.
(Fig. 10)

Lygosoma modestum Boulenger, 1895, Proc. Zool. Soc. London, p. 535; idem,

1895, Ann. Mag. Nat. Hist., (6), 16, p. 168.
Lygosoma modestum laeviceps (non Peters) Parker, 1932, Proc. Zool. Soc.

London, pp. 358, 360.

As pointed out above, the name laeviceps Peters is not available for
the northern form of modesta, and a new one appears to be necessary;
the name sphenopiforme is, apparently, based on an aberrant specimen
of typical, southern modesta. The northern race is represented in
the British Museum and the Museum d'Histoire Naturelle, Paris, by
the following specimens which are the co-types of the name somalica:

PARKER: LIZARDS OF BRITISH SOMALILAND 91

1937.12.5.746. Juv. 4500 ft. 42°50'E x 10°10'N 6.viii.33.

Capt. Taylor

1937.12.5.747. Ad. 5500 ft. 43°E x 10°05'X 29.viii.32.

Capt. Taylor
95.6.14.25-26. Ad. & juv. Inland of Berbera Lort-Phillips

1905.10.30.97-98. 2 Ads. Near Berbera Bury

1905.11.7.30-31. Ad. & juv. Wagga (3—4000 ft.) Golis Mts. Bury
Mus. Paris 1918.5. Ad. Near Berbera Bury

To judge from this material the subspecies is confined to the moun-
tains and possibly also the littoral zone in British Somaliland, whereas
the typical form occurs in central, eastern and northern Tanganyika
Territory, eastern Kenya and southern Italian Somaliland.1 Whether
the two meet and intergrade in the intervening territory is uncertain
but appears unlikely as this region is occupied by Riopa laeviceps.
Since, however, the differences between the northern and southern
forms, though clear, are slight they are ranked only as subspecies until
further material is forthcoming.

Head shields as in Riopa vwdesta except that the frontal is constantly
as long as or a little longer, instead of shorter, than the fronto-parietals
and parietals together, and in 1937. (juvenile) the prefrontals and
posterior loreals are fused; scales smooth, in 26 or 28 rows at mid-
body; limbs slightly longer than in the typical form, the posterior
contained 3.4 to 4.6 times in the length from snout to vent; outer toe
much longer, with 7 — 9 (vice 5 — 6) subdigital lamellae, extending as
far, or nearly as far, as the second. Largest adult S5 mm. from snout
to vent. Juveniles pale brown, spotted with darker brown above,
the spots forming longitudinal series, and with a broad, white-spotted,
dark, lateral band. With increasing age these markings become less
denned and the adult is pale brown with dark spotting on the head
and indefinite small lighter and darker spots dorsally. Lower surfaces
uniform white.

Riopa vinciguerrae (Parker)

17 specs. Haud 2100 ft. 46°20'E x 8°15'N 26.i.-31 .iii.32.

3 specs. Haud 2400 ft. 46°25'E x 8°30'N 17.ii.32.

1 spec. Haud 2300 ft. 45°58'E x 8°21'X 10.viii.32.

2 specs. Haud 2200 ft. 46°09'E x 8°17'X 13.viii.32.
1 spec. Haud 2100 ft. 46°19'E x 8°14'X 17.viii.32.
1 spec. Haud 2400 ft. 45°49'E x 8°24'N 26.viii.32.
1 spec. Ado 2100 ft. 45°15'E x 7°20'N 23.xi.34.

^oulenger, 1895c records "Lygosoma modestum" from Sheik Hussein (40°50'E x 7°45'N),
in Eastern Abyssinia adjacent to the Ogaden. Possibly it is in this region, to the west of the
area occupied by R. laeviceps, that the two subspecies of R. modesta meet.

92 bulletin: museum of comparative zoology

This large series shows no appreciable differences from the types,
though there may be 22 to 24 scales around the middle of the body.
A common abnormality is the fusion of the prefrontals and posterior
loreals which occurs on both sides in 8 specimens and on one side only
in 3, 15 being normal. The species appears to range over the whole of
the Haud and Ogaden regions from 8°30"N in the north, to Lugh in
the south.

APPENDIX

By Capt. R. H. R. Taylor, O.B.E.

A. Topographical Information

The 'Horn of Africa' (of which British Somaliland is a part) may be
divided into three distinct tracts of country:

(1) The fringe of maritime plain between the mountains and the sea.

(2) The maritime mountains running almost parallel to the coast
and -often intersected by inland plains.

(3) The raised plateau to the south with subsidiary hills lining the
water drainage.

British Somaliland is further divided by the Sbmals into the follow-
ing areas, which have distinctive natural features.

(a) The maritime plain and maritime mountains, known as the
Guban. The maritime plain stretches inland from Zeila to a depth
of 60 miles but soon narrows to an average depth of 7 or 8 miles by
Bulhar and Berbera. Thence to the Italian frontier it extends rarely
more than 2 miles inland. Generally speaking the plain is sandy with
a scanty vegetation of a stunted character. Evergreen bushes, how-
ever, are frequently seen growing on the banks of the dry river beds
which intersect the plain on their way to the coast. A certain amount
of grass is to be found in the Zeila plain away from the coast.

The maritime hills south and west of the Zeila plain are a confused
mass of low table-topped plateaux of black trap rock, with tufts of
grass and occasional bushes growing between the thickly strewn
boulders. South of Bulhar the hills are of greater height and are
broken up into a number of parallel limestone ridges. They are bare,
stony and precipitous, and are cut by deep and narrow river gorges.
These gorges are thickly tree'd with thorn. Within the ranges and
between them and the northern edge of the interior plateau are un-
dulating plains, intersected by broad sand-rivers running between
alluvial banks and a jungle of thick thorn. Between the rivers are

PARKER: LIZARDS OF BRITISH SOMALILAND 93

occasional patches of grass but the ground is usually of a gravelly and
stony nature supporting a few thorn trees.

East of Berbera the maritime ranges merge into one another, re-
taining however the same character. Some distance beyond, they
combine with the northern crest of the interior plateau to form an
irregular range, 145 miles long, of limestone hills, which limit the
maritime plain to a breadth varying from 200 yards to 2 miles.

(b) The northern crest bounding the interior plateau runs from Beyu
Anod in a south-south-easterly direction to Hargeisa, whence it
strikes in a north-easterly direction under the names of the Asa and
Golis ranges. It is then broken by the Huguf plain, and afterwards
runs slightly north of east to the Italian frontier.

From Beyu Anod, 2000 feet, the crest rises rapidly and as far as the
mountains west of Hargeisa the average height is about 5000 feet.
On the northern slopes grow cedar, box, accacia and euphorbia.

Near Hargeisa the crest is lower and the northern slope which is of a
terraced nature, affords good pasturage and is known as the Ogo
Guban.

From the Ogo Guban of Hargeisa, the crest rises gradually through
the Asa, 3000 to 4500 feet, to the Golis range, 6000 to 7000 feet.
After a break in the crest by the Huguf plain, the Golis reappears as
the Warsangli mountains which form a uniform ridge some 6000 to
7000 feet high, although near Erigavo they ascend to nearly 8000 feet
to form the Surud Ad mountains. The Warsangli and Golis moun-
tains have a steep northern face which in places is formed by a series
of precipitous steps. Cedar and box are to be found on the higher
slopes and the whole is thickly covered with vegetation.

(c) The southern slope of the northern crest bounding the interior
plateau consists of a strip from 10 to 30 miles in width of grassy downs
or thorn covered wilderness. The slopes are usually gentle. The
central part of this strip running across British Somaliland is known as
the Ogo.

(d) That portion of the Daror valley (in the north-east of the
country) which lies in British Somaliland, is of an arid nature and the
surface for a great part consists of anhydrite beds (gypsum).

(e) The Nogal valley is formed by the junction of two main afflu-
ents, the northernmost rising in the Golis range and the southernmost
in the southern slopes of Bur Dab. The valleys of these affluents
afford good pasturage and are studded with thorn bush scrub. As the
valleys widen, the bush becomes more sparse although good pasturage
is still afforded for some distance after the junction of the two affluents.

94 bulletin: museum of comparative zoology

Near the Italian frontier, the valley is arid and the surface instead of
being of a limestone formation is of anhydrite beds (gypsum).

(f) The Haud may be said to be that part of the interior plateau
running south of east from the line Jijiga-Jifu Meider to a line joining
the geographical co-ordinates L 49°E x 8°N and L 45°E x 6°N, and
bounded on its northern side by the Nogal valley and on its southern
by the Jerer and Fafan rivers. The area is undulating and is waterless
in the dry season with the exception of. a few scattered wells. The
ground falls evenly from NW to SE, from 5500 feet between Jijiga and
Jifu Meider to 400 feet at Galadi. The surface is either a red clay or
red sandy soil and occasional outcrops of limestone rock occur. The
red colour of the soil is a distinct feature of the area and it is particu-
larly noticeable in the central portion. The type and density of the
vegetation varies considerably throughout the area, from impenetrable
thorn jungle, to sparse thorn scrub with an undergrowth of aloes, to
open grass plains. The bush in the south-eastern half of the Haud is,
on the whole, thicker than that in the north-western half where the
grassy plains appear more often and are larger in size. Along the
boundary with Italian East Africa where this passes through the Haud,
it is noticeable that the dominant bushes and trees tend to be lower in
height east of about L 45°E x 8°40'N than the dominants west of that
point.

(g) The area between the Daror and Nogal valleys, known as the
Sol Haud, is uniformly level and at about 3000 feet above sea level.
Its vegetation is similar to that existing on the Haud.

B. Climatic Information

There are four main seasons in British Somaliland.

(1) The dry season (Jilal) which lasts from November to March.
During this period there is practically no rain. The north-east monsoon
blows during this period but is generally a mild wind. The nights are
comparatively cold up country and the days, even on the coastal
plains, not excessively hot. Several degrees of frost have been re-
corded on the frontier north-east of Jijiga.

(2) April is a hot and sultry month. The principal rainy season
during the year occurs during May and June. These rains are known
as the "Ju".

(3) The "Ju" rains are succeeded by the south-west monsoon, the
wind being called locally the "Kharif". The "Kharif" generally blows
from the beginning of June to the beginning of September. It is a
strong wind and raises dust storms in any area where there is little

PARKER: LIZARDS OF BRITISH SOMALILAND

95

vegetation ; it is particularly bad in the coastal area. The period when
this wind blows is the hot season and is called by the Somals "Haga".

(4) October is a hot and sultry month like April, and the "Dair"
rains fall about this time. These rains are very variable, sometimes
there is a fairly good rainfall and at other times only one or two
showers.

Rainfall: The rainfall in the country is very local and for that
reason it generally happens that there is great variation in local condi-
tions as to grazing, rainpools, etc. Also it does not behave in a strictly
regular manner from year to year. Sometimes the rains start early
and sometimes late. Some years they are heavier and in others very
considerably less than the average. This will be noticed on an examina-
tion of the table of rainfalls at several stations over a period of years
which is given below.

Year

Berbera

Burao

Sheikh

Hargeisa

Zeila

Borama

Erigavo

ins.

ins.

ins.

ins.

ins.

ins.

ins.

1919

2.31

—

—

—

—

—

—

1920

2.29

—

—

—

—

—

—

1921

1.35

3.85

—

—

—

—

— ■

1922

0.79

—

16.64

15.28

—

— ■

—

1923

2.86

9.90

47.14

16.86

—

—

—

1924

0.98

8.96

23.51

18.06

0.56

—

—

1925

0.56

11.89

20.59

16.39

2.04

23.24

12.62

1926

6.54

12.30

27.38

31.91

10.81

27.79

18.34

1927

2.02

9.78

11.70

14.58

8.42

19.79

15.29

1928

1.09

9.07

19.90

12.45

3.08

15.99

10.50

1929

0.37

9.25

25.04

17.07

1.81

16.42

11.90

1930

4.63

6.50

27.02

20.67

8.83

22.55

12.84

1931

1.43

9.79

19.43

25.35

0.66

20.24

10.53

1932

2.20

5.58

18.79

15.28

8.12

25.99

11.07

1933

0.86

5.58

17.07

12.42

2.05

15.90

9.81

1934

1.70

5.67

18.17

15.19

5.46

18.37

12.53

1935

6.20

7.26

16.58

18.26

3.29

25.99

13.53

No. of

years

rainfall

17

14

14

14

12

11

11

was re-

corded

Average

annual

rainfall

2.25

8.24

22.07

17.84

4.59

21.12

12.63

for this

period

96

bulletin: museum of comparative zoology

Temperature

The record given below of maximum and minimum temperatures
at certain stations, shows the conditions that prevail in the different
zones in the country.

The figures shown are the mean of the observations of nine years
(1925 to 1933 both inclusive). These observations vary little from
year to year.

Mean ,

Mean

Absolute

Absolute

Station

Maximum

Minimum

Maximum

Minimum

degs. F.

degs. F.

degs. F.

degs. F.

Maritime region :

Berbera

93

78

Ill

62

Northern edge of the in-

terior plateau :

Sheikh .

80

56

92

36

Interior plateau:

Hargeisa

85

57

95

33

Borama

81

56

92

36

Burao

85

63

95

48

PARKER: LIZARDS OF BRITISH SOMALILAND 97

BIBLIOGRAPHY

Anderson

1898. Zoology of Egypt, I, Rept. & Batr.

Andersson

1910. "A new Species of Hemidactylus from Harrar, Abyssinia". Jahrb.
Nassau Ver. Nat., 63, pp. 200-205.
/ 1912. "Notes on Hemidactylus tropidolepis Mocquard", Op. cit., 65,
pp. 227-231.

Bedriaga

1884. "Die neue Lacertiden-Gattung Latastia und ihre Arten". Ann.
Mus. Genova, 20, pp. 307-324.

BOETTGER, O.

1893. "Ubersieht der von Prof. C. Keller anlasslich der Ruspolischen
Expedition naeh den Somalilandern gesammelten Reptilien and
Batraehier". Zool. Anz., pp. 113-133.

BOULENGER, G. A.

1885. "Description d'une espece nouvelle d Agame". Ann. Mus.
Genova, (2), 2, pp. 127-128.

1891. "On some Reptiles collected by Sig. L. Brichetti Robecchi in
Somaliland". Op. cit., (2), 12, pp. 5-15, pi. i.

1895. "Esplorazione del Giuba e dei suoi Affluenti compinta dal Cap- V.
Bottego durante gli anni 1892-93 sotto gli auspicii della Societa
Geografica Italiana. Risultati Zoologici II; Rettili e Batraci".
Op. cit., (2), 15, pp. 9-18, pis. i-iv.

1895b. "On the Reptiles and Batrachians obtained by Mr. E. Lort-

Phillips in Somaliland". Ann. Mag. Nat. Hist., (6), 16, pp. 165-

169, pi. vii.
1895c. "An Account of the Reptiles and Batrachians collected by Dr. A.

Donaldson Smith in Western Somaliland and the Galla Country.

Proc. Zool. Soc. London, pp. 530-540, pis. xxix, xxx.

1896. "A List of the Reptiles and Batrachians collected by the late Prince
Eugenio Ruspoli in Somaliland and Gallaland in 1893". Ann.
Mus. Genova, (2), 17, pp. 5-14.

1896b. "Report on Capt. Bottego's second collection of Reptiles and
Batrachians from Somaliland". Op. cit., (2), 17, pp. 15-23, pi. i.

1896c. "A List of the Reptiles and Batrachians Collected by Dr. Ragazzi
in Shoa and Eritrea". Op. cit., (2), 16, pp. 545-554.

1896d. "Second Report on the Reptiles and Batrachians collected by
Dr. A. Donaldson Smith during his Expedition to Lake Rudolf."
Proc. Zool. Soc. London, pp. 212-217, pis. vii-viii.

1898. "Concluding Report of the late Capt. Bottego's Collection of
Reptiles and Batrachians from Somaliland and British East
Africa". Op. cit., (2), 18, pp. 715-723, pis. ix-x.

98 bulletin: museum of comparative zoology

1898b. "On a second Collection of Reptiles made by Mr. E. Lort-Phillips

in Somaliland". Ann. Mag. Nat. Hist., (7), 2, pp. 130-133.
1901. "A List of the Batrachians and Reptiles obtained by Dr. Donald-
- son Smith in Somaliland in 1899". Proc. Zool. Soc, London,

pp. 47-49, pi. vii.
1907. "Descriptions of Two new African Lizards of the Genus Latastia".

Ann. Mag. Nat. Hist., (7), 19, pp. 392-394.
1909. "On the Reptiles and Batrachians collected by the Tancredi

Expedition to Lake Tsana, Abyssinia." Ann. Mus. Civ. Genova,

(3), 4, p. 193.
1909. "List of the Reptiles collected by Capt. G. Ferrari at Jumbo,

Lower Juba." Op. cit., (3), 4, pp. 308-309.
1909. "List of the Reptiles and Batrachians collected by Capt. U.

Ferrandi at Bardera". Op. cit., (3), 4, pp. 310-311.
1912. "Missione per la Frontiera Italo-Etiopica sotto il commando del

Capitano Carlo Citerni". Op. cit. (3), 5, pp. 329-332.
1917. "On the Lizards of the Genus Philochortus Matschie". Proc. Zool.

Soc. London, pp. 145-157, pis. i-ii.
1920'l. Mon. Lacertidae, land 2 (London).

Brooks

1922. Evolution of Climate. (London).

Calabresi

1915. "Contribute alia conoscenza dei Rettili della Somalia". Mon.
Zool. Ital., 26, pp. 234-237.

1923. "Anfibi e Rettili dell' Africa Orientale raccolti durante le spedi-
zioni Franchetti e Zammarano". Atti Soc. Ital. Sci. Nat., 62, 2,
pp. 145-163.

1927. "Anfibi e Rettili raccolti nella Somalia dai Proff. G. Stefanni e
N. Puccioni". Op. cit. 66, pp. 14-60.

Carpenter

1935. "The Rhopolocera of Abyssinia". Trans. R. Ent. Soc. London,
78, 3, pp. 313-440.

COTT

1934. "The Zoological Society's Expedition to the Zambesi, 1927, No. 5.
On a Collection of Lizards mainly from Portuguese East Africa,
with descriptions of new Species of Zonurus, Monopehis and
Chirindia." Proc. Zool. Soc. London, (1), pp. 145-173, pis. i-iii.

Flower, S. S.

1933. "Notes on the Recent Reptiles and Amphibians of Egypt, with a
list of the Species recorded from that Kingdom." Proc. Zool.
Soc. London, (3), pp. 735-851.

parker: lizards of british somaliland 99

Fuchs

1939. "The Geological History of the Lake Rudolf Basin, Kenya
Colony". Phil. Trans. R. Soc. London, (8), 229, pp. 219-274.

Gestro and Vinciguerra

1932. in Luigi, Duca degli Abruzzi, "Esplorazione della Uabi Uebi
Scebeli".

Gregory

1921. The Rift Valleys and Geology of East Africa (London).

Loveridge

1923. "A List of the Lizards of the British Territories in East Africa".

Proc. Zool. Soc. London, pp. 841-863.
1929. "East African Reptiles and Amphibians in the U.S. National

Museum". Bull. U.S. Nat. Mus., 151.

1935. "New Geckoes of the Genus Lygodactylus from Somaliland, Sudan,
Kenya and Tanganyika". Proc. Biol. Soc. Washington, 48, pp.
195-200.

1936. "African Reptiles and Amphibians in the Field Museum of Natural
History". Publ. Field Mus. Nat. Hist., Zoology, 22, 1, pp. 1-111.

1936. "Scientific Results of an Expedition to Rain Forest Regions in
Eastern Africa, V. Reptiles". Bull. Mus. Comp. Zool. Harvard,
79, 5, pp. 209-337, pis. i-ix.

1937. "Scientific Results of an Expedition to Rain Forest Regions in
Eastern Africa, IX, Zoogeography and Itinerary". Op. cit., 79,
9, pp. 481-541, pis. i-iv.

Macfadyen

1932. "The late Geological History of British Somaliland". Nature,
130, pp. 433-434.

1933. The Geology of British Somaliland. I. Geology and Palaeontology.
(London).

Meek

1897. "List of Fishes and Reptiles obtained by Field Columbian Museum
East African Expedition to Somaliland in 1896". Field Col. Mus.,
Zool. Ser., 1, 8, poibl. 22, pp. 165-183. (For corrections see
Loveridge 1936.)

MOCQUARD

1888. "Sur une Collection de Reptiles et de Batraciens rapportee des
Pays Comalis et de Zanzibar par M. G. Revoil". Mem. Soc.
Philom. Centennaire, pp. 109-134, pis. xi-xii.

Moreau

1933. "Pleistocene Climatic Changes and the Distribution of Life in
East Africa". Journ. Ecology, 21, pp. 415-435.

100 bulletin: museum of comparative zoology

NlNNI

1933. "L'Aporoscelis princeps (O'Shaugh). a Ras Fafun Migiurtuna-
Somalia Italiana". Natura, 24, pp. 134-135.

Neumann

1905. "tJber nordost-afrikanische und arabische Kriechtiere". Zool.
Jahrb., Syst., 22, pp. 389-404.

Parker

1930. '''Three new Reptiles from Somaliland". Ann. Mag. Nat. Hist.,

(10), 6, pp. 603-606.
1932. "Two Collections of Reptiles and Amphibians from British

Somaliland". Froc. Zool. Soc. London, pp. 335-367.
1932. "Scientific Results of the Cambridge Expedition to the East

African Lakes, 1930-1931. 5. Reptiles and Amphibians". Journ.

Linn. Soc. London, Zool., 38, pp. 213-229.

1935. "Two new Lizards from Somaliland". Ann. Mag. Nat. Hist.,
(10), 16, pp. 525-529.

1936. "Reptiles and Amphibians collected by the Lake Rudolf Rift
Valley Expedition, 1934". Op. cit., (10), 18, pp. 594-609.

1938. "A new Gecko from Sokotra". Op. cit., (11), 1, pp. 305-307.

Parkinson

1932. "Climatic Changes in British Somaliland". Nature, 129, p. 651.

Peel

1900. Somaliland. (London).

Peters

1882. "IJber eine neue Art and Gattungder Amphisbaenoiden, Agamodon
anguliceps, mit eingewachsenen Zahnen, aus Barava (Ost Afrika)
und liber die zu den Trogonophides gehorigen Gattungen".
Sitz. Ber. Ak. Wiss. Berlin, 26, pp. 579-584, pl.x.

Roux

1935. Mission Scientifique de l'Omo, III, Zool., Rept. and Amph.,
pp. 157-190.

SCORTECCI

1928. "Rettili dell' Eritrea esistenti nelle collezioni del Mus. Civ. di
Milano". Atti Soc. Ital. Sci. Nat., 67, pp. 290-336.

1929. "Frimo Contributo alia Conoscenza dei Pettili e degli Anfibi della
Somalia Italiana". Op. cit., 68, pp. 245-279, pi. xii.

1930. "Contributo alia conoscenza dei Rettili e degli Anfibi della
Somalia, dell' Eritrea e dell' Abissinia". Boll. Mus. Torino, (3),
41, no. 10.

1931. "Secundo Contributo alia conoscenza dei Rettili della Somalia
Italiana". Atti Soc. Ital. Sci. Nat., 70, pp. 127-152, pi. iii.

PARKER: LIZARDS OF BRITISH SOMALILAXD 101

1933. "Osservazione su di uno strano Agamide della Migiurtinia".

Natura, 24, pp. 92-97.
1935. "Richerche Zoologiche e Questioni Zocgeographice nella Somalia

italiana". Rassegua Economica dell Colonie, 9-10, (Roma).
1935. "II Genere Pristurus nella Somalia italiana". Atti Soc. Ital.

Sci. Nat., 74, pp. 118-156.

Steixdachxer

1891. "Uber neue and seltene Laeertiden aus den Herpetologischen

Sammlungen des K.K. Xaturhistorischen Hofmuseums". Ann.

Nat. Mus. Wien, pp. 371-378, pis. xi-xii.
1906. "liber Homopholis erlangeri (n. sp.) aus Abessinien" etc. Ann.

Nat. Hist. Hofmus. Wien, 21, pp. 149-155.

Stejxeger

1894. "On some Collections of Reptiles and Batrachians from East
Africa" etc. Froc. U.S. Nat. Mus., 16, 1893, pp. 711-741.

1916. "Notes on Amphisbaenian Nomenclature". Proc. Biol. Soc.
Washington, 29, p. 85.

Torxier

1905. "Eidechsen Ausbeuten einer Forschungsreise von Oscar Neumann
and Carlo von Erlanger in Nordost-Afrika". Zool. Jahrb., Syst.,
23, pp. 365-388.

Vaillaxt

1882. Mission G. Revoil aux Pays Comalis. Faune et Flore; Reptiles et
Batraciens. (Paris).

Versluys

1898. "Die mittlere and aussere Ohrsphare der Lacertilia und Rhyncho-
cephalia". Zool. Jahrb., Anat., 12, pp. 161-406, pis. x-xvii.

VrXCIGUERRA

1930. "Spedizione di S.A.R. il Duca degli Abruzzi alia sorgenti dell'
Uebi Scebeli — Eisultati Zoologici; Rettili e Pesci". Ann. Mus.
Genova, 55, pp. 40-42.

Werner

1913. "Neue oder seltene Reptilien and Frcsche des Naturhistorischen
Museums in Hamburg". Jahrb. Hamb. Wiss. Anstalt, 30, 2,
pp. 1-51.

1917. "tiber einige neue Reptilien und einen neuen Frosch des Zool-
ogischen Museums in Hamburg". Mitt. Zool. Mus. Hamburg,
34, pp. 31-36.

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE

Vol. XCI, No. 2

A SUMMARY OF THE
IGUANID GENUS UROSAURUS

By M. B. MlTTLEMAN

With Sixteen Plates

CAMBRIDGE, MASS., U. S. A.

PRINTED FOR THE MUSEUM

September, 1942

PUBLICATIONS
OF THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

The Bulletin and Memoirs are devoted to the publication of
investigations by the Staff of the Museum or of reports by spec-
ialists upon the Museum collections or explorations.

Of the Bulletin, Vols. I to LXXXIX have appeared and of
the Memoirs, Vol. I to LVI.

These publications are issued in numbers at irregular intervals.
Each number of the Bulletin and of the Memoirs is sold separately.
A price list of the publications of the Museum will be sent upon
application to the Director of the Museum of Comparative
Zoology, Cambridge, Massachusetts.

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE

Vol. XCI, No. 2

A SUMMARY OF THE
IGUANID GENUS UROSAURUS

Bv M. B. MlTTLEMAN

With Sixteen Pl\tes

CAMBRIDGE, MASS., U. S. A.

PRINTED FOR THE MUSEUM

September, 1942

To

My wife, Mary Elizabeth, to express in a small
measure my appreciation of her help and un-
flagging encouragement.

No. 2. — A Summary of the Iguanid Genus Urosaurus

By M. B. MlTTLEMAN

CONTENTS

PAGE

Introduction and acknowledgments 106

Part I. The Taxonomy and Relationships of Urosaurus and its

Allies 107

Part II. Summary of the Genus Urosaurus 117

Methodology and Scope of the Study 117

Genus UROSAURUS Hallowell 124

Key to the Lizards of the Genus Urosaurus 128

Urosaurus omatus omatus Baird and Girard 133

Urosaurus omatus schmidti Mittleman 135

Urosaurus omatus caeruleus Smith 136

Urosaurus omatus linearis Baird 137

Urosaurus omatus chiricahuae Mittleman 139

Urosaurus omatus symmetricus Baird 142

Urosaurus omatus graciosus Hallowell 144

Urosaurus omatus wrighti Schmidt 145

Urosaurus omatus lens Stejneger 147

Urosaurus clarionensis Townsend 149

Urosaurus omatus schottii Baird 149

Summary of the omatus Complex 151

Urosaurus gadovi Schmidt 154

Urosaurus irregularis Fischer 156

Urosaurus nigricaudus Cope 157

Urosaurus microscutahts Van Denburgh 159

Summary of the nigricaudus Complex 162

Urosaurus auriculatus Cope 163

Urosaurus bicarinatus bicarinatus Dumeril 164

Urosaurus bicarinatus anonymorphus Mittleman 166

Urosaurus bicarinatus nelsoni Schmidt 168

Urosaurus bicarinatus tuberculatus Schmidt 169

Urosaurus unicus Mittleman 170

Summary of the bicarinatus Complex 173

Literature Cited 176

106 bulletin: museum of comparative zoology

INTRODUCTION

Among the several peculiarly North American genera of lizards,
there remain but few which have not been subjected to an exhaustive
monographic treatment. These have chiefly been small groups con-
taining but relatively few species. One of the notable exceptions is the
highly ramose genus Urosaurus Hallowell, 1854, which in itself is not
to be found in current usage.

The present study had its genesis in the summer months of 1938,
when I was afforded the opportunity to observe and collect numbers
of specimens of the genus Urosaurus, in my capacity as herpetologist
to the 1938 Rainbow Bridge - Monument Valley Expedition, under
the able leadership of Dr. Angus M. Woodbury. The original effort
to settle the contentious status of the species native to the Navajo
country of northeastern Arizona and southeastern Utah, led to this
study.

Many persons have contributed in a very real sense towards the ful-
fillment of the plan of investigation, and to all of these, grateful and
sincere acknowledgment is made. In this connection, particular men-
tion must be made of Dr. Herschel Thomas Gier, of Ohio University,
to whom I am indebted for constant interest, support, and many
fruitful suggestions. Drs. Leonhard Stejneger, Doris M. Cochran, and
Alexander Wetmore, of the United States National Museum, have
kindly made the great resources of that institution constantly avail-
able to me, and have offered much in the way of help and encourage-
ment. Dr. Thomas Barbour must surely be accorded a more than
ordinary word of grateful thanks for his generous help in the publica-
tion of this paper, as well as for the loan of very many specimens.
Benjamin Shreve and Arthur Loveridge, of the Museum of Compara-
tive Zoology; the late Dr. G. Kingsley Noble and Charles M. Bogert,
of the American Museum of Natural History; M. Graham Netting, of
the Carnegie Museum; Joseph R. Slevin, of the California Academy of
Sciences; Dr. Laurence M. Klauber, of the San Diego Society of Natural
History; Drs. E. Raymond Hall, Alden H. Miller, and Jean M. Lins-
dale, of the Museum of Vertebrate Zoology; Drs. Henry Fowler and
Emmett Reid Dunn, of the Philadelphia Academy of Sciences; Dr.
John Hack, of the Peabody Museum; Dr. Angus M. Woodbury, of the
University of Utah; Dr. George F. Knowlton, of Brigham Young Uni-
versity; Dr. Karl P. Schmidt and Clifford Pope, of the Field Museum
of Natural History; Dr. Howard K. Gloyd, of the Chicago Academy of
Sciences; Dr. Raymond B. Cowles, of the University of California at

mittleman: the iguanid genus urosaurus 107

Los Angeles; Dr. George S. Myers, of Leland Stanford University; Dr.
Edward H Taylor, of the University of Kansas; Dr. Helen T. Gaige,
of the University of Michigan; and H. W. Parker, of the British Mu-
seum (Natural History), have each, through personal kindnesses and
the facilities of their several institutions, immeasurably aided this
study. A very special word of grateful appreciation must be given to
Dr. Hobart Muir Smith, of the University of Rochester. Dr. Smith
has witnessed and abetted the researches involved herein with the
same contagious enthusiasm which marks his own studies, and has
done much and gone far to offer help by freely giving advice, loaning
specimens and literature, and making available his intimate knowledge
of Mexican faunistics, particularly in relation to the Iguanidae.

Part I. The Taxonomy and Relationships
of UROSAURUS and its Allies

In 1852, Baird and Girard erected the genus Ufa for a rather small
lizard, prominently characterized by gular folds, auricular openings,
and a fine, homogeneous dorsal scalation. Somewhat later in the same
year these authors described another new form (1852:126), presumably
also referable to the same genus, which was named Uta omata. How-
ever, this latter form differed in several respects from the first-named
form (stansburiana); most noticeable difference being the dorsal scala-
tion, which consisted of rather fine scales everywhere, save along the
median line, where they were abruptly enlarged, rather strongly keeled,
and prominently imbricate; these enlarged scales were divided into two
parallel series on either side of the median line by a series of somewhat
smaller, vertebral scales. In 1854, Hallowell encountered another new
form, generally similar to Baird and Girard's Uta omata, save that the
enlarged dorsals extended the length of the dorsum in a broad, unin-
terrupted band, lacking the presence of the smaller, dividing series of
scales. On the basis of this scalation^ Hallowell set up the genus
Urosaurus. A somewhat similar move was next made by Dumeril,
who described the genus Phymatolcpis (1856:548) for a Mexican lizard
generally similar to the Uta omata, but possessed of only a single
series of enlarged dorsals on either side of the smaller vertebrals. The
attempts of Hallowell and Dumeril were shortlived, however, for in
1858 Baird described the species Uta symmetrica, a close ally of his
previously named Uta omata, and in the next year (1859:7) definitely
placed Hallowell's Urosaurus in the synonymy of Uta. And in this
synonymous category Urosaurus has remained, save for a brief use

108 bulletin: museum of comparative zoology

as a subgeneric appellation by Van Denburgh (1922:182). Dumeril's
Phymatolepis has fared only slightly better, for Cope soon placed it in
synonymy (1864:177), although on at least two occasions thereafter
the name was used (Fischer, 1882:232; Boulenger, 1883:342), but
was soon relegated to the oblivion of synonymy (Boulenger, 1885:214,
216), where it has since remained.

In addition to these attempts at generic restrictions, one other one
was made in the long history of Uta. In 1863, Cope described the very
distinctive Uta thalassina, a rather large iguanid differing in several
respects from any previously known members of the then-recognized
genus Uta. The general habitus of this lizard was sufficiently different
from other known forms, so that Boulenger (1885:205) considered it
generically distinct, and proposed the generic name Petrosaurus.
Boulenger's attempt was promptly ignored by Cope (1887:35), who
retained thalassina in Uta as originally described. Van Denburgh has
used Petrosaurus subgenerically (1922:181), but other than this usage,
the name has been considered a trite synonym of Uta.

A brief description of Uta (auct.) follows:

Size small to quite large; auricular openings present; dorsal scalation
homogeneous, either keeled or smooth, or else with an abruptly en-
larged series of scales along the median line; dorsal scales imbricate
or pavemented; caudals greatly enlarged, spinose, heavily keeled, or
else very minute, smooth, barely imbricate; ventrals rounded and
smooth, or submucronate to mucronate, and keeled ; gular fold heavily
denticulate, or else with an even margin of small scales; supraoculars
in one, two, or three principal rows ; post! emoral dermal pocket present,
or postfemoral dermal pocket absent; posterior maxillary teeth tri-
cuspid; a sternal fontanelle present; xiphisternal abdominal ribs pre-
sent; ventrum with or without blue patches in males; a bright blue
postaxillary blotch present or absent; dorsal pattern either of short,
broken cross bars extending from neck to sacrum, or else with three
or four heavy black bars on the anterior portion of back; a distinctive
dark neck or shoulder band present or absent.

Accepting such a loose definition of Uta makes necessary the inclu-
sion of approximately thirty-seven species and subspecies exhibiting
enormous gaps in structure, pattern, size, and distribution. It is ex-
ceedingly difficult to conceive of such a heterogeneous assemblage as
having arisen from a single, even greatly generalized, primitive ances-
tor. Even the extremely diverse and multitudinous genus Sceloporus
does not present as vicarious a group as is now recognized under the
single all-encompassing heading of Uta.

mittleman: the iguanid genus ukosaurus 109

With quite some comfort, I point out the following statement,
"... genera are groups for convenience, ... no two people are ever
likely to agree for long on generic limitations . . . any means of break-
ing up large genera by setting off particularly well differentiated species
or groups of species is justifiable. . . ." (Stejneger and Barbour, 1933 :vi.)
I cannot reconcile myself to the thought that the many forms included
in Uta are congeneric, and therefore recognize a division of these many
forms; a redefinition of Uta, revival of Urosaurus and Petrosaurus, and
the erection of a fourth genus, seems to adequately and more logically
categorize these animals.

Genus uta Baird and Girard

1852 Uta Baird and Girard, Stansbury's Expl. Surv. Vail. Great Salt Lake,
p. 345 -1

Genotype, stansburiana.

Diagnosis. Small iguanid lizards (maximum size, snout to vent,
approximately 75 mm.); prominently denticulated auricular openings
and gular folds; dorsal scalation homogeneous, consisting of small,
lightly keeled, rounded, imbricate scales; ventrals rounded, smooth,
imbricate; caudals greatly enlarged, heavily keeled and spinose,
strongly imbricate; supraoculars in one principal series; interparietal
(occipital, auct.) large; postfemoral dermal pocket present; superciliar-
ies imbricate ; labials segmental ; no distinctive blue abdominal patches
in males; postaxillary and/or preaxillary dark blotches present; dorsal
pattern consisting of small, pale maculations, or else principally of
larger blotches in parallel series which may or may not be joined to
form longitudinal bands; xiphisternal abdominal ribs present; a sternal
fontanelle present; posterior maxillary teeth tricuspid.

Distribution. North America, from Texas to Washington and south
to the southern portion of Baja California ; the islands bordering Cali-
fornia and Baja California in the Pacific Ocean and Gulf of California;
northern Mexico, as far south as southern Sonora.

Remarks. As defined here, Uta includes stansburiana and its sub-
species, as well as the species taylori, concinna, mannophorus, martinen-
sis, nolascemis, palmeri, squamata, and stellata.

Concerning the question as to whether this name was published as cited here, or in the
Proc. Acad. Nat. Sci. Phila., 1852, 6:69, Dr. Stejneger advises me, "Stansbury's Explorations
has the priority of publication. It was 'published early in 1852, probably late March or early
April.' The paper in the Proc. Phila. Acad, was read on April 27, consequently published
considerably later." Contrary-minded are referred to Taylor, 1935:411, et seij.

110 bulletin: museum of comparative zoology

Genus urosaurus1 Hallowell
1854 Urosaurus Hallowell, Proc. Acad. Nat. Sci. Phila., 7: 92.

Genotype, graciosus.

Diagnosis. Similar in many respects to Uta, but differing principally
as follows : dorsal scalation not homogeneous, but consisting of minute
scales except in the vertebral region, where the scales become abruptly
enlarged, usually strongly carinate and imbricate, occasionally spinose;
enlarged dorsal scales either separated into two parallel series by a
median line of smaller scales, or else extending in a broad band for the
length of the dorsum; ventrals often mucronate and keeled, especially
laterally; postfemoral dermal pocket variable, regularly present in
some forms, variable occasionally, absent in others; males with dis-
tinctive blue abdominal patches; no post- or preaxillary blotches; dor-
sal pattern of short, lateral bars, usually broken on the mid-line, occa-
sionally a pattern aberration of longitudinal stripes; never wholly
maculated above with small, light flecks.

Distribution. Texas west to California, and north to Utah; south
throughout Baja California; Mexico along the west coast principally,
as far south as Chiapas; islands bordering Baja California in the Pacific
and Gulf of California; the Revillagigedo Archipelago.

Remarks. As defined here, Urosaurus includes the species and sub-
species known heretofore as Uta ornata and its subspecies, as well as the
numerous forms recently treated by myself (1941); also the species
microscutatus and nigricaudus. See pages following for a fuller dis-
cussion of the forms of Urosaurus.

Genus petrosaurus Boulenger
1885 Petrosaurus Boulenger, Cat. Liz. Brit. Mus., 2: 205.

Genotype. Uta thalassina Cope, Proc. Acad. Nat. Sci. Phila., 1863:
104.

Diagnosis. Large iguanid lizards (maximum size, snout to vent,
approximating 175 mm. or more); dorsal scalation homogeneous, con-
sisting of small, smooth scales, which are usually pavemented; venter
with slightly larger, smooth, pavemented scales; caudals small, weakly
keeled, barely imbricate, smaller than ventrals; supraoculars in three

1 I have followed the example of the Check-List in dropping the hyphen; cf. Dipsosaurus.

mittleman: the iguanid genus urosaurus 111

principal series ; gular fold barely or not at all denticulate ; supraoculars
smooth (rather than rugose), as are the other cephalic scales; post-
femoral dermal pockets present; venter lacking distinctive abdominal
blotches in males; no pre- or postaxillary blotches; shoulders with
three or four heavy transverse blackish bars.

Distribution. The southern portions of Baja California, and a few
of the adjacent islands in the Gulf of California.

Remarks. As defined here, Petrosaurus includes besides the type
species, the form which should properly be known as Petrosaurus repens
Van Denburgh.

streptosaurus1, gen. nov.

Genotype. Uta mearnsi Stejneger, Proc. U. S. Nat. Mus., 1894: 17,
589.

Diagnosis. Medium-sized iguanid lizards (approximating 100 mm.,
snout to vent), closely related to Petrosaurus, and bearing a super-
ficial resemblance to Uta also, but distinguished from these two genera
as follows: Differs from Petrosaurus in the smaller ventrals ; the en-
larged, strongly keeled, spinose caudals; the presence of two supraocular
rows; the lack of a well developed anterior gular fold; the smaller size;
the well developed lateral dermal fold; greater number of femoral
pores; much larger preauricular spines; the absence of a bold pattern
of three or four transverse bars in the scapular region; the presence
of a distinctive neck band which is dark, and bordered behind with a
lighter hue; the presence of a fairly heavy blue abdominal wash (ap-
proximating somewhat the blotches found in Urosaurus); and the
prominent dorsal peppering of light flecks. Differs from Uta in the
much smaller smooth and pavemented dorsal and ventral scales;
the possession of two principal series of supraoculars; the lack of an
anterior gular fold; the lack of denticulation on the gular fold;
greater number of femoral pores; cephalic scales smooth, rather than
rugose; the presence of a strong neck band of blackish; the absence
of any small blotches dispersed in parallel series, or fused to form
longitudinal bands; the presence of several well-marked dark cross-
bands dorsally; the prominent caudal pattern of dark and light rings.

Distribution. Extreme southern California and the adjacent portion

1 Streptosaurus =arptirTo% (wreathed, banded, or twisted) -\-ax»pa (lizard), in reference to
the prominent neck band, or wreath, of this genus.

112 bulletin: museum of comparative zoology

of Baja California; Angel de la Guardia and Mejia islands, in the Gulf
of California.

Remarks. As defined here, Streptosaurus includes besides the type
species, the form which should be properly known as Streptosaurus
slevini Van Denburgh.

It is my belief that not only is the genus Uta (s.l.) divisible as given
in the preceding paragraphs, but that two of the genera, Petrosaurus
and Streptosaurus, are more closely allied to the section of the Iguan-
idae characterized by the genus Crotaphytus,thanto Uta and Urosaurus,
which are obvious derivatives of the Sceloporus stock. I base this
assumption on several premises. First, Uta and Urosaurus are con-
ceivably derivable only from Sceloporus, or perhaps some sceloporid
form now extinct. Secondly, these two genera are very close to
Sceloporus in their general habitus, and differ in only a very few points.
On the other hand, we find that the general body form and details of
structure characteristic of Petrosaurus and Streptosaurus are most
nearly duplicated in Crotaphytus, of living genera. The genera Cro-
taphytus and Petrosaurus are in fact almost identical on superficial
examination ; however, they differ in the greatly reduced interparietal
of Crotaphytus, as well as the very small and numerous supraoculars,
and the other reduced and multiplied cephalic scales ; they are further
distinguished by the presence of palatines in the former genus, and the
absence of these in Petrosaurus. Sternal fontanelles are apparently
variable in Crotaphytus (fide Cope, 1900:247), but constantly present
in Petrosaurus. Of the two genera, Crotaphytus is probably older in
view of its more widespread distribution, and the closer similarity in
osteology of this genus with that of primitive groups such as Dipso-
saurus and Ctenosaura. Streptosaurus is most closely allied to Petro-
saurus of living genera, and in the main probably best characterized
as a fairly recent derivate. Properly speaking, neither the Crotaphytus-
Petrosaurus-Streptosaurus stock, nor the Uta-Urosaurus-Sceloporus
stock can be particularly considered as being either older or more prim-
itive than the other. An accompanying diagram illustrates the prob-
able derivations and positions of these genera amongst North American
Iguanidae. It will be noticed that I have included several other genera
in addition to those discussed above, for the sake of completeness.
Dipsosaurus is probably the most primitive of the North American
Iguanidae (excepting Ctenosaura, which is properly a Central and
South American form), and possesses several points in common with
Ctenosaura, most easily observed of which is the dorsal crest; the genera
further show their relationship in the similarity of the cephalic scutel-

mittleman: the iguanid genus urosaurus 113

lation which is essentially simple, and shows no particular degree of
differentiation. Sanromalus is considered a specialized offshoot of the
Crotaphytus, or more properly, pve-Crotaphytus stock, by reason of its
solid sternum, as well as the five-lobed teeth; the simple type of cephalic
scalation indicates its affinity with the more primitive Dipsosaurus-

STREPTOSAURUS UTA IIDnQAIIDIIQ

PETROSAURUS^ \ UROSAURUS

CROTAPHYTUS? SCELOPORUS

CALL ISA URUS\/ SA TOR

UMA*~^
HOLBROOKIA-

PHRYNOSOMA

DIPSOSAURUS

CTENOSAURA

PRIMITIVE IGUANID TYPE

Fig. 1. The phylogeny and relationships of North American iguanid
genera.

Ctenosaura stock. The three genera Uma, Callisaurus and Holbrookia
form a compact group, presumably derived within comparatively
recent times from a Crotaphytus-\ike stock. They resemble each
other in the common type of dorsal scalation, and the imbricate super-
ciliaries and supralabials. Uma is easily distinguished by reason of the
enlarged digital fringes which Callisaurus lacks; Holbrookia does not
possess auricular openings, which are present in the other two genera.

114 bulletin: museum of comparative zoology

The other branch of the North American Iguanidae presents first
Phrynosoma, a primitive but highly specialized genus, at once dis-
tinguishable by the squat and flattened habitus, as well as enormously
developed cephalic scalation. Since the relationships of Uta, Urosaurus
and Sceloporus, and that of this latter genus and Sator, have not been
satisfactorily presented previously, I give a brief resume based on
available information.

In 1935, Smith discussed at some lengths the relationships of Scelo-
porus and Uta (s.l.), and concluded that the two genera, as then under-
stood, were most nearly allied by Sceloporus couchii and Uta (= Uro-
saurus) levis. Smith later emended this somewhat, and postulated the
belief that Uta "arose from the forms now extinct which closed the
present gap between couchii and mcrria^^ni.,, Shortly thereafter I sub-
scribed to this point of view (1941), and presented further observations
on the phylogeny of the Mexican Urosaurus. Since that time, further
investigation of Uta, Urosaurus, and Sceloporus shows that these
premises are only partly true.

Uta is considered as very nearly allied to Urosaurus, as indicated by
the available evidence. Both genera apparently have sprung from an
early progenitor which soon diversified sufficiently to produce the
bifurcate branches we recognize today as Uta and Urosaurus. The
latter genus is possibly the older of the two, or at least differentiated
and spread more rapidly, judging by its more widespread distribution
in the outlying Pacific Islands and southern Mexico. It appears that
Uta probably did not become dispersed or diversified specifically until
the beginning of the Miocene, for it is largely restricted to the conti-
nental portions of the United States and Mexico, and the adjoining
islands, while being absent from some of the older Gulf and Pacific
islands which had appeared prior to this time.

The closest relationship between Uta and Urosaurus occurs between
Uta squamata and Urosaurus microscutatus. In these forms, the genera
have apparently produced parallel lines, for squamata possesses the
largest dorsals in Uta, and is fairly large, whereas 7nicroscutatus is one
of the smallest of its genus, and possesses the least development of the
dorsals.

As I have pointed out, while Uta and Urosaurus have doubtless
shared a common ancestor if the evolutionary lines be carried back
sufficiently, this ancestral type probably differentiated at an early date.
Thus, as Smith points out, "in couchii . . . dorsal scales are extremely,
small for the genus, the laterals are minute, and the size of the species
itself is small." In addition to these premises which serve to link Uta

mittleman: the iguanid genus urosaurus 115

and Sceloporus, it may be said that stejnegeri, most primitive of the
Utas, and couchii, occupy overlapping areas. I would agree with Smith
(1939:239) that Uta (through the medium of stejnegeri) is closer to
couchii than to any existing Sccloporus; however, the derivation has
probably taken place through the medium of a pre-cowc/m' form, rather
than through this latter species. Whether such a form would close the
gap between couchii and merriami, in Sccloporus, I cannot say.

Urosaurus apparently most closely approaches Sccloporus through
the maculosus and merriami groups of this latter genus, being not par-
ticularly closer to one or the other of these groups, but occupying a
somewhat intermediate position. Somewhat as in the case of Uta, the
origin of Urosaurus appears to have stemmed from an early precursorial
stock which gave rise to the two groups of Sceloporus, as well as Uro-
saurus. In the merriami and maculosus groups are found small Seelo-
pori, with larger dorsals than in couchii (and hence more similar to
Urosaurus), more rugose dorsals than in couchii; in addition, the fron-
tals are usually divided transversely, and rarely if ever longitudinally
as in couchii; postfemoral dermal pockets present in maculosus, absent
in merriami (variable in Urosavrus), constant in couchii (and also Uta).
It will be seen therefore that Uta and Urosaurus may be considered as
very nearly biological equivalents, for they are widely distributed,
highly prolific, of about {he same age, successful (as adjudged by the
multiplicity of individuals and species), and derived from closely re-
lated progenitors.

The genus Sator Dickerson, 1919, is in some respects unique, and
because of its relationship with Sceloporus, of interest here. Although
Dickerson (op. cit.) mentioned certain osteological differences supposed
to obtain in Sator, I have not been able to ascertain any constant osteo-
logical variations within the genera Sator, Sceloporus, Uta, and Uro-
saurus. Insofar as I am able to determine, distinguishing features of
Sator are those of lepidosis and form. Thus, this genus may be recog-
nized by virtue of its vertically compressed body and tail, rudimentary
gular folds, poorly differentiated dorsal scales, and the rudimentary
postanals in the males. Although some authors have considered Sator
intermediate between Uta and Urosaurus, or between the former genus
and Sccloporus, I am of the opinion that it is actually a direct deriva-
tive of the primitive pyrocephalus group of Sceloporus, and is not
closely related to any other known lizards.

The pyrocephalus complex includes pyrocephalus, gadoviae, and
nelsoni, all three of which exhibit compressed tails to some degree, at
least in the males. Further, in these species, the dorsals are reduced;

116

bulletin: museum of comparative zoology

in one, gadoviae, a postfemoral dermal pocket is present (as it is in
Sator); a high femoral pore count exists; there is only a very gradual
transition from the larger dorsals to the smaller laterals; and finally,

UR OS A UR US

UTA

couch ii

macu

pyrocepho/us

merr /ami

SATOR

(lower case leffers denote
species or groups of spe-
cies--copita/s denote genera)

--modified after Smith, I939--

Fig. 2. The derivation and intergeneric relationships of Urosaurus and
allied iguanid genera.

the postanals of the males are either greatly reduced, or absent. Sator
closely resembles these lizards in all the characters noted. Finally,
Sator possesses dark lateral cross-bars on the male abdomen, a pattern

MITTLEMAX: THE IGUANID GENUS UROSAURUS 117

which is reproduced in S. pyrocephalus, and in no other known Scelo-
porid. Dickerson (op. cit.) postulated a close relationship between
Sator and Sccloporus utiformis because of the rudimentary gular fold
in this latter species, as well as certain other minor details. I have dis-
cussed this point with Dr. H. M. Smith, and we agree that on the
whole, the derivation of Sator through the medium of the pyrocephalus
group, seems much more likely. It is probably worthy of note to add
that the formation of Sator has seemingly been of the most fortuitous
nature, and almost certainly due primarily to insular isolation. This
fact is borne out by the remarkable ability of the stock which produced
Sator, to again emulate the same trend, even in a mainland population
of Sceloporus, where in the relatively isolated gadociae there exists a
remarkable parallelism in color, pattern, ecology, and even scutellation
to some degree.

Part II. Summary of the Genus UROSAURUS
Methodology and Scope of the Study

Approximately 6,500 specimens have been critically examined in the
course of investigation of this problem. ^Yhile ample numbers of some
forms have been available, others are but poorly represented in collec-
tions. Specimens have often been haphazardly collected when and
where opportunities permitted; necessarily therefore, many critical
regions are not represented in available series, or else but scantily. The
dynamic aspect of the study is thus heightened in view of the many
problems which still present themselves. As Dr. E. H. Taylor so aptly
termed Mexico, in relation to his study of Eumcces, it is the terra incog-
nita, and insofar as Urosauriis is concerned, it is to this general region
that future workers must turn for the answers to many questions.

I have not attempted to present all of the available data on the
natural history of the numerous forms in the genus, but rather has the
effort been restricted to the taxonomy involved, ecology being for the
most part utilized only where it serves to explain problems of isolation,
speciation, and kindred matters. It will be observed that each form is
treated principally from the viewpoints of taxonomy, structure, and
distribution. In essence therefore, this report resolves itself into a col-
lection of taxonomic and other data of practical interest to the labora-
tory worker principally. Thus, long and detailed synonymies have

118 bulletin: museum of comparative zoology

been omitted, and only those references which I have considered im-
portant to the proper understanding of the form in question have been
given.

Herpetological taxonomy has seen several excellent uses of statistics
in the study of genera and other natural associations, especially where
large numbers of specimens have been available. In view of the multi-
tude of specimens extant for this study, it was at first thought that
these would lend themselves admirably to the statistical approach.
Unfortunately, extensive measurements, scale counts, and the applica-
tion of numerous formulae have not borne out this hope. True, men-
sural limits of variation and proportion have been ascertained for most
of the forms, but qualitative data reduced to quantitative terms have
proved all but useless. This is chiefly due to the fact that the distin-
guishing features of most forms are of a qualitative, rather than quan-
titative nature. Throughout this study, insofar as available specimens
permit, populations have been examined, and populational trends
have been given the most consideration, with relatively little or no
emphasis on the individual. This has been necessitated by extremes in
metachrosis, as well as an inordinate variation in individuals, particu-
larly in minor details of sealation.

Throughout its range, Urosaurus exhibits a remarkable propensity
for the proliferation of local trends, perhaps even unique genie strains.
For the most part, these have been the result of the extreme ecological
limitation of Urosaurus, which will only under the greatest duress leave
its vertical habitat of rocks, trees, cliffs, and fences, and traverse the
desert floor to a similar, nearby habitat. As a result of this ecological
isolation of many populations, genetic involutions attributable to
forced inbreeding are the mode, and result in many minor strains which
characterize these small groups. The problem then, for the taxonomist,
is to decide which of these trends is clearly definable, which population
possesses these features to a distinguishing degree, and whether appar-
ent morphological distinction is correlated with definite geographic
and/or ecological niches. For the most part, these local trends are
negligible other than for their academic interest. But in a few cases,
closely related forms are separated (and apparently restricted) by only
a scant dozen miles or so of desert floor. Quite possibly because of this
marked propensity for isolation and the proliferation of superficial
details, authors in the past who have dealt with this genus and have
had only small numbers of specimens, or else specimens from widely
separated localities, have in despair considered several distinct species
synonymous with each other, and thus obscured the true relationships

mittleman: the iguanid genus urosaurus

119

of several forms, as well as making practical identification a matter of
clairvoyance in many instances.

Certain details of structure and proportion have been found to be of
prime significance, and are utilized throughout this report in the defini-
tions and diagnoses of the species and subspecies. These are listed and
annotated below:

Enlarged dorsal scales (fig. 3). These constitute the most readily
observable, and in most instances important feature in nearly every

Fig. 3. Details of vertebral, primary, and secondary enlarged scales.

form. Enlarged dorsal scales are of two general types: vertebrals
and the larger bordering scales. Thus, in several forms the smaller
series of vertebrals extends from the nape to the basal portion of the
tail, bordered on either side with a varying degree of constancy, by
one or more series of larger scales. In numerous instances, the verte-
brals are lacking, so that the enlarged scales are present as a more or
less homogeneous band, with the largest scales on the midline, and the
remainder progressively decreasing as they extend laterally. When the
vertebrals are present the first parallel bordering series of enlarged
dorsals is referred to as the "primary series"; if others are present too,
they are termed "secondary series". Particular attention is given to
the distribution and shape of the enlarged dorsal scales, as, whether

120 bulletin: museum of comparative zoology

they are evenly dispersed, whether the individual scales are rounded
posteriorly, or mueronate, submueronate, spinose, smooth, or carin-
ated. Unless used to the contrary, the term "enlarged dorsals" refers
specifically to the series bordering the smaller vertebrals.

Ventral scales. Reference is made in several cases to the structure
and dispersal of these, generally in the same terms as those used for
the enlarged dorsals.

Dermal folds. These are present in most forms, absent in a few.
When present, they usually extend from the cervical region to the
groin or further, and are usually on the dorsolateral or lateral line, or
both. Since these are sometimes of diagnostic importance, especial
attention should be paid to determine whether apparent dermal folds
are actually true folds, or simply superficial wrinkles due to preserva-
tion.

Posifcmoral dermal pocket (fig. 4). This is a slight dermal invagina-
tion within the angle formed by the posterior part of the femur and the

Fig. 4. Location and form of postfemoral dermal pocket.

body. It is lined with scales which are much smaller than those else-
where on the body. The presence or absence of this structure is used
diagnostically, and although variable in some forms, is usually highly
constant. When the term "rudimentary" is employed to describe the
pocket, it infers that it is represented by a small whorl of scales in a
minor depression.

Tubercles. Tuberculation is for the most part referred to in com-
parative terms, but in a few instances the absence or presence of these
modified scales on certain portions of the body is of diagnostic value.

Femoral and tibial enlarged scales (fig. 5). On the anterodorsal sur-
face of the thigh, and in a band of varying width around the tibia, are

MITTLEMAX: THE IGUANID GENUS UROSAURUS

121

dispersed greatly enlarged, heavily keeled, and usually strongly spinose
scales. The comparative size of these is often referred to. They are
regularly present, save in one species.

Fig. 5. Enlarged femoral and tibial scales.

Fig. 6. Enlarged postanal scales of the male.

Enlarged postanal scales (fig. 6). These are characteristic of the
males of all known forms, but are subject to a wide variation in size in
the several forms. They are usually flat, subcircular, and somewhat
depressed.

122 bulletin: museum of comparative zoology

Caudal scales. In all known forms the caudal scales are greatly
enlarged, heavily keeled, strongly spinose, and imbricate; the scales
become smaller as they progress laterally and terminally. The type
of transition from the dorsal to the lateral caudal scales is important,
as are the numbers of scales in each whorl. In several species, the
extent to which the vertebrals encroach upon the basal portion of the
tail is used diagnostically.

Cephalic scutcllation (fig. 7). The following description is typical
of the genus: cephalic plates fairly smooth, large; temporal region
swollen; nostril above the canthus rostralis; frontal plate variable,
transversely divided in some forms, entire in others, occasionally split
vertically ; a principal series of supraoculars, and immediately next to
these another series scarcely smaller; supraoculars separated from
frontal, frontoparietals, and parietals by one or two series of small
granules; interparietal largest; superciliaries elongated, projecting,
strongly imbricate; rostral wider than high; labials segmental; mental
variable, in contact with both sub- and infralabials; gular region vari-
able, scales granular or flat, imbricate or pavemented. Other cephalic
scales are too variable to be of significance in most cases; occasionally
the numbers of prefrontals and frontonasals are diagnostically em-
ployed.

Color. But two facts need be mentioned at this time. First, that
with no known exception, the lizards of Urosaurus are very prone to
exhibit a uniformly melanistic appearance. This appears in all known
forms, and is often, but not always, correlated with habitat. When
melanism occurs, it usually effaces all but the barest traces of pattern.
Secondly, all Urosaurus males possess brightly colored abdominal
blotches, usually of some shade of blue. The extent and intensity of
these is used in the diagnosis of some forms. In one case, at least, the
females also possess blue abdominal patches, but these are irregularly
defined and not so intense as in the males of the species. The usual
condition however, presents the female abdomen unicolor, usually
grayish or white, and sometimes lightly maculated with a darker pig-
ment in haphazard fashion.

Mensural data. With but a few exceptions, only four measurements
are given: head length (tip of snout to posterior border of the auricular
opening), head width (usually taken immediately anterior to the
meatus auditorius), body length (tip of snout to vent), and hind leg
length (measured from the anterior angle of insertion to the tip of the
4th toe, exclusive of the nail). With but a few exceptions, tail lengths
are not given, principally because the tail is so often lacking in pre-

MITTLEMAX: THE IGUAXID GENUS UROSAURUS

123

served specimens that such a feature would be of little use in many
instances. Secondly, for the most part, a gross evaluation of the tail

Fig. 7. Cephalic scutellation and nomenclature in Urosaurus. ro — rostral;
sn — subnasal; na — nasal; in — internasal; fn — frontonasal; pf — prefrontal; fr —
frontal; si — supralabial; so — subocular; suo — supraocular; fp — frontoparietal;
pa — parietal; ip — interparietal.

length, as, whether it is more than twice the combined head and body
length, or less, is all that is needed. All measurements are given in

124 bulletin: museum of comparative zoology

millimeters. For practically every form, the measurements of a sample
of several adults of both sexes (unless specifically stated otherwise),
usually fifty or more specimens, are given. In these cases, the weighted
arithmetic mean is used for each figure; this "mean" is simply a more
accurate representation of the commonly used "average".

The accumulation of large series of data has produced several facts
of interest. From a practical viewpoint as regards these lizards, it
has been noted that measurements and ratios obtained from them,
based on the gross sampling of populations, are often misleading for
the reason that such measurements and/or ratios are not constant in
an individual, and hence population. Or this is better expressed per-
haps, if it is said that these mensural data progress dynamically, and
unless a specific age group and sex is employed, the results are dis-
torted. Thus, throughout Urosaurus innumerable cases of positive
and negative heterogony (depending on the organ under consideration)
have been observed. Several of these heterogonal changes in size and
proportion of age groups of the same and opposite sexes have been
found to fit the formula Y=bXk (Simpson and Roe, 1939:367, et seq.).
To present as much uniformity as possible, all measurements and ra-
tios obtained from series of specimens have been taken from equal
numbers of males and females (unless specifically stated to the con-
trary), which have been sexually mature animals, as adjudged by the
condition of the gonads, or gross size when this has been known to be
indicative of sexual maturitv.

Genus urosaurus Hallowell

1852 Uta Baird and Girard, Proc. Acad. Nat. Sci. Phila., 6, 126; Baird, Proc.
Acad. Nat. Sci. Phila., 1858, December, p. 253; Baird, U. S. Mex.
Bound. Surv., 1859, 2, 7; Cope, Rept. U. S. Nat. Mus. 1898 (1900):
299; Van Denburgh, Occ. Pap. Calif. Acad. Sci., 1922, 10, (1), 180;
Mittleman, Jour. Wash. Acad. Sci., 1941, 31, (2), 66.

1854 Urosaurus Hallowell, Proc. Acad. Nat. Sci. Phila., 7, 92 (type, Uro-
saurus graciosus Hallowell); Van Denburgh, Occ. Pap. Calif. Acad. Sci.,
1922, 10, (1), 182 (subgenus).

1856 Phymatolepis Dumeril, Arch. Mus. Hist. Nat. Paris, 8, 548 (type,
Phymatolepis bi-carinatus Dumeril); Fischer, Abh. Nat. Ver. Bremen,
1882, 7, 232; Boulenger, Ann. Mag. Nat. Hist., 1883, 5, (11), 342
(subgenus).

Diagnosis. Small to medium-sized iguanid lizards with well devel-
oped, pentadactyl limbs adapted for climbing and running; digits with

mittleman: the iguanid genus urosaurus 125

well developed claws; head large with prominent eyes and auricular
openings; a strong gular fold, often immediately preceded by one or
more additional folds; gular fold and auricular orifices denticulated
prominently; interparietal large; superciliaries imbricate; labials seg-
mental; enlarged supraoculars in two principal rows; dorsal scalation
homogeneous except in the median region where the scales become
enlarged, prominently keeled, and usually imbricate; enlarged dorsals
occasionally separated into two or more parallel series by a smaller,
vertebral series of irregularly arranged, weakly carinated scales; ven-
tral scales large, imbricate, occasionally spinose and/or carinate; tail
long, slender, covered with greatly enlarged, keeled, spinose, imbricate
scales; limbs with keeled, imbricate scales of moderate size, except on
the antero-dorsal surfaces of the thighs and a narrow band on the
tibiae, where the scales become greatly enlarged, and usually mucron-
ate to spinose; lateral and dorsolateral dermal folds usually present,
and often with a continuous or broken crest of enlarged tubercles;
cervical region similarly with dermal folds and tubercles; males with
enlarged postanal plates; post-femoral dermal pockets present in some
species, absent in others, occasionally variable; a sternal fontanelle;
xiphisternal abdominal ribs present; lateral and/or posterior teeth
tricuspid; body and tail depressed; no dark post- or preaxillary
blotches; males with prominent blue abdominal patches.

The following nomenclatorial changes are herewith proposed, and
the respective species and subspecies recognized as valid :

Urosaurus ornatus ornatus Baird and Girard, 1852.
omatus graciosus Hallowell, 1854.
ornatus symmetricus Baird, 1858
ornatus sehottii Baird, 1858
ornatus linearis Baird, 1859
ornatus lewis Stejneger, 1890
ornatus wrighti Schmidt, 1921
ornatus caeruleus Smith, 1935
ornatus schmidti Mittleman, 1940
ornatus chiricahuae Mittleman, 1941
nigricaudus Cope, 1864
auriculatus Cope, 1871
irregularis Fischer, 1882
clarionensis Townsend, 1890
microscutatus Van Denburgh, 1894
gadovi Schmidt, 1921

126

bulletin: museum of comparative zoology

unicus Mittleman, 1941
bicarinatus bicarinatus Dumeril, 1856
bicarinatus tuberculatus Schmidt, 1921
bicarinatus nelsoni Schmidt, 1921
bicarinatus anonymorphus Mittleman, 1940

The genus is remarkably free of synonymous forms, only four of these
having been proposed:

Uta gratiosa Coues, Surv. W. 100th Merid., 1879, 5, 596. Synonym
of Urosaurus ornatus graciosus.

Uta (Phymatolepis) lateralis Boulenger, Ann. Mag. Nat. Hist., 1883,
5, (11), 342. Synonym for Urosaurus ornatus schottii, see Mittle-
man (1941:136, etseq.).

Uta gularis Cragin, Bull. Washburn Lab. Nat. Hist., 1884, 1, 7.
Synonym of Urosaurus ornatus schottii, see Mittleman {supra cit.).

Uta parviscutata Cope, Rept. U. S. Nat. Mus. 1898 (1900) : 324.
Synonym for Urosaurus microscutatus.

Within the genus, three major subdivisions, or complexes, are readily
discernible :

ornatus

symmetricus

linearis

graciosus

levis

wrighti

caeruleus

schmidti

chiricahuae

clarionensis

schottii

nigricaudus
microscutatus
gadovi
irregularis

bicarinatus

tuberculatus

unicus

nelsoni

anonymorphus

auriculatus

ornatus Complex

nigricaudus Complex

bicarinatus Complex

mittleman: the iguanid genus urosaurus 127

Distribution of the genus. Southern Texas west through New Mexico
and Arizona to the desert portion of California bordering the Colorado
River, north to Utah (and Wyoming ?) and southwestern Colorado;
The Baja California peninsula; south through Mexico from the Rio
Grande to the Mexican Plateau, and south along the western slopes
of the Sierra Madre del Occidental, terminating at Tonola, Chiapas;
also the islands of Socorro, Clarion, Maria Madre, Maria Magdalena,
Tiburon, Magdalena, Espfritu Santo, San Francisco, San Marcos,
Coronado, Carmen, Danzante, Ballena, and San Jose, in the Pacific
Ocean and Gulf of California.

Key to the Lizards of the Genus Urosaurus

1. Enlarged antero-dorsal femoral scales smooth (Socorro Island,

Revillagigedo Archipelago).

auriculatus
Enlarged antero-dorsal femoral scales strongly keeled 2

2. Enlarged dorsals in a single broad band, uninterrupted by an

intervening series of smaller scales 3

Enlarged dorsals separated into two or more parallel series by the
presence of a vertebral series of smaller scales 7

3. Tail two or more times the length of head and body combined

(the Colorado River valley from extreme southern Nevada to
the Gulf of California; probably extreme northwestern Sonora;
northeastern Baja California as far south as San Felipe).

graciosus
Tail less than twice the length of head and body combined .... 4

4. Dermal folds, when present, not heavily crested with tubercles;

blue abdominal patches only in males; enlarged dorsals com-
paratively small 5

Dermal folds present, always crested with tubercles of fairly
large size; abdominal blue patches sometimes in females as well
as males 6

5. Enlarged dorsals larger, from 17 to 24 in the length of head from

tip of snout to posterior edge of interparietal; gular region in
males deep yellow or orange (Cape region of Baja California;
Magdalena, Espiritu Santo, and Ballena islands).

nigricaudus

128 bulletin: museum of comparative zoology

Enlarged dorsals smaller, 32 to 36 in the length of head from tip
of snout to posterior edge of interparietal; gular region in males
usually blue (Baja California north of the Cape region, north
to Borego Valley, California; the islands of San Marcos, Coron-
ado, Carmen, Danzante, San Jose, San Francisco and Magda-
lena).

microscutatus

6. Four to seven rows of enlarged dorsal scales; abdomen of both

sexes with a blue wash and/or blue patches; dorsolateral folds
not converging in the sacral region (Jalisco and Michoacan).

gadovi
About three rows of enlarged dorsal scales ; only males with a blue
abdomen; dorsolateral folds converging in the sacral region to
form prominent ridges ("Mexican Plateau").

irregularis

7. Enlarged dorsal scales in two or more prominent, nearly equal

series on either side of the smaller vertebral series; frontal
transversely divided; postfemoral dermal pocket almost invar-
iably present; ventrals rounded posteriorly and never carinate

12

Enlarged dorsal scales in a single principal series on either side of
the intervening vertebrals; frontal almost always entire; post-
femoral dermal pocket often absent, occasionally rudimentary ;
ventrals often mucronate to spinose, often carinate in the latero-
ventral areas 8

8. Enlarged dorsals commencing caudad of a line joining the anterior

points of insertion of the fore-limbs; dorsals weakly keeled,
rounded posteriorly, prominently pavemented; general habitus
not at all rugose (Batopilas, Chihuahua).

unicus

Enlarged dorsals commencing craniad of a line joining the anterior

points of insertion of the fore-limbs, or else equal with such a

line; dorsals prominently keeled, usually mucronate or spinose,

imbricate ; ventrals imbricate ; general appearance rugose .... 9

9. Form rugose; enlarged dorsals strongly carinate and prominently

mucronate; tubercles of lateral and dorsolateral folds well de-
veloped; ventrals mucronate; gular surface generally stippled,
with a light median area; blue abdominal patches of males
quite extensive (Michoacan, Morelos, Puebla, Guerrero as far
as Acapulco). bicarinatus

mittleman: the iguaxid genus urosaurus

129

o or n a/ us

a schmidti

0 linearis

+ wrighti

o symmetricus

* graciosus

mfergrading areo

Fig. 8. Distribution of Urosaurus in the United States.

graciosus * symmetricus

J

A schotiii

■ iubercuioius

Ubicarinafus

Mononymorpnus

+ne/soni

umcus

gadovi
+ c aeru/eus
oaur/cu/o/us
®c/or/onens's

Fig. 9. Distribution of Urosaurus in Mexico.

130 bulletin: museum of comparative zoology

General appearance somewhat less rugose; enlarged dorsals not
so often mucronate; tubercles smaller, sometimes absent;
ventrals less mucronate, occasionally rounded; gular surfaces
evenly stippled; abdominal blue of males sometimes restricted
to small sternal patches 10

10. Ventrals rounded; dorsolateral and lateral tubercles very poorly

developed; enlarged dorsals commencing on the nape (Cuicat-
lan, Oaxaca).

nelsoni

Ventrals sub-mucronate to mucronate; dorsolateral and lateral

tubercles well developed; enlarged dorsals commencing on the

shoulders just craniad of a line joining the anterior points of

insertion of the fore-limbs 11

11. Ventrals mucronate, prominently carinated laterally; gular scales

with a tendency toward pavementation, especially anteriorly;
gular surfaces evenly stippled; abdominal blue of males re-
stricted to small sternal patches (Guerrero east of Tierra
Colorada; Oaxaca, except the north-central portion; western
Chiapas).

anonymorphus
Ventrals sub-mucronate (occasionally rounded), only faintly
keeled on the lateral portions of the belly, or else not at all;
gular scales imbricate; gular region with an even blue wash,
and only barely flecked if at all ; abdominal blue of males evenly
distributed (Colima; Jalisco; southern Sinaloa; extreme south-
western Sonora).

tuberculatus

12. Enlarged dorsals commencing on the nape 13

Enlarged dorsals commencing on the shoulders, or caudad of

them 14

13. Enlarged dorsals strongly keeled, scales of primary and secondary

series almost equal in size; postfemoral dermal pocket absent
or rudimentary; no prominent lateral pattern of dark whorls
(Clarion Island, Revillagigedo Archipelago).

clarionensis
Enlarged dorsals not so rugose, scales of primary series promi-
nently larger than those of the secondary series ; postfemoral
dermal pocket regularly present; a distinct lateral pattern of
dark whorls (Tres Marias Islands; Tiburon Island; Sinaloa ?;
Sonora, south of the line Caborca-Magdalena).

schottii

mittleman: the iguanid genus urosaurus 131

14. Tail two or more times the length of head and body combined

(The Colorado River valley from extreme southern Nevada
to the Gulf of California; probably extreme north-western
Sonora; northeastern Baja California as far south as San
Felipe).

graciosus
Tail less than twice the length of head and body combined ... 15

15. Enlarged dorsals extending onto the basal portion of the tail for a

distance equal to the length of the femur, or more ; entire gular
region, including the sublabials, a uniform bright blue (the
northeastern quarter of Chihuahua).

cacrulcus

Enlarged dorsals extending onto the basal portion of the tail for

a distance equal to less than the length of the femur; entire gular

region including the sublabials never completely blue 16

16. Three or four series (rarely two) of enlarged dorsals on either side

of the vertebrals; enlarged dorsals not strongly keeled, not
prominently imbricate; dorsolateral and lateral tubercles not
affecting a diagonal arrangement; dorsolateral and lateral
folds variable, but usually absent or poorly developed; general

appearance not at all rugose 17

Two principal series of enlarged dorsals on either side of the verte-
brals; enlarged dorsals strongly keeled and prominently im-
bricate ; dorsolateral and lateral tubercles usually well developed
and affecting a diagonal arrangement; dorsolateral and lateral
dermal folds present; general appearance rugose 18

17. Enlarged dorsals quite flat, very weakly keeled, slightly imbricate,

or just as often pavemented; dermal folds absent or rudimen-
tary; dorsal basal tail scales barely or not at all differentiated
from the lateral basal scales (Rio Arriba and Sandoval Counties,
New Mexico).

I re is
Enlarged dorsals more convex, more strongly keeled, more imbri-
cate; dermal folds often present, with a fair degree of develop-
ment; dorsal scales on the basal portion of tail abruptly differ-
entiated from the much smaller, lateral basal scales (extreme
southwestern Colorado; northwestern New Mexico; eastern
and southern Utah; possibly extreme southern portions of
Nevada and Wyoming; Arizona, to the south and east of the
Colorado River in the northern portions of the state).

wrighti

132 bulletin: museum of comparative zoology

18. Enlarged dorsals often irregularly arranged; lateral tubercles not

affecting diagonal arrangements; average size less than 45 mm.

snout to vent 19

Enlarged dorsals regularly arranged in parallel series on either side
of the vertebrals; tubercles in parallel diagonal series; average
size greater than 45 mm. from snout to vent 20

19. Scales of primary series not twice as large as those of the secondary

series; largest of the dorsals inferior in size to the enlarged
femorals and tibials; ventral interhumeral and interfemoral
areas immaculate, or but lightly stippled (Texas, west of the
Pecos River in Brewster, Jeff Davis, and Presidio Counties;
adjacent Chihuahua).

schmidti
Scales of primary series almost twice as large as those of the second-
ary series; largest of the dorsals equal to, or larger than, the
enlarged femoral and tibial scales; ventral interhumeral and
interfemoral areas heavily maculated (Texas, east of the Pecos
River; probably adjacent Coahuila and Nuevo Leon).

omaius

20. Largest of the dorsals equal to, or larger than the enlarged fem-

orals ; vertebrals extending onto the basal portion of the tail for
a distance equal to half, or slightly more, of the length of the
femur; the entire gular region in males, except the sublabials,
an intense blue; head length/head width ratio averaging 81%
(Chiricahua and Dos Cabezos Mountains, Cochise County,
Arizona).

chiricahuae

Largest of the dorsals inferior in size to the enlarged femorals;

vertebrals extending only onto the rump, or but slightly further;

no uniform intense blue color present in male gular region ; head

length/head width ratio averaging 75% or less 21

21. Enlarged dorsals separated into two parallel series by the width

of the vertebral series, which is greater in width than the broad-
est of the enlarged dorsals ; prefrontals and frontonasals usually
three each; general coloration pallid, light tan above, whitish
below, males with bright blue abdominal patches ; average head
length/head width ratio 75.4%; average length, snout to vent,
55.1 mm. (Colorado River valley and deserts of southwestern
Arizona and southeastern California; northeastern Baja Cali-
fornia, and northwestern Sonora).

symmetricus

MITTLEMAN: THE IGUANID GENUS UROSAURUS 133

Enlarged dorsals separated by a vertebral series whose width is
less than that of the largest of the dorsal scales; prefrontals two,
rarely three (by the inclusion of an azygous) ; frontonasals five ;
general color variable, but usually dark brown or gray with
dark cross-bands, and heavily stippled, spotted, or blotched
ventrally; abdominal patches in males dark blue to indigo;
average head length/head width ratio 70.6%; average length,
snout to vent, 46.4 mm. (Arizona, south of Lat. 3o°30', except
in the Colorado River valley and desert; west to south-central
New Mexico, and south to Sonora and Chihuahua as far as

Lat. 31°).

linearis

Urosaurtjs ornatus ornatus Baird and Girard

1852 Uta ornata Baird and Girard, Proc. Acad. Nat. Sci. Phila., 6, 126 (part) '
Cope, Rept. U. S. Xat. Mus. 1898 (1900): 315 (part).

1921 Uta ornata ornata Schmidt, Amer. Mus. Nov., 22, 6 (part); Van Den-
burgh, Occ. Pap. Calif. Acad. Sci., 10 (1), 207, 1922 (part); Stejneger
and Barbour, Check List X. Amer. Amph. Rept., 1923: 52 (part); ibid,
1933:57 (part); ibid, 1939:62 (part); Mittleman, Herpetologica, 1940,
2, (2), 33; Mittleman, Jour. Wash. Acad. Sci., 1941, 31, 77.

Type locality. Rio San Pedro ( = Devil's River), Val Verde County,
Texas.

Cotypes. USNM 2750 (male and female).

Diagnosis. One or two rows of enlarged, keeled, imbricate, irregu-
larly arranged vertebral scales, extending from a point slightly caudad
of the insertions of the fore-limbs posteriorly to the basal portion of the
tail on which they continue for a distance equal to less than half the
length of the femur; vertebrals bordered on either side by two series
of larger scales which are keeled and prominently imbricate ; enlarged
dorsals not regularly dispersed ; scales of primary series approximately
twice as large as those of the secondary series; largest of the dorsals
superior in size to the enlarged femorals and tibials; other dorsal
scales very small, granular, and largely pavemented, or only but
slightly imbricate, except for a series on the dorsolateral line which
extends from a point anterior to the axilla posteriorly to the groin;
anteriorly, this series is continuous and forms a dorsolateral ridge, but
posteriorly it is broken up into small clusters of slightly enlarged, tuber-
cular scales around a central, much larger, mucronate scale; diagonally
arranged clusters of tubercles absent from lateral areas; on the lateral

134 bulletin: museum of comparative zoology

line a few sparse clusters of barely enlarged tubercles ; a fairly distinct
lateral fold; -two prominent series of cervical tubercles which extend
posteriorly from the ear and join the dorsolateral series of enlarged
scales immediately anterior to the axilla, although rarely they do not
so merge, and instead retain their individual identity; ventral to the
cervical tubercles and dorsolateral tubercles is usually a third, and
often a fourth series of tubercles, which are almost ventral in position;
posterior surfaces of thighs and arms covered with small granules sim-
ilar to those on the sides and dorsum of body, while the antero-dorsal
portions of the thighs, and a band around the tibiae are covered with
large, mucronate, keeled, imbricate scales; ventrals smooth, rounded
to submucronate, about equal in size to the scales forming the posterior
edge of the gular fold ; thirteen of the largest dorsal scales equal to the
length of the head from snout to posterior border of occipital; frontal
transversely divided; femoral pores 10-11; enlarged postanal plates
in males conspicuous; a postfemoral dermal pocket regularly present.
Coloration (alcoholic) : dorsal surfaces of body and tail grayish to
brown, with cross-bars of light brown which are edged with pale blue;
head tan with a few pale blotches of blue or gray; limbs dorsally blue-
gray with cross-bars of brown; venter of limbs and body whitish,
venter of digits tan; tail mottled with light brown anteriorly, and uni-
formly shaded with the same posteriorly; chin and gular region mottled
with blue and brown, labials a dark grayish blue which suffuses onto
the lateral portions of the head; venter of limbs and tail, as well as
interhumeral and interfemoral areas, heavily stippled with brown or
gray; bright blue abdominal patches in males. Description from
USNM 83117, male; 7 milts south of Babyhead, Llano County, Texas.
Measurements of fifty adults, both sexes; head length, 11.15 mm;
head width, 8.08 mm; snout to vent, 42.25 mm; hind leg, 27.60 mm.

Distribution. TEXAS: Llano, Burnet, Sutton, Kerr, Edwards,
Hays, Comal, Bexar, Uvalde, Guadalupe, and Val Verde (east of the
Pecos River) Counties. MEXICO: Probably Coahuila and Xuevo
Leon.

Remarks. This subspecies is unaccountably absent from most col-
lections of Mexican Urosauri. I have examined several specimens in
the United States Xational Museum which bear only the noncom-
mittal data "Border" or "Mexico"; these are undoubtedlv referable
to this race, but since they lack additional information, are useless
in further delineating the range of ornatus. Baird (1859:7) mentions
a specimen, USNM 2764, from Eagle Pass, Texas, which in itself
would prove the certainty of the occurrence of ornatus in Coahuila

MITTLEMAN : THE IGUANID GENUS UROSAURUS 135

at least. The specimen from Sonora, USNM 2737, also mentioned
by Baird (loc. cit.) is probably referable to linearis if the provenance
is accurate; I have not examined it. A single specimen in the United
States National Museum, USNM 78541, while an excellent example
of this form, is not included here, since it purportedly was taken in
Victoria County, Texas, which is considerably out of the known range
of other available material. Until further collecting substantiates
this record, I consider it best to temporarily delete it.

The type locality is here restricted to the Rio San Pedro, Texas,
and the reference to Sonora omitted, on the grounds that Baird and
Girard had a single specimen from this latter Mexican state, and at
the time confused it with their "Uta ornata". The Sonoran form is
linearis, which was later recognized and so named. The cotypes were
both taken at the Rio San Pedro, and are so catalogued. Since they
agree with other ornatus from the same general region, and since the
Sonoran specimen available at the time of original description was
not of the same race, the type locality is therefore restricted.

Urosaurus ornatus schmidti Mittleman
1940 Uta ornata schmidti Mittleman, Herpetologica 2, 2, 33, pi. 3, fig. 1

Type locality. Fort Davis, Jeff Davis County, Texas.

Type. USNM 32929, male.

Diagnosis. From original description : "Closely related to Uta ornata
ornata, but differing from that race as follows: Enlarged dorsal scales
arranged more regularly ; the inner series of enlarged dorsal scales not
twice as large as those of the outer series; largest of the dorsal scales
occasionally equal to, but more often smaller than, the enlarged,
keeled scales of the antero-dorsal surfaces of the tibia and femur;
enlarged dorsal scales commencing well caudad of a line joining the
anterior points of insertion of the fore-limbs. Lateral fold usually
incomplete when present. Dorsolateral series of tubercles and en-
larged scales usually quite prominent. Elongated series of tubercles
on neck somewhat more prominent. Coloration of both sexes similar,
in most respects, to that of ornata, save that the heavy ventral mot-
tling of the interhumeral and interfemoral areas found in ornata is
regularly lacking in schmidti, or at best is represented by a light
flecking of blue-gray. Measurements of holotype: Snout to posterior
border of ear, 11.5 mm; head width, 9 mm; snout to vent, 44 mm;
hind-limb (insertion to tip of 4th toe, exclusive of nail), 31 mm;
tail, 70 mm." The measurements of twenty-five adults of both sexes

136 bulletin: museum of comparative zoology

are as follows: head length, 10.65 mm; head width, 8.16 mm; snout
to vent, 43.20 mm; hind leg, 28.70 mm.

Distribution. Texas: Brewster, Jeff Davis, and Presidio Counties.
Mexico : North of Lat. 29° in Chihuahua and possibly Coahuila.

Remarks. Mexican records for schmidti, like in the case of ornatus,
are based principally on specimens in the United States National
Museum, bearing only the data "Border" or "Mexico". Apparently
the only true record for this subspecies from Mexico, is that of Smith
(1935:178), who reports "TJta ornate ornata" from near Samalayuca,
Chihuahua. Since this is definitely out of the range of ornatus, and
further, since the specimen bears no resemblance to Smith's caeruleus,
and finally, since it is entirely within the normal range of schmidti,
it is accepted as a bona fide record. As I have mentioned previously
(1940:34), Gadow's record (1905:194) for TJta elegans from Juarez,
Chihuahua, which Smith interprets as probably being referable to
ornatus ( = schmidti) , is more probably TJta stansburiana stejnegeri.

Two of the original paratypes (USNM 32932-3) from El Paso
County, Texas have been reexamined, and deleted from the range
of this form, since they have been found to be intergrades with linearis,
as some newly available material from southern New Mexico further
indicates. However, schmidti may still be taken in the southern por-
tion of this county. Probably it will also be disclosed in Hudspeth
and southern Culberson Counties.

Urosaurus ornatus caeruleus Smith

1935 Uta caerulea Smith, Univ. Kan. Sci. Bull., 12, (7), 172, pi. 26; Mittle-
man, Jour. Wash. Acad. Sci., 1941, 31, 76.

Type locality. Thirty miles north of Chihuahua City, Chihuahua,
Mexico.

Type. David H. Dunkle— Hobart M. Smith Coll. No. 132, now
deposited in the Kansas University Museum.

Diagnosis. Two vertebral rows of enlarged, irregularly arranged,
weakly carinated scales, extending from a point slightly craniad of a
line joining the anterior points of insertion of the fore-limbs, posteriorly
onto the base of the tail for a distance equal to the length of the femur;
vertebrals bordered on either side by two series of enlarged, imbri-
cate, weakly carinated scales, the primary series slightly larger than
the secondary; largest of the dorsals inferior in size to the largest of
the tibials; dorsolateral tubercles but slightly enlarged, and dis-
persed in irregular little clusters; ventrals rounded, smooth, imbricate;

mittleman: the iguanid genus urosaurus 137

frontal transversely divided; a post-femoral dermal pocket. Colora-
tion of male (from original diagnosis, loc. cit.): "Entire ventral sur-
faces of body and tail, except chest, base of tail, and an area between
the hind legs, sky blue; dorsum with1 about seven transverse black
bars on each side; bars usually blue-Sdged." Measurements of type
(Smith, loc. cit.): "Snout to anterior border of ear, 10.0 mm; head
width 10.0 mm; snout to vent, 49.5 mm; hind leg, 30.0 mm."

Distribution. Northeastern Chihuahua, south of Lat. 29°.

Remarks. I have recently pointed out (1941 :76) the close relation-
ship between schmidti and caeruleus, and mentioned the existence of
intergrades between these two forms. At the time, I hesitated to
formally designate caeruleus as a subspecies, of the ornatus complex,
because the extant material which indicated this intergradation bore
no more explicit data than "Mexico", for the most part. I had hoped
in the interim to acquire or locate specimens which would more pre-
cisely indicate the distribution of both these subspecies in Chihuahua.
Unfortunately, such specimens have not yet become available. How-
ever, since some specimens at least have been seen which do indicate
this relationship, and some definite action must be taken, caeruleus
is here accorded the trinomial.

The subspecies schmidti and caeruleus are quite easily separated.
Thus, in caeruleus, the enlarged dorsals extend onto the basal portion
of the tail for a distance equal to the length of the femur, or slightly
more, whereas in schmidti the distance traversed is rarely equal to
half the length of the femur; largest of the dorsals inferior to the
femorals in caeruleus, larger than the femorals in schmidti; caeruleus
with the gular blue wash including the sublabials, while schmidti
possesses the sublabials white or gray, but always distinct from the
remainder of the gular region.

Urosaurus ornatus linearis Baird

1859 Uta ornata var. linearis Baird, U. S. Mex. Bound. Surv., 2, 7; Cope,

Rept. U. S. Nat. Mus. 1898 (1900) :315.
1921 Uta ornata linearis Schmidt, Amer. Mus. Nov., 22, 6; Mittleman, Jour.

Wash. Acad. Sci., 1941, 31, 68; Mittleman, Proc. Biol. Soc. Wash.,

1941, 54, 165.
1875 Uta symmetrica Yarrow, Surv. W. 100th Mer., 5, 569; Boulenger, Cat.

Liz. Brit. Mus., 1885, 2, 213 (part); Cope, Rept. U. S. Nat. Mus. 1898

(1900) :317 (part); Van Denburgh, Occ. Pap. Calif. Acad. Sci., 1922,

10 (1), 202 (part).

Type locality. Los Nogales, Sonora, Mexico.

138 bulletin: museum of comparative zoology

Type. USNM 2759, lost or destroyed.

Neoti/pe. USNM 62077, female, Los Nogales, Sonora, Mexico. F. J.
Dyer, Collector.

Diagnosis. Superficially similar to U. o. ornatus and U. o. schmidti,
but principally differing as follows: enlarged dorsals usually com-
mencing just craniad of a line joining the anterior points of insertion
of the fore-limbs ; enlarged dorsals in two very regular series on either
side of the vertebrals; scales of the secondary series approximating
those of the primary series in size; enlarged dorsals larger, more
strongly carinated; cervical, dorsolateral, and lateral tubercles more
strongly developed; lateral tubercles usually affecting a diagonal
arrangement from axilla to groin; dorsolateral and lateral dermal
folds prominent; blue abdominal patches of males more often fused
medially; general appearance much more rugose and bristling; average
snout to vent size larger, maximum size attained greater (56 mm.
snout to vent, in largest recorded linearis; 51.5 mm. largest schmidti;
46 mm. largest ornatus).

Distribution. Arizona: "Widespread throughout the state south of
Lat. 35°30', and east of Long. 114°, except in Yuma County, where it
occurs only as far west as Long. 113°30'. New Mexico: Generally
south of Lat. 35°, except for the southeastern quarter of the state
bounded by Lat. 34° and Long. 106°. Mexico : Sonora and Chihuahua,
in the area bounded on the west by long. 113°30', on the south by
Lat. 31° (the line Reforma — Cananea — Sabinal — Lucero), and on the
east by Long. 106°30'.

Remarks. The present subspecies has unduly suffered at the hands
of various workers. Undoubtedly this is in a large measure due to
the original description (loc. eit.), which merely said "Similar in mark-
ings to the described character of U. ornata, but with four linear
interrupted black stripes instead of transverse bands." The error
being in many cases, the presumption that the linear striping
was supposedly the only specific character to be considered. The
linear striping can be shown in nearly every subspecies of ornatus in
the United States, and is a mere aberrant pattern variant. It is un-
fortunate that Baird rested the distinction of his species on such a
character. None the less, Schmidt (1921) pointed out the distinct-
ness of the central Arizonan and northern Sonoran form, and cor-
rectly applied the name linearis to this subspecies, thus distinguish-
ing it from the Texas and Californian races (ornatus and symmetricus).
His attempt to revive linearis from the synonymy was short-lived,
unfortunately, for Van Denburgh (1922:207) did not see fit to recog-

MITTLEMAN: THE IGUANID GENUS UROSAURUS 139

nize this distinction, and instead used "symmetrica" to designate the
several races inhabiting the Colorado River valley, most of Arizona,
northern Sonora and New Mexico; authors since Van Denburgh have
uniformly accepted his nomenclature, the exception being Stejneger
and Barbour (1923, 1933, 1939) who have properly restricted symmetri-
ca to the Colorado River valley and desert, but accord ornatus a
range extending from Texas to California.

The type of linearis was lost or destroyed some time prior to 1890;
in the interests of future workers I have designated a neotype.

For a summary of certain mensural data of linearis see the discus-
sions of the following two subspecies.

Throughout its range, despite a wide distribution and some degree
of superficial variation, linearis remains remarkably constant in its
chief diagnostic characters as given above. Occasionally local popu-
lations differ to a certain degree from the majority of available speci-
mens, but in nearly every case, these aberrations can be shown to be
of an incipient nature, and within the limits of variation of the sub-
species. In one case, however, distinctive morphological trends are
correlated with geographic isolation, such that the population is
worthy of formal designation. This population has recently (Mittle-
man, 1941:165) been named, and is discussed below.

Urosaurus ornatus chiricahuae Mittleman
1941 Uta ornata chiricahuae Mittleman, Proc. Biol. Soc. Wash., 64, 165.

Type locality. Pinery Canon, Chiricahua Mountains, 6000 ft.,
Cochise County, Arizona.

Type. MVZ 7751, male.

Diagnosis. "Resembling Uta ornata linearis superficially, but dif-
fering in the greater size of the enlarged dorsal scales ; the extension of
the vertebral series of enlarged scales onto the basal portion of the
tail for a greater distance; greater proportionate width of the head;
and a different arrangement of colors and pattern."

"Description of type. Two, occasionally three, rows of enlarged,
keeled, imbricate, irregularly arranged vertebral scales, extending
from a point a trifle craniad of the insertions of the fore-limbs pos-
teriorly onto the basal portion of the tail for a distance equal to half
the length of the femur; vertebral scales bordered on either side by
two series of regularly arranged, imbricate, and prominently keeled

140 bulletin: museum of comparative zoology

scales which are larger than the vertebrals, equal in size to the en-
larged femoral scales, and larger than the enlarged tibial scales; scales
of the inner series of enlarged dorsals not much larger in size than
those of the outer series, or else scales of both series approximately
equal in size; other dorsal scales very small, granular, or flattened,
lightly keeled and barely imbricate; on the dorsolateral line a series
of enlarged scales which extends from a point just anterior to the
axilla, posteriorly to a point just caudad of the groin; dorsolateral
enlarged scales tubercular, and disposed around other larger, spinose,
strongly carinated scales to form small clusters; distinct dorsolateral
and lateral folds present; lateral areas with prominent series of en-
larged tubercles diagonally dispersed; two short, prominent cervical
series of tubercles, and below these, a lateral series of the same, and
a ventrolateral series; lateral cervical tubercles merging with the
series of tubercles of the dorsolateral line; posterior surfaces of thighs
and arms covered with small granules, while the superior and anterior
surfaces are covered with large, keeled, mucronate, imbricate scales;
ventrals smooth, submucronate, about equal in size to the scales
posteriorly bordering the gular fold; eleven of the largest dorsal scales
equal to the length of head from snout to posterior border of occipi-
tals; frontal transversely divided; femoral pores 12-12; postanal plates
conspicuously enlarged; a postfemoral dermal pocket present. Col-
oration (alcoholic): Head light brown dorsally, with fine spots and
streaks of a darker brown; dorsum of body, limbs, and tail varying
from dark brown to a blue-gray, faintly splotched with light blue;
the five irregular cross bands which extend transversely from the
lateral fold to the enlarged dorsals and break on the median line of
the back, dark brown edged with light blue; venter of limbs with a
suffused blue-gray; chin, from anterior gular fold up to but not in-
cluding the sublabials, a bright sky blue; two large, brilliant, light
blue abdominal patches which are fused medially, and sprinkled with
dark gray laterally; interhumeral and interfemoral areas uniformly
shaded with dark gray, and a few flecks of blackish; preanal region
with a light blue wash. Measurements of type: Snout to posterior
edge of ear, 12 mm; head width 11.5 mm; snout to anus, 51 mm;
hind leg (insertion to tip of 4th toe, exclusive of nail), 33.5 mm; tail
(tip partially regenerated), 70 mm. Measurements of entire type
series (thirty-seven adults, both sexes) : Snout to posterior edge of ear,
11.32 mm; head width 9.16 mm; snout to anus, 47.70 mm; hind leg
(insertion to tip of 4th toe, exclusive of nail), 30.80 mm. (these figures
represent the weighted arithmetic means)."

MITTLEMAX: THE IGUANID GENUS UROSAURUS 141

"Distribution. Restricted to the type locality and the Dos Cabezos
Mountains, Cochise County, Arizona."

"Remarks. The present form is accorded a subspecific designation
on the basis of certain specimens from nearby localities in Cochise
County, which exhibit characteristics that must be considered inter-
mediate between chiricahuae and linearis. In the main, however,
such specimens are few, and chiricahuae is essentially different from
the linearis population of Cochise County, as well as from other
points in the distribution of this latter, parental form.

"The new subspecies exhibits certain mensural differences which
are best illustrated by various ratios. Thus, the type series shows an
average head length — head width ratio of 81 percent, the range being
from 75 percent to 96 percent, with three specimens having a ratio of
75 percent, twenty-six specimens with ratios varying from 79 percent
to 83.5 percent, and eight specimens with ratios between 8-1 percent
and 96 percent. By comparison, a test sample of thirty-seven adult
linearis of both sexes selected at random from a large series of speci-
mens taken at Ramsey Canyon, Huachuca Mountains, Cochise
County, Arizona, shows an average head length — head width ratio
of 70.8 percent, the range being from 58.25 percent to 83 percent;
only four specimens have ratios greater than 77.5 percent, while the
great majority are in the quartile sector of the mean. Certain other
ratios differ in linearis and chiricahuae, but none so markedlv as this
one. Test samples of linearis from Pima and Yavapai Counties, Ari-
zona, as well as some from certain counties in New Mexico, agree
well with the Ramsey Canyon sample in their morphological as well
as mensural details, and exhibit about the same differences towards
chiricahuae.

"The uniform blue color of the gular region in chiricahuae is very
rare in linearis, and is usually replaced by a medial blotch of yellow
or orange; similarly, the relative uniformity of color in the inter-
humeral and interfemoral regions in chiricahuae is, in the great majority
of linearis, replaced by a heavy stippling or mottling of dark gray
or brown.

"The subspecies linearis and chiricahuae are further differentiated
by the nature of the enlarged dorsal scales; in the former race, these
are usually smaller than the enlarged femoral and tibial scales, and
only rarely equal them in size. In chiricahuae, the enlarged dorsals
at least equal the femorals in size, and are consistently larger than
the tibials. Further, the vertebral series of enlarged scales in linearis
does not extend onto the basal portion of the tail for a distance equal

142 bulletin: museum of comparative zoology

to much more than one quarter the length of the femur; in chiricahuae,
the distance is equal to at least half the length of the femur, often
more."

Urosaurus ornatus symmetricus Baird

1858 Uta symmetrica Baird, Proc. Acad. Nat. Sci. Phila., December, p. 253;
Boulenger, Catl. Liz. Brit. Mus., 1885, 2, 213; Cope, Rept. U. S. Nat.
Mus. 1898 (1900) :317 (part); Stejneger and Barbour, Check List N.
Amer. Amph. Rept., 1917:52.

1921 Uta ornata symmetrica Schmidt, Amer. Mus. Nov., 22, 6; Van Denburgh,
Ccc. Lap. Calif. Acad. Sci., 1922, 10 (1), 202 (part); Stejneger and
Barbour, Check List N. Amer. Amph. Rept., 1923:52; ibid, 1933:57;
ibid, 1939:62.

Type locality. Fort Yuma, Imperial County, California.

Type. USNM 2760, lost or destroyed.

Neotype. USNM 2744a(1), male; Fort Yuma, Imperial County,
California; M. Thomas, collector.

Diagnosis. Closely related to Urosaurus ornatus linearis, but differ-
ing from that subspecies as follows: vertebral series of enlarged scales
wider than the breadth of the widest enlarged dorsal scale; usually
three prefrontals and three frontonasals, although occasionally there
is present an additional anterior pair of much reduced frontonasals
whose presence is due to the fission of the frontonasals proper; average
size larger, maximum size attained greater; coloration very pallid
above and below.

Distribution. The Colorado River desert and valley in Yuma and
Mohave Counties, Arizona; and Imperial, Riverside, and San Berna-
dino Counties, California; the Colorado River valley and desert to
the delta, in Baja California and the Gran Desierto of Sonora.

Remarks. The type locality for symmetricus has been generally
accepted (Stejneger and Barbour, 1939:62) as Fort Yuma, Arizona.
Although it makes no practical difference which side of the river the
type was actually collected on, there being no difference in the popu-
lations of either side, I am of the opinion that the proper point of
origin of the type was Fort Yuma, Imperial County, California. The
type locality, as originally stated (Baird, loc. cit.), was "Ft. Yuma,
Cal."; Dr. Stejneger advises me in a letter that the change to the

1 USNM 2744, sixteen specimens with the same data; neotype has been designated by "a"
scratched on the tag.

mittleman: the iguanid genus urosaurus 143

Arizona locality was influenced by the fact that Baird subsequently
(1859:7) referred this form to the "Gila River", which is in Arizona.
It seems to me that this in no way changes the provenance of the
type, for in addition to the original citation of the type locality, it
may be that Baird meant the valley of the Gila River where it be-
comes an affluent of the Colorado, which could be construed to logi-
cally include old Fort Yuma, California. Further, in several other
instances, much of Baird's data in the United States and Mexican
Boundary Survey Report is erroneous; in some cases due to typo-
graphical error, in others apparently due to errata in information or
data. Certainly, since there is little difference either way, and the
point is purely technical, it seems best to adhere to the type locality
as originally given.

A thorough investigation to find the type specimen, in 1890, by
Dr. Stejneger, failed to disclose it. Since it has yet to be found, and
was in all probability lost or destroyed, I have herewith designated a
neotype.

The most northerly point from which I have seen symmetricus, is
Needles, San Bernadino County, California. Mr. Klauber advises me
in a letter that the specimens of "Uta omata symmetrica" which he has
recently recorded from Willow Beach, Mohave County, Arizona
(1939 :S9), are actually referable to U. o. graciosus. I have examined
some of the specimens of "symmetrica!'' reported by Cowles and Bogert
(1936:37) from Black and Eldorado Canons, Clark County, Nevada,
and from Black Canon and Travertine Springs, Mohave County,
Arizona; I have similarly examined the "Uta omata" reported by
Yarrow (1883:56) from "Nevada"; I find that all of these specimens
are intergrades of U. o. linearis x U. o. wrigkti. Intergrades of sym-
metricus x linearis are apparently common from the Castle Dome,
Chocolate, Dome Rock, and Plomosa Mountains, all in Yuma County,
Arizona. This race is like many others of its genus, chiefly a boulder
and cliff dweller. Although it will occasionally resort to small trees in
the absence of its favorite medium, the ideal niche is on rocks, boulders,
and cliffs.

The usual arrangement of prefrontals and frontonasals demands
three of each; however, as noted above, occasional splitting will pro-
duce an additional, minute pair of frontonasals. Similarly, fusion is
often encountered. The present race is most easily distinguished from
linearis by the character of the enlarged dorsals which are widely
separated on either side of the vertebrals, the distance separating the
parallel rows of enlarged dorsals being greater than the width of the

144 bulletin: museum of comparative zoology

widest enlarged scale; in linearis the distance between the opposite
enlarged dorsal series is much less, due to the extreme crowding and
imbrication of the vertebrals. Typically, symmetricus is a pale sandy
color, with bright, but lightly tinted blue abdominal patches in the
males. However, as previously noted, melanism is frequent, and many
specimens, especially from the mountains, are a uniform blue-black
dorsally and ventrally. This latter condition obtains in every known
form of Urosaurus.

That symmetricus is a substantially and significantly larger form
than linearis is brought out by the average snout to vent measurement
of a series of 350 specimens of both sexes of the former race, which is
55.10 mm; the same measurement in a similar series of linearis aver-
ages 46.45 mm. A greater size is also attained by symmetricus, the
maximum recorded in over 450 specimens being 63 mm. snout to vent,
while the maximum size recorded for more than 1800 linearis is 56 mm.
A somewhat wider head is also characteristic of symmetricus, the aver-
age head length/head width ratio of the series of 350 specimens noted
above, being 75.4%; the same ratio in a random sample of 350 adult
linearis of both sexes is 69.23%.

Urosaurus ornatus graciosus Hallowell

1854 Urosaurus graciosus Hallowell, Proc. Acad. Nat. Sci. Phila., 7, 92.

1859 Uta graciosa Baird, U. S. Mex. Bound. Surv., 2, 7; Cope, Rept. U. S.
Xat. Mus. 1898 (1900) :325; Stejneger and Barbour, Check List N.
Amer. Amph. Rept., 1917:50; ibid, 1923:50; ibid, 1933:56; ibid, 1939:61;
Van Denburgh, Occ. Pap. Calif. Acad. Sci., 1922, 10 (1), 212.

1875 Uta gratiosa Coues, Surv. W. 100th Mer., 5, 596.

Type locality. Lower California ( = Southern California).

Cotypes. AXSP 8550-1, both males.

Diagnosis. A Urosaurus of the ornatus complex, closely resembling
symmetricus in scalation, size, and color, but distinguished by the tail
length which is two or more times the length of the head and body com-
bined, and by the absence of vertebral scales separating the enlarged
dorsals into two parallel series.

Distribution. The Colorado River desert and valley in Nevada:
Clark County; Arizona: Mohave and Yuma Counties; California:
San Bernadino, Riverside, Imperial and San Diego (fide Van Den-
burgh, 1922:214) Counties; Mexico: Baja California along the Gulf of

mittlemax: the iguanid genus urosaurus 145

California as far as San Felipe; probably the portions of Sonora adja-
cent to the Colorado River.

Remarks. The close alliance, yet differentiation existing between
graeiosus and symmctricus is a remarkable attestation to the sharp
ecological preferences and consistencies characteristic of Urosaurus
as a whole. The two subspecies are undeniably very close, and I can
only consider them as equivalent offshoots of the same stock, of about
equal age. Yet despite this superficial and possibly genetic nearness'
graeiosus and symmetricus remain distinct throughout the greater por-
tion of their mutual range. Insofar as I can determine, although doubt-
less there are other contributing factors, the chief isolating mechanism
seems to be simply a matter of ecological preference — graeiosus being
consistently a tree and shrub dweller, and symmctricus displaying an
obvious preference for boulders and cliffs. In the southern portion of
the range these ecological distinctions merge, and intergradation takes
place, having been observed chiefly from the vicinity of Yuma, Arizona.
Yet in itself this is further confirmation of the ecological postulation,
for in the intergrading region there is not the abundance of distinctive
ecological niches available for both forms which are found throughout
the greater portion of the range. Yet even in the vicinity of Yuma, both
graeiosus and symmetricus to a large extent remain distinct, and inter-
grades are notably few.

When intergradation occurs between graeiosus and symmetricus, the
intergrades possess the distinctively long tails of the former race, but
indicate their aberration by the nature of their enlarged dorsal scales,
which tend to become smaller along the median line, and simulate the
vertebral series characteristic of symmctricus and typically lacking in
graeiosus. Measurements and proportionate data for the two races
are about the same, save for a somewhat narrower head in symmetricus
and longer tail in graeiosus. In coloration and pattern, both forms are
nearly identical.

Urosaurus ornatus wrighti Schmidt

1921 Uta wrighti Schmidt, Amer. Mus. Nov., 22, 3.

1922 Uta levis Van Denburgh, Occ. Pap. Calif. Acad. Sci., 10 (1), 208;
Stejneger and Barbour, Check List X. Amer. Amph. Rept., 1923:51;
ibid, 1933:56; ibid, 1939:61.

Type locality. Grand Gulch, 4000-5000 ft., San Juan County, Utah.
Type. AMNH 18097, male.

146 bulletin: museum of comparative zoology

Diagnosis. Enlarged dorsal scales commencing well caudad of a line
joining the anterior points of insertion of the fore-limbs; usually two
vertebral series, the scales of which are very small, weakly keeled, and
bordered on either side by a primary and two or three secondary series
of enlarged dorsals, which are weakly to moderately keeled, and slightly
imbricate ; enlarged dorsals very gradually merging with the remainder
of the dorsal scales; cervical tubercles moderately developed; dorso-
lateral and lateral tubercles and dermal folds absent or rudimentary;
lateral scales at base of tail abruptly smaller than dorsal basal scales.
Coloration (alcoholic) of male : Dorsum of head, body, and limbs and
tail usually gray-blue, occasionally light tan or blue-black; limbs, body
and tail barred with narrow cross-bands of dark blue or grayish; supra-
labials and infralabials suffused with white, which diffuses through
part of the sublabials and gular region; median gular region with a
yellow, orange, or whitish light spot; abdominal blue patches fused
medially for the greater part of their length; interhumeral and inter-
femoral regions pale gray, mottled or not, but usually with dark gray
or blue; venter of limbs and tail lighter than dorsal surfaces. Measure-
ments of 200 adults, both sexes: Head length, 12.05 mm; head width
8.62 mm; snout to vent, 48.30 mm; hind leg, 30.70 mm.

Distribution. Colorado: Montezuma County; New Mexico: San
Juan and McKinley Counties; Utah: Uintah, Duchesne, Carbon,
Emery, Grand, San Juan, Wayne, Garfield, Kane and Washington
Counties; Arizona: Apache, Navajo, Coconino, and probably Mohave
Counties, north of Lat. 35° 30'.

Remarks. Intergradation with U. o. linearis is common throughout
the northern portion of Mohave County, Arizona, as well as in the
vicinity of San Francisco Mountain, Coconino County, Arizona.
Other cases of intergradation of linearis x wrighti have been previously
discussed. Specimens from the vicinity of St. George, Washington
County, Utah, are aberrant, and indicate the trend towards linearis.
This subspecies may also be expected to occur in extreme southwestern
Wyoming, as I have seen specimens from Vernal, Uintah County,
Utah, which is but forty miles south of the Utah-Wyoming line. How
far west it ranges is open to speculation ; it seems quite possible that it
will be found in northern Clark and southern Lincoln Counties, Nevada,
at least.

The present subspecies has suffered unmerited oblivion at the hands
of most authors, who have uniformly followed Van Denburgh (1922:
209) in regarding this form as being synonymous with the older levis,
from which it is quite definitely distinguishable.

mittleman: the iguanid genus urosaurus 147

This form is consistently a dweller of large boulders and open cliffs.
In the Navajo Country of Arizona and Utah, it was only in these asso-
ciations that wrighti could be found, showing an extreme predilection
for the vertical habitat. Indeed, we found this lizard extremely com-
mon along the sheer and overhanging cliffs which jut to meet the
water, along the upper reaches of the Colorado River, from the point
of affluence with the San Juan River in Utah, to Lee's Ferry, Arizona.
In the Navajo Country it is one of the commonest of all reptiles, and
is the first to appear in the morning, and one of the last to retire at dusk
Knowlton (1934:11; 1938:236) has shown that this subspecies is an
important insect pest control -agent in Utah. Eaton (1935:10) has
reported that courtship and copulation took place between a pair
observed on June 14th, at an elevation of 6000-6300 ft., in northern
Navajo County, Arizona.

Urosaurus ornatus levis Stejneger

1890 Uta levis Stejneger, N. Amer. Fauna, 3, 108; Cope, Rept. U. S. Nat.
Mus. 1898 (1900) :313; Stejneger and Barbour, Check List. N. Amer.
Amph. Rept., 1917:50; ibid, 1923:51 (part); ibid, 1933:56 (part); ibid,
1939:61 (part); Van Denburgh, Oce. Pap. Calif. Acad. Sci., 1922, 10
(1), 208 (part); Schmidt, Amer. Mus. Nov., 1921:6.

Type locality. Tierra Amarilla, Rio Arriba County, New Mexico.

Type. USNM 11474, male.

Diagnosis. Most closely related to Urosaurus ornatus wrighti, but
distinguished as follows: enlarged vertebrals and the three or four series
of enlarged dorsals bordering them on either side, extremely flat, very
weakly keeled, barely imbricate, or often pavemented; cervical, dorso-
lateral, and lateral tubercles absent or very poorly developed; lateral
fold occasionally present, but then very weakly differentiated; dorsal
scales of basal portion of tail very gradually merging to the lateral
basal scales, without any noticeable demarcation between dorsal and
lateral scales. Coloration (alcoholic) of male: olivaceous tan above,
slightly darker beneath; dorsum with six to eight greenish brown
wavy cross-bands, which break on the vertebral line; abdominal
patches bright blue, usually not fused medially. Measurements of ten
adults (six males, four females): head length, 10.7 mm; head width
7.3 mm; snout to vent, 42.8 mm; hind leg, 28.6 mm.

148 bulletin: museum of comparative zoology

Distribution. New Mexico: Rio Arriba and extreme northern
Sandoval Counties.

Remarks. The present subspecies is accorded the trinomial on the
basis of specimens from extreme southern Sandoval and Torrance
Counties, New Mexico, which are intermediate between this form and
linearis. Doubtless too, collecting in southwestern Colorado will reveal
further intergradation between levis and wrighti.

Urosaurus o. levis has had a peculiar history which I believe is
principally attributable to the rarity of this form in collections. Prior
to Schmidt's naming of ivrighti and Van Denburgh's synonymizing
of this race with levis, levis was properly considered the rare lizard it is.
When, however, the form which ranges so widely and commonly
over much of Colorado, Utah, and Arizona (wrighti) was believed
conspeeific with levis, this latter name crept into the literature quite
rapidly. Thus, Van Denburgh (1922), Knowlton (1934, 1938), Wood-
bury (1931), Smith (1935) and others, have variously discussed and
figured supposed levis; the fact remains that very few of these references
more than touch in part on any phase of the natural history or tax-
onomy of levis, nearly all the references bearing actually on wrighti.

Despite diligent inquiry and investigation, I have been able to dis-
cover only eleven specimens of levis in various collections. Ten of these
specimens are adults, and one is a juvenile. All of these specimens
have been taken in the comparatively rigidly proscribed area composed
of Rio Arriba and northern Sandoval Counties, New Mexico. These
specimens serve to show that while the alliance between levis and
wrighti is unquestionably close, the distinguishing characters given are
actually present and constant. Further, levis which is an end form of
wrighti (again through ecological isolation), is apparently closely
restricted to its own distinctive region, and is not a common animal.
Whether the rarity of this subspecies in collections indicates also its
rarity in nature, or simply a lack of collecting in the region it inhabits,
I cannot say; none the less, the distinctions between levis and wrighti
are real and correlated with geographic distribution.

W hile I hesitate to offer mensural data based on ten specimens of
lizards as highly variable as the Urosauri, it may be said that levis is
seemingly a smaller animal, with a proportionately narrower head,
and proportionately longer hind leg than wrighti.

The type of levis, as indicated previously, is actually a male. Cope
(1900:314) has correctly figured the type specimen, but in his table
on the same page has reversed the data; thus, USNM 11474, the type,
actually is a male, and USNM 8554, a paratype, is a female.

MITTLEMAN: THE IGUANID GENUS UROSAURUS 149

Urosaurus clarioxexsis Townsend

1890 Uta clarionensis Townsend, Proc. U. S. Nat. Mus., 13: 143.

Type locality. Clarion Island, Revillagigedo Archipelago, Mexico.
Type. USNM 15904, male.

Diagnosis. Enlarged dorsals strongly carinate, prominently imbri-
cate, and very regularly dispersed in two parallel series on either side
of a series of very small vertebrals which commence on the nape;
scales of primary series well developed, and slightly larger than those
of the secondary series; largest of the dorsals considerably larger than
the enlarged femorals anchor tibials; dorsolateral tubercles very well
developed, as are those of the lateral fold; between the dorsolateral
tubercles and the lateral fold are short series of tubercles which occa-
sionally present a diagonal arrangement; frontal transversely divided;
lateral abdominal scales mucronate and carinate. Coloration (alco-
holic) of male: general over-all color blue above and below; limbs and
tail lightly ringed with gray or black; dorsum with longitudinal dark
streaks. Measurements of six adults (three of each sex) : head length,
13.78; head width, 10.55 mm; snout to vent, 54.50 mm; hind leg 41 mm.

Distribution. Restricted to the type locality.

Remarks. This species is most closely allied to schottii, of existing
forms, and shows no close affinity for the neighboring auriculatus of
Socorro Island. Because of this, it assumes quite some importance in
the phylogenetic aspects of the Urosauri, and will be dealt with more
thoroughly in connection with this problem.

It can be confused with no other known Urosaurus, bv reason of its
distinctive morphology and greatly restricted habitat. It should be
added that although the postfemoral dermal pocket is absent in most
specimens, a few examples possess a very small whorl of reduced
scales which may be interpreted as a rudimental pocket.

I have not been able to learn anything of the natural history of this
species.

Urosaurus ornatus schottii Baird

1858 Uta schottii Baird, Proc. Acad. Nat. Sci. Phila., 10: 253.

1883 Uta (Phymatolepis) lateralis Boulenger, Ann. Mag. Nat. Hist., 5, (11):
342.

1884 Uta gularis Cragin, Bull. Washburn Lab. Xat. Hist., 1: 7.

1885 Uta lateralis Boulenger, Cat. Liz. Brit. Mus., 2: 214; Schmidt, Amer.
Mus. Nov., 1921:6.

1922 Uta ornaia lateralis Van Denburgh, Occ. Pap. Calif. Acad. Sci., 10 (1):

199; Mittleman, Jour. Wash. Acad. Sci., 1941, 31: 66.
1941 Uta ornata schottii Mittleman, Copeia, 3: 138.

150 bulletin: museum of comparative zoology

Type locality. Magdalena, Sonora, Mexico.

Type. USNM 2761, destroyed sometime prior to 1889. (See Mittle-
man, 1941:138).

Diagnosis. One to three vertebral rows of enlarged, imbricate, car-
inate, irregularly arranged scales, extending from the nape posteriorly
onto the base of the tail for a distance equal to about half the length
of the femur, and bordered on either side by two series of enlarged
dorsals, of which the primary series is considerably the larger; largest
of the dorsals greater in size than the enlarged femorals and tibials;
two or three elongated series of thoracic tubercles ; a dorsolateral series
of enlarged, mucronate tubercles, extending from the supra-axillary
or thoracic region to the basal portion of the tail; several lateral series
of enlarged, spinose granules; ventrals abruptly differentiated from
the granular laterals; abdominal and gular scales prominently imbri-
cate and submucronate ; frontal typically divided transversely; post-
femoral dermal pocket regularly present. Coloration (alcoholic) of
male: six to nine dark spots on the dorsolateral line from axilla to groin;
a vertebral series of smaller, alternating spots extending from the
nape to the basal portion of the tail; dorsolateral and lateral spots on
both sides usually joined by undulated brown bands, which are occa-
sionally broken medially; dorsal coloration of body and limbs gen-
erally light brown or gray, or occasionally a uniformly rufescent dark
brown which completely obliterates any semblance of pattern; limbs
barred above with dark brown; dorsum of tail similar to dorsum of
body, and lightly ringed with pale brown; lateral areas a light blue-
gray, and lightly streaked with irregular brown patches; abdomen
with two elongate, light blue patches, which may or may not be fused
medially; rostral and supralabials white, this color extending poster-
iorly in a narrow streak to the insertion of the fore-limbs; infralabials
flecked with gray; gular region anterior to the fold, light blue; under-
side of limbs, tail, and interhumeral and interfemoral areas, whitish.
Measurements of fifty adults, both sexes, insular and mainland: head
length, 12.6 mm; head width, 9.6 mm; snout to vent, 49.5 mm; hind
leg 35.9 mm.

Distribution. Tres Marias Islands; Tiburon Island; Sinaloa (Bou-
lenger, 1883 :342) ; Sonora, south of the line Caborca-Magdalena.

Remarks. I have recently shown (1941:138) that the long-standing
Uta lateralis Boulenger, 1883, is synonymous with the older Uta
schottii Baird, a name for many years considered of puzzling identity.
Since lateralis was demonstrated to intergrade extensively with linearis
(Van Denburgh, 1922:199; Mittleman, 1941:68), subspecific designa-

mittleman: the iguanid genus urosaurus 151

tion was made; similarly, therefore, schottii assumes the trinomial of
the name it has replaced.

The extent of the southerly distribution of schottii remains entirely
uncertain. Similarly, the status of Urosauri from Sinaloa generally, is
an unknown quantity. There is an utter dearth of material from this
Mexican state, and until the gap is filled, we are confronted with sev-
eral inexplicable problems. For instance, Boulenger (loc. cit.) has re-
corded schottii (under lateralis) from Presidio de Mazatlan, Sinaloa,
which is several hundred miles south of other known records for this
subspecies, except for those from the Tres Marias Islands. On the other
hand, U.b.tubcrculatus has been taken at the Presidio, which previously
constituted a new northerly record for the subspecies, until I reported
it from extreme southern Sonora, which again moved the distribution
of tuberculatus several hundred miles northward. Such a distribution is
paralleled by Sceloporus nelsoni (Smith, 1939:364), but is not dupli-
cated by any of the Urosauri. It is entirely possible that further col-
lecting in Sinaloa will explain these queer problems of distribution, and
possibly show that schottii and tuberculatus are simultaneously dis-
tributed throughout this state and extreme southern Sonora, but are
effectively separated by ecological and/or physiological barriers.

Comparisons made between large series of insular and mainland
schottii reveal only slight mensural differences, which are neither con-
stant nor marked enough to warrant any distinctions being made
between the two populations.

U. o. schottii and U. o. linearis are easily separated. Thus, schottii
possesses the enlarged dorsals on the nape, whereas in linearis they
commence about equal with a line joining the anterior points of inser-
tion of the fore limbs; schottii possesses the scales of the primary series
about twice as large as those of the secondary series, while linearis
has the scales of both series about equal in size. In linearis there is no
prominent pattern of dorsolateral spots in series, while in schottii
such a pattern is distinctive and always present, save in melanistic
individuals.

Summary of the ornatus Complex

The ornatus complex is that group within the genus Urosaurus char-
acterized by two or more principal series of enlarged dorsals on either
side of the vertebral series, or else with a uniform band of enlarged
scales which is not longitudinally separated by a vertebral series; in
addition to these premises, members of the ornatus complex regularly

152 bulletin: museum of comparative zoology

possess a postfemoral dermal pocket, with the exception of one form
in which it is occasionally present in a rudimentary condition; further,
the frontal plate is always transversely divided.

A critical study of Urosaurus indicates that schottii is at present
most nearly akin to the primitive form which largely gave rise to the
modern genus. To further corroborate this view, available data con-
form to several of Adams' criteria (1902), and indicate the general
region composed of western Sonora, the adjacent portion of Baja
California, and northern Sinaloa as the likely center of origin and dis-
persal of Urosaurus. It is notable therefore, that schottii largely popu-
lates this region, almost to the complete exclusion of other known
forms.

The form schottii seems to have developed two very distinctive
genetic lines; one characterized by the ornatus complex, the other by
the bicarinatus complex. Further, this genetic divergence must have
taken place at a relatively early date, for on the neighboring Clarion
and Socorro islands of the Revillagigedo Archipelago, we find the very
dissimilar clarionensis and auriculatus respectively, the former belong-
ing to the ornatus complex, and the latter to the bicarinatus complex.
Through the agency of the specific distinctness of these two forms, and
available knowledge concerning the geological history of west-coast
Mexico, the period of development of the two diverse genetic lines in
Urosaurus can be placed as having occurred sometime between the late
Oligocene and the early Miocene, for in this era the solid mass of land
which composed modern wes.t-coast Mexico and extended as far west
as the Revillagigedo Archipelago became immersed, and formed the
modern Californian peninsula as well as the numerous Gulf and Pacific
islands.

Probably during the early Miocene schottii, or more properly per-
haps, the pre-schottii form, commenced to successively spread and in-
vestigate the land masses to the north and south of the center of origin.
To the north, the proliferation of the line linearis-u-righti-levis formed,
while on the east-west axis, schrnidti-ornatus-carnrfrus developed to the
east, and symmetricus-graciosus to the west; chiricahuae is an end form
of linearis, and only owes its morphological distinctness to a forced in-
breeding necessitated by ecological limitations and restrictions. The
species clarionensis developed early in the history of the group, and
bears a closer resemblance to schottii than does any existing Urosaurus
known. The ornatus complex will therefore be seen to follow a pattern
of multiplication of species through isolation and mutation of older,
pre-existing species (Dunn, 1934), and to present an orthogenetic line
of dynamic status.

^ L

* 6

mor
■ do,

<b

-*-~ Cb

<~r> o,

^ -^

o o

^^

<b <-

k *?

Ch ^

enh
usuo

C3 «>

-1

~<1

Q ^

^Jo o

k o

C a

3

O

o

5

*0

£

.^

1

0)

^§

o

a

-*^'

c

<w

"^

■w

<o

k

0)

<ij

<l>

-*>

<o

10

o

CO

-~

c

_o
-13

a

eg
'3

v
a

00

—
cu

r3

5

CO
bt,

J2

o

o

154 bulletin: museum of comparative zoology

The trend in the ornatus complex has been, from south to north,
and from the central portion to peripheral sectors, towards a decrease
in the size of the enlarged dorsals, an increase in the size of the verte-
brals to ultimately produce a band of dorsal scales of uniform size
(notably attained in graciosus), a loss of heavy melanins dorsally, and
a greater degree of gular pigmentation. Further, peripheral forms tend
to develop proportionately wider heads, longer legs, and a lesser degree
of carination on the dorsal scales.

Urosaurus gadovi Schmidt
1921 Uta gadovi Schmidt, Amer. Mus. Nov., 22: 3.

Type locality. Cofradia, Jalisco, Mexico.

Type. AMNH 20355, male.

Diagnosis. Frontal usually entire, rarely split transversely into a
large anterior portion and a much smaller posterior section ; four to
seven rows of enlarged dorsal scales along the median line, of which the
median row is largest, and the remainder progressively diminish in
size; no vertebrals separating the enlarged dorsals into parallel longi-
tudinal series; enlarged dorsals extending posteriorly from the shoulders
to the basal portion of the tail, or often ending in the sacral region;
enlarged dorsals smaller than the enlarged femorals, and equal to or
smaller than the tibials; external to the enlarged dorsals are several
poorly defined, slightly enlarged clusters of tubercles, extending from
axilla to groin ; on the dorsolateral line of the neck and body a dermal
fold crested with numerous small series of enlarged tubercles inter-
spersed with larger, flat, mucronate scales; usually another, short fold,
restricted to the supra-axillary region; along the lateral areas are three
to five longitudinal series of clustered tubercles; lowest series of tuber-
cles often in contact with the ventrals; ventrals abruptly differentiated
from the lateral granules; gular scales flat, and largely pavemented;
scales of the gular fold elongate, mucronate, imbricate, and laterally,
faintly keeled; ventrals imbricate, mucronate to spinose, and faintly
keeled especially laterally, males with rather small postanals; post-
femoral dermal pocket absent. Coloration (alcoholic): Dorsal color
ranging from slatey gray (in melanistic specimens) to a lighter brown-
ish gray, or brown; head, tail, and median line of back slightly lighter;
three to six blackish cross-bands, about evenly distributed from
cervical region to sacrum, these being about as wide as three to five of
the largest dorsals; in all except the darkest specimens, the dorsolateral,
lateral, and ventrolateral surfaces are maculated with whitish areas,

mittleman: the iguanid genus urosaurus

155

these usually being restricted to small clusters of enlarged, tubercular
scales which dot these areas; entire labial and gular region, save for a
rounded area immediately anterior to the gular fold, spotted and irreg-
ularly streaked with black, these being heavier in the males; venter of
limbs, basal portion of tail, and interhumeral and interfemoral areas
whitish and lightly maculated with dark flecks; abdomen, from axilla
to groin, covered with an overlay of blue in both sexes, but anteriorly
more prominent and assuming an ovoid outline in males; tail faintly
circled with narrow brown bands. Because there is a greater quantita-
tive dimorphism between the sexes of this species, than in any other
form yet observed, the mensural data have been correlated as given
below, based on forty specimens (twenty-two males, eighteen females)
of adults :

mean

Measurement or ratio

males

females

of all
specimens

Head length

10.5—11.7—13.0

9.0—10.4—12.5

11.17

Head width

8.0— 9.4—10.0

7.0— 8.3—10.0

8.92

Head length /head width

75.0—80.5—87.0

75.0—79.4—87.0

79.25

Snout to vent

40.0—46.5—52.0

34.0—40.2—51.0

43.35

Hind leg

23.0—28.1—30.5

21.0—24.6—29.0

26.22

Tail length

68.0—73.5—80.0

51.0—61.4—77.0 |

67.46

The left and right figures

represent the lowest and highest measi

irements

and /or ratios c

•btained; bold figun

js are the means.

Distribution. Jalisco and Michoacan, Mexico.

Remarks. Urosaurus gadovi has been almost unknown since the time
of its original description; save for the type and five other specimens
(two in very poor condition), no other material has been extant until
very recently. In 1939, Dr. Hobart M. Smith had the good fortune to
collect thirty-seven specimens of this species, of which thirty-five were
sexually mature adults, at Apatzingan, Michoacan. These have bene-
fited the present study enormously, and have made it possible to pre-
sent a more complete picture of the variation of this form.

This species possesses smaller postanals (in males) than any other
form in the genus, these scales being so small in immature specimens
that determination of sex must be made by dissection. It is further
unique in that it is the only species in the genus wherein both males
and females possess the blue abdominal coloration normally restricted
to the males alone.

Urosaurus gadovi cannot be confused with any species of the genus
occurring on mainland Mexico. It is geographically remote from any

156 bulletin: museum of comparative zoology

related species of the genus; so much so, in fact, that several pertinent
facts in the phylogeny of Urosaurus remain a matter of conjecture and
uncertainty. In the main, however, gadovi appears most closely allied
to U. nigricaudus of Baja California, and is probably a mainland rem-
nant of the primitive Urosaurus stock which gave rise principally to
certain peninsular species.

Urosaurus irregularis Fischer

1882 Phymatole-pis {Uta) irregularis Fischer, Abh. Nat. Ver. Bremen, 7: 232,

pi. 17, figs. 1-4.
1885 Uta irregularis Boulenger, Cat. Liz. Brit. Mus., 2: 216; Schmidt, Amer.

Mus. Nov., 1921, 22:6; Smith, Zool. Ser. Field Mus. Nat. Hist., 24,

(4): 23.

Type locality. "... aus dem Hochlande von Mexico. ..."

Type. Municipal Natural History Collection of Bremen no. 437.

Diagnosis, (from original description) : dorsum covered with carin-
ated scales; along the dorsal median line larger keeled scales not ar-
ranged in regular rows; two longitudinal ridges converging posteriorly
on the back anterior to the pelvic region; the dorsal median series of
irregular, larger scales diffusing more and more towards the base of the
tail on both sides; after the fifth verticil of the tail these scales become
smaller, more mucronate, and less noticeable, and, commencing at this
point they arrange themselves with those of the ventral surface into
complete verticils which surround the tail up to the tip. Ground color
greenish-gray, ventrally lighter; on the back, three narrow, black, ob-
lique stripes, the laterad-most of which diverge outward posteriorly;
gular region speckled and marbled with yellowish gray and black;
ventral side light gray.

Distribution. "The plateau of Mexico."

Remarks. The above description is condensed from the original one
of Fischer's, and only those passages which are of any practical use are
given. I have utilized a literal translation of Fischer's phraseology; for
his kind help in the translation of the original paper, I am indebted to
Dr. Eugen H. Mueller, of Ohio University.

The status of irregularis is indeed a puzzling one, and I have given
the problem much thought, attacked it from every angle, and have
gotten a uniform result : that no opinion whatsoever can be offered con-
cerning this animal. Fischer's description is a lengthy mass of verbiage,
replete with minute data, but lacking such information as the sex of his

mittleman: the iguanid genus urosaurus 157

type, measurements, etc. It seems certain, however, if the description
and plate are at all accurate, that the type was a female, for Fischer's
notes conspicuously lack any reference to distinctive abdominal color-
ing, in an otherwise complete color description. Further, the ventral
figure in the plate does not show enlarged postanals.

Other than the original description and plate no other information
on this species is available. I have not seen the type (if it is indeed still
in existence), nor have the few museum catalogues which bear entries of
"irregularis" borne fruit. Smith (1939:23) reported irregularis from
Laguna Coyuca, near Acapulco, Guerrero (FMNH 25884), but this
specimen was found to be an intergrade of bicarinatus x anonymorphus.
Similarly, a specimen catalogued as irregularis in the Museum of Com-
parative Zoology collection, was actually an example of U. gadori. I
do not know from whence Schmidt's brief description (1921:6) of
irregularis was drawn, but I suspect that the original description of
Fischer was used. Although Boulenger's specimens (1885:216) have
not been available, I believe that they are also referable to bicarinatus
or one of its allies, with the original description furnishing the source
material for Boulenger's descriptive notes.

Although the extensive Mexican explorations and collections of
Nelson and Goldman, and more recently Taylor and Smith, have failed
to bring to light any specimens that can be considered conspecific with
irregularis as I know it from its original description, I prefer to retain
this form in name at least, as valid. I take this action on the basis of
Fischer's notes and plates, which if accurate to any degree, picture a
very distinct animal and one not to be considered synonymous with
any known form. Further, great regions which quite possibly are pop-
ulated with Urosauri, as the Durangian and Zacatecan uplands, are
very poorly known today, save that they possess a highly distinctive
endemic fauna. It is quite possible that irregularis will be found to
occupy a niche somewhere in this portion of "der Hochlande von
Mexico."

Urosaurus nigricaudus Cope
1864 Uta nigricaada Cope, Proc. Acad. Nat. Sci. Phila., :176.

Type locality. Cape San Lucas, Baja California.

Cotypcs. USNM 5307 (twelve specimens).

Diagnosis. Enlarged dorsals in seven to ten rows, commencing about
equal to the insertions of the fore-limbs, and extending posteriorly to
the sacrum; median rows of dorsals largest, and progressively diminish-

158 bulletin: museum of comparative zoology

ing in size as they extend laterally ; enlarged dorsals prominently keeled,
imbricate, rounded posteriorly; seventeen to twenty-four of the largest
dorsals equal to the length of head from tip of snout to posterior border
of the interparietal; dorsolateral and lateral folds usually present,
nearly always crested with enlarged, spinose scales; usually several
lateral clusters of enlarged tubercles; frontal usually entire, sometimes
transversely divided; enlarged femorals and tibials larger than any of
the dorsals ; ventrals mucronate on the gular fold, pectoral, and lateral
areas, but rounded elsewhere. Coloration (alcoholic) of male topo-
type: dorsum of body, limbs, head and tail ranging from grayish to
dark brown ; limbs and tail ringed with narrow bands of dark brown to
black; head finely lined with dark brown; body with nine alternating
short bars which extend from the dorsolateral fold to about the median
line of the back; dorsal bars about two or three scales wide, and of a
dark brown color, edged with pale blue posteriorly; labial regions
flecked with gray, as is also the gular region save for a light central
area which is a pale tan; pectoral region flecked rather heavily with
gray, as are also the undersides of the limbs and tail; abdomen with
two elongate sky blue patches which are partially fused medially; pre-
anal region with a blue wash; abdominal and lateral areas necked with
numerous individual scales which are a paler blue than the remainder
of the body. Measurements of fifty specimens, both sexes: head
length, 10.35 mm; head width, 7.75 mm; snout to vent, 42.0 mm;
hind leg, 28.0 mm; tail, 66.0 mm.

Distribution. South of Lat. 24°30' on Baja California peninsula;
also the islands of Espiritu Santo, Ballena, San Jose and Magdalena (?).

Remarks. This species seems closely restricted to the southern end
of the Californian peninsula, for with the exception of its possible oc-
currence on the Isla Magdalena (fide Van Denburgh, 1922:218), from
whence I have not seen any specimens, all available records are closely
clustered in the region extending from La Paz to Cape San Lucas.
The range of nigricaudvs, insofar as its insular distribution is con-
cerned, is in the main exclusive of microscutatus. However, this latter
species has been taken on San Jose Island by Linsdale (1932:362), and
I have similarly reported (supra) the occurrence of nigricaudus on this
island too, based on two specimens in the National Museum (USNM
24413-4). However, San Jose offers a multitude of ecological niches,
as I judge from Nelson's description (1921:92), and it seems entirely
possible that both forms can exist there without undue competition.
At any rate, the nigricaudus from this island do not differ in any way
from peninsular examples. I have not seen any nigricaudus from Mag-

mittleman: the iguanid genus urosaurus 159

dalena Island, although I have examined microscutatus from this lo-
cality. Here again it seems quite probable for both species to occur,
in view of the two suitable localities which are to be found on the
island, and again, both of these well separated (Nelson, 1921:89).
Linsdale (1932:361) mentions the occurrence of nigricaudus in a var-
iety of situations, notably all vertical, with no specimens having been
found on the ground.

Urosaurus microscutatus Van Denburgh

1894 Uta microscutata Van Denburgh, Proc. Calif. Acad. Sci., 2 (4), 298.
1900 Uta parviscutata Cope, Rept. U. S. Nat. Mus. 1898:324, fig. 45.

Type locality. San Pedro Martir Mountains, Baja California.
Type. Stanford University Museum no. 1221, male.

Diagnosis. A fairly small Urosaurus, characterized by the diminu-
tive size of the enlarged dorsal scales ; enlarged dorsals extending from
the shoulders to the sacrum, in about ten rows; dorsals keeled, rounded
posteriorly, pavemented or semi-imbricate, about thirty-two to thirty-
six in the length of the head from snout to posterior border of inter-
parietal ; dorsolateral and lateral dermal folds and tubercular clusters
present; otherwise similar to nigricaudus in structure and dorsal color-
ation; ventral coloration usually a uniform deep blue over the entire
gular and abdominal areas, except for the pectoral region, which is
grayish, and maculated with darker gray or black; central portion of
gular surfaces usually most intense blue; venters of limbs and tail
blue-gray to brown ; ventrolateral and lateral regions often with fleck-
ing of light blue which is usually restricted to a single scale; abdominal
blue patches in males usually fused for their entire length. Measure-
ments of fifty adults, both sexes: head length, 10.45 mm; head width,
8.05 mm; snout to vent, 39.75 mm; hind leg, 28.2 mm; tail 72.45 mm.

Distribution. Borego Palm Cafion, San Diego County, California,
south through the San Pedro Martir district and Lower Sonoran zone
of Baja California to Medano Amarillo at Lat. 24° {fide Linsdale,
1932:362); also the islands of San Marcos, Coronado, Carmen, Dan-
zante, San Jose, San Francisco and Santa Magdalena.

Remarks. This and the preceding species are certainly closely allied,
with microscutatus apparently a derivative of the older nigricaudus.
There is little to separate them, in a qualitative sense, unless it be the
ventral coloration. In other respects, they are chiefly distinguishable
by the nature of certain quantitative features. Possibly the simplest

160

bulletin: museum of comparative zoology

microscuiotus
nigr/caudus

Fig. 11. Distribution of Urosaurus mieroscutatvs and Urosaquus nigricaudus
in Baja California and the United States.

mittleman: the iguanid genus urosaurus 161

feature to verify in microscutatus is the size and number of the en-
larged dorsal series. However, these increase in size, and decrease
correspondingly in number, as the species progresses in a southerly
direction. I believe that intergradation between microscutatus and
nigricaudus will be shown when a more complete series of specimens of
both forms, from the region between Comondu and Baie Magdalena,
is available. In three specimens of microscutatus from Comondu
(USNM 65822-4), there is a definite tendency toward nigricaudus;
thus, the dorsals are larger than in more northerly microscutatus, and
the heavy blue wash so typical of this species is replaced by two ab-
dominal patches and a gular pigmentation quite akin to that of nigri-
caudus. While these specimens seem to definitely indicate an incipient,
if not actual intergradation, I hesitate to accord the species involved
the trinomials on such slim evidence.

Despite their close alliance and the certain derivation of one from
the other, microscutatus and nigricaudus are apparently everywhere
distinct geographically and ecologically (unless in the Comondu-Baie
Magdalena region previously discussed). Linsdale (loc. cit.) points out
that while microscutatus is often found on the ground, nigricaudus is
always restricted to the vertical habitat on trees, rocks, and fences.
Even so, available specimens indicate that there is no overlap of ranges,
unless it is in the hypothetical intergrading area. Where the two species
occur together, as on San Jose and Magdalena islands, they are even
then distinct ecologically. Both of these islands are only recently
isolated from the mainland of the peninsula, and represent two or more
leveled mountains, with intervening areas filled by the deposition of
sand by the sea. As such, they present diverse ecological niches which
are filled by the two species in question, each to the exclusion of the
other.

Mensural data indicate that while microscutatus is a somewhat
smaller form than nigricaudus, it possesses a proportionately longer
and wider head, a longer hind leg and a longer tail. Klauber (1939:89)
says that the blue gular color usually present in males of microscutatus
in the southerly portions of the range, is replaced by orange, with a
yellow central spot, in the northern regions. A few specimens from
Jacumba, San Diego County, California (USNM 75338) have the
same blue coloration of the gular region that is to be seen elsewhere in
the range of microscutatus. The only region where the blue gular color
is replaced by any other hue, as far as I know, is in the extreme southern
end of the range, in the vicinity of Comondu, where the few available
specimens display the orange (tan in preservative) throat and gular

162 bulletin: museum of comparative zoology

region characteristic of nigricaudus. If indeed microscutatus does
exhibit this color aberration in the northern part of its range, it would
be interesting to determine if it undergoes intergradation with any
other species in the genus. Specimens from the north are still too few
to determine this.

Summary of the nigricaudus Complex

The nigricaudus complex presents a small, compact group of four
more or less closely allied forms, of somewhat uncertain origins. They
compose a homogeneous group by reason of the usually entire frontal,
and uniform band of enlarged dorsals which are not divided into par-
allel series by the presence of vertebral, smaller scales; there is also a
certain uniformity of color and pattern, and a relative uniformity of
size.

While the uncertainty surrounding the status of U. irregularis has
been previously discussed, it may be said that this form (judging solely
from the original description and figures) seems to be the most primi-
tive one of the complex, and apparently the only one allied to any spe-
cies outside the complex. The derivation of microscutatus from nigri-
caudus has been indicated elsewhere ; similarly, the very close relation-
ship between nigricaudus and gadovi has been discussed. I presume
that nigricaudus is directly derived from gadovi in view of the greater
age of the Jalisco-Michoacan region which this latter species inhabits,
and the known history of the formation of the Californian peninsula.
Retrogressive reasoning, therefore, would show that the complex
arose somewhere near or on the west coast of Mexico, possibly in the
Jalisco-Michoacan region, and spread westward. Continuing in this
vein, it follows that some sort of relationship, perhaps not so tenuous,
exists between gadovi and the much-discussed irregularis. By the
process of elimination, it has been shown that irregularis would hypo-
thetically inhabit some portion of the Mexican highlands, or plateau,
a region of comparatively great age. Since this region is older than any
other known which supports any of the members of the complex;
and further, since the derivative nature of microscutahis and nigri-
caudus is known, as is also the relatively recent character of their
habitat, it would follow, I assume, that irregularis is the primitive
member of the nigricaudus complex. I have pointed out that in the
ornatus complex, there is a distinct trend toward an increase in size
of the vertebrals, which ultimately has resulted in the uniform size
of the enlarged dorsals in graciosus. It does not seem far-fetched at all

mittleman: the iguanid genus urosaurus 163

to suppose that a similar condition has produced the nigricaudus
group. Occasional specimens of U. b. Hcarinatus have the parallel
rows of enlarged dorsals interrupted by an inordinate increase in size
of a few of the vertebrals. A successive development such as this
would result in a form which we might today identify as U. irregularis.
From irregularis to the highly evolved mieroscutatus we can establish

a fairly certain line.

The trend in the line irrcgularis-gadon-nigrieaudus-mieroscutatus is
from a large, strongly imbricate, heavily keeled, mucronate enlarged
dorsal scale, to a much smaller one, which is only partly keeled, occa-
sionally pavemented, and distinctly rounded posteriorly. Similarly,
in the same line, the ventrals change from sharply mucronate, spinose,
keeled scales, to rounded, smooth ones. There is also a gradual in-
crease in the number of rows of enlarged dorsals; thus in irregularis
the number is two or three, in gadovi four to seven, in nigricaadus
seven to ten, and in mieroscutatus, at least ten, and occasionally twelve
to fourteen. The heavily crested dorsolateral folds which form actual
ridges, progressively diminish, from irregularis to mieroscutatus, until
in this latter form," the folds are only topped with small clusters of
slightly enlarged tubercles. In addition to these premises, from irregu-
laris to mieroscutatus there is a tendency to produce a smaller body and
a longer tail.

Urosaurus auriculatus Cope
1871 Uta auriculata Cope, Proc. Boston Soc. Nat. Hist., 14: 303.

Type locality. Socorro Island, Revillagigedo Archipelago.

Type. USNM 7027, female.

Diagnosis. A fairly large member of the genus, unique in the pos-
session of smooth enlarged femorals; two rows of vertebrals, which are
rather weakly carinate, imbricate, and regularly arranged, bordered
on either side by a single series of enlarged dorsals, which are relatively
flat and weakly keeled, and not very regularly dispersed; frontal
divided or entire; external to the primary series of enlarged dorsals
there are often one or two additional series of irregularly arranged
scales which do not continue in even rows, and are of variable size;
along the dorsolateral fold a variable series of enlarged scales, some
tubercular and others flat and mucronate, these in irregular groups;
post-femoral dermal pocket variable; coloration ranging from grayish
to bright blue, with six to eight short cross-bars along the dorsum in
alternating position, which are light-edged (usually with pale blue) on

164 bulletin: museum of comparative zoology

their posterior borders ; light lateral and ventral neckings of pale blue,
restricted to single scales, or small groups of tubercular scales; males
with extensive deep blue abdominal patches, females with an uneven
light blue abdominal wash. Measurements of seven adults, both
sexes: head length, 16.5 mm; head width, 12.45 mm; snout to vent,
65.5 mm; hind leg, 44.0 mm.

Distribution. Restricted to the type locality.

Remarks. As indicated under the discussion of U. clarionensis as
well as the remarks dealing with the ornatus complex as a whole,
auriculatus is well separated from the nearby clarionensis of Clarion
Island, and rather than exhibiting any morphological similarity with
this latter species, the alliance is with the bicarinatus group.

This species is one of the largest in the genus, and appears to have
adapted itself to its insular habitat very well, if size and the number of
specimens inhabiting the island (fide Cope, loc. cit.; Van Denburgh,
1922:199) may be considered sound criteria on which to base such a
supposition. The smooth enlarged femorals are unique in the genus,
and of uncertain phyletic importance and derivation. As a whole,
the species is not a rugose one, and seems to have lost much of the
bristling appearance common to other Urosauri; it is likely that a
chance mutation, or series of mutations, has caused the appearance
of these peculiar characters, and have become prominently evidenced
by the enforced inbreeding of the island population.

Urosaurus bicarinatus bicarinatus Dumeril

1856 Phymatolepis bi-carinatus Dumeril, Arch. Mus. Hist. Nat. Paris, 8:
549, pi. 23, figs. 2, 2a, 2b.

1864 Uta bicarinata Cope, Proc. Acad. Nat. Sci. Phila., 16: 177; Boulenger,
Cat. Liz. Brit. Mus., 1885, 2:215; Schmidt, Amer. Mus. Nov., 1921,
22: 6; Smith, Univ. Kan. Sci. Bull., 1935, 12 (7): 169.

1941 Uta bi-carinata bi-carinata Mittleman, Jour. Wash. Acad. Sci., 31: 70.

Type locality. "Mexico".

Type. Not designated; if in existence, probably in the Museum
d'Histoire Naturelle de Paris.

Diagnosis. Two or three vertebral series of enlarged scales extend-
ing in a continuous or often broken line from the nape to the basal
portion of the tail; on either side of the vertebrals is a single series of
greatly enlarged, strongly carinate, imbricate scales, which are often
irregularly dispersed, and frequently interrupted by the intrusion of
small heterogeneous scales, or the aberrantly large size of a vertebral

mittleman: the iguanid genus urosaurus 165

scale; largest of the dorsals subequal to the enlarged femorals and
tibials, which are strongly carinate; external to the enlarged dorsal
series, a series of enlarged scales which almost equal the primary dor-
sals in size and rugosity ; these latter scales sometimes in contact with
the dorsals, but more often separated from them by the varying width
of two to four of the small granular scales of the dorsum; these en-
larged scales bordering the primary dorsals are also irregular in size
and disposition, but usually occur in a line which commences anterior
to the primary dorsals; two or three series of thoracic tubercles; a
prominent series of enlarged tubercles, and mucronate and spinose
flatter scales on the dorsolateral line ; ventral to the dorsolateral series
of enlarged scales, are four other series of enlarged tubercles, the lowest
of which is in contact with the ventrals; ventrals mucronate, and pos-
teriorly and laterally, they become spinose, as well as keeled; gular
scales granular and pavemented, except along the gular fold, where
they are elongated, mucronate, and heavily imbricate; frontal most
often entire ; postfemoral dermal pocket absent, but occasionally repre-
sented by a minor whorl of underdeveloped scales. Coloration (alco-
holic) of male: dorsum of head, body and tail grayish, with four to
six dark cross-bars, which are occasionally obliterated by a heavy de-
position of dark melanins throughout the skin; the cross-bars are
usually broken medially, and often alternate; dorsum of body and
limbs often flecked with varying shades of gray or brown, lateral areas
similar, but usually tinted with a bluish wash; venter of limbs, and
interhumeral and interfemoral areas of- varying shades of gray, but
most often heavily flecked with darker gray or brown; ventral basal
portion of tail gray to brown, flecked with darker; long, deep blue
abdominal patches, which may or may not be overlaid with a heavy
stippling of brown or dark gray; except for a median light area occa-
sionally, the entire gular surface including the infralabials, is heavily
mottled with black, gray, or brown. Smith (1935:170), reporting on
freshly collected specimens from Morelos and Guerrero, says "the
entire gular region is orange, coarsely reticulated or diagonally barred
with black except in a large, round median area just anterior to the
gular fold." Measurements of fifty adults, both sexes: head length

12.5 mm; head width, 10.6 mm; snout to vent, 52.5 mm; hind leg,

29.6 mm.

Distribution. Michoacan, Morelos, Puebla, and Guerrero west of
Acapulco.

Remarks. Because Dumeril designated only "Mexico" as the type
locality for bicarinaius, and this only by inference, subsequent authors

166 bulletin: museum of comparative zoology

have reported many specimens ostensibly of this species, from widely
divergent points in Mexico (see Smith, 1935:171). Actually, as
Schmidt (1921) pointed out, at least two of the populations of what had
previously been accepted as bicarinatus, were referable to separate
subspecies. It is therefore obvious that the majority of previously
published records for bicarinatus are actually referable not only to
this form, but to several others, which do not occur within the range
of bicarinatus. Schmidt (op. cit.), Smith (loc. cit.), and I, have re-
viewed the populations of bicarinatus, and find, insofar as available
material indicates, that bicarinatus is distributed as given above.

I have pointed out (1940:34) that specimens from the southern peri-
phery of the range of bicarinatus, notably in the area between Acapulco
and Tierra Colorada, Guerrero, differ from the more northerly popula-
tion of this subspecies. Ultimately, in extreme southeastern Guerrero,
and continuing eastward through most of Oaxaca to southwestern
Chiapas, these lizards become clearly recognizable as a separate
biological entity, and have been so designated (Mittleman, loc. cit.).

Smith (1935:171) reports "The species is apparently entirely arbor-
eal. Some specimens were found on some of the large species of cactus
of the genus Opuntia. Their coloration is extremely protective; they
were frequently discovered only by striking likely -looking trees with a
shovel or heavy stick. Usually two or more occurred together on the
same tree or cactus."

Urosaurus bicarinatus anonymorphus Mittleman

1940 Uta anonymorpha Mittleman, Herpetologica 2, (2): 34, pi. 3, fig. 2.

1941 Uta bi-carinata anonymorpha Mittleman, Jour. Wash. Acad. Sci., 31: 71.

Type locality. Tehuantepec, Oaxaca, Mexico.

Type. USNM 46988, male.

Diagnosis. Enlarged vertebrals and the single series of enlarged
dorsals commencing on the shoulders only slightly craniad of a line
joining the anterior points of insertion of the fore-limbs; enlarged dor-
sals in a continuous series, or only barely encroached upon by a few
of the small granules of the back; enlarged dorsals regularly arranged,
forming parallel rows on opposite sides of the vertebrals, and not too
strongly carinated; external to the enlarged dorsals is a sparse series
of large scales, which are a trifle larger than the vertebrals but never
approach the enlarged dorsals in size; this series of scales never in con-
tact with the enlarged dorsals, but in contact at several points with the
rather poorly developed dorsolateral series of tubercles and mucronate

mittleman: the iguanid genus urosaurus 167

scales, through the medium of small, elongate clusters of slightly
enlarged, granular scales; thoracic tubercles not well developed, and
not prominent; dorsolateral and lateral tubercles often not enlarged,
and similarly often not in clusters, instead there is usually a single
enlarged, tubercular scale, surrounded by a few mucronate, spinose,
flat scales; ventrals mucronate and spinose, especially posteriorly;
laterally, the ventrals become carinate and quite distinctly pave-
mented; gular scales granular, flat, and pavemented, except in the
region of the fold, where they are elongate, spinose, lightly carinate,
and imbricate; general appearance not very rugose; frontal always
entire ; postfemoral dermal pocket regularly lacking. Coloration (alco-
holic) of male holotype: quite similar to that of bicarinatus, save that
the ventral blue, or blue-black patches are very abbreviated and re-
stricted to the pectoral area. Chin usually not so heavily maculated as
in bicarinatus; occasional specimens are uniformly suffused with a deep
blue-gray which completely obliterates any traces of the dorsal pat-
tern. Measurements of holotype: head length, 11.5 mm; head width,
9.0 mm; snout to vent, 50.0 mm; hind leg, 27 mm.

Distribution. Guerrero, east of Tierra Colorada; Oaxaea, except the
north-central portion; western Chiapas (Tonola).

Remarks. At the time of original description of anonymorphus I
postulated a probable subspecific relationship between this form and
bicarinatus. Somewhat later (1941:72) I was able to report that such
a relationship actually existed, as borne out by a series of 98 specimens
taken by Dr. Hobart M. Smith from eastern Guerrero to Chiapas, as
well as others from Morelos, western Guerrero, and elsewhere in the
range of bicarinatus. Dr. Smith's fine series of specimens show that
bicarinatus apparently terminates in the vicinity of Acapulco, Guer-
rero, and that in the relatively short distance from this city to Tierra
Colorada, Guerrero, the transition from this form to anonymorphus
takes place, with this latter subspecies becoming clearly recognizable
from the vicinity of Tierra Colorada, and continuing southward and
eastward. The extension of the range of anonymorphus to include
Chiapas is on the basis of two specimens taken by Dr. Smith at Tonola.
While these two individuals possess certain aberrations which to a
degree remove them from anonymorphus, they resemble this form most
closely, and are considered to be such, at least until further specimens
will have been taken from more easterly points in Chiapas.

Generally speaking, anonymorphus is quite easily separated from
bicarinatus. In males, the abbreviated blue abdominal patches are
quite distinctive in anonymorphus; while the ventral coloration tends

168 bulletin: museum of comparative zoology

to take on the appearance of an evenly diffused wash in bicarinatus
males. In specimens of either sex, anonymorphus can be identified at
once by its much less rugose appearance, and the definitely weaker
carination of the enlarged dorsals. More often too, anonymorphus will
possess an evenly mottled chin, whereas bicarinaius has a tendency to
possess a light median area, as has been noted by Smith (1935:170).
The holotype and the paratypes on which anonymorphus is based,
agree very well with the series of 98 specimens taken by Smith.

Urosaurus bicarinatus nelsoni Schmidt

1921 Uta nelsoni Schmidt, Amer. Mus. Nov., 22: 4.

1941 Uta bi-carinata nelsoni Mittleman, Jour. Wash. Acad. Sci., 31: 72.

Type locality. Cuicatlam (=Cuicatlan), Oaxaca, Mexico. ■

Type. USNM 46836, male.

Diagnosis. Most closely allied to bicarinaius and anonymorphus,
from which races it differs only as follows: ventrals not mucronate;
dorsolateral and lateral series of tubercles poorly developed; head
narrower proportionately than in bicarinatus, and proportionately
broader than in anonymorphus; enlarged dorsals smaller. Measure-
ments of type: head length, 13.5 mm; head width, 10.5 mm; snout to
vent, 58.0 mm; hind leg, 33.0 mm.

Distribution. Restricted to the type locality.

Remarks. I am familiar with nelsoni only through the type specimen,
which is distinct enough from other subspecies of bicarinatus, as given
above. I hesitate to postulate further on it, since it is so little known.
My reasons for considering nelsoni a subspecies of bicarinatus have
been previously (1941 :73) given as follows: "U. b. nelsoni is designated
as a subspecies of bi-carinata for the following reasons: The marked
similarity in structure to the typical form and anonymorpha; the con-
tinuity and contiguity of its distribution with the bi-carinata — anony-
morpha stock, the ranges of all three being juxtaposed; and the possi-
bility that nelsoni represents an intermediate population in position
between the bi-carinata — anonymorpha stock, and some form, as yet
undescribed, from extreme northwestern Oaxaca and possibly southern
Veracruz." Since the writing of the above passage, neither additional
specimens nor information have come to hand, so that further investi-
gation of this form must await additional, comparative material.

mittleman: the iguanid genus urosaurus 169

Urosaurus bicarinatus tuberculatus Schmidt

1921 Uta tuberculata Schmidt, Amer. Mus. Nov., 22: 4; Smith, Univ. Kan.
Sci. Bull., 1935, 12 (7): 171.

1941 Uta bi-carinata tuberculata Mittleman, Jour. Wash. Acad. Sci., 31: 73.

Type locality. Coliina, State of Colima, Mexico.

Type. AMNH 13737, male.

Diagnosis. Most closely related to bicarinatus, from which it differs
as follows: enlarged dorsals larger, more regularly arranged, and equal
to or greater than the enlarged femorals and tibials; external to the
enlarged dorsals, but in contact with them or separated by only one
or two granules, is a series of slightly enlarged scales which are visibly
keeled, but neither so large nor so prominent as the primary dorsals;
dorsolateral and lateral tubercles and enlarged spinose scales very
regularly arranged, but not so prominent as in bicarinatus; lowest
series of tubercles in contact with the ventrals which are not sharply
differentiated from the granular scales of the sides; ventrals rounded,
occasionally submucronate ; slightly keeled laterally; gular scales
elongate and imbricate, except those immediately adjacent to the
infralabials, where they are granular and pavemented; frontal variable,
usually divided; postfemoral dermal pocket variable, occasionally
present; coloration similar to bicarinatus. Measurements of type:
head length, 12.0 mm; head width, 9.0 mm; snout to vent, 45.0 mm;
hind leg, 27.0 mm.

Distribution. Discontinuous, recorded from Colima and Jalisco
(Schmidt, loc. cit.); Presidio de Mazatlan, Sinaloa (Smith, 1935:171)
20 miles southeast of Alamos, Sonora (Mittleman, 1941:73).

Remarks. "Other than some slight variation in color and pattern,
the specimens I have seen agree rather well with the type, differing
only in a few minor points.

"Because of a dearth of Utas from southern Sonora to central
Jalisco, the distribution of tuberculata is imperfectly known. First
known from Jalisco and Colima, the type series remained unique until
Smith (loc. cit.) reported a specimen taken by him just south of Presidio
de Mazatlan, Sinaloa, which extended the range northward for about
two hundred miles. In the course of an examinati'on of Mexican Utas
in the collection of the Museum of Comparative Zoology I came upon
two specimens, MCZ nos. 37856-7, collected near Guirocaba, 20
miles southeast of Alamos, Sonora. These two specimens are quite
typical of the subspecies, and on the basis of their locality, the range

170 bulletin: museum of comparative zoology

of tuberculoid is extended northward again for another two hundred
and eighty miles. Dr. Smith tells me in a letter that this closely corre-
sponds to the distributional pattern of Sceloporus nelsoni.

"U. b. tuberculoid is obviously a member of the neotropical bi-
carinata stock; just what its relationships with the nearactic lateralis
might be must await the discovery of further specimens from Sinaloa,
southern Sonora, and northern Jalisco." (Mittleman, 1941:73-74).

Urosaurus unicus Mittleman

1879 Uta bicarinata Cope, Proc. Amer. Phil. Soc, 18: 261; Bull. U. S. Nat.
Mus., 1887, 32: 35; Amer. Nat., 1896, 30: 1020; Rept. U. S. Nat. Mus.
1898 (1900) :320.

1941 Uta unica Mittleman, Jour. Wash. Acad. Sci., 31 (2): 74, figs, lc, 2, 3.

Type locality. (?) Batopilas, Chihuahua, Mexico.

Type. USNM 14248, female.

Diagnosis. "Cephalic scales comparatively smooth; frontal entire,
separated behind from the interparietal by a pair of frontoparietals;
rostral much wider than high; supralabials 5-5, the fourth and fifth
subocular in position; infralabials 7-7; auricular opening anteriorly
denticulated by several enlarged, granular scales; a few scattered en-
larged scales on the nape and shoulders, extending caudad from a point
just posterior of a line joining the anterior insertions of the fore limbs,
along the vertebral line onto the base of the tail for a distance sub-
equal to the length of the femur, is a series of enlarged scales, bordered
on each side by a single series of much larger scales, which are, how-
ever, inferior in size to the enlarged scales of the femur, but larger than
the enlarged tibial scales; enlarged dorsal scales only weakly carinated,
and prominently pavemented; external to the enlarged scales and in
contact with them, or more often separated by the width of the
vertebral series, is another series of enlarged scales, spaced about two
scale lengths apart; these latter equal to or slightly smaller than the
enlarged scales bordering the vertebral series; the outer enlarged scales
often surrounded by minutely enlarged tubercular scales; on the
dorsolateral, lateral, and ventrolateral areas are evenly dispersed four
longitudinal series of small clusters of slightly enlarged, somewhat
convex scales, which are not at all rugose; the lowermost of these
rows of clusters barely in contact with the ventrals; ventral scales
imbricate and mucronate anteriorly, but medially, laterally, and

mittleman: the iguanid genus urosaurus 171

posteriorly, they become rounded and quite pavemented, again be-
coming spinose and imbricate as they approach the anal region;
ventrals abruptly diminishing in size to meet the lateral scales; gular
scales pavemented and rounded anteriorly, but mucronate and imbri-
cate posteriorly, and noticeably increasing in this tendency, until in
the region of the gular fold the scales are longer than wide and dis-
tinctly spinose; gular fold extending laterally and dorsally around the
anterior edge of the insertions of the fore limbs, and met by a heavy
postauricular fold; caudal scales large, prominently keeled, spinose,
and at least basally, in irregular whorls of three verticils, of which the
first is always prominently largest; postfemoral dermal pocket absent.
Coloration of holotype (alcoholic) : Dorsum of head and body greenish
gray, the head finely reticulated with light brown, and the body with
two light brown bands which are narrow on the vertebral line and
widen as they progress laterally; dorsum of the body irregularly
flecked and barred with dark brown; axillary, inguinal, lateral, pre-
humeral, posthumeral, and postanal regions washed with dark brown;
an irregular dark brown pectoral blotch ; gular area and the remainder
of body and tail venter a very pale greenish gray; limbs narrowly
barred with light brown. Cope (loc. cit.) [1900:322] describes the
specimen which was then fresh, as having "limbs and tail shaded with
reddish brown," and says further that the "inferior regions tinted
yellow lightly stippled with brown; males have the entire abdominal
region bluish gray." Measurements of holotype: Snout to posterior
border of ear, 11.5 mm; head width, 9.0 mm; snout to vent, 50.0 mm;
hind leg (insertion to tip of 4th toe, exclusive of nail), 26.5 mm; tail
52.0 mm." (Mittleman, Loc. cit.)

Distribution. At present known only from extreme southwestern
Chihuahua.

Remarks. This very unique species is quite different from other
known Urosaurus, and as far as I am able to determine, it represents a
dwarf offshoot of a probable pre-tubcrculatus stock. Its distinctness
indicates a considerable age, and long separation from other members
of the genus.

When I recently described the species on the basis of the sole speci-
men known, I presented all of the available data on its provenance,
simply that it was collected in Chihuahua, Mexico, by Edward Wilkin-
son. Since then, certain data have accumulated, to some degree cir-
cumstantial, which appear to indicate that the type locality of unicus
is properly Batopilas, Chihuahua. Because several other genera and
species of reptiles were probably collected at the same locality and

172 bulletin: museum of comparative zoology

time by Wilkinson, and because data on these forms have been
similarly lacking, I present a brief summary of what is known con-
cerning these animals, as well as unicus.

The first published reference to a herpetological collection made by
Edward Wilkinson, is that of Cope's (1879:261). Here Cope lists eight
species of lizards, and seven of snakes, all taken at Batopilas, Chi-
huahua. Included in this list is "Uta bicarinata" with which form Cope
later (1900:320) confused the specimen now designated as the type of
U. unicus. Somewhat later, Cope again made reference to the Batopilas
collections of Wilkinson, and said (1887 :35), " Uta bicarinata Dumeril . . .
Batopilas, Chihuahua, Wilkinson. City of Chihuahua, Wilkinson . . ."
Still another reference to the "bicarinata" and the Batopilas collection
was made (1896:1020), when Cope mentioned "A small collection made
by Wilkinson in southern Chihuahua at Batopilas has the character of
the Chihuahuan fauna, with the following species not otherwise found
in it . . . Uta bicarinata Dum." Finally, Cope says (1900:322) of
"Uta bicarinata", "This species occurs throughout Mexico, as far
north as the City of Chihuahua, where it was obtained by Mr. Edward
Wilkinson." However, on the page preceding, the figure of Cope's
"bicarinata" (fig. 43), based on USNM 14248, now the type of unicus,
bears only the legend "Chihuahua, Mexico."

I think the preceding quotations and references show that there
can be little doubt that the origin of unicus is Batopilas. Cope's last
reference to "bicarinata" from the City of Chihuahua is certainly
attributable to one of his careless oversights. Either he mistakenly
considered a specimen of another Urosaurus from the City of Chihua-
hua to be one of his "bicarinata" ', which does not seem likely, or else
he confused the locality data of his specimen, which was probably
shipped to the National Museum from this city, although collected
at Batopilas. Furthermore, in Cope's original paper on Wilkinson's
Batopilas collections (1879:262) appears the description of a new
species, Procinura aemula, which Cope later refers to as Scolecophis
aemulus (1900:1109), and which, significantly enough, he says was
collected at Batopilas. Since the Procinura or Scolecophis was origin-
ally listed with the "Uta bicarinata" as having been taken at Bato-
pilas, and the reference again given at a later date, Cope's other
references to "bicarinata" from as far north as the City of Chihuahua
are plainly explicable as stated above. In view of extensive collections
throughout the greater part of Mexico, which indicate that bicarinatus
does not range anywhere near the City of Chihuahua, or even in the
state itself, and further, since Chihuahua is known to possess several

mittleman: the iguanid genus urosaurus 173

forms of Urosaurus, except in the southwestern corner from whence
no specimens have been available (other than unicus if it was actually
taken there), in the light of the evidence adduced, I consider Batopilas
as the type locality of unicus.

As a further matter of record, it may be stated that through some
sort of agreement between Cope and Wilkinson, the collections of the
latter went to Cope through the National Museum, probably to avoid
payment of duties. Also, Wilkinson made at least two other collections,
in the vicinity of the City of Chihuahua, for Trimorphodon vilkinsonii
was taken there, and described by Cope in 18S5, while a single bird egg
is catalogued in the National Museum as having been taken in the
Sierra Eulalia, near the City of Chihuahua, Wilkinson, collector, 1886.
Naturally, there is the possibility of Wilkinson's having spent several
years collecting throughout Chihuahua, and sent several collections
from different points within the state. Cope apparently kept all the
Wilkinson specimens in his personal collection for some time, and then
donated them to the National Museum, and catalogued them all at
once. This is corroborated by the fact that all Wilkinson specimens
are catalogued in the National Museum under the numbers 14222-
14310. Since the "Uta bicarinata" and other specimens reported by
Cope in 1879, as well as the type of Trimorphodon vilkinsonii which was
collected several years later are all catalogued together, it bears out
the contention that the numerous specimens were deposited in the
National Museum simultaneously. Incidentally, Cope, by previous
arrangement with Wilkinson apparently, sent him twenty of the speci-
mens after they had been catalogued (Nos. 14257, two specimens;
14263-4; 14270*; 14274; 14276; 14281-2; 14296, six specimens; 14303,
five specimens). For none of these is there any information as to
provenance or identity. They have never been located, although there
is some reason to believe that they may be housed with some of Wil-
kinson's personal effects (if they are indeed still in existence) in the
Mansfield (Ohio) Memorial Museum.

Summary of the bicarinatus Complex

The several species and subspecies which are here included in the
bicarinatus complex have common bonds by reason of the frontal
which is nearly always entire, the usual lack (or only rudimentary de-
velopment) of the postfemoral dermal pocket, and chiefly, the single
principal row of enlarged dorsals on either side of the vertebrals.

174 bulletin: museum of comparative zoology

With the exception of unicus, whose phyletic relationships and deriva-
tion can only be hazarded hypothetically, the constituent forms of the
complex are closely united, and differ only in degree, rather than kind.
They occur, more or less continuously, from southern Sonora along
western Mexico to western Chiapas.

The neotropical bicarinatus complex appears to have been directly
derived from schottii, or a pre-schottii form, at about the time of the
early Miocene, and must have then invaded the newly emerged land
lying to the south of what is now Sinaloa, for prior to this era, most of
west-coast Mexico from northern Sinaloa to Jalisco, and south to the
Yucatan Peninsula, was inundated. The divergence of the complex
has been hampered only by the high Sierra Madre del Occidental to
the east, and the Pacific Ocean on the west. The sole member of the
complex which occurs anywhere other than on the Mexican mainland
(auriculatus , of Socorro Island), is an ancient relict form, whose ap-
pearance has been most fortuitous (see discussion of clarionensis and
the ornatus complex).

With the exception of unicus, which for the present can only be dis-
missed as a relict end-form, the bicarinatus group has resulted through
direct orthogenetic activity, and the constant drive into previously
unoccupied territory, accompanied by genie changes within successive
end populations. There is little apparent difference in the ecology of
the several forms, so that it can only be assumed that constant differ-
ences and the comparative lack of wide intergradation are due to
physiological barriers, rather than ecological and/or geographic ones.

The complex has formed in the line auriculatus-tuberculatus-unicus-
bicarinatus-nelsojii and anonymorphus, the last two possibly having
occurred simultaneously. In this line, we find that the postfemoral
dermal pocket is variable in auriculatus and tuberculatum, absent in
unicus, nearly always absent or at best rudimentary in bicarinatus,
and always lacking in nclsoni and anonymorphus. Similarly, from
north to south, the trend is from a fairly rounded, smooth ventral, to a
mucronate or spinose, and carinated one. In the progression from
north to south, the frontal assumes a more stable character, so that
ultimately in nelsoni and anonymorphus only an occasional aberrant
possesses a divided frontal.

It seems entirely within reason to expect the occurrence of still an-
other Urosaurus in eastern Chiapas. I have mentioned that specimens
of anonymorphus from Tonola are aberrant. In addition, it may be
pointed out that the general region including eastern Chiapas was
separated from the Tehuantepecan area by a post-Miocene immersion,

mittleman: the iguanid genus urosaurus 175

and did not emerge and become continuous again until the Pliocene,
thus affording ample time for the differentiation of another race,
especially if an endemic population remained. A similar case in point
has been found in Cncmidophorus (Burt, 1931:73).

176 bulletin: museum of comparative zoology

Literature Cited
Adams, C. C.

1902. Southeastern United States as a center of geographical distribution
of flora and fauna. Biol. Bull., 3: 115-131.

Baird, S. F.

1858. Description of new genera and species of North American lizards
in the museum of the Smithsonian Institution. Proc. Acad. Nat.
Sci. Phila., December: 253-256.

Original descriptions of Uta symmetrica and Uta schottii.

1859. Reptiles of the boundary. U. S. Mex. Bound. Surv., II, 2: 1-25,
pis. 1-41.

Original description of Uta ornata var. linearis, as well as locality
data and redescriptions of ornata, symmetrica, schottii, and graciosa.

Baird, S. F. and C. Girard

1852. Reptiles (pp. 336-365, 8 pis.) in HOWARD STANSBURY,
Exploration and survey of the valley of the Great Salt Lake of Utah,
including a reconnaisance of a new route through the Rocky Moun-
tains. Philadelphia, Lippincott, Grambo & Co. 487 pp., pis., maps.
Original description of UTA .

1852. Characteristics of some new reptiles in the museum of the Smith-
sonian Institution. Proc. Acad. Nat. Sci. Phila., 6, 125-129.
Original description of Uta ornata.

BOULEXGER, G. A.

1883. Descriptions of new species of lizards and frogs collected by Herr
A. Forrer in Mexico. Ann. Mag. Nat. Hist. (ser. 5), 11: 342-344.
Original description of Uta (Phymatolepis) lateralis.

1885. Catalogue of the lizards in the British Museum (Natural History).
London, printed by order of the trustees. 2, xiii + 497 pp., 23 pis.
Original description of PETROSAL7 RUS. Redescriptions and
locality data for Uta auriculata, Uta lateralis, Uta ornata, Uta
symmetrica, Uta gratiosa, Uta nigricauda, Uta irregxdaris, and
Uta bicarinata.

Burt, C. E,

1931. A study of the teiid lizards of the genus Cnemidophorus, with spe-
cial reference to their phylogenetic relationships. Bull. U. S. Nat.
Mus., 154: 1-286, figs. 1-38.

Cope, E. D.

1863. Descriptions of new American squamata in the museum of the
Smithsonian Institution. Proc. Acad. Nat. Sci. Phila.. 1863:
100-106.

MITTLEMAX: THE IGUANID GENUS UROSAURUS 177

1864. Contributions to the herpetology of tropical America. Proc. Acad.

Xat. Sci. Phila,, 16: 166-181.

Original description of Uta nigricauda.
1871. Description of the common lizard of Socorro. Proc. Boston Soc.

Nat. Hist., 14: 303.

Original description of Uta auriculata.
1879. Eleventh contribution to the herpetology of tropical America.

Proc. Amer. Phil. Soc, 18: 261-277.
1887. Catalogue of batrachians and reptiles of Central America and

Mexico. Bull U. S. Xat. Mus., 32: 1-98.
1896. The geographical distribution of batrachia and reptilia in North

America. Amer. Xat., 30: 886-902, 1003-1026.
1900. The crocodilians, lizards and snakes of North America. Rept.

U. S. Xat. Mus. for 1898:153-1270, figs. 1-347, pis. 1-35.

Redescriptions and locality data for Uta ornata, Uta levis, Uta

symmetrica, Uta schottii (description only), Uta bicarinata, Uta

nigricauda, Uta parviscutata, and Uta graciosa.

Cotjes, E.

1875. Synopsis of the reptiles and batrachians of Arizona; with critical
field notes, and an extensive synonomy. U. S. Geog. Surv. Tf".
100th Mer., 5:585-633.

Principally locality data for Uta gratiosa, Uta ornata, and Uta
symmetrica.

Cowles, R. B. and C. M. Bogert

1936. The herpetology of the Boulder Dam region (New, Ariz., Utah).
Herpetologica, 1, (2): 33^3.
Locality data for Uta graciosa and Uta ornata symmetrica.

Cragin, F. W.

1884. Notes on some southwestern reptiles in the cabinet of Washburn
College. Bull. Washburn Lab. Xat. Hist., 1: 6-8.
Original description of Uta gularis.

Dickerson, M. C.

1919. Diagnoses of twenty-three new species and a new genus of lizards
from Lower California. Bull. Amer. Mus. Xat. Hist., 41, (10):461~
477.
Original description of SATOR, and discussion of its relationships.

Dtjmeril, A.

1856. Description des reptiles nouveaux ou imparfaitement connus de la
collection du Museum d'Histoire Naturelle et remarques sur la
classification et les carecteres des reptiles. Arch. Mus. Hist. Xat-
Paris, 8: 439-588, pis. 17-24.
Original description of Phymatolepis bi-carinatus.

178 bulletin: museum of comparative zoology

Dunn, E. R.

1934. Systematic procedure in Herpetology. Copeia, 4: 167-172.

Eaton, Jr., T. H.

1935. Report on amphibians and reptiles of the Navajo Country. Bull.
3, Rainbow Bridge — Monument Valley Expedition, Berkeley,
California, pp. 3-19, 3 pis.

Distributional and ecological data for Uta ornata symmetrica
( = U. o. wrighti).

Fischer, J. G.

1882. Herpetologische bemerkungen vorzugsweise iiber stucke der
sammlung des Naturhistorischen Museum in Bremen. Abhand.
Naturwiss. Ver. Bremen, 7: 225-238, pis. 14-17.
Original description of Phymatolepis {Uta) irregularis.

Gadow, H.

1905. The distribution of Mexican amphibians and reptiles. Proc. Zool.
Soc. London, 2: 191-244, figs. 29-32.

Hallowell, E.

1854. Description of new reptiles from California. Proc. Acad. Nat. Sci.
Phila., 7: 91-97.
Original description of URO-SAURUS (graciosus).

Klauber, L. M.

1939. Studies of reptile life in the arid southwest. Bull. Zool. Soc. San
Diego, 14: 1-100, tables 1-19.

Locality and ecological data for Uta microscutata, Uta graciosa,
and Uta ornata symmetrica.

Knowlton, G. F.

1938. Lizards in insect control. Ohio Jour. Sci., 38, (5): 235-238.
Food habits of Uta lens.

Knowlton, G. F. and M. J. Janes

1934. Distributional and food habits notes on Utah lizards. Copeia, 1:
10-14, fig. 1.
Locality data and food habits of Uta levis.

LlNSDALE, J. M.

1932. Amphibians and reptiles from Lower California. Univ. Calif.
PxM. Zool., 38, (6): 345-386.

Locality and ecological data for Uta microscutata, Uta nigricauda,
Uta graciosa, and Uta ornata symmetrica.

1940. Amphibians and reptiles in Nevada. Proc. Amer. Acad. Arts and
Sci., 73, (8) : 197-257, figs. 1-29.

Principally locality data for Uta graciosa and Uta ornata sym-
metrica.

mittleman: the iguanid genus urosaurus 179

MlTTLEMAN, M. B.

1940. Two new lizards of the genus Uta. Herpetologica, 2, (2): 33-38,
pi. 3.
Original descriptions of Uta ornata schmidti and Uta anonymorpha.

1941a. A critical synopsis of the Mexican lizards of the Uta ornata com-
plex, and a description of a new species from Chihuahua. Jour.
Wash. Acad. Sci., 31: 66-81, figs. 1-3.

Original description of Uta unica. Redescriptions and locality
data for Uta ornata lateralis, Uta auriculata, Uta clarionensis, Uta
bi-carinata bi-carinata, Uta bi-carinata tubercidata, Uta bi-carinata
nelsoni, Uta bi-carinata anonymorpha, and Uta caerulea.

1941b. The status of Uta schottii Baird. Copeia, 3: 136-138.

Revives Uta ornata schottii Baird, to take precedence over Uta
lateralis Boulenger and Uta gularis Cragin.

1941c. A new lizard of the genus Uta from Arizona. Proc. Biol. Soc. Wash.,
54: 165-168.
Original description of Uta ornata chiricahuae.

Nelson, E. W.

1921. Lower California and its natural resources. Mem. Nat. Acad. Sci.,
16: 1-194, pis. 1-35.
Lists Uta graciosa, Uta microscutata, and Uta nigricauda.

Schmidt, K. P. •

1921. New species of North American lizards of the genera Holbrookia
and Uta. Amer. Mus. Nov., 22: 1-6.

Describes Uta gadovi, Uta wrighti, and Uta tubercidata. Also key
to known forms of Uta.

Simpson, G. G. and A. Roe

1939. Quantitative zoology. McGraw-Hill, New York, xv-f 1-414,
figs. 1-52, tables 1-13.

Smith, H. M.

1935. Descriptions of new species of lizards from Mexico of the genus

Uta, with notes on other Mexican species. Univ. Kan. Sci. Bull.,

12, (7): 157-183, pi. 26.

Original description of Uta caerulea. Notes on Uta tubercidata,

Uta bicarinata, Uta ornata ornata, and Uta ornata lateralis. Notes

on the phylogeny and relationships of Uta.
1939. Notes on Mexican reptiles and amphibians. Zool. Ser. Field Mus.

Nat. Hist., 24, (4) : 15-35.

Lists Uta irregularis from Laguna Coyuca, Acapulco, Guerrero.
1939. The Mexican and Central American lizards of the genus Scelo-

porus. Zool. Ser. Field Mus. Nat. Hist., 26: 3-397, figs. 1-59, pis.

1-31.

Some brief mention of the relationships and derivation of Uta.

180 bulletin: museum of comparative zoology

Stejneger, L.

1890. Annotated list of reptiles and batrachians collected by Dr. C.
Hart Merriam and Vernon Bailey on the San Francisco Mountain
Plateau and Desert of the Little Colorado, Arizona, with descrip-
tions of new species. N. Amer. Fauna, 3: 103-118.
Original description of Uta levis.

1894. Description of Uta mearnsi, a new lizard from California. Proc.
U. S. Nat. Mus., 17: 589-591.
Original description of Uta mearnsi.

Stejneger, L. and T. Barbour

1917. A check list of North American amphibians and reptiles. Cam-
bridge, Harvard University Press, iv+5-125.

1923. A check list of North American amphibians and reptiles. 2nd ed.
Cambridge, Harvard University Press, x+1-171.

1933. A check list of North American amphibians and reptiles. 3rd ed.
Cambridge, Harvard University Press. xiv + 1-185.

1939. A check list of North American amphibians and reptiles. 4th ed-
Cambridge, Harvard University Press, xvi+1-207.

Taylor, E. H.

1935. A taxonomic study of the cosmopolitan scincoid lizards of the
genus Eumeces. Univ. Kan. Sci. Bull., 23, (1): 1-643, pis. 1-43.

Townsend, C. H.

1890. Reptiles from Clarion and Socorro Islands, and the Gulf of Cali-
fornia, with descriptions of a new species. Proc. U. S. Nat. Mus.,
13: 143.
Original description of Uta clarionensis .

Van Denburgh, J.

1894. Descriptions of three new lizards from California and Lower
California, with a note on Phrynosoma blainvillii. Proc. Calif.
Acad. Sci., (ser. 2), 4: 296-301.
Original description of Uta microscutata.

1922. The reptiles of western North America. I. Lizards. Occ. Pap.
Calif. Acad. Sci., 10: 1-611, pis. 1-57.

Redescriptions and locality data of Uta auriculata, Uta clarionen-
sis, Uta ornata lateralis, Uta ornata symmetrica, Uta levis, Uta
graciosa, Uta nigricauda, and Uta microscutata.

Woodbury, A. M.

1931. A descriptive catalog of the reptiles of Utah. Bull. Univ. Utah, 21,
(5), x+1-129, figs. 1-58.
Description and discussion of Uta levis.

mittleman: the iguanid genus UROSAURUS 181

Yarrow, H. C.

1875. Report on the collections of batrachians and reptiles made in por-
tions of Nevada, Utah, California, Colorado, New Mexico, and
Arizona, during the years 1871-72-73 and 74. U. S. Geog. Sum.
W. 100th Mer., 5: 511-584, pis. 16-25.
Principally locality data for Uta ornata and Uta symmetrica.

1883. Check list of North American reptilia and batrachia. Bull. U. S-
Nat. Mus., 24: 3-249.

EXPLANATION OF PLATES

PLATE 1

Mittleman — The Iguanid Genus Urosaurus

PLATE 1

Urosaurus ornatus ornatus Baird and Girard, cotypes, USNM 2750, Rio San
Pedro, Val Verde County, Texas.

BULL. MUS. COMP. ZOOL.

Mittleman: Iguanid Genus Urosaurus. PlateI

PLATE 2

Mittleman — The Iguanid Genus Urosaurus

PLATE 2

Urosaurus ornatus schmidti Mittleman, type, USNM 32929, Fort Davis,
Jeff Davis County, Texas.

BULL. MUS. COMP. ZOOL.

Mittleman: IguanioGenus Urosaurus. Plate 2

PLATE 3

Mittleman — The Iguanid Genus Urosaurus

PLATE 3

Urosaurus ornatus linearis Baird, neotype, USNM 62077, Los Nogales,
Sonora, Mexico.

BULL. MUS. COMP. ZOOL.

Mittleman: IguanidGenus Urosaurus. Plate3

*-"K -.. '

wX

. >

PLATE 4

Mittleman — The Iguanid Genus Urosaurus

PLATE 4

Urosaurus ornatus symvietricus, neotype, USNM 2744a, Fort Yuma, Im-
perial County, California.

BULL. MUS. COMP. ZOOL.

Mittleman: IguanidGenus Urosaurus. Plate4

PLATE 5

Mittleman — The Iguanid Genus Urosaurus

PLATE 5

Urosaurus ornatus chiricahuae Mittleman, type, MVZ7751, Pinery Canon,
Chiricahua Mountains, 6000 ft., Cochise County, Arizona.

BULL. MUS. COMP. ZOOL.

Mittleman: IguanidGenus Urosaurus. Plate 5

i

• HB •

PLATE 6

Mittleman — The Iguanid Genus Urosaurus

PLATE 6

Urosaurus ornatus levis Stejneger, type, USNM 11474, Tierra Amarilla, Rio
Arriba Countv, New Mexico.

BULL. MUS. COMP. ZOOL.

Mittleman: I guani d Genus Urosaurus. Plate 6

'*S^3 ''

'^'-ikJi^

PLATE 7

Mittleman — The Iguanid Genus Urosaurus

PLATE 7

Urosaurus ornatus graciosus Hallowell, cotypes, ANSP 8550-1, Lower
California (= Southern California).

BULL. MUS. COMP. ZOOL.

Mittleman: IguanidGenus Urosaurus. Plate 7

J'

f«~

%*S?j

PLATE 8

Mittleman — The Iguanid Genus Urosaurua

PLATE 8

Urosaurus clarionensis Townsend, type, USNM 15904, Clarion Island,
Revillagigedo Archipelago, Mexico.

BULL. MUS. COMP. ZOOL.

Mittleman: Iguanid Genus Urosaurus. Plate8

PLATE 9

Mittleman — The Iguanid Genus Urosaurus

PLATE 9

(Top) Urosaurus omatus wrighti Schmidt, CAS 3759, Moab, Grand
County, Utah.

(Bottom) Urosaurus omatus caervleus Smith, type, David H. Dunkle —
Hobart M. Smith Coll. No. 132, 30 miles north of Chihuahua City, Chihuahua,
Mexico.

BULL. MUS. COMP. ZOOL.

Mittleman: Iguanid Genus Urosaurus. Plate 9

TO

-— ■ v_- - ' - -r

.

*...*.

*S r»; 5 \ *^ sK '**

<*#;

SS£?fltt!?

u^?^*^*

" ": Hrl't-v^-®

-v1-

vzim*?*

3f^

PLATE 10

Mittleman — The Iguanid Genus Urosaurus

PLATE 10

Urosaurus nigricaudus Cope, cotype, USNM 5307, Cape San Lucas, Baja
California, Mexico.

BULL. MUS. COMP. ZOOL.

Mittleman: IguanidGenus Urosaurus. PlateIO

PLATE 11

Mittleman — The Iguanid Genus Urosaurus

PLATE 11
Urosaurus gadovi Schmidt. USNM 113421, Apatzingan, Michoacan.

BULL. MUS. COMP. ZOOL.

Mittleman: Iguanid Genus Urosaurus. Plate 11

v

PLATE 12

Mittleman — The Iguanid Genus Urosaurus

PLATE 12

(Top) Urosaurus irregularis Fischer, type, Municipal Natural History
Collection of Bremen No. 437, "... aus dem Hochlande von Mexico." (from
Fischer, 1882, pi. 17)

(Bottom) Urosaurus bicarinatus tuberculatus Schmidt, Edward H. Taylor
Coll. No. 552, Presidio de Mazatlan, Sinaloa, Mexico.

BULL. MUS. COMP. ZOOL.

Mittleman: IguanidGenus Urosaurus. Plate12

-^ij&g

*>vr>-~

V>

-•■A

<-

*

PLATE 13

Mittleman — The Iguanid Genus Urosaurus

PLATE 13

Urosaurus bicarinatus bicarinatus Dumeril, USNM 10242, Tupataro,
Michoacan, Mexico.

BULL. MUS. COMP. ZOOL.

Mittleman: IguanidGenus Urosaurus. Plate 13

-

-■'

'

PLATE 14

Mittleman — The Iguanid Genua Urosaurus

PLATE 14

Urosaurus bicarinatus anonymorphus Mittleman, type, USNM 46988,
Tehuantepec, Oaxaca, Mexico.

BULL. MUS. COMP. ZOOL.

Mittleman: Iguanid Genus Urosaurus. Plate 14

PLATE 15

Mittleman — The Iguanid Genus Urosaurus

PLATE 15

Urosaurus bicarinatus nelsoni Schmidt, type, USNM 46836, Cuicatlan,
Oaxaca, Mexico.

BULL. MUS. COMP. ZOOL

Mittleman: IguanidGenus Urosaurus. Plate 15

PLATE 16

Mittleman — The Iguanid Genus Urosaurus

PLATE 16

Urosaurus auriculatus Cope, type, TJSNM 7027, Socorro Island, Revillagi-
gedo Archipelago, Mexico.

BULL. MUS. COMP. ZOOL

Mittleman: IguanidGenus Urosmjrus. Plate 16

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE

Vol. XCI, No. 3

SCIENTIFIC RESULTS OF A FOURTH EXPEDITION
TO FORESTED AREAS IN EASTERN AFRICA

III
DECAPOD CRUSTACEA

By Fenner A. Chace, Jr.

CAMBRIDGE, MASS., TJ. S. A.

PRINTED FOR THE MUSEUM

October, 1942

PUBLICATIONS
OF THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

The Bulletin and Memoirs are devoted to the publication of
investigations by the Staff of the Museum or of reports by spec-
ialists upon the Museum collections or explorations.

Of the Bulletin, Vols. I to XC, and Vol. XCI, No. 1 and
No. 2 have appeared and of the Memoirs, Vol.1 to LVI.

These publications are issued in numbers at irregular intervals.
Each number of the Bulletin and of the Memoirs is sold separately.
A price list of the publications of the Museum will be sent upon
application to the Director of the Museum of Comparative
Zoology, Cambridge, Massachusetts.

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE

Vol. XCI, No. 3

SCIENTIFIC RESULTS OF A FOURTH EXPEDITION
TO FORESTED AREAS IN EASTERN AFRICA

III
DECAPOD CRUSTACEA

By Fexner A. Chace, Jr.

CAMBRIDGE, MASS., U. S. A.

PRINTED FOR THE MUSEUM

October, 1942

No. 3. — Scientific Results of a Fourth Expedition to Forested Areas

in Eastern Africa

III

Decapod Crustacea

By Fexxer A. Chace, Jr.

The decapod Crustacea collected by Mr. A. Loveridge consist of
389 specimens belonging to 20 species, nine of which are coastal and
marine species and the remainder fresh-water crabs. Two of the latter,
referred to the subgenus Geothelphusa of the genus Potamon, are here
described as new. As the number of known species of Potamonidae
from Africa increase, their determination becomes more and more
difficult, and so it was decided to draw up a list of the species which
have been recorded from Africa in the hope that such a summary will
save future workers a little of the time spent in locating species scat-
tered through the literature. Although Parisi (1923, p. 334) has pub-
lished a list of the species of Potamonidae of the world described since
Miss Rathbun's monograph (1904-1906), it can do no harm to bring
this list up to date for the African forms. The structure of the first
abdominal appendage of the male is becoming increasingly important
among the Brachyura as a diagnostic character, so the liberty has
been taken here to figure this appendage from a few of the type speci-
mens of African potamonids in the Museum of Comparative Zoology;
the determination of the species of this group, as in most of the families
of fresh-water Crustacea, is so difficult that it is only by frequent ref-
erence to the types that synonymies can be satisfactorily straightened
out.

PENAEIDAE

Penaeus indicus H. Milne Edwards

Penaeus indicus H. Milne Edwards, 1837, Hist. Nat. Crust., 2, p. 415.
Peneus indicus Alcock, 1906, Catal. Ind. Dec. Crust., pt. 3, fasc. 1, p. 12, pi. 1,

figs. 3, 3a.
Peneus indicus Schmitt, 1926, Biol. Res. Fish. Exp. F.I.S. "Endeavour", 5,

pt. 6, p. 359.

1 cf 6 9 (M. C. Z. 11220) Kilindini, Mombasa Id., Kenya Colony.
25. vii. 39.

In all of these adult specimens in which the rostrum is intact it
extends beyond the antennal scales, but in at least one case it reaches
only slightly beyond.

186 bulletin: museum of comparative zoology

PAGURIDAE

Clibanarius longitarsus (de Haan)

Pagurus longitarsus de Haan, 1849, Faun. Japon. Crust., p. 211, pi. 50, fig. 3.
Clibanarius longitarsis Alcock, 1905, Catal. Indian Dec. Crust., pt. 2, fasc. 1,

p. 158.
Clibanarius longitarsus Buitendijk, 1937, Temminckia, 2, pp. 253 and 266.

lc?l 9 (M.C.Z. 11221) Mikindani, Tanganyika Territory, iv. 39.

POTAMONIDAE

POTAMON (POTAMONAUTES) HILGENDORFI (Pfeffer)

Telphusa Hilgendorfi Pfeffer, 1889, p. 32.

Telphusa suprasulcata Hilgendorf, 1898, p. 8, pi., figs. 5-5d.

Potamon (Potamonautes) Hilgendorfi Rathbun, 1905, p. 171.

Potamon (Potamonautes) suprasulcatus Rathbun, 1905, p. 172.

Potamon [Potamonautes) suprasulcatum Colosi, 1924, p. 4.

Potamonautes hilgendorfi Balss, 1929b, p. 344.

Potamon (Potamonaxdes) hilgendorfi Rathbun, 1933, p. 256.

Potamon (Potamonautes) hilgendorfi Rathbun, 1935, p. 26.

3 immature 9 (M. C. Z. 11222) Mt. Magrotto, near Muhesa, Tan-
ganyika Territory, vii. 39.

Type localities. Stream near "Nekonda" and "Hanaha" stream
near "Mangaala", both in Ungiiu, Tanganyika Territory.

These three specimens, all of which are young females as indicated
by the narrow abdomen, have the carapace ranging in breadth from
34 mm. to 47.6 mm. Comparison of these specimens with those col-
lected by Mr. Loveridge at Amani in the Usambara Mts. and identi-
fied as P. hilgendorfi by Miss Rathbun, after examination of speci-
mens of that species from the Hamburg Museum, discloses no differ-
ences of importance; in xVmani specimens of similar size there is often
a broad tooth at the end of the post-frontal crest. It is remarkable
how similar in appearance young specimens of P. hilgendorfi are to
P. dybowskii from the Belgian Congo; in fact, if the present speci-
mens had come from the Congo, they could easily have been identified
as the latter species. Compared with an adult female of P. dybowskii,
the carapace differs only in the rougher post-frontal crest and ante-
rolateral margins and the coarser transverse striae on the lateral por-
tions of the carapace as well as the more granulated side walls; how-

chace: decapod Crustacea 187

ever, the carpal spine in P. dybowskii is followed by a series of denticles
rather than a distinct spine as in P. hilgendorfi. It is very likely, as
Balss (1929b) has pointed out, that P. mrogoroensis is synonymous
with this species but the latter species has been retained in the list
below in the hope that more conclusive proof of its identity with
P. hilgendorfi will be forthcoming.

POTAMON (POTAMONAUTES) DYBOWSKII Rathbun

Potamon (Potamonautes) Dybowskii Rathbun, 1904, pi. 15, fig. 3; 1905, p. 177.
Potamon (Potamonautes) ambiguus Lenz, 1910b, p. 121. (Not P. ambiguus

Rathbun, 1904. See Balss, 1929b, p. 345.
Potamonautes Dybowskii Halss, 1914a, p. 103.

Potamon (Potamonautes) dybowskii Rathbun, 1921, p. 410, text-fig. 7, pi. 24.
Potamon (Potamonautes) dybowskii Parisi, 1925, p. 99.
Potamonautes dybowskii Balss, 1929b, p. 345.
Potamon (Potamonautes) dybowskii Balss, 1936, p. 187, text-fig. 23 (map).

1 9 with young (M. C. Z. 11223) Goma, north end of Lake Kivu,
Belgian Congo. 13-14. ii. 39.

Type locality. Bangui, French Equatorial Africa.

This specimen has a carapace breadth of 58.8 mm. The carapace
appears rather worn so that all of the usually sharp angles have be-
come blunt; the post-frontal crest consequently appears lower than
is generally the case in this species and the teeth at the ends of that
crest are indicated only by a concavity just inside the lateral margins.
The specimen has been compared with a smaller female collected by
Dr. J. C. Bequaert at the village of Malela (Chief of Kasende),5° 40' S.,
23° 45' E., Belgian Congo, and identified by Miss Rathbun. Despite
, the worn appearance of the Kivu specimen, there is little doubt that
both individuals belong to the same species.

I find it impossible to agree with Dr. Balss (1936) that P. stan-
leyensis is a synonym of this species. There are at my disposal four
paratypes of the latter species from Stanleyville, Belgian Congo.
Aside from its smaller size (mature specimens compared), P. stanley-
ensis differs from P. dybowskii in having the carapace more convex
antero-posteriorly, a much less prominent furrow dividing the branch-
ial region, a distinct though smaller spine rather than a row of denti-
cles following the carpal spine, the propodite of the last ambulatory
leg more slender and very differently formed first abdominal append-
ages of the male as shown by Miss Rathbun (1921, text-figs. 7d and 9h).

188

bulletin: museum of comparative zoology

Potamox (Potamoxautes) lirraxgexsis Rathbun
Text-figure 1

Potamon {Potamonautes) lirrangensis Rathbun, 1904, pi. 14, fig. 8; 1905, p. 169

Potamondutes lirrangensis Balss, 1914b, p. 404.

Potamon {Potamonautes) lirrangensis Rathbun, 1921, p. 413, text-fig. 8, pis. 25

and 26, fig. 3.
Potamonautes lirrangensis Balss, 1929b, p. 347.
Potamon {Potamonautes) lirrangensis Balss, 1936, p. 188, text-fig. 24 (map).

1 d" (M. C. Z. 11224) Idjwi Island, Lake Kivu, Belgian Congo, ii. 39.

Type locality. Lirranga, at the junction of the Congo and Ubangi.
Rivers, Belgian Congo.

This single specimen agrees more closely with the figure of the type
from Lirranga than with specimens identified by Miss Rathbun from
Stanlevville with which it was directly compared. The front in the

A

B

Fig. 1. Potamon {Potamonautes) lirrangensis; A, dorsal view of carapace
of male from Idjwi Island, x 0.9; B, posterior view of end of first right ab-
dominal appendage of the same specimen, x 9.

Idjwi specimen, and apparently also in the type, is nearly transverse
anteriorly whereas in the Stanleyville specimens the front is markedly
bilobate in both dorsal and frontal view. Rathbun (1921), in discussing
the Stanleyville specimens, fails to mention this character except to say,
"margin of frontal lobes rather regularly arched", but it is obvious
from her figures that these specimens have the frontal margin deeply
concave in the median portion. In all other particulars, including
the form of the first abdominal appendage of the male, the present
specimen agrees with the Stanleyville specimens ; apparently the latter

chace: decapod Crustacea

189

represent but a local variety of the species. Balss (1929b and 1936)
has recorded P. lirrangensis from Lake Kivu, but he fails to mention
the form of the front in either paper.

Potamox (Potamoxautes) usambarae Rathbun

Text-figure 2

Potamon (Potamonautes) usambarae Rathbun, 1933, p. 257, pi. 6.

4 d" 2 9 (M. C. Z. 11225) Mt. Magrotto, near Muhesa, Tanganyika
Territory, vii. 39.

Type localities. Amani and Kizerui, both in the Usambara Moun-
tains, Tanganyika Territory.

These specimens agree very well with the male described and
figured by Miss Rathbun from Amani. The male and ovigerous
female from Kizerui, although probably the same species, have the

A B C

Fig. 2. Potamon ((Potamonautes) usambarae; A, posterior view of first
right abdominal appendage of male cotype (carapace breadth 22.6 mm.) from
Amani, x7.3; B, posterior view of tip of same appendage, x 36.5; C, abdo-
men of same specimen, x 2.9.

carapace much less convex antero-posteriorly ; this is particularly true
of the female. This is probably not the species called T. hilgendorfi
by Hilgendorf (1898) to which Miss Rathbun (1904) has already
given the name P. ambiguns and which Balss (1929b) refers to P. john-
sto7ii. P. usambarae is a much smaller species that P. johnstoni and
an examination of the first abdominal appendages of the male dis-
closes that, in this respect at least, it is not even closely related to
the latter species (cf. Balss, 1936, text-fig. 17).

190 bulletin: museum of comparative zoology

Potamon (Potamonautes) obesus (A. Milne Edwards)

Thelphusa obesa A. Milne Edwards, 1868, p. 86, pi. 20, figs. 1-4.

Potamon (Potamonautes) obesus Rathbun, 1904, pi. 15, figs. 8 and 9; 1905,

p. 180, text-fig. 45.
Potamon (Potamonautes) obesus Sendler, 1912, p. 199.
Potamon (Potamonautes) obesus Bouvier, 1921, p. 49.
Potamonautes obesus Balss, 1929b, p. 348.
Potamonautes obesus Barnard, 1935, p. 484.

1 a* 1 9 (M.C.Z. 11226) Siga Caves, near Tanga, Tanganyika Ter-
ritory, vi. 39.

Type locality. Zanzibar.

Potamon (Potamonautes) aloysii-sabaudiae Nobili
Text-figure 3

Potamon (Potamonautes) Aloysii Sabaudiae Nobili, 1906b, p. 1. (Corrected to

Aloysii-Sabaudiae under "errata").
Potamon (Potamonautes) Aloysii Sabaudiae Nobili. 1909, p. 357.
Potamon (Potamonautes) johnstoni Caiman, 1909, p. 51, text-figs. 9, 10 and 12

(not text-fig. 11).
Potamon (Potamonautes) Johnstoni Colosi, 1920, p. 32.
Potamon (Potamonautes) Johnstoni Colosi, 1924, p. 21, text-fig. 15.

3 d" 9 9 (1 ovigerous) (M. C. Z. 11227) Kibale Forest, Uganda,
xii. 38.

14 cf 28 9 (M. C. Z. 11228) Bundibugyo, Bwamba Forest, north-
western part of Mt. Ruwenzori, Uganda. 20. xii. 38.

16 cf 17 9 (2 ovigerous) 40 young (M. C. Z. 11229) Mihunga, Mt.
Ruwenzori, Uganda, i. 39.

Type localities. Colosi (1920) gives Ibanda and "Bijunga (3505 m.)",
Mt. Ruwenzori, as the localities of Xobili's types.

In spite of Caiman's conviction that this species is identical with
P. johnstoni from Mt. Kilimanjaro, it seems to me best at present
to recognize both species. When more material is available from
other localities, intermediate forms may be discovered which will
prove without doubt that P. aloysii-sabaudiae is but a local race of
P. johnstoni; in the meantime little is lost by considering the species
as distinct. Caiman has shown that in P. johnstoni the post-frontal
crest is sharp even in males larger than any known Ruwenzori speci-
mens, whereas in P. aloysii-sabaudiae it is broadly rounded in adult
males. In females the crest is somewhat sharper than in males, par-
ticularly toward the lateral margins, but it becomes sharp for its

chace: decapod Crustacea

191

entire length, as in Kilimanjaro specimens, only in very immature
individuals; this tendency for the post-frontal crest to become much
sharper in young specimens is noticed in practically all species of the
genus, even the most extreme forms of the subgenus Geothelphusa in
which the crest is nearly absent in adults. Of even greater importance
is the difference between the first abdominal appendages of the adult
male. This difference can best be realized by comparing the accom-
panying figures with similar figures of a Kilimanjaro specimen of P.
johnstoni in Balss (1936, text-fig. 17). In P. aloysiisabaudiae the tip

B «-

Fig. 3. Potamon (Potamonantes) aloysiisabaudiae; A, larger chela of male
(carapace breadth 50 mm.) from Bundibugyo, x 1.1; B. posterior view of
end of first right abdominal appendage of a male (carapace breadth 50.8 mm.)
from Mihunga, x 15; C, antero-median view of end of same appendage, x 15.

curves outward much more strongly than in P. johnstoni and, as far as
curvature is concerned, it is intermediate between P. johnstoni and P.
orbitospiniis as figured by Balss (1936, text-fig. 18). In addition, the
tip of the appendage in dorsal, or antero-median, view differs consider-
ably from the same surface of P. johnstoni. The specimens from the
three localities listed above agree almost perfectly with the figure
of the Ruwenzori specimen in Caiman (text-fig. 9), but it is very
difficult to believe that they also agree with the cotype of P. johnstoni
(text-fig. 11). The present specimens, with due regard for characters
which obviously change with increase in size, are similar to one an-
other, except that the three largest males from Bundibugyo have
the fingers of the larger chela widely gaping as in the accompanying
figure; the others agree with Caiman's figure of the chela (text-fig. 9).

192 bulletin: museum of comparative zoology

Potamon (Geothelphusa) berardi (Audouin)?

Potamons or crabes fluviatiles Savigny, 1817, pi. 2, fig. 6.

Thelphusa Berardi Audouin, 1826, p. 82.

Potamon (Geothelphusa) Berardi Rathbun, 1904, pi. 18, figs. 3 and 10; 1905,

p. 203.
Potamon (Geotelphusa) Berardi Lenz, 1910b, p. 124. (According to Balss,

1929b, these specimens are P. emini).
Potamon (Geothelphusa) Berardi de Man, 1914, p. 126, pi. 2, figs. 3-3a.
Potamon (Geothelphusa) Berardi Colosi, 1919, p. 50.
Potamon (Geotelphusa) Berardi Colosi, 1920, p. 34.
Geotelphusa berardi Balss, 1929b, p. 350.
Potamon berardi Flower, 1931, p. 732.
Potamon (Geothelphusa) berardi Rathbun, 1935, p. 25.

1 9 (M.C.Z. 11230) Idjwi Island, Lake Evu, Belgian Congo, ii. 39.

Type locality. Egypt.

This specimen is distinguished from the two other species of the
subgenus Geothelphusa taken at Idjwi Island, P. (G.) idjiciensis and
P. (G.) emini, by the much less vaulted carapace. As the carapace is
only 17.8 mm. broad and the individual is obviously nearly or quite
mature, as indicated by the broadened abdomen, it apparently be-
longs to one of the smaller species of the genus; wThether or not it
should be assigned to P. (G.) berardi is open to argument, however.
Although the form and areolation of the carapace is very like that of
P. (G.) berardi, there are a few points of difference which should be
mentioned. The carapace is like that figured by de Man (1914)
from an Egyptian specimen, even to the dimple on either protogastric
region, but the front is more strongly deflexed in the Idjwi specimen so
that the margin is hidden from dorsal view. A comparison with
Egyptian and Mount Elgon specimens available in the Museum of
Comparative Zoology discloses the following differences in addition
to the more deflexed front; the Idjwi specimen has the surface of the
carapace much less grossly punctate and consequently it appears
more glossy; the post-frontal crest is slightly more removed from the
dorsal margin of the orbit than in the typical P. (G.) berardi; in
frontal view7, the orbits are more strongly convex below the outer
angle and so appear to extend a little beyond this angle in the present
specimen; finally, the ischium of the third maxillipeds of this speci-
men has a stronger longitudinal groove than in any of the specimens
of P. (G.) berardi examined. In all other particulars the Idjwi speci-

chace: decapod Crustacea 193

men agrees with that species. Apparently the most closely related
Congo species is P. ((?.) congocnsis, but our specimen belongs to a
somewhat smaller species and the post-frontal crest curves slightly
forward at the outer ends rather than running obliquely backwards
in nearly a straight line. It is likely that this female will prove to
belong to a varietal form of P. ((?.) berardi, but relationships may
be so masked in female specimens that more material, including
males, must be collected before its true position can be ascertained.

Potamox (Geothelphusa) emixi (Hilgendorf)

Telphusa emini Hilgendorf, 1892, p. 11.

Potamon {Geothelphusa) Emini Rathbun, 1904, pi. 18, fig. 9; 1905, p. 209.

Potamon (Geothelphusa) emini Rathbun, 1909, p. 102.

Potamon (Geotelphusa) Emini Lenz, 1910b, p. 125.

Potamon {Geothelphusa) Emini Bouvier, 1921, p. 50, text -fig. 4.

Potamon (Geothelphusa) Emini Rathbun, 1922, p. 35.

Potamonautes emini Balss, 1929b, p. 345.

Potamon (Geothelphusa) emini (var. Bouvier) Rathbun, 1933, p. 258, pi. 4,

figs. 1 and 2, pi. 5.
Potamon (Geothelphusa) emini Balss, 1936, p. 193, text -fig. 28 (map).

ltfl? (M.C.Z. 11231) Idjwi Island, Lake Kivu, Belgian Congo,
ii. 39.

Type locality. Near Bukoba, northwestern Tanganyika Territory.

These specimens have been compared with those collected by Mr.
Loveridge in the Uzungwe Mountains, Tanganyika Territory, and
identified by Miss Rathbun (1933) as a variety of this species. The
present specimens are very similar to that lot, differing only as fol-
lows: the outer orbital angles are more obtuse, less rectangular; the
notch between the outer orbital angle and the end of the post-frontal
crest is less deep, not V-shaped; the sharp portion at the outer end
of the post-frontal crest runs back obliquely in an almost straight
line rather than being concave forward; the diagonal grooves on the
male sternum are similar but slightly stronger; and the fifth segment
of the male abdomen has the sides slightly concave rather than
slightly convex. The male has the carapace 17.4 mm. broad, the
female 17.3 mm. broad.

194

BULLETIN*: MUSEUM OF COMPARATIVE ZOOLOGY

Potamon (Geothelphusa) mutandensis, spec, now
Text-figures 4 and 5

Type & (M. C. Z. 11232) Mushongero, Lake Mutanda, southwestern

Uganda. 1. ii. 39.
14 d" 3 9 (1 ovigerous) (M. C. Z. 11232), same data.

Carapace relatively narrow, nearly three-fourths as long as broad,
moderately convex in both directions and with the regions fairly
well delimited. Surface with a few scattered punctae but more or
less smooth and polished to the naked eye. Post-frontal crests dis-
cernible, but rounded off so that they are more or less indistinct in the

Fig. 4. Potamon (Geothelphusa) mutandensis; male holotype, x 1.3.

adult except for the protogastric prominences; the crests are more
prominent with a tendency to become sharp-edged in females and
young. Anterior mesogastric region roof-shaped and marked off by a
very faint line which may extend nearly to the frontal margin, but is
usually almost invisible. A rather deep, H-shaped depression delimits
the posterior gastric and cardiac regions from the branchial region, and
there is a broad depression separating the branchial region into two
parts; this last depression usually, but not always, curves down onto
the side wall sufficiently to make the postero-lateral margin slightly
concave. Front typically strongly bilobed in dorsal view and turned

chace: decapod Crustacea

195

D

E

Fig. 5. Potamon (Geothelphusa) mutandensis; A, right chela of holotype,
x 1.5; B, left chela of holotype, x 1.5; C, outer maxilliped of holotype,
x 3.8; D, abdomen of holotype, x 3; E, anterior portion of sternum of holo-
tj'pe, x 3; F, posterior view of first right abdominal appendage of male
paratvpe, x 7.5; G. posterior view of tip of same appendage, x 30.

196 bulletin: museum of comparative zoology

down so that the frontal margin is just visible; in a few of the speci-
mens, however, the frontal outline is nearly straight, and the down-
ward curvature is subject to some variation. Viewed from in front,
the frontal margin is either slightly convex or strongly bilobed.
Upper margin of orbit slightly sinuous and nearly transverse, but
tending to run obliquely outward and forward; outer angle obtuse
and not prominent; no sinus below outer angle. Antero-lateral line
of carapace slightly raised only near the orbital angle in adults, but
more prominent for its entire length in the very young; it is slightly
constricted at the juncture with the post-frontal crest. In the adult
the carapace extends but little, if at all, laterally beyond the antero-
lateral line at the widest point of the carapace.

Mandibular palp two-jointed, with the terminal joint simple.
There is a faint oblique groove on the ischium of the third maxillipeds ;
outer margin of merus broadly rounded; exognath with a well-devel-
oped, plumose palp.

Chelipeds unequal, usually strikingly so in adult males in which
the larger chela has the fingers widely gaping and crossing each other
to such a degree in certain cases that they have a deformed appear-
ance ; the smaller chela is slender with very long fingers ; in the females,
and occasionally in males, the chelae are sub-equal. Both chelipeds
have the merus ornamented on the anterior border with a row of
tubercles, the most distal of which is enlarged and placed on a slightly
lower level than the rest; on the proximal portion of the dorsal sur-
face there may be a few scattered tubercles. Carpus with a short,
stout inner spine followed by two or three tubercles.

Anterior of the sternal grooves complete and well marked; poste-
rior groove distinct in its lateral portions, interrupted centrally.
The margin of the sternum at the point of insertion of the chelipeds
is slightly, but not abnormally, raised.

Sixth somite of male abdomen fully twice as broad at base as long
and with very slightly sinuous margins. Terminal somite distinctly
longer that the preceding, its base extending laterally slightly be-
yond the distal margins of the sixth and its margins very slightly
concave. Extremities of first abdominal appendages of male turned
strongly outward.

Measurements. Male holotype, length of carapace 19.0 mm.,
breadth 25.6 mm.; ovigerous female, length of carapace 15.0 mm.,
breadth 20.9 mm.

There is so much variation in this species as regards the areolation
of the carapace, the form of the front and the degree of dissimilarity

chace: decapod Crustacea 197

between the chelae that it is not only difficult to determine the rela-
tionship between it and other species, but it is even possible that the
species has been previously described from one of its forms. The
species to which it might be most closely related include P. neumanni,
P. laetabilis, P. amalerensis, P. berardi, P. emini, P. congoensis, P. per-
parvus, P. granriki and P. odhncri. It is a smaller species than P. neu-
manni and the sixth abdominal somite in the male is not more than
half as long as broad at the base as in that species. This latter charac-
ter, the length of the penultimate abdominal somite, serves to distin-
guish P. mutandensis from P. laetabilis, P. amalerensis and P. congoen-
sis. The narrower carapace, blunter post-frontal crest and different
form of the major chela of the male further distinguish this species
from P. amalerensis. The carapace is less noticeably punctate and
more strongly areolated and the form of the smaller chela and the first
abdominal appendages of the male are different from those of P. be-
rardi. It is a somewhat larger species than P. emini and the carapace
is less swollen and more sharply areolated. The post-frontal crest is
blunter than in P. congoensis and the chelae are different. The
transverse branchial furrow is more pronounced than in P. perparvus
and the post-frontal crest is less distinct laterally. The first abdomi-
nal appendages of the male are different from those of P. granviki
and, finally, the post-frontal crest is less distinct than that of
P. odhneri.

Potamon (Geothelphusa) idjwiexsis, spec, now
Text-figures 6 and 7

Type d" (M. C. Z. 11234) Idjwi Island, Lake Kivu, Belgian Congo.

ii. 39.
89 c? 106 9 (11 ovigerous, 1 with young) (M. C. Z. 11235), same data.

Carapace very convex antero-posteriorly and very broad, being
about two-thirds as long as broad in the adult. Branchial regions
swollen. Surface grossly punctate. Post-frontal crests nearly obso-
lete, being represented by faint swellings denoting the epigastric
lobes and obscure depressions behind the orbits which delimit very
faint convexities in back of them. Depression behind gastric region
well marked and cardiac region outlined. No cervical groove except
for an almost invisible row of elongate pits. There are four irregular
pits on either side of the widest portion of the mesogastric region
and an indistinct pit behind each orbit in line with the extremity
of the cornea. Median frontal groove well marked; anterior meso-

198

bulletin: museum of comparative zoology

gastric region very narrow and obscurely delimited. Edge of front
hidden from dorsal view, sinuous with a slight notch in the
midline. Upper margin of orbit nearly transverse in dorsal view;
outer angle obtuse and not prominent ; there is no outer sinus ; orbital
margin raised and nearly smooth for its entire extent. Anterolateral
line of carapace raised and crenulate in the young, scarcely distin-
guishable in the adult; it forms an obtuse angle at the end of the in-
distinct post-frontal crest. Postero-lateral border short and very
faintly concave. The side wall of the carapace is so swollen that it
extends laterally far beyond the antero-lateral line in the adult.

Fig. 6. Pot anion (Geothelphusa) idjwiensis; male holotype, x 1.

Mandibular palp two-jointed with the terminal joint simple. There
is a groove on the ischium of the outer maxilliped, nearer the inner
than the outer margin; outer margin of merus angulate and forming
a similar obtuse angle with the anterior margin; exognath with a
long, plumose palp.

Chelipeds unequal; anterior margin of merus decorated with a row
of small tubercles, one of which near the anterior end is more promi-
nent ; carpus with a well developed spine followed by a low denticulate
ridge. Larger chela broadly gaping in the larger males.

Transverse groove on sternum between the bases of the outer
maxillipeds very deep except near the margin where it becomes

chace: decapod Crustacea

199

-

D

E

ill

fit

i km

H

Fig. 7. Potamon (Geothelphusa) idjwiensis; A, fronto-orbital region of holo-
type, x 1.4; B, outer maxilliped of holotype, x3.4; C. larger chela of holo-
type, x 1.4; D, abdomen of holotype, x 1.4; E, anterior portion of sternum
of holotype, x 1.4; F, posterior view of first right abdominal appendage of
male paratype, x 3.4; G, posterior view of tip of same appendage, x 17;
H, carapace of young paratype (13.8 mm. broad), x 1.4; I, carapace of
male paratype (21.0 mm. broad), x 1.4.

200 bulletin: museum of comparative zoology

shallower. Behind the transverse groove are a pair of similar furrows
running from in front of the bases of the ehelipeds obliquely back to
the ridge bounding the terminal segment of the abdomen; there is a
tubercle near the anterior end of each of these oblique grooves. The
margin of the sternum at the insertion of the ehelipeds is raised to
form a broad ridge.

Sixth somite of male abdomen with slightly concave lateral margins ;
sides of terminal segment more strongly concave. Extremities of first
abdominal appendages of male more or less parallel to the midline,
curving inwards slightly at the tips.

The following are the most noticeable changes which accompany
growth in this species: the carapace becomes broader and extends
proportionately farther beyond the antero-lateral line due to the
greater inflation of the sub-branchial regions; the antero-lateral line
becomes less and less distinct; the carapace becomes slightly more
convex antero-posteriorly, but the edge of the front is concealed from
dorsal view even in the young; and the eyes decrease in size. Young
specimens may be distinguished from the two other Idjwi Island
species of the subgenus Geothelphusa by the fact that the front is
evenly convex in dorsal view — not bilobed.

Measurements. Male holotype, length of carapace 25.8 mm.,
breadth 37.5 mm.; ovigerous female, length of carapace 22.8 mm.,
breadth 32.0 mm. Old males in which the fingers are widely gaping
have the carapace from 34.1 to 38.7 mm. broad. In females, the
abdomen becomes broadened for the reception of the eggs when the
carapace is about 25 mm. broad and ovigerous specimens have the
carapace from 27.7 to 36.2 mm. broad; the specimen carrying young
beneath the abdomen has a carapace breadth of 30.8 mm.

This species agrees very closely with the original description of
P. neumanni, but the figures of the cotype of that species given by
de Man (1914) leaves little doubt that it is not the same species.
P. idjwiensis has the carapace broader, protruding farther beyond
the antero-lateral line, the front curved downward more strongly so
that the edge is hidden from dorsal view, the sub-hepatic and sub-
branchial regions are less areolated, the terminal segment of the male
abdomen is more concave along the lateral margins and the larger
chela of the male is higher in proportion to its length.

chace: decapod Crustacea 201

Deckenia mitis Hilgendorf

Deckenia mitis Hilgendorf, 1898, p. 24, fig. 8.

Deckenia mitis Rathbun, 1905, pi. 21, fig. 7; 1906, p. 71, text-fig. 123.

Deckenia imitatrix, var. mitis Bouvier, 1921, p. 57.

Deckenia mitis Rathbun, 1921, p. 434, text-fig. 16, pi. 34.

Deckenia mitis Colosi, 1924, p. 19.

Deckenia mitis Balss, 1929b, p. 353.

Deckenia mitis Rathbun, 1933, p. 259.

1 d* 1 9 (M. C. Z. 11236) Siga Caves, near Tanga, Tanganyika Ter-
ritory, vi. 39.

Type localities. Wembere Steppe (north of Tabora) and Dar-es
Salaam, Tanganyika Territory, and Mombasa, Kenya Colony.

GRAPSIDAE

Grapsus strigosus (Herbst)

Cancer strigosus Herbst, 1799, Krabben u. Krebse, Bd. 3, Hft. 1, p. 55, pi. 47,

%. 7.
Grapsus strigosus Tesch, 1918, Siboga-Exped., Monogr. 39°, p. 71, pi. 4, figs.

1 and 4.

2 & (M. C. Z. 11237) Mikindani, Tanganyika Territory, iii. 39.

Sesarma (Sesarma) meixerti de Man

Sesarma meinerti de Man, 1887, Zool. Jahrb., Bd. 2, pp. 648 and 668.
Sesarma (Sesarma) meinerti Tesch, 1917, Zool. Med. Mus. Leiden, Deel 3, pp.

171 and 246.
Sesarma meinerti Cott, 1930, Proc. Zool. Soc, London, for 1929, pp. 679-692,

4 text-figs., 1 col. pi.
Sesarma (Sesarma) meinerti Miyake, 1938, Annot. Zool. japon., p. 108.

3 cf (M. C. Z. 11238) Mbanja, near Lindi, Tanganyika Territory.

iv. 39.

A chela taken from the stomach of a crocodile (Crocodylus niloticus)
captured at Mbanja has been identified as belonging to this species.
Mr. Loveridge informs me that the stomach of this crocodile was
filled with the remains of this crab.

202 bulletin: museum of comparative zoology

OCYPODIDAE

OCYPODE CERATOPHTHALMA (Pallas)

Cancer ceratophthalmus Pallas, 1772, Specilegia Zool., 9, p. 83, pi. 5, figs. 7-8.
Ocypoda ceratophthahna Cott, 1930, Proc. Zool. Soc, London, for 1929, pp.
755-765, 1 text-fig., 1 pi.

2 cf 2 9 (M. C. Z. 11240) Lindi, Tanganyika Territory. 31. v. 39.

All of these specimens are small and the prolongation of the eye-
stalk is as yet little more than an acute tubercle.

Ocypode kuhlii de Haan

Ocypode (Ocypode) kuhlii de Haan, 1835, Fauna Japon. Crust., p. 58.
Ocypode kuhlii Rathbun, 1935, Bull. Mus. Comp. Zool., 79, no. 2, p. 26.

3 & 2 9 (M. C. Z. 11241) Lindi, Tanganyika Territory. 31. v. — 4.

vi. 39.

Uca annulipes (H. Milne Edwards)

Gelasimus annulipes H. Milne Edwards, 1837. Hist. Nat. Crust., 2, p. 55, pi. 18,

figs. 10-13.
Gelasimus annulipes Alcock, 1900, Journ. Asiat. Soc. Bengal, 69, pt. 2, no. 3,

p. 353.

2 d1 (M. C. Z. 11242) Lindi, Tanganyika Territory. 31. v. 39.

Uca inversa (Hoffmann)

Gelasimus inversus Hoffmann, 1874, Crust. Echinod. Madagascar, p. 19, pi. 4,

figs. 23-26.
Uca inversa Rathbun, 1935, Bull. Mus. Comp. Zool., 79, no. 2. p. 27.

1 tf (M. C. Z. 11243) Lindi, Tanganyika Territory. 31. v. 39.

Uca marionis (Desmarest)

Gelasimus Marionis Desmarest, 1825, Consid. Gen. Crust., p. 124, pi. 13, fig. 1#
Uca marionis McNeill, 1920, Rec. Austr. Mus., 13, p. 105, pi. 19, text-figs. 1-5.

3 d" (M. C. Z. 11244) Lindi, Tanganyika Territory. 31. v. 39.
All three of these specimens belong to the variety nilida of Dana.

chace: decapod Crustacea 203

List of African Fresh-water Crabs (Potamonidae)
Known up to January 1, 1938

In the following list the subgenus has been omitted except in the
synonymies. Although the definitions of the subgenera as generally
recognized may be open to question, there is little doubt that sub-
genera of some sort are useful in clarifying the relationships of the
Potamonidae; but, since several species have been referred to more
than one subgenus by different authors and since this list makes no
attempts to solve such problems except when material of a particular
species is at hand, it is hoped that less confusion will result from re-
taining the subgenera only in the synonymy than by arbitrarily
choosing one of the two or three that may have been employed by
previous authors. For much the same reason, all forms are treated as
full species even though they were described or subsequently referred
to as subspecies or varieties. The state of our knowledge of specific
relationships is so incomplete at present that a form now considered
a subspecies may eventually be given full specific rank, while the
species to which it was thought to be closely related may become a
subspecies or variety of some entirely different form.

The original plan for this list called for the inclusion of the known
range of each species, but it soon became apparent that the task of
checking all of the localities was hardly worth-while when one con-
sidered the many inaccuracies unavoidably introduced into the results
through the numerous misidentifications in the literature. In order to
furnish some idea of the region in which each species is found and to
aid future collectors in obtaining desirable topotypic material, the
type locality has been noted wherever possible. I wish to take this
opportunity to sincerely thank Mr. A. Loveridge and Dr. J. C.
Bequaert for their able assistance in checking many of these localities.

Although many workers since the publication of Miss Rathbun's
monograph have considered Potamon a neuter noun, it is here con-
sidered masculine. In its original spelling the word must have been
derived from -kotolixlov, a masculine proper noun, even though its
connotation must remain obscure.

The references given below include, in addition to the original one,
only those which have appeared since Miss Rathbun's monograph
(1904-1906); all other earlier references may be obtained from that
work.

204 bulletin: museum of comparative zoology

POTAMONIDAE

POTAMONINAE

POTAMON

Potamon africanus (A. Milne Edwards)

Thelphusa africana A. Milne Edwards, 1869, p. 186, pi. 11, figs. 2, 2a.
Potamon (Potamonautes) africanus Rathbun, 1904, pi. 16, fig. 6; 1905, p. 188,

text-fig. 47.
Potamon {Potamonautes) africanum Colosi, 1920, p. 34.
Potamon (Potamonautes) africanum Colosi, 1924, p. 21, text-fig. 16.
Potamon (Potamonautes) africanum Roux, 1927, p. 237.
Potamonautes africanus Balss, 1929a, p. 124, text-figs. 5-7.
Potamonautes africanus Balss, 1936, p. 166.

Type locality. Gabon, French Congo.

Potamon alluaudi Bouvier
Potamon (Potamonautes) Alluaudi Bouvier, 1921, p. 46, text-figs. 1-3.

Type locality. Amboni River (alt. 1,800 m.), western part of Mt.
Kenya, Kenya Colony.

This species is considered a synonym of P. neumanni by Balss
(1929b).

Potamon aloysii-sabaudiae Nobili

See page 190.

Potamon amalerensis Rathbun
Text-figure 8

Potamon (Geothelphusa) amalerensis Rathbun, 1935, p. 25, pi. 2.

Type locality. Amaler River (5,000 ft.), Mt. Debasien, Uganda.

Potamon ambiguus Rathbun

Telphusa Hilgendorfi Hilgendorf, 1898, p. 9, pi., fig. 3. Not T. hilgendorfi Pfef-

fer (1889).
Potamon (Potamonautes) ambiguus Rathbun, 1904, pi. 14, fig. 7; 1905, p. 171.
Potamon (Potamonautes) ambiguus Lenz, 1910b, p. 121. (These specimens

should be referred to P. dybovcskii according to Balss (1929b).

chace: decapod Crustacea

205

Miss Rathbun's original specimens came from the Lumi (Loumi)
River, Mt. Kilimanjaro, Tanganyika Territory, and Bura (Boura) in
the Taita Mountains and Kibwezi, Kenya Colony.

Bouvier (1921), after comparing the types of P. ambiguus with a
figure of the type of P. johnstoni, concluded that this is a synonym of
the latter species and Balss (1929b) has concurred in this.

Fig. 8. Potamon {Geothelphusa) amalerensis. Posterior view of tip of first
abdominal appendage of male holotype. x 40.

POTAMOX AXCHIETAE (Capello)

Tdphusa Anchietae Capello, 1871, p. 132, pi. 2, fig. 11.
Potamon (Potamonautes) Anchietae Rathbun, 1905, p. 166.
Potamon (Potamonautes) anchietae Sendler, 1912, p. 199.
Potamonautes anchietae Balss, 1929a, p. 117.

Capello's original specimens came from Dondo, Pungo-Andongo
and Ambaca, Angola.

Doflein (1904) and Colosi (1920 & 1924) considered this species
synonymous with P. perlatus, but Balss (1929a) retains it as a distinct
species and synonymizes P. regnieri with it.

206 bulletin: museum of comparative zoology

POTAMON ANKARAHARAE NobiH

Potamon (Geothelphusa) ankaraharae Nobili, 1906a, p. 1, text-figs. A-C.
Parathelphusa (Bary thelphusa) Ankaraharae Colosi, 1920, p. 22.
Potamon (Geotelphusa) ankaraharae Balss, 1929b, p. 356.

Type locality. "Ankarahara", Madagascar.

P. methueni is a synonym of this species according to Colosi (1920)
and Balss (1929b). '

Potamon antjieus Colosi

Potamon {Geothelphusa) Anthem Colosi, 1920, p. 35.

Potamon (Geothelphusa) Anteus Colosi, 1924, p. 17, text-fig. 12, pi. 1, fig. 6.

Geotelphusa antheus Balss, 1929b, p. 351, text-fig. 1.

Type locality. Southwestern Ethiopia.

Potamon antongilensis Rathbun

Potamon (Parathelphusa) antongilensis Rathbun, 1905, p. 265, pi. 14, fig 5.
Potamon (Geotelphusa) antongilensis Balss, 1929b, p. 355, text-fig. 2.

Type locality. Antongil Bay, Madagascar.

Potamon aubryi (H. Milne Edwards)

Thelphusa Aubryi H. Milne Edwards, 1853, p. 210.

Potamon (Potamonautes) Aubriji Rathbun, 1904, pi. 17, figs 3, 4 & 7; 1905,

p. 191.
Potamon (Potamonautes) Aubryi Doflein, 1904, p. 105.
Potamonautes aubryi Stebbing, 1910, p. 294.
Potamonautes aubryi Balss, 1914a, p. 104.
Potamonautes aubryi Balss, 1914b, p. 405.
Potamon (Potamonautes) aubryi Roux, 1927, p. 237.
Potamonautes aubryi Balss, 1929a, p. 122, text-figs. 2-3.

Type locality. Gabon, French Congo.

Potamon ballayi (A. Milne Edwards)

Thelphusa Ballayi A. Milne Edwards, 1886, p. 149.

Potamon (Potamon) Ballayi Rathbun, 1904, p. 294, pi. 12, fig. 9.

Potamon (Potamon) ballayi Rathbun, 1921, p. 419, text-fig. 10, pis. 27 and 28,

fig. 1.
Potamonautes ballayi Balss, 1936, p. 174, text-fig. 9-13.

Type locality. Nganchu ("Ngancin"), French Congo.

chace: decapod Crustacea 207

POTAMON BAYONIANUS (Capello)

Telphusa Bayoniana Capello, 1864, p. 2, pi., fig. 3.

Potamon (Potamonautes) Bayonianus Rathbun, 1904, pi. 15, fig. 1; 1905, p. 178.

Potamonautes bayonianus Balss, 1922, p. 72.

Type locality. Duque de Braganca district, Angola.

Potamon berardi (Audouin)

Potamons or crabes fluviatiles Savigny, 1817, pi. 2, fig. 6.

Thelphusa Berardi Audouin, 1826, p. 82.

Potamon (Geothelphusa) Berardi Rathbun, 1904, pi. 18, figs. 3 & 10; 1905,

p. 203.
Potamon (Geotelphusa) Berardi Lenz, 1910b, p. 124. (These specimens are

P. emini according to Balss, 1929b, p. 345.)
Potamon (Geothelphusa) Berardi de Man, 1914, p. 126, pi. 2, figs. 3-3a.
Potamon (Geotelphusa) Berardi Colosi, 1919, p. 50.
Potamon (Geothelphusa) Berardi Colosi, 1920, p. 34.
Geotelphusa berardi Balss, 1929b, p. 350.
Potamon berardi Flower, 1931, p. 732.
Potamon (Geothelphusa) berardi Rathbun, 1935, p. 124.

Type locality. Egypt.

Potamon biballensis Rathbun

Telphusa Anchietae var. ? Capello, 1871, p. 132, pi. 2, fig. 11a.
Potamon (Potamonautes) biballensis Rathbun, 1905, p. 176.
Potamonautes biballensis Balss, 1936, p. 169, text-figs. 4-5.

Type locality. Biballa, Angola.

Potamon bipartitus (Hilgendorf)

Telphusa bipartita Hilgendorf, 1898, p. 15.

Potamon (Potamonautes) bipartitus Rathbun, 1905, p. 174.

Hilgendorf's original specimens came from the Belgian Congo at
Alibuaki, west of the Issango River; Ali stream, Undussuma district;
Bundeko; and Koganos.

Potamon bombetokensis Rathbun

Potamon (Potamon) bombetokensis Rathbun, 1904, p. 298, text-fig. 30, pi. 12,

fig. 6.
Potamon (Potamon) bombetokensis Balss, 1929b, p. 354.

Type locality. Near "Bombetok", Madagascar.

208 bulletin: museum of comparative zoology

Potamon bottegoi de Man

Potamon (Potamonautes) Bottegoi de Man, 1898, p. 262, pi. 3.
Potamon {Potamonautes) Bottegoi Rathbun, 1905, p. 180.
Potamon (Potamonautes) Bottegoi Colosi, 1925, p. 2.
Potamon [Potamonautes) Bottegoi Parisi, 1925, p. 98.
Potamon (Potamonautes) bottegoi Rathbun, 1933, p. 258.
Potamon (Potamonautes) bottegoi Rathbun, 1935, p. 26.

Type locality. "Matagoi Bool", between Brava and Lugh, Italian
Somaliland.

Comparison of a male specimen of P. obesus from Zanzibar, the type
locality of that species, with the several specimens from Kenya and
Uganda identified as P. bottegoi by Miss Rathbun leads me to agree
with Balss (1929b) that this species should be synonymized with
P. obesus.

Potamon bouvieri Rathbun

Potamon (Potamon) Bouvieri Rathbun, 1904, p. 293, pi. 12, fig. 5.

Type locality. "Pools of Velantanguel, Southarkot", India. Also
taken at Mauritius.

Potamon calcaratus Gordon

Potamon (Potamonautes) calcaratum Gordon, 1929, p. 405, text-figs. 1-5.

Type locality. Charre and Caia, Lower Zambezi Valley, Mozambi-
que.

Potamon campi (Rathbun)

Parathelphusa campi Rathbun, 1894, p. 25.

Potamon (Parathelphusa) Campi Rathbun, 1905, p. 256, pi. 14, fig. 1.

Potamon (Parathelphusa) Campi Lenz, 1912, p. 7.

Potamonautes campi Balss, 1936, p. 186, text-fig. 22 (map).

Type locality. Stanley Pool, Belgian Congo.

Potamon capelloanus Rathbun

Telphusa Bayoniana var. a Capello, 1871, p. 131, pi. 2, fig. 10.
Potamon (Potamonautes) Capelloanus Rathbun, 1905, p. 179.

Capello's original specimens were found at Kakonda ("Caconda")
and Huilla, Angola.

chace: decapod Crustacea 209

Potamon chaperi (A. Milne Edwards)

Parathelphusa Chaperi A. Milne Edwards, 1887, p. 144, pi. 8, fig. 4.
Potamon (Parathelphusa) Chaperi Rathbun, 1905, p. 262, pi. 14, fig. 6.

Type locality. Assini, Ivory Coast.

Potamon chavanesii (A. Milne Edwards)

Thelphusa Chavanesii A. Milne Edwards, 1886, p. 150.
Potamon (Parathelphusa) Chavanesii Rathbun, 1905, p. 232, pi. 13, fig. 1.
Potamon (Parathelphusa) chavanesii Sendler, 1912, p. 200.
Potamonautes chavanesii Balss, 1929a, p. 127.

Type locality. Franeeville Lake, Gabon, French Congo.

Potamon congoensis Rathbun

Potamon (Geothelphusa) congoensis Rathbun, 1921, p. 422, text-fig. 11, pis. 28,

fig. 3, and 29.
Potamon (Geotelphusa) congoensis Parisi, 1925, p. 97.
Geothelphusa congoensis Balss, 1936, p. 192, text-fig. 27 (map).

Type locality. Nepoko River, above Gamangui, Belgian Congo.

Potamon decazei (A. Milne Edwards)

Thelphusa Decazei A. Milne Edwards, 1886, p. 150.

Potamon (Potamonautes) Decazei Rathbun, 1904, pi. 16, fig. 3; 1905, p. 197.

Potamon (Potamonautes) decazei Sendler, 1912, p. 200.

Potamonautes Decazei Balss, 1914a, p. 103.

Potamonautes decazei Balss, 1914b, p. 405. (Part of these specimens were later

referred by Balss (1929a) to P. pobeguini).
Potamonautes decazei Balss, 1929a, p. 118, pi., fig. 2.

Type locality. Franeeville (Alima River), Gabon, French Congo.

Potamon depressus (Krauss)

Thelphusa depressa Krauss, 1843, p. 38, pi. 2, figs. 4-4c.
Potaman (Potamonautes) depressus Rathbun, 1905, p. 169.
Potamonautes depressus Stebbing, 1910, p. 294.
Potamon (Potamonautes) depressus Lenz, 1912, p. 7.
Potamonautes depressus Barnard, 1935, p. 484.

Type locality. Near Pietermaritzburg, Natal.

210 bulletin: museum of comparative zoology

Potamon didieri Rathbun

Potamon {Potamonautes) Didieri Rathbun, 1904, pi. 14, fig. 9; 1905, p. 170.
Potamon {Potamonautes) didieri Sendler, 1912, p. 198.
Potamon {Potamonautes) Didieri Colosi, 1924, p. 5.
Potamonautes emini didieri Balss, 1929b, p. 246.
Potamon {Potamonautes) didieri Rathbun, 1935, p. 26.

Type locality. "Le Kibali (embouchure), 1015 metres d'altitude;
L. Didier, 1903, Mission du Bourg de Bozas." I have been unable to
check this locality; it probably refers to an Ethiopian locality since I
doubt that Didier collected in the Belgian Congo near the mouth of
the Kibali River.

Potamon dubius (Capello)

Telphusa dubia Capello, 1873, p. 254, pi. 1, figs. 1-2.
Potamon {Potamonautes) dubius Rathbun, 1905, p. 179.
Potamon {Pota?nonautes) dubius Colosi, 1918, p. 106.
Potamon {Potamonautes) dubium Colosi, 1919, p. 51.
Potamon {Potamonautes) dubium Colosi, 1920, p. 32.
Potamonautes dubius Balss, 1922, p. 71.

Type locality. Kunene (Cunene) River, interior of Mossamedes,
Angola.

Potamon dybowskii Rathbun
See page 187.

Potamon ecorssei (Marchand)

Potamon {Potamonautes) Ecorssei Marchand, 1902, p. 334, pi. 13, figs. 2 and 6
Potamon {Potamonautes) Ecorssei Rathbun, 1905, p. 180.
Potamon {Potamonautes) ecorssei Roux, 1935a, p. 32.

Type locality. Lake Tele, west of Timbuktu, French Sudan.

Potamon edulis (Latreille)

Potamophilus edulis Latreille, 1818, pi. 297, fig 4.

Potamon {Potamon) edidis Rathbun, 1904, p. 254, pi. 9, fig. 1.

Potamon {Potamon) edule var. africanum Colosi, 1920, p. 31.

Type locality. ?.

Potamon emini (Hilgendorf)
See page 193.

chace: decapod Crustacea 211

POTAMON FARADJENSIS Rathbun

Potamon (Acanthothelphusa) faradjensis Rathbun, 1921, p. 428, text-fig. 13,

pi. 31.
Potamonautes faradjensis Balss, 1929a, p. 126, text-fig. 8.
Potamonautes faradjensis Balss, 1936, p. 166, fig. 1 (map).

Type locality. Dungu River at Faradje, Belgian Congo.

Potamon floweri de Man

Potamon (Potamonautes) Floweri de Man, 1901, p. 94, pi. 10.

Potamon (Potamonautes) Floweri Rathbun, 1904, pi. 17, figs. 2 and 6; 1905,

p. 193.
Potamon (Potamonautes) floweri Rathbun, 1921, p. 406, text-fig. 6, pi. 20, fig. 2.
Potamon (Potamonautes) Floweri Parisi, 1925, p. 99.
Potamonautes floweri Balss, 1929b, p. 347.
Potamon floweri Flower, 1931.
Potamonautes floweri Balss, 1936, p. 171, text-fig. 6 (map).

Type locality. Bahr el Gebel, Anglo-Egyptian Sudan.

Potamon goudoti (H. Milne Edwards)

Thelphusa Goudoti H. Milne Edwards, 1853, p. 212.

Potamon (Potamon) Goudoti Rathbun, 1904, p. 305, pi. 13, fig. 10.

Potamon Goudoti Lenz, 1910a, p. 557.

Potamon (Potamon) goudoti Caiman, 1913, p. 920.

Potamon (Potamon) goudoti Balss, 1929b, p. 920.

Type locality. Madagascar.

Potamon grandidieri Rathbun
Potamon (Potamon) Grandidieri Rathbun, 1904, p. 298, pi. 12, fig. 11.
Type locality. Near "Bombetok", Madagascar.

Potamon granulata (Balss)

Potamonautes decazci granulata Balss, 1929a, p. 119.

The original specimens were taken at Misahohe and Bismarck-
burg, Togo.

Potamon granviki Colosi

Potamon (Geothelphusa) Granviki Colosi, 1924, p. 16, text-fig. 11, pi. 1, fig. 5.

Type locality. Mount Elgon, Uganda (7,000 feet).
Balss (1929b) and Roux (1935) consider this species a synonym of
P. loveni.

212

bulletin: museum of comparative zoology

Potamon harvardi Rathbun
Text-figure 9.

Potamon (Geothelphusa) harvardi Rathbun, 1935, p. 23, pi. 1.

A B

Fig. 9. Potamon {Geothelphusa) harvardi. A, posterior view of first and
second right abdominal appendages of male holotype, x 6.9; B, posterior view
of tips of same appendages, x 21.4.

Type locality. Sipi (6,500 feet), western part of Mount Elgon,
Uganda.

Only in the holotype does the second abdominal appendage run
up through the first; in all other males these two appendages are
quite separate from one another.

Potamon hilgendorfi (Pfeffer)

See page 186.

chace: decapod Crustacea 213

Potamon humbloti Rathbun
Potamon (Potamon) Humbloti Rathbun, 1904, p. 297, pi. 12, fig. 10.
Type locality. Madagascar.

Potamon idjwiensis Chace
See page 197.

Potamon ignestii Parisi

Potamon (Geotelphusa) Ignestii Parisi, 1923, p. 332. text-fig. 1, pi. 8.
Potamon (Geotelphusa) Ignestii Parisi. 1925, p. 98.

Type locality. Gondar, from stream emptying into Lake Tana,
Ethiopia.

Potamon inflatus (H. Milne Edwards)

Thelphusa inflata H. Milne Edwards, 1853, p. 210.

Potamon (Potamonautes) inflatus Rathbun, 1904, pi. 15, fig. 2; 1905, p. 174.

Potamon (Potamonautes) inflatus Cunnington, 1907, p. 259.

Potamonautes inflatus Stebbing, 1910, p. 294.

Potamonautes inflatus Barnard, 1935, p. 484.

Type locality. Durban (Port Natal), Natal.

Potamon infravallatus (Hilgendorf)

Telphusa infravallata Hilgendorf, 1898, p. 12, pi., figs. 2, 2a.
Potamon (Potamonautes) infravallatus Rathbun, 1905, p. 174.

Hilgendorf's specimens came from Buloa and Derema in the Usam-
bara Mts., Tanganyika Territory.

Potamon jallae (Nobili)

Thelphusa dubia var. Jallae Nobili, 1896.

Potamon (Potamonautes) dubius Jallae Rathbun, 1904, pi. 15, fig. 6; 1905,

p. 179.
Potamonautes dubius jallae, Balss, 1922, p. 72.

Potamon (Potamonautes) dubium Jallae Colosi, 1924, p. 4, text-fig. 2.
Potamonautes dubius var. jallae Barnard, 1935, p. 486, text-figs. 1 (k-1).
Potamonautes dubius jallae Balss, 1936, p. 177, text-figs. 14-16.

Type locality. Kazungula, Northern Rhodesia.

214 bulletin: museum of comparative zoology

POTAMON JEANNELI Bouvier

Potamon (Geothelphusa) Jeanneli Bouvier, 1921, p. 51, text-figs. 5 and 6.
Potamon (Geothelphusa) Jeanneli Colosi, 1924, p. 15, text-fig. 10.

Type locality. Mount Kenya (2,700 meters), Kenya Colony.

Potamon johnstoni (Miers)

Thelphusa depressa var. johnstoni Miers, 1885, p. 237.

Potamon (Potamonautes) Johnstoni Rathbun, 1905, p. 170.

Potamon (Potamonautes) johnstoni Caiman, 1909, p. 51, text-figs. 9-12. (The

Ruwenzori specimens are P. aloysii-sabaudiae; see page 190.)
Potamon (Potamonautes) Johnstoni Sjostedt, 1910, p. 1.
Potamon (Potamonautes) johnstoni Lonnberg and Budde-Lund, 1912, p. 1.
Potamon (Potamonautes) johnstoni ? Sendler, 1912, p. 198.
Potamon (Potamonautes) Johnstoni Colosi, 1920, p. 32. (These specimens

should be referred to P. aloysii-sabaudiae; see page 190.)
Potamon (Potamonautes) Johnstoni Bouvier, 1921, p. 44.
Potamon (Potamonautes) Johnstoni Colosi, 1924, p. 21, text-fig. 15. (These

specimens should be referred to P. aloysii-sabaudiae; see page 190.)
Potamonautes johnstoni Balss, 1929b, p. 343.
Potamonautes johnstoni Balss, 1936, p. 180, text-fig. 17.
Potamonautes johnstoni (f. typica) Pesta, 1937, p. 157.

Type locality. Mount Kilimanjaro, Tanganyika Territory.

Bouvier (1921) refers P. ambiguus to this species; Balss (1929b)
concurs in this and adds P. infravallatus and P. reichardi; and the
same author (1936) considers P. usambarae a synonym. See page 189
for remarks on the latter species. P. aloysii-sabaudiae is also con-
sidered synonymous by Colosi (1920 and 1924) and Balss (1929b)
on the basis of Caiman's discussion; this matter is discussed above
on page 190. It is difficult to determine whether the specimens iden-
tified as P. johnstoni by Lonnberg and Budde-Lund and by Sendler
really belong to this species, but I believe that the other references
(with the possible exception of some of the specimens listed by Balss
(1929b)) are correct unless otherwise noted.

Potamon laetabilis de Man

Potamon (Geothelphusa) Neumanni var. laetabilis de Man, 1914, p. 122, pi. 2,

figs, 1-lb.
Potamonautes emini laetabilis Balss, 1929b, p. 346.

Type locality. Let Marefia, Choa ("Schoa"), Ethiopia.

chace: decapod Crustacea 215

Potamon langi Rathbun

Potamon (Acanthothelphusa) langi Rathbun, 1921, p. 430, text-fig. 14, pi. 32.
Potamonautes langi Balss, 1936, p. 189, text-fig. 25 (map).

Type locality. Congo River at Stanleyville, Belgian Congo.

Potamon latidactylus de Man

Telphusa africana de Man, 1881, p. 121. Not Thelphusa africana A. Milne

Edwards, 1869.
Potamon {Potamonautes) latidactylum de Man, 1903, p. 41, pi. 9. figs. 1-6.
Potamon (Potamonautes) latidactylus Rathbun, 1904, pi. 16, fig. 7; 1905, p. 190.
Potamonautes latidactylus Balss, 1914b, p. 405.

Potamon (Potamonautes) latidactylum Colosi, 1924, p. 12, text-fig. 8.
Potamon (Potamonautes) latidactylus Roux, 1935a, p. 31.

Type locality. Prah River, Ashanti, Gold Coast.

Potamon lindblomi Colosi

Potamon (Potamonautes) Lindblomi Colosi, 1924, p. 5, text-fig. 3, pi. 1, fig. 2.

Type locality. Machako's, southeast of Nairobi, Kenya Colony.

Potamon lirrangensis Rathbun
See page 188.

Potamon longimerus Roux
Potamon (Geotelphusa) Loveni longimerus Roux, 1935, p. 244, text-figs. 1-3.
Type locality. Mount Elgon (3,900 to 4,000 meters), Kenya Colony.

Potamon loveni Colosi

Potamon (Geothelphusa) Loveni Colosi, 1924, p. 13, text-fig. 9, p. 1, fig. 4.

Geotelphusa loveni Balss, 1929b, p. 351.

Potamon (Geotelphusa) Loveni Roux, 1935, p. 243.

Type locality. Mount Elgon, Uganda.

P. granviki is a synonym of this species according to Balss (1929b)
and Roux (1935).

Potamon loveridgei Rathbun

Potamon (Potamonautes) loveridgei Rathbun, 1933, p. 251, pi. 1, pi. 2, fig. 1.

Type locality. Luiche River, Ujiji, Tanganyika Territory.
P. stappersi is a synonym of this species.

216 bulletin: museum of comparative zoology

Potamox lueboexsis Rathbun

Potamon {Potamonautes) lueboensis Rathbun, 1904, pi. 14, fig. 2; 1905, p. 166.
Potamonautes lueboensis Balss, 1936, p. 172, text-figs. 7 and 8.

Type locality. Luebo, Belgian Congo.

Colosi (1924) includes this species in the synonymy of P. perlatum.
Balss (1929a) considered it a synonym of P. anchietae but later
(1936) decided that it was distinct.

Potamox macropus Rathbun
See Cylindrotelphusa macropus (p. 226).

Potamox madagascariexsis (A. Milne Edwards)

Thelphusa madagascariensis A. Milne Edwards, 1872, p. 1.

Potamon {Potamon) madagascariensis Rathbun, 1904, p. 264, pi. 9, fig. 7.

Patamon madagascarensis Lenz, 1910a, p. 557.

Potamon {Potamon) madagascariense Caiman, 1913, p. 916.

Potamon {Potamon) madagascariensis Balss, 1929b, p. 354.

Type locality. On route between "Bombetok" and Antananarivo
(Tananarive), Madagascar.

P. pittarellii was considered a synonym of this species by Balss
(1929b), but later (1934) he decided the two were distinct.

Potamox marchei Rathbun

Potamon {Parathelphusa) Marchei Rathbun, 1902, p. 187.
Potamon {Parathelphusa) Marchei Rathbun, 1905, p. 264, text-fig. 70, pi. 14,
fig. 4.

Type locality. Samkita, Ogowe River, Gabon, French Congo.

Potamox margaritarius (A. Milne Edwards)

Thelphusa maragaritaria A. Milne Edwards, 1869, p. 185, pi. 9, figs. 4-4b.
Potamon {Potamonautes) margaritarius Rathbun, 1904, pi. 14, fig. 10; 1905,

p. 168, text-fig. 41.
Thelphusa margaritaria Osorio, 1905, p. 149.
Potamonautes margaritarius Balss, 1914a, p. 102.
Potamon {Potamonautes) margaritarius de Man, 1914, p. 135.

Type locality. St. Thomas Island, west of Gabon, French Congo.

chace: decapod Crustacea 217

Potamon methueni Caiman

Potamon {Potamon) methueni Caiman, 1913, p. 920, pi. 91.

Type locality. "Imerimandrosa", Madagascar.

This species is a synonym of P. ankaraharae according to Colosi
(1920) and Balss (1929b).

Potamon monodi (Balss)

Potamonautes aubryi monodi Balss, 1929a, p. 123, text-fig. 4.

The original specimens were taken at the following localities in
Gabon, French Congo: Garua, between Tschamba and Laro, Satsche,
Benue and Moba.

Potamon mrogoroensis (Hilgendorf)

Telphusa mrogoroensis Hilgendorf, 1898, p. 10.

Potamon {Potamonautes) mrogoroensis Rathbun, 1905, p. 173.

Type locality. Morogoro, Tanganyika Territory.
This species is a synonym of P. hilgendorft according to Balss
(1929b).

Potamon mutandensis Chace
See page 194.

Potamon neumanni (Hilgendorf)

Telphusa neumanni Hilgendorf, 1898, p. 18, pi., fig. 6.

Potamon {Geothelphusa) Neumanni Rathbun, 1905, p. 210.

Potamon {Geothelphusa) Neumanni de Man, 1914, p. 122, pi. 2, figs. 2-2b.

Potamon {Geothelphusa) Neumanni Colosi, 1920, p. 34.

Potamon {Geothelphusa) Neumanni Colosi, 1924, p. 18, text-fig. 13, pi. 1,

fig. 7.
Geotelphusa neumanni Balss, 1929b, p. 350.
Potamon {Geotelphusa) Neumanni Roux, 1935, p. 246.

Type locality. Ngare Rongai ("Xgare Longai"), Kenya Colony.
P. alluaudi is a synonym of this species according to Balss (1929b).

Potamon nigrensis Rathbun

Potamon {Potamon) nigrensis Rathbun, 1904, p. 295, pi. 12, fig. 8.
Potamonautes nigrensis Balss, 1936, p. 200.

Type locality. Niger River between Timbuktu and Say, French
Sudan.

218 bulletin: museum of comparative zoology

Potamon niloticus (H. Milne Edwards)

Thelpheusa nilotica H. Milne Edwards, 1837, p. 12.

Potamon (Parathelphusa) niloticus Rathbun, 1905, p. 263, pi. 14, fig. 15.

Parathelphusa nilotica Nobili, 1906b, p. 1.

Parathelphusa nilotica Nobili, 1909, p. 357.

Potamon (Parathelphusa) niloticus Lenz, 1912, p. 3.

Potamon {Acanthotelphusa) niloticum Colosi, 1919, p. 52.

Potamon (Acanthothelphusa) niloticum Colosi, 1920, p. 27.

Potamon (Potamonautes) niloticum Colosi, 1924, p. 12, text-fig. 7.

Potamonautes niloticus Balss, 1929b, p. 348.

Potamon nilotica Flower, 1931, p. 733.

Potamon (Acanthothelphusa) niloticus Rathbun, 1933, p. 258.

Potamon (Acanthothelphusa) niloticus Rathbun, 1935, p. 25.

Type locality. Nile River.

Potamon obesus (A. Milne Edwards;
See page 190.

Potamon odhneri Colosi

Potamon (Potamonautes) perlatum Colosi, 1920, p. 33.

Potamon (Potamonautes) Odhneri Colosi, 1924, p. 7, text-fig. 4, pi. 1, fig. 3.

Type locality. Limuru, Kenya Colony.

Potamon orbitospinus Cunnington

Potamon (Potamonautes) orbitospinus Cunnington, 1907, p. 259, pi. 16, fig. 1.
Potamonautes orbitospinus Balss, 1929b, p. 349.
Potamonautes orbitospinus Balss, 1936, p. 182, text-fig. 18.

Cunnington's specimens were taken along the western shore of
Lake Nyasa.

Potamon paecilei (A. Milne Edwards)

Thelphusa Paecilei A. Milne Edwards, 1886, p. 149.

Potamon (Parathelphusa) Paecilei Rathbun, 1904, pi. 17, fig. 5; 1905, p. 257,
text-fig. 67.

Type locality. Alima River, Ubangi-Shari Province, French Congo.

Potamon pelii (Herklots)

Thelphusa Pelii Herklots, 1861, p. 13.

Potamon (Potamonautes) Pelii Rathbun, 1905, p. 193.

Potamon (Potamonautes) pelii Sendler, 1912, p. 198.

Type locality. Elraina ("St. George del Mina"), Gold Coast.

chace: decapod Crustacea 219

Potamon perlatus (H. Milne Edwards)

Thelpheusa perlata H. Milne Edwards, 1837, p. 13.

Potamon (Potamonautes) perlatus Rathbun, 1904, pi. 14, fig. 4; 1905, p. 163.

Potamon (Potamonautes) perlatum Doflein, 1904, p. 105.

Thelphusa perlata Osorio, 1905, p. 149.

Potamonautes perlatus Stebbing, 1905, p. 33.

Thelphusa perlata Stimpson, 1907, p. 113.

Potamonautes perlatus Stebbing, 1910, p. 293.

Potamon [Potamonautes) perlatus Lenz, 1910a, p. 558.

Potamon {Potamonautes) perlatus Lenz, 1910b. p. 124.

Potamon (Potamonautes) perlatum Colosi, 1920, p. 33. (These specimens were

later (1924) made the types of P. odhneri by Colosi).
Potamonautes perlatus Balss, 1922, p. 71.

Potamon (Potamonautes) perlatum Colosi, 1924, p. 2, text-fig. 1.
Potamonautes perlatus Barnard, 1935, p. 482, text-figs, la-lb.
Potamonautes perlatus Balss, 1936, p. 184, text-figs. 20 and 21.

Type locality. Cape of Good Hope.

Colosi (1924) included P. anchictae, P. lueboensis, P. regnieri,
P. reichardi and P. sidneyi among the synonyms of this species.
P. anchietae had previously been synonymized by Doflein (1904) and
Colosi (1920). Balss (1929a) followed Colosi in synonymizing P. lue-
boensis but later (1936) considered the latter a distinct species.

Potamon perparvus Rathbun

Potamon (Geothelphusa) perparvus Rathbun, 1921, p. 425, text-fig. 12, pis.

28, fig. 2, and 30.
Potamon (Geothelphusa) perparvus Rathbun, 1935, p. 24.

Type locality. Stanleyville, Belgian Congo.

Potamon perrieri Rathbun
See Cylindrotclphusa perrieri (page 226).

Potamon pilosus (Hilgendorf)

Telphusa pilosa Hilgendorf, 1898, p. 19.

Potamon (Geothelphusa) pilosus Rathbun, 1905, p. 210.

Potamonautes emini var. pilosus Balss. 1929b, p. 347.

Type locality. Rain forest near Marangu (at base of Mt. Kili-
manjaro), Tanganyika Territory.

220 bulletin: museum of comparative zoology

POTAMON PITTARELLII Nobili

Potamon (Potamon) Pittarellii, Nobili, 1905, p. 1, text-fig.
Parathelphusa (Oziothelphusa) Pittarellii Colosi, 1920, p. 25.
Potamon pittarellii Balss, 1934, p. 520, pi. 1, fig. 1.

Type locality. Moramanga, Madagascar.

Balss (1929b) considered this a synonym of P. madagascariensis
but later (1934) retained it as a distinct species

Potamon platycentron (Hilgendorf)

Telphusa platycentron Hilgendorf, 1897, p. 81.

Potamon (Potamonautes) platycentron Rathbun, 1905, p. 173.

Potamonautes platycentron Balss, 1929b, p. 349.

Type locality. Lake Chala ("Tschala"), Kenya Colony.

Potamon platynotus Cunnington

Potamon (Potamonautes) platynotus Cunnington, 1907, p. 264, pi. 17, figs.

1 and 3.
Potamonautes platynotus Balss, 1929b, p. 349.
Potamonautes platynotus Balss, 1936, p. 185.

The original specimens came from Lake Tanganyika.

Potamon pobeguini Rathbun

Potamon {Potamonautes) Pobeguini Rathbun, 1904, pi. 16, fig. 8; 1905, p. 195.

Potamonautes pobequini Stebbing, 1910, p. 295.

Potamonautes pobeguini Balss, 1929a, p. 120, text-fig. 1, pi., fig. 1.

Type locality. Bata (Batah), Gabon, French Congo.

Potamon potamios (Olivier)

Cancer potamios Olivier, 1804, p. 240, pi. 30, fig. 2.

Potamon {Potamon) potamios Rathbun, 1904, p. 257, text-fig. 2, pi. 9, fig. 5.

Type locality. ?

Potamon pseudoperlatus (Hilgendorf)

Telphusa suprasulcata var. pseudoperlata Hilgendorf, 1898, p. 9.

Potamon (Potamonautes) suprasulcatus pseudoperlatus Rathbun, 1905, p. 173.

Type locality. Usambara region, Tanganyika Territory.

chace: decapod Crustacea 221

Potamon regnieri Rathbun

Potamon {Potamonautes) Regnieri Rathbun, 1904, pi. 14, fig. 3; 1905, p. 168,
text-fig. 40.

Type locality. Sanga River, French Congo.

This species was referred to P. perlatus by Colosi (1924) and to
P. anchietae by Balss (1929a).

Potamon reichardi (Hilgendorf)

Telphusa Reichardi Hilgendorf, 1898, p. 13.
Potamon {Potamonautes) Reichardi Rathbun, 1905, p. 166.
Potamonautes reichardi Balss, 1914b, p. 404.

Potamon {Potamonautes) reichardi Rathbun, 1933, p. 254, pi. 3, pi. 4, figs 3
and 4.

Type locality. South of Tabora (?), Tanganyika Territory.
Colosi (1924) synonymized this species with P. perlatum and Balss
(1929b) referred it to P. johnstoni.

Potamon rodolphianus Rathbun

Potomon {Potamonautes) rodolphianus Rathbun, 1909, p. 102.
Potamon {Potamonautes) rodolphianus Rathbun, 1922, p. 35, text-fig. 1, pi.
C3, figs. 1-3.

Type locality. South of Lake Rudolf, Kenya Colony.

Potamon rothschildi Rathbun

Potamon {Potamonautes) Rothschildi Rathbun, 1909, p. 103.
Potamon {Potamonautes) Rothschildi Rathbun, 1922, p. 37, text-fig. 2, pi. C3,
figs. 4-9.

Type locality. Kenya Colony.

Potamon schubotzi (Balss)
Geotelphusa schubotzi Balss, 1914a, p. 103, figs. 7-12.
Type locality. Duma, Belgian Congo.

222 bulletin: museum of comparative zoology

Potamon sidneyi Rathbun

Potamon {Potamonautes) Sidneyi Rathbun, 1904, pi. 14, fig. 5; 1905, p. 165,

text-fig. 38.
Potamonautes sidneyi Stebbing, 1910, p. 295.
Potamon {Potamonautes) Sidneyi Lenz, 1912, p. 7.
Potamonautes sidneyi Balss, 1922, p. 71
Potamonautes perlatus {sidneyi form) Barnard, 1935, p. 483, text-fig. lc.

Type locality. Natal.

Colosi (1924) refers this species to P. perlatus.

Potamon socotrensis (Hilgendorf)

Telphusa socotrensis Hilgendorf, 1883, p. 171.

Potamon {Geothelphusa) Socotrensis Rathbun, 1905, p. 212.

Potamon socotrense Balss, 1929b, p. 342.

Type locality. Kerignigi, Soeotra Island.

Potamon stanleyensis Rathbun
Text-figure 10

Potamon {Potamonautes) stanleyensis Rathbun, 1921, p. 415, text-fig. 9, pi. 26,
figs. 1-2.

Type locality. Affluents of the Chopo (Tshopo) River at Stanley-
ville, Belgian Congo.

Balss (1936) considers this species synonymous with P. dybowskii;
for a discussion of this, see remarks above under the latter species
(page 187).

Potamon stappersi (Balss)
Potamonautes johnstoni stappersi Balss, 1936, p. 182, text-figs, 19 and 20.

The original specimens were taken near Lake Tanganyika.

As Balss suggested, this species should be synonymized with
P. loveridgei; comparison of the first male abdominal appendage of
that species with Balss' figures proves the two species to be identical.

Potamon suprasulcatus (Hilgendorf)
See P. hilgendorfi (page 186).

chace: decapod Crustacea

223

POTAMON UNISULCATUS Rathbun

Potamon (Potamonautes) johnstoni unisulcatus Rathbun, 1933, p. 255, pi. 2,
figs. 1, 2 and 4.

Type locality. Bagilo, Uluguru Mountains, Tanganyika Territory.

Fig. 10. Potamon (Potamonautes) stanleyensis; posterior view of end of
first right abdominal appendage of male paratype (carapace breadth 342.
mm.) from Stanleyville, Belgian Congo, x 25.

Potamon usambarae Rathbun
See page 189.

Potamon vandexbrandeni (Balss)
Potamonautes vandenbrandeni Balss, 1936, p. 190, text-fig. 26.
Type locality. Leopoldville, Belgian Congo.

Potamon walderi Colosi

Potamon (Potamonautes) Walderi Colosi, 1924, p. 8, text-fig. 5.
Potamonautes Walderi Balss, 1936, p. 167, text-figs. 2 and 3. •

Type locality. "Kingoyi", Lower French Congo.

224 bulletin: museum of comparative zoology

Potamon warreni Caiman

Potamon (Potamonautes) warreni Caiman, 1918, p. 234.

Potamon (Potamonautes) Warreni Colosi, 1924, p. 9, text-fig. 6, pi. 1, figs. 1-la.

Potamonautes warreni Barnard, 1935, pp. 483-484, text-figs. Id— lj .

Type locality. Potchefstroom, Transvaal.

Potamon, sp. ? Cunnington
Potamon {Potamonautes) sp. ? Cunnington, 1907, p. 262.
Kondowe to Karonga, Nyasaland.

Potamon, sp. ? Cunnington
Potamon {Potamonautes) sp. ? Cunnington, 1907, p. 266.
Lake Tanganyika.

HYDROTHELPHUSA

Hydrothelphusa agilis (A. Milne Edwards)

Hydrothelphusa agilis A. Milne Edwards, 1872, p. 2.

Hydrothelphusa agilis Rathbun, 1905, p. 266, text-fig. 72, pi. 17, fig. 7.

Hydrothelphusa agilis Nobili, 1906b, p. 1.

Hydrothelphusa agilis Caiman, 1913, p. 922.

Hydrothelphusa agilis Colosi, 1920, p. 21.

Hydrotelphusa agilis Balss, 1929b, p. 357.

Type locality. Sakaleone River, Madagascar.

PLATYTHELPHUSA

Platythelphusa armata A. Milne Edwards

Platythelphusa armata A. Milne Edwards, 1887, p. 147, pi. 9.
Platythelphusa armata Rathbun, 1905, p. 269, pi. 21, fig. 4.
Platythelphusa armata Cunnington, 1907, p. 268, text-fig. 84.
Platytelphusa armata Balss, 1929b, p. 352.
Platythelphusa armata Balss, 1936, p. 196.

Type locality. Lake Tanganyika.

chace: decapod Crustacea 225

Platythelphusa conculcata Cunnington

Platythelphusa conculcata Cunnington, 1907, p. 273, pi. 17, figs. 2 and 4.
Platythelphusa conculcata Balss, 1936, p. 196, text-fig. 29.

Type locality. South end of Lake Tanganyika (10-15 fathoms).

Platythelphusa maculata (Cunnington)

Limnothelphusa maculata Cunnington, 1899, p. 698, pi. 38.
Limnothelphusa maculata Rathbun, 1905, p. 269.
Platythelphusa maculata Cunnington, 1907, p. 271, pis. 5-6.
Platythelphusa maculata Balss, 1936, p. 196.

Type locality. Kituta Bay, Lake Tanganyika.

ERIMETOPUS

Erimetopus brazzae (A. Milne Edwards)

Telphusa Brazzae A. Milne Edwards, 1886, p. 148.

Erimetopus Brazzae Rathbun, 1905, p. 270, text -fig. 73, pi. 21, fig. 8.

Erimetopus Brazzae Lenz, 1912, p. 9.

Erimetopus Brazzae Colosi, 1920, p. 27.

Erimetopus brazzae Rathbun, 1921, p. 433, text-fig. 15, pi. 33.

Erimetopus brazzae Balss, 1936, p. 195.

Type locality. Nganchu (Ngancin), Belgian Congo.

DECKENIINAE
DECKENIA

Deckenia alluaudi A. Milne Edwards and Bouvier

Deckenia Alluaudi A. Milne Edwards and Bouvier, 1893, p. 325, pi. 8.
Deckenia Alluaudi Rathbun, 1905, pi. 21, fig. 5; 1906, p. 72, text-fig. 124.
Deckenia alluaudi Borradaile, 1907, p. 63.

Type locality. Praslin Island, Seychelle Islands.

226 bulletin: museum of comparative zoology

Deckenia imitatrix (Hilgendorf)

Deckenia imitatrix Hilgendorf, 1869a, p. 2.

Deckenia imitatrix Rathbun, 1905, pi. 21, fig. 6; 1906, p. 69.

Deckenia imitatrix Colosi, 1918, p. 107.

Deckenia imitatrix Colosi, 1919, p. 53.

Deckenia imitatrix Colosi, 1925, p. 3.

Deckenia imitatrix Parish 1925, p. 99.

Deckenia imitatrix Balss, 1929b, p. 353.

Type locality. "Kudiano", Kenya Colony.

Deckenia mitis (Hilgendorf)
See page 201.

GECARCIXUCINAE
CYLINDROTELPHUSA

Cylindrotelphusa macropus (Rathbun)

Potamon (Geothelphusa) macropus Rathbun, 1898, p. 29, pi. 2, figs. 1-4.
Potamon (Geothelphusa) macropus Rathbun, 1904, pi. 18, fig. 1; 1905, p. 221.
Geotelphusa macropus Balss, 1914b. p. 406.
Cylindrotelphusa macropus Rathbun, 1921, p. 385.
Potamon (Geothelphusa) macropus Balss, 1936, p. 200.

Type locality. Mouth of the Mesurado River, Monrovia, Liberia.

Cylindrotelphusa perrieri (Rathbun)

Potamon (Geothelphusa) Perrieri Rathbun, 1904, pi. 18, fig. 11; 1905, p. 222,

text-fig. 53.
Cylindrotelphusa perrieri Rathbun, 1921, pp. 385 and 386.

Type locality. "Congo".

PARATHELPHUSA

Parathelphusa afzelii Colosi

Parathelphusa (Barythelphusa) Afzelii Colosi, 1924, p. 19, text-fig. 14, pi. 1,
fig. 8.

Type locality. Sierra Leone. Balss (1936, p. 200) doubts that this
locality is correct.

chace: decapod Crustacea 227

BIBLIOGRAPHY 1

Axcock, A.

1910a. Catalogue of the Indian Decapod Crustacea in the Collection of
the Indian Museum. I. Brachyura, II. Indian Freshwater Crabs,
Potamonidae. Calcutta, pp. 1-135, pis. 1-14.
1910b. On the classification of the Potamonidae. Rec. Indian Mus., 5,
pp. 253-261.

Audouin, V.

1826. Explication sommaire des planches de Crustaces de l'Egypte et de
la Syrie, publiees par Jules-Cesar Savigny, membre de l'lnstitut;
offrant un expose des caracteres naturels des genres, avec la
distinction des especes, in; Description de l'Egypte. Histoire
naturelle. Memoires, 1, pt. 4, pp. 77-98.

Balss, H.

1914a. Decapode Crustaceen von den Guinea-Inseln, Siid-Kamerun und

dem Congogebiet. Ergeb. zweiten Deutschen Zentral-Afrika-Exp.

1910-1911, 1, Zool., pp. 97-108, text-figs. 1-12.
1914b. Potamonidenstudien. Zool. Jahrb., Sj-st., 37, pp. 401-410, pi. 15,

text-figs. A-F.
1922. Crustacea Decapoda, in : Beitrage zur Kenntnis der Land- und

Siisswasserfauna Deutsch-Siidwestafrikas. by W. Michaelson.

Ergebn. Hamburger deutsch-siidwestafrik. Studienreise, 2, pp.

71-72.
1929a. Crustacea. V. Potamonidae, in : Contribution a l'Etude de la

Faune du Cameroun. Faun. Colon, franc, 3, pp. 115-129,

1 pi., text-figs. 1-8.
1929b. liber Ostafrikanische Potamonidae (Decapoda). Mit Anhang:

Potamoniden von Madagascar. Zool. Jahrb., Syst., 58, pp.

339-358, text-figs. 1-2.

1934. Sur quelques Decapodes brachyoures de Madagascar. Faun.
Colon, franc, 5, pp. 501-528, pi. 1, 1 text-fig.

1936. Beitrage zur Kenntnis der Potamonidae (Siisswasserkrabben)
des Kongogebiets. Rev. Zool. Bot. afr., 28, pp. 165-204, text-figs.
1-29.

Barnard, K. H.

1935. Scientific Results of the Vernay-Lang Kalahari Expedition, March
to September, 1930, Crustacea. Ann. Transv. Mus., 16, pp. 481-
492, pis. 26-27, text-figs. 1-2, 1 chart.

BORRADAILE, L. A.

1907. Percy Sladen Trust Expedition. Xo. III. Land and Freshwater
Decapoda. Trans, Linn. Soc, London, ser. 2, 12, pt. 1, pp. 63-68.

1 These references deal only with the fresh-water crabs (Potamonidae) and include only those
papers included in the text.

228 bulletin: museum of comparative zoology

Bouvier, E. L.

1917a. Sur la classification des Eupotamonea, Crabes d'eau douce de la

famille des Potamonides. C.R. acad. sci., 165, pp. 615-621.
1917b. Sur la classification des Parapotamonea, Crabes d'eau douce de la

famille des Potamonides. C.R. acad. sci., 165, pp. 657-659.
1921. Decapoda, in Voyage de Ch. Alluaud et R. Jeannel en Afrique

orientale (1911-1912). Resultats scientifiques. Crustac6s, III,

Paris, pp. 23-62, text-figs. 1-8.

Calman, W. T.

1909. Zoological Results of the Ruwenzori Expedition, 1905-1906.

5. Crustacea. Trans. Zool. Soc, London, 19, pt. 1, pp. 51-56,

text-figs. 9-12.
1913. On Freshwater Decapod Crustacea (Families Potamonidae and

Palaemonidae) collected in Madagascar by the Hon. Paul A.

Methuen. Proc. Zool. Soc, London, 1913, pp. 914-932, pis. 91-92,

text-fig. 161.
1918. A new River-crab from the Transvaal. Ann. Mag. Nat. Hist.,

(9) 1, pp. 234-236, fig.

Capello, F. de B.

1864. Descripcao de tres especies novas de Crustaceos da Africa occi-
dental e Observacoes acerca do Penoeus Bocagei Johnson, especie
nova dos mares de Portugal. Lisboa, pp. 1-11, 1 pi.

1871. Algumas especies novas ou pouco conhecidas de crustaceos
pertencentes dos generos "Calappa" e "Telphusa". J. Sci. Math.,
Phys. Nat., 3, pp. 128-134, pi. 2.

1873. Descripcao d'uma nova especie de "Telphusa" d'Africa occidental.
J. Sci. Math., Phys. Nat., 4, pp. 254-257, pi. 1, figs. 1-2.

Colosi, G.

1918. Crostacei Decapodi raccolti nella Somalia dai Dottori Stefanini e
Paoli. Monit. Zool. Ital., 29, no. 7, pp. 100-108.

1919. I Potamonidi conservati nel R. Museo Zoologico di Firenze. Bull.
Soc. entom. ital., 50 (1918), pp. 39-62.

1920. I Potamonidi del R. Museo Zoologicao di Torino. Boll. Mus.
Zool. Anat. Comp., Torino, 35, pp. 1-39.

1924. Potamonides africains du Museum de Stockholm. Ark. Zool.,
Stockholm, 16, no. 1, pp. 1-24, pi. 1, text-figs. 1-16.

1925. Crostacei raccolti nella Somalia dalla Missione della R. Societa
geografica (1924). Boll. Mus. Zool. Anat. Comp., Torino, 39, no.
32 (1924), pp. 1-4.

CUNNINGTON, W. A.

1899. On a new Brachyurous Crustacean from Lake Tanganyika.
Proc. Zool. Soc, London, 1899, pp. 697-704, pi. 38.

chace: decapod Crustacea 229

1907. Zoological Results of the Third Tanganyika Expedition, conducted
by Dr. W. A. Cunnington, 1904-1905. — Report on the Brachyur-
ous Crustacea. Proc. Zool. Soc, London, 1907, pp. 258-276, pis.
16-17, text-fig. 84.

1920. The Fauna of the African Lakes: a Study in Comparative Lim-
nology with special reference to Tanganyika. Proc. Zool. Soc,
London, 1920, pp. 507-622.

DOFLEIN, F.

1904. Wiss. Ergebn. Deutschen Tiefsee-Exped. "Valdivia". VI. Brach-
yura. Pp. 1-314.

Flower, S. S.

1931. Notes on Freshwater Crabs in Egypt, Sinai, and the Sudan.
Proc. Zool. Soc, London, 1931, pp. 729-735.

Gordon, I.

1929. A new River-Crab of the Subgenus Potamonautes from Portuguese
East Africa. Ann. Mag. Nat. Hist. (10) 3, pp. 405-411, 5 figs.

HlLGENDORF, F.

1869a. Ueber eine neue Gattung der Kurzschwanzigen Krebse aus den

Sammlungen des Baron von der Decken, Deckenia imitatrix.

Sitz.-Ber. Ges. naturf. Freunde, (1868) 1869, p. 2.
1869b. Crustaceen, in von der Decken, Reisen in Ost-Afrika, Zool., 3,

pp. 69-116, pis. 1-6.
1883. In Taschenberg, Beitrage zur Fauna der Insel Sokotra, vorzuglich

nach dem von Herrn Dr. Emil Riebeck aus Halle zusammen-

gestellt. Zeits. f. Naturw. (4) 2, pp. 171-173.
1892. Ueber eine neue ostafrikanische Siisswasserkrabbe (Telphusa

emini). Sitz.-Ber. Ges. naturf. Freunde, 1892, no. 1, pp. 11-13.

1897. Eine neue Landkrabbe (Telphusa platycenlron) aus Ostafrika.
Sitz.-Ber. Ges. naturf. Freunde, 1897, no. 6, pp. 81-85.

1898. Die Land- und Susswasser-Dekapoden Ostafrikas. Die Thierwelt
Deutsch Ost-Afr., 4, Lief. 7, pp. 1-37, 1 pi. 3 text-figs.

Krauss, C. F. F.

1843. Die Stidafrikanischen Crustaceen. Stuttgart, pp. 1-68, pis. 1-4.

Latreille, P. A.

1818. Tableau Encyclopedique et Methodique des Trois Regnes de la
Nature. Part 24. Crustaces, Archnides et Insectes. Paris.

Lenz, H.

1910a. Crustaceen von Madagaskar, Ostafrika und Ceylon, in: Voeltz-
kow, Reise in Ostafrika in den Jahren 1903-1905, 2, pp. 539-576,
text-figs. 1-4.

230 bulletin: museum of comparative zoology

1910b. Dekapode Crustaceen Aquatorial-afrikas, in: Wiss. Ergebn.

deutsch. Zentral-Afrika-Exped. 1907-1908 unter Flihrung

Adolph Friedrichs, Herzogs zu Mecklenburg, 3 (Zool, 1), pp. 121-

134, pi. 3.
1912. Afrikanische Crustaceen aus Schwedischen Sammlungen. Ark.

Zool., Stockholm, 7, no. 29, pp. 1-10.

Lonnberg, E. and Budde-Ltjnd, G.

1912. Crustacea collected by the Swedish Zoological Expedition to
British East Africa 1911. Ark. Zool. Stockholm, 7, no. 26, pp. 1-9,
text-figs. 1-5.

Man, J. G. de

1881. Carcinological Studies in the Leyden Museum, No. 1. Notes

Leyden Mus., 3, pp. 121-144.
1898. Description d'une espece nouvelle du genre Potamon Saw pro-

venant du pays des Somalis. Ann. Mus. Civ. Stor. Nat.. (2) 19

(39), pp. 262-270, pi. 3.

1901. Description of a new Freshwater Crustacean from the Soudan;
followed by some Remarks on an allied Species. Proc. Zool. Soc,
London, 1901, pp. 94-104, pi. 10.

1903. On Potamon (Potamonautes) latidactylum, a New Freshwater

Crab from Upper Guinea. Proc. Zool. Soc, London, 1903, pp.

41-47, pi. 9.
1914. Note sur quelques Crustaces Decapodes Brachyures terrestres et

d'eau douce appartenant au Musee civique de Genes. Ann. Mus.

Civ. Stor. Nat., 46, pp. 122-138, pis. 2-3.

Marchand, E.

1902. Description de Deux Arthropodes nouveaux Provenant du
Soudan frangais (Trox Borgognoi [Coleopt.] et Pot. [Potamonautes]
Ecorssei [Crust.]). Bull. Soc. Sci. Nat. Ouest France, (2), 2, pp.
331-342, pi. 13, figs. 2-6.

MlERS, E. J.

1885. Description of a new Variety of River-Crab, of the genus Thel-
phnsa, from Kilimanjaro. Proc. Zool. Soc, London, 1885, pp.
237-239.

Milne Edwards, A.

1868. Description de quelques Crustaces nouveaux provenant des
voyages de M. Alfred Grandidier a Zanzibar et a Madagascar.
Nouv. Arch. Mus. Hist. Nat., 4, 1868, pp. 69-92, pis. 19-21.

1869. Revision du Genre Thelphuse et description de quelques especes
nouvelles faisant partie de la collection du Museum. Nouv. Arch.
Mus. Hist. Nat., 5, pp. 161-191, pis. 8-11.

1872. Note sur les Crabes d'eau douce de Madagascar. Bibliotheque de
l'Ecole Hautes Etudes (Sec. Sci. Nat.), 5, art. 8, pp. 1-3.

chace: decapod crust: cea

231

1S86. La description de quelques Crustaces du genre Thelphusa re-
cueillis par M. de Brazza dans les regions du Congo. Bull. Soc.
Philom., Paris, (7), 10, pp. 148-151.

1887. Observations sur les Crabes des eaux douces de 1'Afrique. Ann.
Sci. Nat., Zool., (7), 4, pp. 121-149, pis. 7-9.

Milne Edwards, A. and Botjvier, E. L.

1893. Sur une espece nouvelle du genre Deckenia (Hilgendorf) re-
cueillie par M. Alluaud aux iles Seychelles. Ann. Sci. Nat., Zool.,
(7), 15, 1893, pp. 325-336, pi. 8.

Milne Edwards, H.

1837. Histoire Naturelle des Crustaces. Paris, 2, pp. 1-532.

1853. Observations sur les affinites zoologiques et la classification

naturelle des Crustaces. Memoire sur la Famille des Ocypodiens.

Ann. Sci. Nat., Zool., (3), 20, pp. 163-228, pis. 6-11.

Nobili, G.

1896. Di una nuova varieta della Thelphusa dubia B. Capello raccolta

dal Rev. Luigi Jalla a Kazungula. Boll. Mus. Zool. Anat. Comp.,

Torino, 11, no. 262, pp. 1-2.
1905. Descrizione di un nuovo Potamonide di Madagascar. Boll. Mus.

Zool. Anat. Comp., Torino, 20, no. 507, pp. 1-4, 1 text-fig.
1906a. Una nuova Telfusa di Madagascar. Boll. Mus. Zool. Anat.

Comp., Torino, 21, no. 532, pp. 1-4, text-figs. A-C.
1906b. Spedizione al Ruwenzori di S. A. R. Luigi Amedeo di Savoia.

Duca Degli Abruzzi. IX. Crostacei. [Nota preventiva]. Boll.

Mus. Zool. Anat. Comp., Torino, 21, no 544, pp. 1-2.
1909. Crostacei, in: II Ruwenzori, S. A. R. il Principe L. Amedeo di

Savoia. Parte Scientifica. I. Milano (Hoepli), pp. 357-358.

Olivier, G. A.

1804. Voyage dans l'Empire Ottoman, 4 and atlas.

Osorio, B.

1905. Uma nova Lista de Crustaceos Africanos. J. Sci. Math., Phys.,
Nat., (2), 7, pp. 149-150.

Parisi, B.

1923. Un nuovo Potamonide dell' Abissinia. Atti Soc. Ital. Sci. Nat.,

61, pp. 332-339, pi. 8.
1925. Su alcuni Potamonidi Africani. Atti Soc. Ital. Sci. Nat., 64,

pp. 97-99.

Pesta, O.

1937. Siisswassercrustaceen aus Deutsch-Ostafrika (Tanganyika-Ter-
ritorium). Zool. Anz., 117, pp. 157-160.

232

bulletin: museum of comparative zoology

Pfeffer, G.

1889. Uebersicht der von Herrn Dr. Franz Stuhlmann in Agypten, auf
Sansibar und dem gegenuberliegenden Festlande gesammelten
Reptilien, Amphibien, Fische, Mollusken und Krebse. Jahrb.
Hamburg Wiss. Anst., 6, hft. 2, pp. 1-36.

Rathbun, M. J.

1894. Descriptions of a new genus and two new species of African
Fresh-Water Crabs. Proc. U. S. Nat. Mus., 17, pp. 25-27.

1898. Descriptions of three new species of Fresh-Water Crabs of the
Genus Potamon. Proc. Biol. Soc, Washington, 12, pp. 27-30,
pis. 1-2.

1902. Description des nouvelles especes de Parathelphusa appartenant
au Museum de Paris. Bull. Mus. Hist. Nat., no. 3, pp. 184-187.

1904-06. Les Crabes d'Eau Douce (Potamonidae). Nouv. Arch. Mus.
Hist. Nat.: part 1 in vol. 6, 1904, pp. 225-312, pis. 9-18, text-
figs. 1-37; part 2 in vol. 7, 1905, pp. 159-321, pis. 13-22, text-figs.
38-105; part 3 in vol. 8, 1906, pp. 33-122, text-figs. 106-124.

1909. Collections recueillies par M. Maurice de Rothschild dans l'Af-
rique oriental anglaise. Crabes d'eau douce nouveaux. Bull.
Mus. Hist. Nat., 15, no. 3, pp. 101-105.

1921. The Brachyuran crabs collected by the American Museum Congo
expedition, 1909-1915. Bull. Amer. Mus. Nat. Hist., 43, pp.
379-474, pis. 15-64, text-figs. 1-22, map.

1922. Crustaces; Crabes d'eau douce, en: Voyage Baron M. de Roth
schild en Ethiopie et Afr. Or. Anglaise (1904-1905). Resultat
Sci. Anim. Artie. Part 1, Paris, pp. 35-39, pi. C3, text-figs. 1-2. -

1933. Reports on the Scientific Results of an Expedition to the South-
western Highlands of Tanganyika Territory. V. Crabs. Bull.
Mus. Comp. Zool., Harv., 75, pp. 251-260, pis. 1-7.

1935. Scientific Results of an Expedition to Rain Forest Regions in
Eastern Africa. II. Crustacea. Bull. Mus. Comp. Zool., Harv.,
79, pp. 23-28, pis. 1-2.

Roux, J.
1927

Note sur une collection de crustaces decapodes du Gabon. Bull.

Soc. Vaud. Sci. Nat., 56, pp. 237-244, text-figs. 1-7.
1935a. Voyage de Ch. Alluaud et P. A. Chappuis en Afrique Occidentale

francaise (Dec. 1930-Mars 1931). VII. Crustaces Decapodes

d'eau douce. Arch. Hydrobiol., 28, pp. 21-34.
1935b. Crustacea. II, Decapoda, in: Mission Sci. de l'Omo, 2 (Zool.)

fasc. 13. Mem. Mus. Hist. Nat. (nouv. ser.), 2, pp. 241-248,

text-figs. 1-3.

chace: decapod Crustacea 233

Savigny, J.-C.

1817. Description de l'Egypte, ou Recueil des observations et des re-
cherches qui ont ete faites en Egypte pendant l'expedition de
l'armee francaise. Hist. Nat., Paris, pi. 2, Crustaces.

Sendler, A.

1912. Zehnfusskrebse aus dem Wiesbadener Naturhistorischen Museum.
Jahrb. Nassau. Ver. Naturk., 65, 189-207, text-figs. 1-7.

Sjostedt, Y.

1910. Decapoda. In: Sjostedts Kilimandjaro-Meru Expedition, Stock-
holm, 21, 1, 1 p.

Stebbing, T. R. R.

1905. South African Crustacea. Part III. in: Mar. Invest. S. Afr., 4,

pp. 21-123, pis. 17-26.
1910. General Catalogue of South African Crustacea. Ann. S. Afr.

Mus., 6, 281-593.

Stimpson, W.

1907. Report on the Crustacea (Brachyura and Anomura) Collected
by the North Pacific Exploring Expedition, 1853-1856. Smithson.
Misc. Coll., 49, pp. 1-240, pis. 1-26.

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE
Vol. XCI, No. 4

4T f*

* DEC 23 !£«

SCIENTIFIC RESULTS OF A FOURTH EXPEDITION
TO FORESTED AREAS IN EAST & CENTRAL AFRICA

IV
REPTILES

By Arthur Loveridge

With Six Plates

The Library

f

Rhiseum of Comparat_ve

Harvard UniA

CAMBRIDGE, MASS., U.S.A.

PRINTED FOR THE MUSEUM

December, 1942

PUBLICATIONS
OF THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

The Bulletin and Memoirs are devoted to the publication of
investigations by the Staff of the Museum or of reports by spec-
ialists upon the Museum collections or explorations.

Of the Bulletin, Vols. I to XC, Vol. XCI, Nos. 1, 2 and
3, and Vol. XCII, No. 1 have appeared and of the Memoirs,
Vol. I to LVI.

These publications are issued in numbers at irregular intervals.
Each number of the Bulletin and of the Memoirs is sold separately.
A price list of the publications of the Museum will be sent upon
application to the Director of the Museum of Comparative
Zoology, Cambridge, Massachusetts.

After 1941 no more Memoirs are to be published.

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE

Vol. XCI, No. 4

SCIENTIFIC RESULTS OF A FOURTH EXPEDITION
TO FORESTED AREAS IN EAST & CENTRAL AFRICA

IV
REPTILES

By Arthur Loveridge

With Six Plates

The Libra-
r Museum of Comparat
"^s. Harvard Uni^

CAMBRIDGE, MASS., U.S.A.

PRINTED FOR THE MUSEUM

December, 1942

^ ^1 Con»*i;.>

DEC 23 1842 J

«• I B R A R *

■ *-' J <-*-■-

Museum of Comparat
Harvard Unix

No. 4. — Scientific Results of a Fourth
Expedition to Forested Areas in East and Central Africa

IV

Reptiles

By Arthur Loveridge

CONTENTS

Page

Introduction 237

Acknowledgments 239

Summary of Taxonomic Alterations 240

Index to List of Species collected 241

Systematic Discussion 246

Bibliography 372

INTRODUCTION

The collection on which the following report is based, .was made
by the author while investigating the fauna of certain forested regions
of East and Central Africa. The enquiry was carried out on behalf
of the Museum of Comparative Zoology with a fellowship granted
by the John Simon Guggenheim Memorial Foundation of New York.

A synopsis of the itinerary is given in the caption accompanying
Plate 1 — a map showing the position of the principal collecting locali-
ties. Altitudes and detailed information regarding the various camps
will be furnished in the final report of this series which will deal
with the general conclusions arrived at.

The period of collecting reptiles was from October 27, 1938, to
July 25, 1939, during which time 1,862 reptiles, representing 144
species, were secured. This total, which does not include seven addi-
tional species received as a gift, comprises 1 species of crocodile,
6 of tortoises, 74 (+ 7 donated) of snakes, and 61 of lizards, includ-
ing chameleons. In all 17 of these were new to the collection of the
Museum of Comparative Zoology exclusive of 7 other races not pre-
viously recognized, but now considered valid as a result of this study
of additional material.

238 bulletin: museum of comparative zoology

Seven forms are here described as new, some (in parentheses) are
based on material from earlier collections. The new forms are:

Typhlops tettensis rondoensis, Nchingidi, Rondo Plateau, T. T.
Hemidactylus tropidolepis barbouri, Changamwe, near Mombasa, K. C.

^Aniphisbaena rondoensis, Nchingidi, Rondo Plateau, T. T.

Melanoseps at,er rondoensis, Nchingidi, Rondo Plateau, T. T.

(Melanoseps ater matengoensis, Ugano, Matengo Highlands, T. T.)

(Melanoseps ater uzungwensis, Kigogo, Uzungwe Mountains, T. T.)

Chamaeleo dilepis idjwiensis, Idjwi Island, Lake Kivu, B. C.

In addition to these new forms, the undermentioned races or
species are recorded for the first time for certain countries.

New for Kenya Colony
Lycophidion capense ornatum Parker

New for Uganda

Dipsadoboa unicolor Giinther
Miodon gabonensis collaris (Peters)
Brookesia spectrum boidengeri (Steindachner)

New for Tanganyika Territory

Lycophidion capense ornatum Parker
Lygosoma tetradactylus hemptinnei (Witte)

New for Belgian Congo

Typhlops blanfordii lestradei Witte
Miodon gabonensis graueri Sternfeld

Several species, heretofore regarded as rarities, might be singled
out for special mention, among them: Cycloderma frenatum, Rham-
nophis a. elgonensis, Miodon g. graueri, Amphisbaena phylofiniens,
A. orientalis, A. ewerbecki, Algiroides vauercselli, A. africanus, Lygo-
soma g. graueri, L. meleagris, L. blochmanni, and Chamaeleo xenor-
hinus.

1 The description of this species recently (1941) appeared in connection with a revision of the
Amphisbaenidae (Bull. Mus. Comp. Zool., 87, p. 394, fig. 23).

loveridge: African reptiles 239

ACKNOWLEDGMENTS

The opportunity is taken of thanking Dr. Thomas Barbour, Direc-
tor of the Museum of Comparative Zoology, for his ever ready
encouragement in furthering the prosecution of this work, and to
the John Simon Guggenheim Memorial Foundation without whose
generous aid this expedition would not have been possible.

Lt. Col. C. R. S. Pitman of the Game Department, not only wel-
comed us to Uganda and smoothed our path by many helpful deeds,
but donated sundry snakes. His name follows in parenthesis after
the locality of such of them as are included in this report.

In appreciation of the action of His Excellency the Governor of
the Congo Beige in granting permission to collect on Idjwi Island,
a selection of duplicates of such species as were collected in Belgian
territory are being set aside for dispatch to the Congo Museum,
Tervueren, after the German evacuation of Belgium.

Messrs C. M. Bogert (American Museum) and V. FitzSimons
(Transvaal Museum) have patiently answered all manner of ques-
tions which entailed lengthy examination of much material in the
herpetological collections of which they are in charge.

A number of my colleagues have aided by identification of para-
sites or prey in their particular field. Among those to whom I am
indebted are: Dr. J. C. Bequaert (ticks), Dr. Fenner A. Chace Jr.
(crabs), Dr. J. P. Chapin (nestlings), Dr. P. J. Darlington Jr. and
Floyd G. Werner (insects in certain stomach contents), Dr. H. R.
Hill (linguatulids), and Drs. B. Schwarz and J. T. Lucker (nema-
todes and trematodes), of the United States Department of Agri-
culture.

The photographs illustrating this report were taken by my son,
Brian A. Loveridge, and for permission to use the blocks of plates
2-6 we are indebted to the Editor of the Scientific Monthly, in which
journal (June and July, 1940) they appeared as illustrations to a
popular account of the safari.

240

BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

SUMMARY OF TAXONOMIC ALTERATIONS

The undermentioned change in generic status is made:
Lacerta vauereselli Tornier becomes Algiroides vauereselli (Tornier)

The following forms are accorded subspecific rank:
Dipsas medici Bianconi revived as Dasypeltis scaber medici (Bianconi)
Dasypeltis fasciatus A. Smith revived as Dasypeltis scaber fasciatus A. Smith
Coluber palmarum Leach revived as Dasypeltis scaber palmarum (Leach)
Mabuia boulengeri Sternfeld revived as Mabuya maculilabris boulengeri Stern-

feld
Melanoseps ater longicauda Tornier revived

Typhlops obtusus Peters becomes Typhlops tettensis obtusus Peters
Calamelaps feae Boulenger becomes Calamelaps unicolor feae Boulenger
Calamelaps warreni Boulenger becomes Calamelaps unicolor warreni Boulenger
Aparallactus uluguruensis B. & L. becomes Aparallactus capensis uluguruensis

Barbour & Loveridge
Sepsina hemptinnei Witte becomes Scelotes tetradactylus hemptinnei (Witte)
Rhampholeon boulengeri Steindachner becomes Brookesia spectrum boulengeri

(Steindachner)

The undermentioned are considered to be synonyms:
Kinixys spekii Gray = Kinixys belliana belliana Gray

Thrasops j. mossambicus Mertens = Dispholidus typus (A. Smith)

Dasypeltis macrops Boulenger =D. scaber fasciatus A. Smith

Calamelaps mellandi Boulenger =C. unicolor warreni Boulenger

Calamelaps pellegrini Angel = Rhinocalamus ventrimaculatus Roux

Aparallactus ubangensis Boulenger = Aparallactus modestus (Giinther)

Aparallactus dolloi Werner
Aparallactus congicus Werner
Aparallactus batesii Boulenger
Aparallactus nigrocollaris Chabanaud
Aparallactus n. roucheti Chabanaud
Aparallactus graueri Werner
Atractaspis schoutedeni Witte
Atractaspis katangae Boulenger
Hemidactylus tasmani Hewitt
H. persimilis Barbour & Loveridge
Hemidactylus mandamus Loveridge
Algiroides boulengeri Peracca
Lygosoma gromieri Angel
?L. (Siaphos) compressicauda Witte
?Siaphos dewittei Loveridge nom. nov.
Lygosoma (Siaphos) burgeoni Witte

= Atractaspis irregularis (Reinhardt)

= Atractaspis bibronii (A. Smith)

= H. mabouia (Moreau de Jonnes)

=H. gardineri Boulenger

=H. gardineri Boulenger

= A. vauereselli (Tornier)

= Lygosoma kilimense Stejneger

= Lygosoma meleagris Boulenger

It might be added that the extensive synonymizing of Aparallactus
results from revisionary studies of that and allied genera now in M S.

loveridge: African reptiles 241

LIST OF SPECIES COLLECTED*

CROCODYLIDAE Page

Crocodylus niloticus Laurenti 246

TESTUDINIDAE

Kinixys belliana belliana Gray 247

Malacochersus tornieri (Siebenrock) 248

PELOMEDUSIDAE

Pelomedusa subrufa olivacca (Schweigger) 24S

Pehisios subniger (Lacepede) 249

Pclusios sinuatus (Smith) 250

AMYDIDAE

Cyelodcnna frenatum Peters 251

TYPHLOPIDAE

Typhlops schlegelii mucruso (Peters) 253

Typhlops blanfordii lestradci ^Yitte 254

Typhlops punctatus punctatus (Leach) 255

Typhlops tettensis rondoensis subsp. nov 250

Typhlops unitaeniatus unitaeniatus Peters 258

Typhlops graueri Sternfeld 258

Typhlops braminus (Daudin) 258

Typhlops lumbriciformis (Peters) 259

LEPTOTYPHLOPIDAE

Lcptotyphlops conjuncta conjuncta (Jan) 259

Lcpioty phlops emini emini (Boulenger) 259

Lcptotyphlops longicauda (Peters) 260

BOIDAE

Python sebac (Gmelin) 260

COLUBRIDAE (COLUBRINAE)

Neusterophis olivaceus olivaceus (Peters) 261

Neusterophis olivaceus uluguruensis (Loveridge) 261

Bothrophthalmus lineatus lineatus (Peters) 262

Boaedon lineatus lineatus Dumeril & Bibron 263

Boacdon olivaceus (Dumeril) 265

* The seven species marked by an asterisk (*) were not collected, but were presented to the
Expedition by Lt. Col. C. R. S. Pitman. The three species in brackets are discussed, though
not collected.

242 bulletin: museum of comparative zoology

Page

Lycophidon meleagris Boulenger 265

Lycophidion capense ornatum Parker 266

Lycophidion capense capense (Smith) 268

Lycophidion capense acntirostre Giinther 269

[Mchelya capensis capensis (Smith)] 269

[Pseudaspis cana (Linnaeus)] 270

Chlorophis carinatus Anderson 270

Chlorophis macrops (Boulenger) 270

Chlorophis hoplogaster (Giinther) 271

Chlorophis neglectus (Peters) 271

Chlorophis irregularis (Leach) 272

Chlorophis heterolepidotus (Giinther) 273

Philothamnus semivariegatus semivariegatus Smith 274

Gastropyxis smaragdina (Schlegel) 275

Hapsidophrys lineata Fischer 275

Rhamnoplus acthiopissa elgonensis Loveridge 276

Thrasops jacksonii jacksonii Giinther 277

Meizodon semiornata (Peters) 278

*Meizodon coronata (Sehlegel) 278

*Grayia smythii (Leach) 279

Duberria tutrix abyssinica (Boulenger) 279

[Duberria tutrix shirana (Boulenger)] 281

Prosymna ambigiia stuhlmanni (Pfeffer ) 281

DASYPELTIXAE

Dasypeltis scaber medici (Bianconi) 282

Dasypeltis scaber fasciatus Smith 284

Dasypeltis scaber palmarum (Leach) 285

*Dasypeltis scaber scaber (Linnaeus) 286

BOIGINAE

Geodipsas vauerocegae Tornier 286

*Boiga blandingii (Hallowell) 287

Boiga pulverulenta (Fischer) 287

Dipsadoboa unicolor Giinther 287

Crotaphopeltis hotamboeia hotamboeia (Laurenti) 288

Trimerorhinus tritaeniatus multisquamis Loveridge 289

Rhamphiophis rubropunctatus (Fischer) 290

Rhaviphiophis oxyrhynchus rostratm Peters 290

*Dromophis lineatus (Dumeril & Bibron) 291

Psammophis sibilans sibilans (Linnaeus) 291

loveridge: African reptiles 243

Page

Psammophis subtaeniatus sudanensis Werner 292

Thclotomis kirilandii kirtlandii (Hallowell) 292

Thclotornis kirtlandii capensis Smith 294

Dispholidus typus (Smith) 294

Calamelaps unicolor warreni Boulenger 295

Miodon and its forms 296

Miodon gabonensis collaris (Peters) 298

Miodon gabonensis graueri Sternfeld 298

Aparallactus modestus (Gunther) 299

Aparallactus jacksonii (Gunther) 300

Aparallactus werneri Boulenger 301

Aparallactus capensis uluguruensis Barbour & Loveridge .... 301

Aparallactus capensis capensis Smith 301

ELAPIXAE

Elapsoidea giintherii Bocage 302

*Xaja haje haje (Linnaeus) 303

Xaja nigricollis nigricollis Reinhardt 303

Naja melanoleuca Hallowell 304

Pseudohaje goldii (Boulenger) 306

Dendroaspis angusticeps (Smith) 306

Dcndroaspis jamesoni kaimosae (Loveridge) 307

VIPERIDAE

*Causus rhombeatus (Lichtenstein) 308

Causus resimus (Peters) 308

Causus defilippii (Jan) 309

Causus lichtensteinii (Jan) 309

Bitis arietans (Merrem) 310

Bitis gabonica (Dumeril & Bibron) 310

Bitis nasicornis (Shaw) 312

Atheris squamigera squamigera (Hallowell) 312

Atheris nitschei nitschei Tornier 313

Atractaspis irregularis (Reinhardt) 315

Atractaspis bibronii Smith 316

GEKKONIDAE

Cnemaspis quattuorscriatus (Sternfeld) 317

Cnemaspis africanus elgonensis Loveridge 318

Cnemaspis africanus africanus ^Verner) 319

Hemidactylus tropidolepis barbouri subsp. nov 320

244 bulletin: museum of comparative zoology

Page

Hemidactylus mabouia (Jonnes) 322

Hemidadylus gardineri Boulenger 323

Lygodactylus grotei grotei Sternfeld 324

Lygodactylus picturatus gutturalis (Bocage) 325

Ly god act y his picturatus mombasicus Loveridge 326

Lygodactylus picturatus picturatus (Peters) 326

AGAMIDAE

Agama mossambica mossambica Peters 326

Agama mossambica montana Barbour & Loveridge 328

Agama atricollis Smith 328

ZONURIDAE

Zonurus tropidostcrnum Cope 330

Chamaesaura tenuior Giinther 330

VARAXIDAE

Varanus occllatus Riippell 330

Varan/us niloticus (Linnaeus) 332

AMPHISBAENIDAE

Amphisbacna phylofiniens Tornier 332

Amphisbaena orientalis (Sternfeld) 332

Amphisbaena ewerbecki (Werner) 332

Amphisbaena rondoensis Loveridge 333

LACERTIDAE

Nucras boulengeri kilosac Loveridge 333

Laccrta versus Algiroides 333

Laccrta jacksoni Boulenger 334

Algiroides vaucreselli (Tornier) 335

Algiroides africanus Boulenger 336

Ichnotropis squamulosa Peters 337

Eremias spekii spekii Giinther 337

Holaspis gucnthcri Gray 338

GERRHOSAURIDAE

Gerrhosaurus nigrolineatus nigrolincatus Hallowell 340

SCINCIDAE

Mabuya maculilabris maculilabris (Gray) 342

Mabuya maculilabris comorensis (Peters) 344

Mabuya maculilabris boulengeri Sternfeld 344

loveridge: African reptiles 245

Page

Mabuya planifrons (Peters) 345

Mabuya megalura (Peters) 346

Mabuya varia varia (Peters) 347

Mabuya striata (Peters) 347

Riopa fernandi (Burton) 348

Riopa sundevallii (Smith) 349

Riopa pembanum (Boettger) 349

Lygosoma kilimense Stejneger 350

Lygosoma graucri graueri Sternfeld 351

Lygosoma mcleagris Boulenger 352

Lygosoma blochmanni Tornier 355

Ablepharus boutonii africanus Sternfeld 356

Ablepharus wahlbergii (Smith) 356

Scelotcs tetradactylus tetradactylus (Peters) 357

Scelotcs tetradactylus hcmptinnei (Witte) 358

Melanoseps ■ — Key to the Races of ater 359

Melanoseps ater rondoensis subsp. nov 360

Melanoseps ater matengoensis subsp. nov 361

Melanoseps ater uzungwensis subsp. nov 361

CHAMAELEONTIDAE

Chamaeleo scnegalensis senegalensis Daudin 362

Chamaeleo dilepis quilensis Bocage 363

Chamaeleo dilepis dilepis Leach 363

Chamaeleo dilepis idjwiensis subsp. nov 363

Chamaeleo bitaeniatus bitaeniatus Fischer 364

Chamaeleo bitaeniatus ellioti Giinther 365

Chamaeleo bitaeniatus ? bergeri Sternfeld 366

Chamaeleo fischeri matschiei Werner 367

Chamaeleo xenorhinus Boulenger 368

Chamaeleo melleri (Gray) 369

Chamaeleo johnstom ' Johnston i Boulenger 369

Brookesia spectrum boulengeri (Steindachner) 370

Brookesia brevicaudata (Matschie) 371

246 bulletin: museum of comparative zoology

CROCODYLIDAE

Crocodtlus xiloticus Laurenti

Crocodylus niloiicus Laurenti (part), 1768, Syn. Rept., p. 53: "India orientali,
et Aegypto."

Juv. ale. (M. C. Z. 48000) Mbanja, nr. Lindi, T. T. 3.V.39.

Distribution. Crocodiles were plentiful at Kitaya on the Ruvuma
River and frequently seen basking on the banks of the Mkulumusi
River, opposite the Siga Caves, near Tanga.

Native names. Ngwena (Kiyao, and Kimakonde at Mbanja);
mbidu (Kimakonde at Mikindani).

Measurements. A stocky 9 from Kitaya measured 8 feet from
snout to anus, but possessed only the stump of a tail 3 feet in length,
the injury had healed long ago. The estimated weight would be be-
tween 250 and 300 pounds.

Breeding. Some of her ova, on April 1, were enlarged to an inch
in diameter, but still spherical.

Diet. Her stomach was empty except for a couple of pounds
weight of pebbles and gravel. That of the Mbanja crocodile was
full of marine crabs (Sesarma meinerti) from the nearby estuary.

Defence. An Mbanja man, going after dark to bathe in a small
pool near his house, was greeted by a loud hissing noise. Returning
home for a pole, he killed the young crocodile which had sought to
scare him.

Migration. A Kitaya man, at daybreak was passing along a native
footpath through a maize plantation when he heard a grunting noise.
Thinking of pigs, he crept to the spot and peering through the stems
of maize, saw an eleven-foot crocodile. After obtaining a heavy pole,
he smashed in the skull so effectively that he rendered the reptile
useless as a specimen.

Folklore. On learning that I had no use for the carcass, one of
my Baganda skinners asked to be allowed to excise the musk glands
and tip of the injured tail. These, said he, if hung on a game net,
would insure that wild pigs and blue duiker became ensnared during
a drive, a form of hunting on which he was something of an authority.

At Mbanja, the chief and his brother requested me most earnestly
to see that the viscera of the young crocodile was buried, and on no
account to allow anyone to have it, "for", said they, "the bile is
intensely poisonous and some years ago, when put into a well here,
caused the death of several people."

loveridge: African reptiles 247

TESTUDINIDAE

KlNIXYS BELLI AN A BELLIAXA Gray]

Kinixys Belliana Gray, 1831, Syn. Rept., p. 69: No locality.
Kinixys Spekii Gray, 1862, Ann. Mag. Nat. Hist, (3), 12, p. 381: Central
Africa.

1 cf 3 9 (M. C. Z. 48001-4) Ujiji, T. T. ll.iii.39.

d" (M. C. Z. 48005) Kitaya, T. T. 27.iii.39.
2^2 9 (M. C. Z. 48006-9) Mikindani, T. T. 21-24.iv.39.

9 (M. C. Z. 48010) Amboni Estate, T. T. 19.vi.39.

Distribution. Seen also at Xchingidi, Rondo Plateau; and near
Lake Rutamba.

Native names. Xgongo (Kiyao) ; nambi (Kimakonde and Kimawiha).

Variation. After further careful study of all the Tanganyika and
Kenya material of this genus in the Museum, I am forced to the con-
clusion that they represent but one highly variable species possessing
no stable characters on which races can be separated. Apparently
the high vaulted type, which I previously thought was characteristic
of the savanna, is the result of vertical growth continuing after longi-
tudinal growth has slowed down. Females are somewhat broader
than males but the whole Mikindani series are distinctly broader
than any other East African specimens.

Height of carapace is included in its length from 2.0 to 2.5 times,
though the total lengths are from 127 to 191 mm., i.e. semiadult or
adult; width of nuchal is included in the width of its adjacent mar-
ginal from 3 (M. C. Z 48004) to 18 (M. C. Z. 48008-9) times in the
whole series, or from to 4 to 18 times in the Mikindani series alone!
Vertebrals 5, except for M. C. Z. 48001 which has 6; marginals 22,
except for M. C. Z. 4S001 which has 24.

Sexual dimorphism. Boulenger (1889a, p. 144) thought that the
length of the thickened anterior lip of the plastron equaled about a
quarter of the plastral length in males, a fifth in females. There is
an average difference but the area of overlap is too great to make
the character of value, thus, length of gulars into plastral length in
9 East African males ranges from 6 to 73^2 times, average 6.5, in 12
females it ranges from 7 to 8^ times, average 8.4, but in a cf and 9
(M. C. Z. 48008-9) from the same locality it may be the same, i.e.
approximately 7 times.

Measurements. Largest cf(M. C. Z. 48005) measures 183 mm. in
length of carapace, 78 mm. in height; largest 9 (M. C. Z. 48010)
measures 191 mm. in length of carapace, 92 mm. in height.

248 bulletin: museum of comparative zoology

Parasite. Ticks (Amblyomma nuttalli) were present on individuals
from every locality.

Malacochersus torxieri (Siebenrock)

Testudo tornieri Siebenrock, 1903, Anz. Akad. Wiss. Wien, 40, p. 185, pi. — :
Busisi, s. end of Lake Victoria, Tanganyika Territory.

1 (M. C. Z. 48001) Between Kiponda & Mitungu, T. T. 8.V.39.

Native name. Kobe (Kimwera).

Habitat. Though only fifty miles south of Lindi, from which the
species has been recorded, the finding of this young, 51 mm., tortoise
on the Rondo Plateau was something of a surprise for no rocks were
visible in the vicinity. As it was lying dead beside a much-used
path, the possibility of its having been transported by native
agency should not be overlooked.

PELOMEDUSIDAE

Pelomedusa subrufa olivacea (Schweigger)

Emys olivacea Schweigger, 1814, Prodromi Mon. Chelon., p. 38: "In Fabulosis

Nigritiae" Adanson coll. = Senegal.
Pentonyx Gehafie Ruppell, 1835, Neue Wirbelth. Fauna Abyss., Amph., p. 2,

pi. i: Massaua, Eritrea.
Pelomedusa Gasconi Rochebrune, 1884, Faune Senegambie, Rept., p. 25, pi. i,

figs. 1-2: Dagana, Senegal (restricted).

1 (M. C. Z. 48012) Mabira Forest, U. 14.xi.38.

History. Heretofore all East African marsh terrapin have been
referred by me to P. galeata, a name which Mertens (1937, Zool.
Anz., 117, p. 139) has shown must give place to subrufa Lacepede,
1788, whose type locality of "Indien" he restricts to the Cape of
Good Hope. It is this typical form which occurs throughout Tan-
ganyika and over the greater part of Kenya and Uganda. It is char-
acterized by the pectoral shields forming a suture on the median
line of the carapace.

From West Africa we have the name olivacea Schweigger, 1814,
and though his description is scanty, the type is still in the Paris
Museum. In 1884 another Senegal terrapin was named gasconi
by Rochebrune, whose plate shows the characteristic separation
of pectorals reported by Parker (1936e, 609) from northern
Gold Coast, Nigeria, etc. In fact this form appears to stretch in a

loveridge: African reptiles 249

belt from Senegal right across to Eritrea from where Riippell de-
scribed gehafie, of which the Museum of Comparative Zoology pos-
sesses a eotype.

Intermediates between the two races occur in a wide area of the
Anglo-Egyptian Sudan, Ethiopia, British and Italian Somaliland,
Kenya (Kaliokwell River, Lake Rudolf, reported by Parker) and
Uganda, and it is solely with a view to getting authors to assist in
defining this area that I employ the name olivacea for this Mabira
terrapin in which the pectorals are separated. One (M. C. Z. 40052)
of the three terrapin from Kirui's, Mt. Elgon, referred by me to
subrufa (as galeata) in 1936, also has separated pectorals. The con-
dition is rare in East Africa, however, for these are the only instances
out of thirty-eight specimens examined.

It is true that the condition of separated pectorals occurs spora-
dically further south, having been reported from South West Africa
(Werner), Angola (Bocage), Transvaal (M. C. Z. 41942) and Mada-
gascar (Mertens), but these do not invalidate the recognition of a
northern race for they form such a minute percentage of the pre-
dominating typical form in South Africa, or race icettsteini in Mada-
gascar.

Native name. Njaba (Luganda).

Coloration. Plastron wholly black like those from Kirui's, in sharp
contrast with the yellow plastrons from the dry savanna areas.

Pelusios subniger (Lacepede)

Testudo subnigra Lacepede, 1789, Hist. Nat. Quadrup. ovip. Serpens, 2, Synop-
sis methodica: (Based on La Noiratre of Lacepede).

9 (M. C. Z. 48013) Butiaba, U. 29.xi.38.

Synonymy. Sternfeld (1912c, pp. 200-201) records two terrapin
from Lake Albert. One from Butiaba (misspelt Rutiala), he refers to
Pelomedusa galeata, the other from Kassanje on the Congo shore he
identified as Sternothaerm sinuatus. Nieden (1913c, Mitt. Zool. Mus.
Berlin, 7, p. 61) reexamined both specimens and found them to be
nigricans, a name which is antedated by subniger Lacepede, though
not by subniger Daudin.

Tornier (1896, p. 4) recorded sinuatus from Sesse Islands (misspelt
Ussi), as did Boulenger (1909b, Ann. Mus. Civ. Stor. Nat. Genova
(3) 4, p. 302). The latter author (1911c, he. cit. (3), 5, p. 162) also
recorded derbianus from Bussu, near Jinja, the only LTganda records
for either species. Having doubts regarding them, I took the oppor-

250 bulletin: museum of comparative zoology

tunity of examining this material — through the courtesy of Dr.
Oscar de Beaux — when passing through Genoa. Both specimens
are referable to subniger, for after careful investigation I have come
to the conclusion that derbianus cannot be recognized as even sub-
specifically distinct.

I (1923g, p. 930) made a similar mistake, repeating (1928d, p. 51)
it when I referred a 360 mm. skeletal carapace from the Ruaha River
to nigricans (i.e. subniger) for the characteristic youthful characters
of a posteriorly serrated carapace and protruberances on the verte-
ral shields are blurred in this very old specimen. All these errors
arose from the literal use of the keys supplied by Boulenger (1889a,
p. 192) whose material was inadequate. We have then in East Africa
only two members of the genus, both characterised by the suture
between the abdominal shields being less, often considerably less,
than the length of the anterior lobe of the plastron. A key for sepa-
rating all the members of the genus will be found in my (1941d, p. 482)
Revision of the Pelomedusidae.

Pelusios sinuatus (Smith)

Sternotherus sinuatus A. Smith, 1838, 111. Zool. S. Africa, Rept., pi. i: In rivers
to the north of 25° S., South Africa.

6 (M. C. Z. 48014-9) Ujiji, T. T. 10.iii.39.
6 (M. C. Z. 48020-5) Mbanja, T. T. 27.iv.39.

Native name. Xgongo (Kimakonde).

Variation. The present material, coming as it does from points
750 miles apart on the western and eastern shores of Tanganyika
Territory, is of interest as both series exhibit the outer edge of the
pectoral usually longer than, sometimes equal to, occasionally shorter
than, the outer border of the humeral; anterior lobe of the plastron
longer than the abdominal suture in all because they are young,
with carapace lengths of from 51 to 173 mm. in length. The height
is included in this length from 2.39 to 2.68 times.

Coloration. All present the characteristic angular black pattern
on the periphery of the yellow (brick red in four youngest) plastron.

Breeding. On March 10 and April 27 I was brought two hatch-
lings from each locality, all had a carapace length of 51 mm., but
the two from Ujiji had a height of 21 mm., those from Mbanja of
19 mm.

Defence? The largest (173 mm.) of these terrapin, on being picked
up and turned over, ejected a fine jet of fluid from its right axilla

loveridge: African reptiles 251

or shoulder, I could not see clearly which, to a distance of one foot,
a second jet followed from the region of insertion of the left fore leg,
then a third from the right hind leg.

Habitat. The day following our arrival at Ujiji I visited the old
cement cistern, an abandoned sugar refinery vat, from which I had
removed two young sinuatus and two large Rana occipitalis on May
28, 1930. After waiting quietly for a few minutes I saw the stagnant
surface broken by the snout of a terrapin which was as quickly with-
drawn. I set the two skinners to work to bail out the 150 gallons of
water in the bottom of the vat. It took them three hours to drain
it and we found only three terrapin the largest of the series. From
this it might be deduced that one terrapin falls into the vat every
three years, escape being impossible: fish being abundant in Ujiji
makes it improbable that a hungry native would take the trouble
to capture an odiferous terrapin. It was interesting to note that
there were no frogs present; though many frogs and toads of
different species were captured in the three adjacent vats which,
being cracked, held only a few inches of water at most. It seems
reasonable to assume, therefore, that those which had fallen into
the vat containing terrapin, had been eaten by the latter.

AMYDIDAE

Cycloderma frenatum Peters

Cycloderma frenatum Peters, 1854, Monatsb. Akad. Wiss. Berlin, p. 216:
Zambezi, Mozambique.

8 adult (M. C. Z. 48026-33) Kitaya, T. T. 24-31.iii.39.
36 eggs (M. C. Z. 48034) Kitaya, T. T. 27-28.iii.39.

Native names. Litetamera (Kiyao) ; nahi (Kimakonde).

Variation. The four alcoholics present a very different appearance
from the four dried carapaces. The longitudinal dermal ridges, so
conspicuous on the carapace of a young alcoholic cotype (AL C. Z.
21901), are absent from the smooth carapaces of the alcoholic adults,
though perhaps indicated laterally. The entoplastral bony callosity
exhibits surprising variability not only in size but in shape; the
hyoplastral and hypoplastral callosities, said to be widely separated
by Boulenger (1889a, p. 265) are rarely so in this series, normally
they are closely in contact.

Coloration. In life. Above, very dark olive; except for one lateral
line, the dark longitudinal lines on head and neck are not conspicuous

252 bulletin: museum of comparative zoology

as figured by Peters (1882a, Reise nach Mossambique, 3, pi. i).
Below, fleshy pink and china white much variegated with dusky
marblings (still the case in alcohol). Perhaps this is the 9 coloring
(vide Peters, 1882a, pi. iiia) while his pi. i may be based on c? color-
ing, it agrees well with the plastral appearance of our young Zam-
bezi cotype.

Measurements. Length of carapace (without dermal margins) of
largest gravid 9 (M. C. Z. 48030) 390 mm., width of carapace (with-
out margins) 310 mm. Weight 25 lbs.

Breeding. This 9 , taken April 27th, held 17 hard-shelled spherical
eggs, each measuring 32 mm. in diameter, and ready for laying: in
addition there were 55 smaller ova in various stages of development.
The following day a second 9 was found to hold 19 hard-shelled
eggs, each approximately 31 mm. in diameter, besides numerous
developing ova.

The fact that all four dissected specimens are females, suggests
that the whole series are of that sex and so had fallen an easy prey
to natives when they had ventured ashore to lay. I surprised one
about twenty feet from the edge of a great hippo-harbouring lagoon
for which it was making in leisurely fashion. On seeing me it speeded
up, I gave chase and we captured it in the water.

Diet. If an inference may be drawn from the fact that each of
these turtles had a powerful, all-pervading fishy odour, it seems
probable that they subsist largely on fish.

Parasites. Leaches were not uncommon on these mud turtles.

Enemies. Dr. J. O. Shircore, C.M.G., when Director of Medical
Services in Tanganyika Territory, related to me how, about 8 a.m.
one morning in 1926, he had encountered two small rufous otters.
They were running around an almost dessicated puddle about a
hundred yards from the right bank of the Kilombero River in Ma-
henge district; then he observed that they were attempting to eat
one of these mud turtles which had withdrawn within its defences,
at least this was the position when he walked up after shooting one
of the otters. Subsequently he sent me a photograph of otter and
turtle, the latter lying on its back, the former posed in much the
same position as when shot.

Temperament. Contrary to my expectations, these turtles appear
timid and inoffensive, withdrawing their heads and enclosing the
hind limbs beneath their protective dermal valves at the slightest
disturbance. When all is quiet, the head is protruded with the utmost
caution by very gradual stages, to be withdrawn precipitately, to the

loveridge: African reptiles 253

accompaniment of a species of snort, at the slightest sound or move-
ment in their vicinity.

Encountering a native on a path one day, I stopped and spoke
to him of my requirements in the way of small mammals, then passed
on without paying any particular attention to the burden on his
head which appeared to be a shallow cooking pot of sorts. He called
after me to know if I had no need of turtles, then I saw that it was
one of these, plastron uppermost, which he was carrying:
the usual ring of twisted grass, commonly employed by porters to
balance their loads, had been used to steady the strongly convex
carapace. The docile reptile had been transported several miles in
this manner.

Habitat. With head held well above the water, one of these turtles
paddled swiftly past me when I was wading thigh-deep in the leech
infested waters of a lake. I was stalking duck at the time and, lest
I shatter it, had to wait till the turtle was twenty feet away before
firing. However it dived and was not seen again.

TYPHLOPIDAE

Typhlops schlegelii mucruso (Peters)

Onychocephalus mucruso Peters (part), 1854, Monatsb. Akad. Wiss. Berlin,
p. 621 : Makanga, Mozambique.

3 (M. C. Z. 48007-9) Mikindani, T. T. 15.iv.39.
2 (M. C. Z. 48090-1) Mbanja, T. T. l.v.39.
2 (M. C. Z. 48092-3) Lake Rutamba, T. T. S.v.39.
1 (M. C. Z. 48094) Lindi town, T. T. l.vi.39.

Native names. Xanumira kuu-iri (Kimakonde); lilenga (Kimawiha.)

Variation. Midbody scale-rows 30-34; diameter included in length
24-40 times.

Coloration. The Mikindani series correspond to the gray lineolatus
type of punctatus; the rest resemble the congestus variety of the same
species, being orange with black blotches at Mbanja, blue gray with
black blotches at Rutamba and Lindi.

Measurements. Largest (M. C. Z. 48090) measures 515 (510 +5)
mm., diameter 21 mm.

Enemies. As I was passing through some native gardens near
Mbanja, I saw a woman with a ten-foot pole belabouring something
on the ground at the back of her hut ; her small daughter, awestruck,
looked on from a safe distance. The object of attack was the large
and helpless blind snake whose measurements are given above.

254 bulletin: museum of comparative zoology

Typhlops blanfordii lestradei Witte
Plate 2, fig. 2.

Typhlops Lestradei Witte, 1933, Revue Zool. Bot. Afr., 23, p. 206, figs. 1-3:
"Rubengera" i.e. Ruhengeri, Belgian Ruanda-Urundi.

Eggs and 8 (M. C. Z. 48070-8) Mushongero, U. l-3.ii.39.
3 (M. C. Z. 48079-81) Nyakabande, U. 5.ii.39.

3 (M. C. Z. 48082-3) Kisenyi, B. R. 10.ii.39.

4 (M. C. Z. 48084-6) Idjwi Id., B. C. 17.ii.39.

Native names. Kirumira habili (Lukiga); kichidachuzi (Lulega).

Synonymy. Witte distinguishes lestradei from blandfordii Boulenger
on four characters, all of which are now shown to be inconstant.
T. blanfordii of Ethiopia had 30 midbody scale-rows, but so does
lestradei occasionally; the eye in the former was distinct whereas in
lestradei it is usually hidden, though sometimes sharply distinct
(M. C. Z. 48071, etc.); the relative proportions of the head shields
exhibit too much variability to be considered, the prefrontal, though
usually larger than the supraoculars in lestradei, is sometimes sub-
equal. There remains then, only the character of the rostral, as seen
from below, being broader in the figure of blanfordii than in lestradei,
in which it narrows rather abruptly towards the buccal border.
Direct comparison, however, of a specimen of blanfordii from Harrar,
Ethiopia, with our lestradei material, reveals no difference. There is,
though, an average difference in midbody scale-rows and diameter,
so I prefer to regard lestradei of western Uganda and Ruanda, as a
race of blanfordii of Eritrea, Ethiopia (and possibly Kenya, fide
Sternfeld).

It seems possible that T. dubius Chabanaud (1916f, Bull. Mus.
Paris, 22, p. 364) may take precedence over lestradei. This snake had 30
midbody scale-rows but its diameter into length was said to be about
thirty times, both of which would put it in the synonymy of punc-
tatus. The snout is allegedly similar to that of blanfordii, but as
Chabanaud's holotype was young — only 147 mm. — the slightly
more angular snout of adult punctatus would not be emphasized.
Much turns on the type locality, said to be "Congo beige: volcans
du Kivori (altitude 1,500 metres)." This certainly appears to be the
Kivu volcanoes, but I am informed that the collector was very care-
less in labeling his material and that insects labeled from the vol-
canoes, never came from there. There is a Kivari in Uganda to the
east of Ruwenzori so perhaps the name is repeated somewhere in
the Congo.

loveridge: African reptiles 255

Chabanaud suggests that dubius may be a synonym of adolfi Stern-
feld (1910e. p. 70), described as from "Fort Blus" but later (1912c,
p. 263) corrected to Fort Beni. T. adolfi, however, seems certainly a
synonym of punctatus for it not only had 30 midbody scale-rows,
but a diameter which was included in its length twenty-five times.
We have typical punctatus from Irumu which is just west of Fort
Beni.

Measurements. Largest (M. C. Z. 4S071) measures 670 (662 + 8)
mm., diameter 15 mm., smallest (M. C. Z. 48081) measures 195
(192 + 3) mm., diameter 6 mm.

Coloration. A very glossy species. Above, gray, plumbeous, black
or rich coppery or bronzy brown, the basal half of each dorsal scale
lighter. Below, belly somewhat lighter (Mushongero), or median
region of belly irregularly white (Kisenyi and Idjwi). When about
to slough, sometimes uniformly opaquely white.

Breeding. At Mushongero, on February 3, 9 eggs, each measuring
about 27 x 14 mm., in a 9 .

Diet. One very fat Idjwi snake held termite nymphs, the stomach
contents of a second have been identified for me as larval and callow
ants by Dr. P. J. Darlington.

Enemies. Remains of one in stomach of a Miodon g. graueri on
Idjwi Id.

Habitat. I captured the Nyakabande series in the course of an
hour by turning over the larger blocks of lava lying scattered about
on the plain in the vicinity of the rest house. The snakes immediately
attempted to slip down holes among the interstices of the lava and
it was necessary to pry up the rocks to extricate the larger snakes,
so tenaciously did they hold on. A Kisenyi snake was hoed up by
a woman working in her garden. The fine Mushongero series were
brought me in a basket by a native who said that he had dug them
from termite hills, capturing them alive and uninjured.

Ttphlops punctatus punctatus (Leach)

Acontias punctatus Leach, 1819, in Bowdich, Miss. Ashantee, p. 493: Fantee,
Gold Coast.

2 (M. C. Z. 48059-60) Mabira Forest, U. 5 & 14.xi.38.
5 (M. C. Z. 48061-5) Magrotto Mtn., T. T. l-12.vii.39.

Distribution. It was curious that this common species was en-
countered only at the first and last camps of the entire trip, and

256 bulletin: museum of comparative zoology

that it is the typical form and not the race gierrai which occurs at
Magrotto.

Native names. Mugova (Luganda); mkonko (Kisambara).

Variation. Midbody scale-rows 26-28; diameter included in length
27-36 times; tail included in length 42-79 times; preocular in contact
with upper labials.

Coloration. All of the lineolatus type, adults bronzy brown or gray,
juveniles gray or black.

Measurements. Largest, a 9 (M. C. Z. 48061), measures 557 (550
4- 7) mm.

Breeding. On July 1, this largest snake held 19 small eggs meas-
uring 10 x 5 mm. On July 12, a 9 only 30 mm. smaller held 10 eggs
of varying sizes, the largest measuring 17x7 mm.

Aestivation. Both the largest snakes, 540 and 557 mm. respectively,
from Mabira and Magrotto, had extensive deposits of fat and were
killed while crossing roads after heavy rain.

Folklore? Apropos previous remarks {vide Barbour & Loveridge,
1928, p. 108), it was a curious coincidence that, within an hour of
my arrival at Magrotto, Mr. C. Clausen, Manager of the Estate,
should produce a bottle containing one of these snakes, saying "Per-
haps you can tell us about this creature. It was brought to me by a
native who found it wriggling unharmed in a procession of siafu
(driver ants) ; he split a bamboo with which he removed it from their
line and brought it to me saying that the creature was their queen."

Typhlops tettensis rondoensis subspec. now

Type. Museum of Comparative Zoology, No. 48,066, from Nching-
idi, 2,700 feet, Rondo Plateau, southeastern Tanganyika Territory,
collected by Arthur Loveridge, May 5, 1939.

Paratypes. Museum of Comparative Zoology, Nos. 48067-9 and
a fifth specimen in the British Museum, all with same data as the
type.

Diagnosis. Midbody scales in 24 (22-24 in obtusus) rows, snout
very prominent, rounded.

Preocular in contact with second, third and fourth labials; distance
between rostral and nostril about equal to distance between latter
and posterior edge of nasal ; midbody diameter 36 to 45 times in
total length.

loveridge: African reptiles 257

Belly bluish gray like back; rostral not extending back to an im-
aginary line connecting the anterior borders of the eyes (Tette,
Zambezi, Mozambique) t. tettensis

Belly pure white; rostral extending back to an imaginary line con-
necting the anterior borders of the eyes (Rondo Plateau, Tan-
ganyika Territory) t. rotidoensis

Preocular in contact with second and third labials only; distance
between rostral and nostril about one third of the distance be-
tween latter and posterior edge of nasal; midbody diameter 43
to 50 times in total length.

Belly pure white; eyes concealed, but rostral apparently not ex-
tending back to an imaginary line level with their probable posi-
tion (Shire Highlands, Nyasaland) t. obtusus

Description. Midbody scale-rows 24 in all; midbody diameters in-
cluded 36 to 45 times in total length; length of tails included 57 to 78
times in total length; agreeing with tettensis in all respects except for
those characters indicated in the diagnosis, and an azygous scale
split off from the second labial on one side of one specimen.

Coloration. In life. Above, blue-gray, thus closely resembling the
young of punctatus and schlegelii mneruso. Below, china white. In
alcohol while the bluish gray effect is retained, a close examination
reveals each dorsal scale as blackish at the tip, white at the base,
resulting in a lineolate appearance. Below, white. Thus they differ
from tettensis which were: "Im leben griinblau, in YYeingeist iiberall
graugriin."

Measurements. Total length of Type, 228 (224 + 4) mm.; mid-
body diameter 5 mm. The paratypes range from 145 to 165.5 mm.
in length, with diameters of from 4 to 5 mm. Probably none are
fullgrown.

Habitat. All taken under logs at the forest edge and within two
hundred yards of my camp, to be described in the final report.

Remarks. T. tettensis is known only from the types in the Berlin
Museum and therefore unavailable at the moment of writing. The
types of T. obtusus are in the British Museum and I am indebted to
Mr. H. W. Parker for making comparison between them and a para-
type of rondoensis, which he considers to be distinct from obtusus.
It was he who pointed out the difference in position of the nostril,
which is more anteriorly situated in the nasal plate in obtusus.

258 bulletin: museum of comparative zoology

Typhlops unitaeniatus unitaenlvtus Peters

Typhlops (Letheobia) unitaeniatus Peters, 1878, Monatsb. Akad. Wiss. Berlin,
p. 205, pi. ii, fig. 5: Teita, Kenya Colony.

1 (M. C. Z. 48058) Amboni, T. T. 14.vi.39.

Variation. Midbody scale-rows 24; diameter included in length 83
times.

Measurements. Total length 332 (307 + 25) mm., midbody diame-
ter 4 mm.

Diet. Termites in its stomach.

Enemies. Taken from the stomach of a lizard-buzzard (Kaupifalco
monogrammica) shot shortly after sunset: the snake could have been
swallowed only very recently.

Typhlops graueri Sternfeld

Typhlops graueri Sternfeld, 1912, Wiss. Ergeb. Deut.-Zentral-Afrika-Exped-
1907-1908, 4, p. 264: Virgin forest behind boundary mountains northwest
shore of Lake Tanganyika, Belgian Ruanda-Urundi.

7 (M. C. Z. 48051-7) Ujiji, T. T. ll-15.iv.39.

Variation. Midbody scale-rows 24; diameter included in length
67-75 times.

Measurements. Largest (M. C. Z. 48051) measures 366 (360 -f- 6)
mm., midbody diameter 5 mm.

Habitat. I took three beneath rotting debris piled about the base
of a mango tree in a shamba between township and shore; the rest in
soil beneath garden refuse in plantations of bananas, etc. in nearby
Ruanda, T.T., as in 1930 (vide Loveridge, 1933h, p. 212).

Typhlops braminus (Daudin)

Eryx braminus Daudin, 1803, Hist. Nat. Rept., 7, p. 279: Bengal, India.

2 (M. C. Z. 48049-50) Lindi, T. T. 31.V.39.

Distribution. These constitute the fourth and fifth records of the
occurence of braminus along the coast of Tanganyika. My colleague,
Mr. Shreve, recently found two (M. C. Z 33602-3) in a collection
from Chilpancingo, Mexico, so that it is incorrect to limit its distri-
bution to Asia and the islands of the Indian Ocean.

Variation. Midbody scale-rows 20; diameter included in length
48-55 times.

loveridge: African reptiles 259

Measurements. Larger (M. C. Z. 48049) measures 145 (141.5 4- 3.5)
mm., midbody diameter 3 mm.

Habitat. Found together beneath bundles of rotting grass stacked
in native town near the Beach Hotel. Diligent search failed to un-
cover any more.

Typhlops lumbriciformis (Peters)

Onychocephalus (Letheobia) lumbriciformis Peters, 1874, Monatsb. Akad. Wiss.
Berlin, p. 377: Zanzibar coast.

4 (M. C. Z. 48045-8) Amboni, T. T. 19.vi.39.

Variation. Midbody scale-rows 18; diameter included in length
61-72 times; tail included in length 55-65 times.

Measurements. Largest (M. C. Z. 48048) measures 360 (354 -4- 6)
mm., midbody diameter 5 mm.

Habitat. Ploughed up by tractor in a sisal plantation on Amboni
Estate, near Tanga.

LEPTOTYPHLOPIDAE

Leptotyphlops conjuncta conjtjncta (Jan)

Stenostama conjuncia Jan, 1861, Arch. Zool. Anat. Fisiol., 1, p. 189: South
Africa.

1 (M. C. Z. 48035) Kitaya, T. T. 5.iv.39.

Native name. Nanumira kuwiri (Kimakonde at Kitaya).

Variation. Midbody scale-rows 14; diameter included in length 45
times; tail included in total length 12 times.

Trinomials are used, following Bogert (1940, p. 13) who regards
distanti as a race.

Measurements. Total length 135 (123 + 12) mm.; midbody diam-
eter 3 mm.

Leptotyphlops emini emini (Boulenger)
Plate 2, fig. 1.

Glauconia emini Boulenger, 1890, Ann. Mag. Nat. Hist. (6), 6, p. 91: Karagwe,
Bukoba, Tanganyika Territory.

2 (M. C. Z. 48036-7) Bundibugyo, U. 20.xii.38.
5 (M. C. Z. 48038-42) Mbanja, T. T. 26.iv.39.

260 bulletin: museum of comparative zoology

Native names. Kikelere (Luamba); keechwa mugongo (Lutoro);
mbitu (Kimakonde at Mbanja where they did not know the name
given me at Kitaya for L. conjuncta, and apply mbitu to amphis-
baenids also).

Variation. Midbody scale-rows 14; diameter included in length
40-45 times; tail included in length total 9-13 times.

Measurements. Largest, a 9 (M. C. Z. 48037), measures 145 (131
+ 14) mm., midbody diameter 3.25 mm.

Habitat. Two of the Mbanja series were taken just below the sur-
face (rainy season) in uprooting rank grass and scraping over tent
site.

Leptotyphlops longicauda (Peters)

Stenostoma longicauda Peters, 1854, Monatsb. Akad. Wiss. Berlin, p. 621: Tete,
Mozambique.

2 (M. C. Z. 48043-4) Mbanja, T. T. 27.iv.39.

Native name. Mbitu (Kimakonde, but see remarks under L. emini).

Variation. Midbody scale-rows 14; diameter included in length
72-81 times; tail included in total length 9-11 times.

Measurements. Larger measures 102 (93 + 9) mm., midbody dia-
meter 1.25 mm.

BOIDAE

Python* sebae (Gmelin)

Coluber sebae Gmelin, 1788, Syst. Nat., 1, p. 1118: No type locality.

2 (M. C. Z. 48095-6) Ujiji, T. T. 10.iii.39.
1 (M. C. Z. 48097) Mikindani, T. T. 10.iv.39.

Distribution. The only other python seen was at Mubango, where
I refused to purchase a heavy eight-foot specimen that had been
dragged for four miles through the Mabira Forest, Uganda.

Native names. Timba (Luganda); mbira (Luamba); nzilamiri (Lu-
toro); satu (Kiyao); hato and ihatu (Kimakonde at Mikindani and
Kitaya); mbidi (Kimakonde at Mbanja).

Variation. Midbody scale-rows 83-91; ventrals 274-278; anal en-
tire; subeaudals 75-77.

Measurements. Largest (M. C. Z. 48095) measures 2500 (2170
+ 330) mm.

Diet. The largest python disgorged a fowl when struck.

Parasites. Three nematodes {0 phidasearis sp.) from its stomach
were preserved.

loyeridge: African reptiles 261

COLUBRIDAE

Xeusterophis olivaceus olivaceus (Peters)

Coronella olivacea Peters, 1854, Monatsb. Akad. Wiss. Berlin, p. 622: Tete,
Mozambique.

2 (M. C. Z. 48098) Mushongero, U. l.ii.39.

1 (M. C. Z. 48099) Idjwi Id., B. C. 3.iii.39.

1 (M. C. Z. 48100) Ujiji, T. T. ll.iii.39.

4 (M. C. Z. 48101-2) Magrotto Mtn., T. T. 3.vii.39.

Native names. Bulifu (Lukiga, but not distinguished from Du-
berria); kashaveri (Lulega).

Variation. Midbody scale-rows 19; ventrals 135-140; anal divided;
subcaudals 46-67; labials 8, the fourth and fifth entering the orbit;
preocular 1, rarely 2 (left side of M. C. Z. 48099 only); postoculars 3,
rarely 2 (right side of M. C. Z 48099 only); 5, rarely 4, infralabials
in contact with an anterior sublingual.

Coloration. One Mushongero snake had the lateral edges of the
ventrals impinged with magenta, the other with gray. The Idjwi
Island reptile was uniform olive above, the olive only slightly im-
pinging on the ventrals which were otherwise paler olive. The Ujiji
9 was distinctly reddish with a broad, dark vertebral band, the red
impinging on the lateral edges of the ventrals which were otherwise
white.

Measurements. Largest, a 9 (M. C. Z. 48162), measures 390
(2S0 + 110) mm. A larger cT, as well as two others of the series,
had lost the tip of its tail.

Habitat. One Mushongero snake was hoed up in grassland near
the lake shore. It is interesting to note that both races occur on
Magrotto, those listed above were brought to me by natives and
may well have come from the long-cultivated areas lower down the
mountain as those taken by myself at the forest edge had all the 17
midbody scale-rows of the montane forest race.

Xeusterophis olivaceus uluguruensis Loveridge

Natrix olivacea uluguruensis Loveridge, 1935, Bull. Mus. Comp. Zool., 79.
p. 7: Xyange, Uluguru Mountains, Tanganyika Territory.

6 (M. C. Z. 48103-5) Magrotto Mtn., T. T. 3-9.vii.39.

9 (Vienna Museum) Ugano, Matengo Highlands, T. T.

Distribution. I have included some material submitted by the
Vienna Museum as it constitutes a southward extension of the range.

262 bulletin: museum of comparative zoology

The race reaches Southern Rhodesia, however, for FitzSimons (1939b,
Ann. Transvaal Mus. 20, p. 20), under the name olivaceus, records
three snakes from the Chirinda Forest, Mt. Selinda, as having 17
midbody scale-rows.

Native name. Nyoka usambia (Kisambara, also for Aparallactus
werneri) .

Variation. Midbody scale-rows 17, except one (Vienna Mus.) with
16 and one with 15 (M. C. Z. 44093), both from Ugano. Heretofore
the only example with 15 rows known to me, was from Pemba Island;
perhaps the island race may not be recognisable, a point which can
be definitely settled by someone on Pemba securing a series of these
snakes. Ventrals 130-140; anal divided; subcaudals 63-83; labials 8,
the fourth and fifth entering the orbit; preoculars 1 or 2 (latter in
four of the Ugano series) ; postoculars 3, rarely 2 (M. C. Z. 48164
only) ; temporals 1 + 2, rarely 1 + 1 and 2 + 2 (Vienna specimens) ;
5, rarely 4, infralabials in contact with an anterior sublingual.

Coloration. A Magrotto c? was noted as having creamy-white
ventrals, those of two 9 9 were as bright yellow as those of
Aparallactus werneri.

Measurements. Largest cf (M. C. Z. 48163) measures 373 (265
+ 108) mm., largest 9 (M. C. Z. 48165) 394 (275 + 119) mm.
This refers to Magrotto series only, largest from Ugano (unsexed)
measures 411 (287 + 124) mm.

Diet. A frog (Arthroleptis xenodactylus) in one.

Parasites. A nematode.

PjOTHROPHTHALMUS LIXEATUS LIXEATUS (Peters)

Elaphis (Bothrophthalmus) lineatus Peters, 1863, Monatsb. Akad. Wiss.
Berlin, p. 287: Guinea.

c? (M. C. Z. 48106) Mabira Forest, U. 14.xi.38.

Distribution. The typical form ranges from French Guinea east-
wards to the Mabira Forest, Kyagwe, where one has already been
taken by Pitman (1936, p. 227). In Fernando Po and on the opposite
mainland of southwestern Cameroon, the uniformly coloured race
brunneus Giinther (with infuscatus Buchholz & Peters, modestus Fis-
cher, and olivaceus Miiller, as possible synonyms) is given off.

Variation. Midbody scale-rows 23; ventrals 192; anal entire; sub-
caudals 72; labials 7, the fourth and fifth entering the orbit.

Coloration, iibove, head white with fine black lines on top and
side, the usual five vertebral and lateral coral red stripes separated

loveridge: African reptiles 263

by the black ground colour. Below, throat white, rest of under-
surface coral red.

Measurements, cf measures 670 (550 + 120) mm.

Boaedox lineatus lineatus Dumeril & Bibron
Boaedon Lineatum Dumeril & Bibron, 1854, Erpet. Gen., 7, p. 363: Gold Coast.

5 (M. C. Z. 48107-9) Mabira Forest, 17. 14-18.xi.38.

1 (M. C. Z. 48110) Fort Portal. U. 19.xii.38.

1 (M. C. Z. 48111) Bugoye, U. 27.xii.38.
12 (M. C. Z. 48112-9) Nyakabande, U. 28-30.L39.
Eggs & 4 (M. C. Z. 48120-4) Mushongero, U. l-2.ii.39.

9 (M. C. Z. 48125-9) Kisenyi, B. R.-tl. 10-13.ii.39.
10 (M. C. Z. 48130-4) Idjwi Id., B. C. 23-28.ii.39.
10 (M. C. Z. 48135-9) Ujiji, T. T. ll-15.iii.39.

5 (M. C. Z. 48140-2) Kitaya, T. T. 25-31.iii.39.

14 (M. C. Z. 48143-53) Mikindani. T. T. 10-20.iv.39.

4 (M. C. Z. 48154-6) Mbanja, T. T. 27-30.iv.39.

15 (M. C. Z. 48157-69) Nchingidi, T. T. 10-19.V.39.
1 (M. C. Z. 48170) Siga Caves, T. T. 15.vi.39.

5 (M. C. Z. 48171-3) Amboni Estate, T. T. 19.vi.39.

14 (M. C. Z. 48174-84) Magrotto Mtn., T. T. l-19.vii.39.

Native names. Kifirfa (Luganda); nyamutane (Lukonjo); nama-
jina (Lukiga); namaragwe (Kimwera); nyika (Kisambara) ; while the
following Kimakonde names were given me: namalutu (at Kitaya),
naliohi (at Mikindani), mambala and gangganguru (at Mbanja) which
suggest that confusion with some other species exists.

Variation. Midbody scale-rows 27-33, there is an interesting tend-
ency to have only 27-29 on the East Coast where only four out of
fifty-eight snakes have 30 or 31, none has 33; on the other hand, in
the Central Lake region only one out of fifty-two snakes has 33 and
none has 27, all the rest having 29-311; ventrals 196-21S, reflect the
same tendency being 187-218 at coast and 196-235 in Lake region;
anal entire; subcaudals 46-70, the coast alone embracing whole
range; upper labials normally S, the fourth and fifth entering the
orbit, rarely 7, the fourth entering (1 ex.), or 8, the third, fourth
and fifth (7 ex.), or 9, the fourth and fifth (2 ex.), or 9, the fourth,
fifth and sixth (2 ex. one side only), or 9, the fifth and sixth (2 ex.
one side only); 3 or 4, very rarely 1 or 2, lower labials in contact
with the anterior sublinguals, in one snake the second sublabial on
either side is excluded from contact with the sublinguals; loreal

1 Cf. Loveridge, 1936J, p. 238, for similar comparison to the north.

264 bulletin: museum of comparative zoology

present except in M. C. Z. 48109 where they are fused with the pre-
frontals; preoculars 1 (65 ex.) or 2 (39 ex.), the rest having an azy-
gous combination, one snake (M. C. Z. 48181) lacking a preocular
on the right side owing to fusion with the prefrontal; postoculars 2;
temporals 1 + 2, rarely 1 + 1 (5 ex. at least on one side) or 1 + 3
(9 ex. at least on one side).

Coloration. In Mabira Forest an adult was olivaceous, a young
one blackish. An adult 9 from Ujiji decidedly greenish, a second
specimen black. Another 9 from Mikindani was pale brown above,
a red bar from nostril through orbit becoming less distinct posteri-
orly, about five, irregular, longitudinal rows of red blotches on the
dorsum also become less distinct posteriorly. Below, pure white.

Measurements. Largest cf (M. C. Z. 4S157) measures 650 (540
+ 110) mm.; largest 9 (M. C. Z. 48135) measures 951 (835 + 116)
mm.

Sexual dimorphism. The <?c? have a lower ventral and higher
subcaudal count, viz. 35 cT c? range from 191-212 ventrals and 57-70
subcaudals, while 19 9 9 range from 205-239 ventrals and 46-53
subcaudals, the tails of the 9 9 being noticeably shorter.

Breeding. The following records of developing eggs were made.
Ujiji, March 11, 1939, 9 held 2 eggs measuring 20 x 10 mm. Kitaya.
March 25, 1939. 9 held 8 eggs measuring 15 x 9 mm. While at
Mushongero, February 2, 1939, 6 eggs measuring 40 x 24 mm. were
found in a termitarium. They contained young measuring 220 (185
-f- 35) mm. which, on hatching a month later, measured 234 (205
+ 29) mm.

Diet. Rodent fur in a Kisenyi snake, that of a jumping rodent in
a Mikindani reptile. Rats (Rattus r. kijabius) recovered from Fort
Portal and Ujiji specimens; pigmy mice (Leggada spp.) from Mabira,
Ujiji and Kitaya snakes; a mouse (Lophuromys a. aqnilns) at Nyaka-
bande; an arboreal rat (Thamnomys s. surdaster) in a Magrotto
snake; and a tree mouse (Dendromus I. kivu) in one of the Idjwi
Island series.

A gecko (Hemidaetylus mabouia) with tail intact in a Kitaya snake;
a young lizard (Gerrhosaurus n. nigrolineatus) and adult skink (Able-
■pharus wahlbergii) in Nchingidi specimens. Of frogs Rana m mas-
carcniensis at Siga Caves, Arthroleptis adolfi-friederiei on Magrotto
Mountain, A. s. stenodaetylus and Hemisus m. marmoratum in Mikin-
dani snakes.

It was noticeable that the pigmy mice, lizards, and frogs were
swallowed by the smaller snakes, the larger rodents, such as the

LOVERIDGE: AFRICAN REPTILES 265

mole rat (in M. C. Z. 48135) measuring 350 (190 + 160) mm., by
the adult reptiles.

Parasites. Ticks on Nyakabande, Ujiji, and Kitaya snakes; worms
(Ophidascaris sp.) in Kisenyi and (Polydelphis sp.) in Amboni specimens.

Habitat. Under palm fronds and other vegetation, as well as under
logs; others were taken in native huts according to their captors.

Boaedon olivaceus (Dumeril)
Holuropholis olivaceus A. Dumeril, 1856, Rev. Mag. Zool. (2), 8, p. 466: Gaboon.

1 (M. C. Z. 48185) Mabira Forest. U. 5.xi.38.
5 (M. C. Z. 48186-90) Bundibugyo, U. 22.xii.38.

Native name. Kiliba (Luamba).

Variation. Midbody scale-rows 27-31; ventrals 208-224; anal en-
tire; subcaudals 41-48; labials 8, the fourth and fifth, or (M. C. Z.
48187) third, fourth and fifth entering the orbit; frontal once to once
and a quarter as long as broad, as long as, or shorter, than its distance
from end of snout.

Measurements. Largest 9 (M. C. Z. 48190) measures 782 (695
-f 87) mm.

Breeding. One Bundibugyo snake held very small eggs, in another
they measured 39 x 18 mm., but several were smashed.

Diet. Mouse fur in stomach of latter.

Lycophidion meleagris Boulenger

Lycophidium meleagris Boulenger, 1893, Cat. Snakes Brit. Mus., 1, p. 337,
pi. xxi: Ambriz and Ambrizette, Angola.

4 (M. C. Z. 48267-70) Magrotto Mtn., T. T. 8-21.vii.39.

Variation. Midbody scale-rows 15; ventrals 149-160; anal entire;
subcaudals 25-36; labials 8, the first in contact with the posterior
nasal, third, fourth and fifth entering the orbit.

Coloration. A young d" reminded me of L. c. uzungwensis in that a
salmon pink band follows the contour of the snout from eye to eye,
being narrowly edged above with buff and below with black. An adult
d\ 9 , and a young 9 exhibited only the buff band. In all, the slightly
bluish white flecks on the glossy black scales of the back and flanks
present a very handsome appearance.

Measurements. Larger a71 (M. C. Z. 48268) measures 301 (260 + 41)
mm.; larger 9 (M. C. Z. 48267) measures 303 (275 4- 28) mm., both

266 bulletin: museum of comparative zoology

being exceeded slightly by Usambara and Uluguru specimens in the
collection.

Diet. The stump of a tail and two eggs of the montane forest skink
(Lygosoma kilimense) .

Defence. On being disturbed these snakes make no attempt to bite
but flatten themselves greatly, often coiling as well.

Habitat. In East Africa this is a truly sylvicoline species and I cap-
tured all the above snakes among drifted leaves between buttress
roots of giant trees in the depths of the forest.

Lycophidiox capexse orxatum Parker

Lycophidion ornatum Parker, 1936, Novit. Zool., 40, p. 122: Congulu, Angola'

1 (M. C. Z. 48191) Bugoye, U. 23.L39.
1 (M. C. Z. 48303) Nyakabande, U. 27.L39.
1 (M. C. Z. 48192) Mushongero, U. 2.ii.39.
57 (M. C. Z. 48193-249) Idjwi Id., B. C. ii.39.
1 (M. C. Z. 48250) Ujiji, T. T. ll.iii.39.

Distribution. New for Tanganyika Territory and Kenya Colony
(see below).

Native name. Busugu (Lulega). Thought to be the young of
Boaedon lineatus by Kigezi natives.

Corrigenda. This snake is readily distinguishable from L. c. capense
in the field by the double series of dark spots along the entire length of
the dorsum, difficult to detect after formalin preservation. I (1936j,
p. 241) commented on this under L. c. capense where I erroneously
stated that: "This difference is not correlated with any scale charac-
ters enabling me to separate them." Parker (1936c, p. 122) with
greater perspicacity detected that the first labial is constantly sepa-
rated from the postnasal in ornatum, whereas it is in contact in capense.
Unfortunately my paper was in press before Parker's reached me.
The specimens which should be removed from capense in the citation
(1936J, p. 241) given, are:

1 (M. C. Z. 39966) Kigezi district, U.

3 (M. C. Z. 40468-70) Sipi, Alt. Elgon, U.

4 (M. C. Z. 40471-3) Kaimosi, K. C.

The other Mount Elgon specimen, from the deforested Sabei region, is
typical capense. I regard L. c. ornatum as a forest form with head-
quarters in the high country of the Central Lake region much of which
has undergone deforestation within living memory. Both forms meet
at Ujiji.

loveridge: African reptiles 267

My reasons for regarding ornatum as a race of capense, rather than
a full species, is on account of the spotted snake from Bugoye, eastern
slopes of Ruwenzori, which has the first labial and post nasal meeting
in a point, as have also occasional capense from Gulu, Acholi, and
elsewhere.

Variation. The uniformity exhibited by this extensive series is re-
markable, leaving little to add to the exhaustive description furnished
by Parker.

Midbody scale-rows 17; ventrals 180-205; anal entire; sub'caudals
34-49; labials 8, the first separated from the postnasal except for
the Bugoye snake, the third, fourth and fifth entering the orbit, or
very rarely 7 labials, the third and fourth entering the orbit (2 ex.);
4 or 5 lower labials in contact with an anterior sublingual ; loreal 1 ;
preocular 1; postoculars 2; temporals 1 + 2, rarely 1 + 1 (1 ex.) or
2 + 2 (4 ex.).

Measurements. Largest cf (M. C. Z. 4S196) measures 405 (346 +
59) mm.; largest 9 (M. C. Z. 48235) measures 475 (406 4- 69) mm.

Sexual dimorphism. Unfortunately there is a slight overlap in the
number of subcaudals, viz.

23 cf c?1 range from 180-198 ventrals and 41-49 subcaudals,
37 9 9 range from 188-205 ventrals and 34-43 subcaudals.

Breeding. At Mushongero, on February 2, a batch of 5 eggs measur-
ing 21 x 15 mm. containing well advanced embryos, were found in a
termitarium (fide native). They hatched on Idjwi Island on March 6,
at which time one of the hatchlings (M. C. Z. 4S192) measured 145
(125 + 20) mm. The rest were released.

On Idjwi Id., February 21, a 9 held 4 eggs measuring

22 " 4

22 " 4

25 " 3

25 " 5

28 " 3

28 " 3

Diet. A lizard (Algiroides boulengeri) in one, a skink (Mabuya m.
maeulilabris) in another, and 22 skinks (Lygosoma blochmanni) in
almost as many other snakes, showing that this little skink constitutes
the principal food of L. c. ornatum on Idjwi Island.

Habitat. I caught this species on paths right in the forest and its

1 Probably one or more eggs had been laid already.

9x

5 mm.

9x

5 mm.

14 x

6 mm.

20 x

9 mm.

10 x

5 mm.

12 x

6 mm.

26 x

10 mm.

10 x

5 mm.

268 bulletin: museum of comparative zoology

prey are principally found on the sunny banks at the sides of such
paths. There is little or no forest, however, in the Uganda and Tangan-
yika localities whence single specimens were taken.

Lycophidiox capense capense (Smith)

Lycodon capense A. Smith, 1831, S. Afr. Quart. Journ., 1, p. 18: Kurrichane,
i.e. Rustenberg district, Transvaal.

•1 (M. C. Z. 48251) Butiaba, U. 5.xii.38.
2 (M. C. Z. 48252-3) Ujiji, T. T. 10.iii.39.

1 (M. C. Z. 48254) Mikindani, T. T. 18.iv.39.

7 (M. C. Z. 48255-9) Mbanja, T. T. 26-30.iv.39.
4 (M. C. Z. 48260-3) Nchingidi, T. T. 18.V.39.

2 (M. C. Z. 48264-5) Amboni Est., T. T. 24.vi.39.
1 (M. C. Z. 48266) Opp. Kilindini, K. C. 25.vii.39.

Native names. Namaluto (Kimakonde at Mikindani) ; lukunguviro
(Kimakonde at Mbanja).

Variation. Midbody scale-rows 17; ventrals 184-211; anal entire;
subcaudals 37-57; upper labials 8, the first in contact with the post-
nasal except on one side of an Mbanja snake, the third, fourth and
fifth entering the orbit; postoculars 2, except on right side of M. C. Z.
48252; temporals 2 + 2 or 1 + 1 (M. C. Z. 48265 only).

Coloration. Agree with the typical form in having the throat more
or less white, except for the Ujiji snakes which are dark, agreeing in
this respect with the Zanzibar race acutirostrix!

Measurements. Largest cf (M. C. Z. 48264) measures 408 (345 +
63) mm.; largest 9 (M. C. Z. 48252) measures 467 (425 + 42) mm.

Sexual dimorphism. Subcaudal range in 9 d71 d* is 46-57, in 8 9 9
it is 37-41.

Breeding.
At Mbanja, April 26, a 9 held 5 eggs measuring 24 x 10 mm.
" 28, " 5 " " 10 x 4 mm.

" 30, " 4 " " 10 x 4 mm.

" 30, " 4 " " 23 x 8 mm.

The last mentioned snake was sloughing and paired with a male which
was so entrapped in the slough that it could not free itself! At Nchin-
gidi, May 18, a very young snake, measuring only 182 (160 + 22) mm.
was taken.

Diet. Young lizards (Gerrhosaurus n. nigrolineattis) in Mikindani and
Amboni snakes; adult skinks (Mabuya maculilabris) in Amboni and
Gulu, Acholi (Pitman coll.) reptiles; an adult M. striata in the largest

loveridge: African reptiles 269

Ujiji female, M. v. varia in Mbanja and Nchingidi snakes and an
Ablepharas wahlbergii in a very young Mbanja specimen.

Habitat. Under palm fronds or garden refuse at Butiaba, Mikindani
and opposite Kilindini.

Lycophidion capense > < acutirostre Giinther

Lycophidion intermediates between capense and acutirostre Loveridge, 1933,
Bull. Mus. Comp. Zool., 74, p. 234: Zanzibar and Bagamoyo, Morogoro
and Kilosa in Tanganyika Territory.

3 (Vienna Mus.) Ugano, T. T. 1935 (H. Zerny).
juv. 9 (M. C. Z. 48271) Mbanja, T. T. 27.iv.39.

Distribution. I have included some recently examined Zerny mate-
tial as these two localities in extreme southern and southeastern
Tanganyika Territory constitute the most southerly records of this
form. Moreover typical L. c. capense occurs in both these localities!

Native name. Lukungwriro (Kimakonde at Mbanja for both forms).

Variation. Midbody scale-rows 17; ventrals 161-172; anal entire;
subcaudals 23-37; labials 8, the first in contact with the posterior
nasal, the third, fourth and fifth entering the orbit. Agreeing in their
low ventral and subcaudal counts and black throat with this race
as defined in the citation given above.

Measurements. The juvenile 9 measures 151 (137 -f- 14) mm.

Diet. A skink (Ablcpharus wahlbergii) in this tiny snake.

[Mehelya capensis capensis (Smith)]

Heterolepis capensis A. Smith, 1847, 111. Zool. S. Africa, Rept., pi. lv: Eastern
parts of Cape Province, South Africa.

Distribution. When writing the revision of this genus, I (1939c,
p. 144) was surprised at the absence of records of this race between
the Usambara Mountains in Tanganyika and Delagoa Bay in Mo-
zambique. At Lindi I met Dr. L. Stirling of the Universities Mission
who described this snake to me beyond the shadow of a doubt as
occurring southwest of Lindi at Lulindi, where there is a remnant of
forest. He encountered one in the church at 7 p.m. while another was
killed in the bed of a native child. The latter, a boy aged 12, felt
something moving about his feet so got up. Both reptiles were large.

270 bulletin: museum of comparative zoology

[Pseudaspis cana (Linnaeus)]

Coluber canus Linnaeus, 1758, Syst. Nat., ed. 10, 1, p. 221 : "Indiis."

Breeding. On South Kinangop, Kenya Colony, Mrs. Nightingale
Jr. asked me the name of a large four-foot olive snake, which had
thirty live chequered young in its oviducts, that they had killed.

Chlorophis carinatus Andersson

Chlorophis carinatus Andersson, 1901, Svenska Vetensk.-Akad. Hand., 27,
No. 5, p. 9: Cameroon.

1 (M. C. Z. 48275) Budongo Forest, U. 29.xi.38.

2 (M. C. Z. 48276) Idjwi Island, B. C. 22.ii.39.

Native name. Lushangabanyeri (Lulega).

Variation. Midbody scale-rows 13; ventrals 141-158; anal entire;
subcaudals 76-86; labials 9, the fourth, fifth and sixth entering the
orbit; temporals 2 -f- 2, rarely 2+1.

Coloration, cf (M. C. Z. 48276) Above, dark olive, head uniform,
the vertebral series of scales edged baso-laterally with very pale
blue, the outer scales on the outer baso-lateral side only also pale
blue, tail uniform. Below, chin white, throat yellowish, rest of under-
surface uniformly pale green.

Measurements. Large d71 (M. C. Z. 48276) measures 556 (410
+ 146) mm.

Chlorophis macrops (Boulenger)

Oligolepis macrops Boulenger, 1895, Ann. Mag. Nat. Hist. (6), 16, p. 171:
Usambara Mountains, Tanganyika Territory.

3 9 9 (M. C. Z. 48272-4) Nchingidi, Rondo Plateau, T. T. 12.V.39.

Corrigenda. In Barbour & Loveridge, 1928, Mem. Mus. Comp.
Zool., 50, p. 117, lines 7, 8, a,nd 23, for ventrals 169 read 149, for
subcaudals 122, read 112. The erroneous figures result from the in-
clusion of three specimens of C. neglectus in the extensive series of
topotypic macrops.

Variation. Midbody scale-rows 13; ventrals 135-141; anal divided;
subcaudals 69-84; labials 8, the fourth and fifth entering the orbit,
or 9, the fifth and sixth entering (on left side of M. C. Z. 48273 only);
temporals 1 + 1 or 1 + 2.

loveridge: africax reptiles 271

Coloration. The largest 9 exhibited a very peculiar coloration for
a member of this genus. Above, uniform brown except for a few
scales which resemble the laterals in being edged with black. Below,
white, each ventral with a dull blood-red blotch above the keel (i.e.
laterally) and a less distinct smaller one below the keel (i.e. abdomi-
nally) ; tail dull white with each subcaudal faintly tinged with blood
red basally.

Measurements. Largest 9 (AL C. Z. 4S272) measures 697 (525
-+- 172) mm.

Diet. A chameleon (Brookesia brecicaudata) in largest, tail of
a skink, also a frog (Arthroleptis s. lonnbergi) in another.

Chlorophis hoplogaster (Gunther)

Ahaetulla hoplogaster Gunther, 1863, Ann. Mag. Nat. Hist. (3), 11, p. 284:
Port Natal, i.e. Durban, Natal.

9 (M. C. Z. 48277) Mabira Forest, U. 12.xi.38.

Native name. Xewandegala (Luganda).

Variation. Midbody scale-rows 15, ventrals without keels 157;
anal divided; subcaudals 92; labials 8, the fourth and fifth entering
the orbit; temporals 1 + 1.

Measurements. 9 measures 880 (627 + 253) mm.

Breeding. Oviducts held six eggs measuring 30 x 10 mm.

Diet. A lizard (Algiroides africanus) in stomach.

Chlorophis neglectus (Peters)

Philothamniis neglectus Peters, 1866, Monatsb. Akad. Wiss. Berlin, p. 890:
Praso Boror, Mozambique.

2 (M. C. Z. 48301-2) Ujiji, T. T. 10.iii.39.

1 (M. C. Z. 48304) Kitaya, T. T. 24.iii.39.

2 (M. C. Z. 48305-6) Mikindani, T. T. 14.iv.39.
c? (M. C. Z. 48307) Lake Rutamba, T. T. 8.V.39.
J* (M. C. Z. 48308) Nchingidi, T. T. 14.V.39.

cf (M. C. Z. 48309) Amboni Estate, T. T. 19.vi.39.
4 (M. C. Z. 48310-2) Magrotto Mtn., T. T. l.vii.39.

Native names. Namalanga (Kimakonde) ; nyaika amani (Kisambara).

Variation. Midbody scale-rows 15; ventrals with lateral keels 146-

152; anal divided; subcaudals 85-99; labials 8, the fourth and fifth

272 bulletin: museum of comparative zoology

entering the orbit; preocular 1, rarely 2 (one side of one snake only);
postoculars 2; temporals 1 + 1, rarely 1 + 2 or 2 + 3 (M. C. Z.
48305).

Coloration. A Mikindani snake was green above with seven short,
dark, transverse bars on nape giving it somewhat the appearance of a
Causus resimus. The young Rutamba snake was rich velvety green
with a brown, transverse bar on nape followed by a series of nine, more
or less paired, brown spots.

Measurements. Largest cf (M. C. Z. 48302) measures 657 + (462 +
185 +) mm., tail tip missing.

Diet. Frogs (Arthroleptis s. stenodactylus) in the Kitaya snakes.

Habitat. A young one in a heap of debris beneath a mango tree at
Mikindani, another was taken near my tent at Kitaya, following
heavy rain.

Chlorophis irregularis (Leach)

Coluber irregularis Leach, 1819, in Bowdich, Miss. Ashantee, p. 494: Ashanti,
Gold Coast.

1 (M. C. Z. 47880) Mabira Forest, U. 8.xi.38.

4 (M. C. Z. 48278-80) Bundibugyo, U. 21-24.xii.38.
3 (M. C. Z. 48281-3) Mihunga, U. 16.i.39.
Eggs 13 (M. C. Z. 48285-95) Mushongero, U. l-4.ii.39.

2 (M. C. Z. 48296) Kisenyi, B. R. 10.ii.39.

6 (M. C. Z. 48297-9) Idjwi Id., B. C. 16-28.ii.39.

Native names. Newandagala (Luganda); salaln (Luamba); nyaru-
babi or nyaruteti (Lutoro); chienzi (Lukonjo); muckenganyi (Lukiga);
lushangabanyeri (Lulega) .

Variation. Midbody scale-rows 15; ventrals 154-167; anal divided;
subcaudals 92-117; labials 9, rarely 8, the fourth, fifth and sixth or
fifth and sixth (M. C. Z. 48298 left side only) entering the orbit; tem-
porals 1 + 1 (on 43 sides) or 1 + 2 (on 15 sides).

Measurements. Largest cf (Idjwi Id.) measures 845 (575 + 270)
mm.; largest 9 (M. C, Z. 48278) measures 1058 (735 + 323) mm.

Breeding.
Bundibugyo, December 21, a 9 held 5 eggs measuring 22 x 7 mm.
Mihunga, January 16, " 5 " 32 x 11 & 28 x 12 mm.

Kisenyi, February 10, " 8 " 25 x 10 mm.

In addition, at Mushongero, on February 1, a native brought me 193
eggs which he allegedly dug from two termitaria. One batch of 8
eggs measuring 25 x 14 mm., another of 8 measured 39 x 17 mm., while

loveridge: afkican reptiles 273

3 eggs selected from different batches measured 43 x 18 mm., 30 x 18
mm., and 29 x 18 mm. respectively, their diameter evidently condi-
tioned by the girth of the parent; each contained an embryo nearly
ready for hatching, such measuring 203 (143 + 60) mm., and on hatch-
ing a few weeks later a cf measured 260 (180 + 80) mm., and a 9
measured 249 (ISO + 69) mm.

Diet. A large lizard (Lacerta jacksoni) two young toads (Bufo regu-
laris) and three yellow sedge frogs {Hyperolius schubotzi) in three
Idjwi Island snakes, a frog (Rana fuscigula) in a Mushongero snake.

Parasites. Nematodes (Amphicaecum sp. and Ascaroidea sp.) in an
Idjwi Island snake.

Temperament. A snake, losing its hold on a palm frond overhanging
a path in the public garden on the Lake shore at Kisenyi, fell at my
feet. As I seized it, the snake gaped till its jaws were in almost a single
plane as it struck, the teeth drew little blood.

Habitat. At Alihunga I captured several in a swamp where they
were undoubtedly hunting frogs.

Chlorophis heterolepidotus (Giinther)

Ahaetulla heterolepidota Giinther, 1863, Ann. Mag. Nat. Hist. (3), 11, p. 286:
Africa.

9 (M. C. Z. 47809) Bukakata, U. 7.ix.38. (C.R.S.Pitman).
4 (M. C. Z. 47841-4) Lira, Lango, U. iv-viii.38. (C.R.S.P.).
c? (M. C. Z. 48284) Nyakabande, Kigezi. U. 28.L39.
d" (M. C. Z. 48300) Kitaya, Ruvuma River. T. T. 24.iii.39.

Variation. Midbody scale-rows 15; ventrals 173-185; anal divided;
subcaudals 107-124; labials 9, rarely 8 (M. C. Z. 47809 left side only),
the fourth, fifth and sixth entering the orbit; temporals 1 + 1, rarely
1 + 2.

This snake, so apt to be confused with other members of the genus,
is recognizable by its extremely slender form anteriorly and higher
ventral count. The highest ventral count is for the Kitaya snake,
the highest subcaudal for the Nyakabande reptile.

Measurements. Largest d" (M. C. Z. 48300) measures 730 (492 +
23S) mm.; largest 9 (M. C. Z. 47842) measures 586 (400 + 186) mm.

Diet. A sedge frog {Hyperolius rossii) in each of the two Lira males,
a yellow frog (Hyperolius sp.) in the Nyakabande snake, a frog in the
Kitaya specimen.

274 bulletin: museum of comparative zoology

Philothamnus semivariegatus semivariegatus Smith

Philothamnns semivariegatus A. Smith, 1849, 111. Zool. S. Africa, Rept., pis.
lix, lx, lxiv: Bushman's Flats and Kurrichane, i.e. Rustenberg district,
Transvaal.

1 (M. C. Z. 47820) Katwe, U. x.38. (C.R.S.Pitman).
1 (M. C. Z. 48313) Budongo Forest, U. 3.xii.38.
1 (M. C. Z. 48314) Bundibugyo, U. 22.xii.38.

1 (M. C. Z. 48315) Ujiji, T. f . 15.iii.39.

8 (M. C. Z. 48316-21) Kitaya, T. T. 25.iii.-3.iv.39.
6 (M. C. Z. 48322-5) Mikindani, T. T. 15-21.iv.39.

2 (M. C. Z. 48326-7) Mbanja, T. T. 1 & 5.V.39.

1 (M. C. Z. 48328) Siga Caves, T. T. 14.vi.39.

3 (M. C. Z. 48329-31) Amboni Est., T. T. 17.vi.39.

2 (M. C. Z. 48332-3) Magrotto Mtn., T. T. 3.vii.39.

Native names. Not distinguished from C. irregularis in Luamba
and Lutoro; kisumera (Kimakonde at Kitaya); namahamba (Kima-
konde at Mikindani and Mbanja); nawirangira (Kimawiha); ngoe
(Kisambara, but supposed to be young of green mamba, D.angusticeps).

Variation. Midbody scale-rows 15; ventrals 164-193; anal divided;
subcaudals 127-157; labials 8, the fourth and fifth entering orbit
(3 odd sides), or 9, the fourth, fifth and sixth (23 sides), or 9, the
fifth and sixth (22 sides), or 10, the fifth, sixth and seventh (3 sides),
or 10, the sixth and seventh (1 side) entering orbit; temporals 2 +
2 (41 sides), or 2 + 1 (5 sides), or 1 -f 2 (5 sides), or 1 + 1 (1 side).

Coloration. At Kitaya snakes with blue heads as well as with green
heads were present, one of the latter had the anterior third of the
body transversely barred with blue. At Mikindani most specimens
were uniformly green, being distinguished from Chlorophis neglectus
only by the subcaudal keel and scale-counts; one had a bluish head.

Measurements. Largest c? (M. C. Z. 48333) measures 1113 (690
+ 423) mm., largest 9 (Mikindani) measures 1162 (745 + 417) mm.

Sexual dimorphism. None in scalation, for the
10 c? c? range from 164-193 ventrals and 127-156 subcaudals, while
16 9 9 range from 167-192 ventrals and 128-157 subcaudals.

Breeding.
Budongo Forest, December 3, a 9 held 4 eggs measuring 30 x 8 mm.

Kitaya,

March 25,

it

7

C(

ft

21 x 7 mm.

it

April 3,

it

6

ft

ft

11x3 mm.

Mikindani,

April 20

It

5

It

ft

28 x 8 mm.

It

April 21

11

3

If

ft

29 x 8 mm.

Mbanja

Mayl,

ft

5

It

ft

20 x 5 mm.

n

May 5,

tt

5

ft

ft

24 x 6.5 mm.

loveridge: African reptiles 275

Diet. Geckos {Lygodaetylus p. gutturalis) in the Katwe and Budongo
snakes, L. g. grotci in Ujiji, Mikindani and Mbanja specimens, Hemi-
dactylus mabouia in Kitaya and Mbanja reptiles; three young tree
frogs (Leptopelis concolor in a Siga snake, Mcgalixalus brachycnemis
in one from Kitaya, Hyperolius parkcri in a juvenile which I captured
in sedges growing from knee-deep water in an Amboni swamp.

Parasites. Nematodes (Thubunaea sp., probably T. asymmetrica)
and immature cestodes in a Mikindani snake.

Habitat. I shot a female with truncated tail as it was basking about
a knot hole of an almost vertical tree trunk at Kitaya.

Gastropyxis smaragdina (Schlegel)
Dendrophis smaragdina Schlegel, 1837, Essai Phys. Serp., 2, p. 237: Gold Coast.
9 (M. C. Z. 48334) Bundibugyo, U. 22.xii.38.

Native names. Not distinguished from Chlorophis irregularis in
Luamba and Lutoro.

Variation. Midbody scale-rows 15; ventrals 160; anal divided; sub-
caudals 144; labials 9, the fifth and sixth entering orbit; postoculars
2-3; temporals 1 + 2.

Measurements. 9 measures 1143 (715 + 428) mm.

Breeding. Oviducts held 4 eggs measuring 23 x 7 mm.

Hapsidophrys lineata Fischer
Plate 2, fig. 3.

Hapsidophrys lineata Fischer, 1856, Abhand. Nat. Ver. Hamburg, 3, p. Ill,
pi. ii, fig. 5: Elmine, Gold Coast.

c? (M. C. Z. 48335) Budongo Forest, U. 22.xi.38.
9 (M. C. Z. 48336) Kibale Forest, U. 14.xii.38.

Variation. Midbody scale-rows 15; ventrals 159-161; anal entire;
subcaudals M and 104; labials 8, the fourth and fifth entering orbit;
postoculars 2-3; temporals 2 + 2.

Coloration. 9 . Above, rich velvety green, each scale heavily edged
with black forming ten black lines on dorsum converging to form five
broad ones on tail. Below, a paler, slightly yellowish, green, the
lateral keels and edges of the ventrals darker, almost bluish, a median
black line along tail.

Measurements, c? measures 985+ (760 + 225+) mm., tail tip mis-
sing; 9 measures 983 (705 + 278) mm.

276 bulletin: museum of comparative zoology

Breeding. 9 held 2 eggs measuring 25 x 6 mm.

Diet. A frog (Phrynobatrachus dendrobates) in stomach of male,
frog's bones in that of female.

Temperament. The female was moving slowly along a branch of a
sapling in deep forest. It was only five feet from the ground and as I
took it by the neck it made no attempt to bite, nor later when sub-
jected to considerable provocation during a quarter-of-an-hour's
posing for its photograph.

Rhamnophis aethiopissa elgonensis Loveridge

Rhamnophis aethiopissa elgonensis Loveridge, 1929, Bull. U. S. Nat. Mus. 151,
p. 24: Yala River nr. Kaimosi, Kakamega, Kenya Colony.

tf1 (M. C. Z. 48343) Mabira Forest, U. ll.xi.38.
cf (M. C. Z. 48344) Kibale Forest, U. 17.xii.38.

Synonymy and a correction. All the Uganda material referred to
ituriensis by Pitman (1936-1938) are really elgonensis, this also applies
to his fine colored plate. To avoid further confusion of the forms the
following key, based on all records in the literature and a detailed
examination of M. C. Z. material, is given.

1. Midjbody scale-rows 13; anal entire; subcaudals less than 116;

maxillary teeth 30 + 3 to 35 -f- 3 batesii

Midbody scale-rows 15-19; anal divided; subcaudals more than
116; maxillary teeth 17 + 3 to 20 + 3 2

2. Midbody scale-rows 17, very rarely 15, 16, or 19; lower postocular

in contact with 3 upper labials; range: French Guinea east to

the western Belgian Congo a. aethiopissa

Midbody scale-rows 15, very rarely 17; lower postocular in contact
with 2 or 3 upper labials ; range : eastern Belgian Congo to western
Kenya Colony 3

3. Subcaudals 134-150; normally upper labials 8, sometimes 7; range:

Ituri region of eastern Belgian Congo a. ituriensis

Subcaudals 117-138; normally upper labials 7, sometimes 6 or 8;
range : western Uganda to western Kenya Colony ... a. elgonensis
Taxonomically batesii, which ranges from Cameroon to the eastern
Belgian Congo, appears to be nearly related to a. elgonensis, thus
paralleling to some extent the situation in the allied genus Thrasops
where the extreme westerly and easterly forms are more nearly related
than the (?) derived forms occupying the Cameroon-Congo region.

The type number of a. elgonensis is M. C. Z. 18198, not 18189 as
printed.

loyeridge: African reptiles 277

Variation. Midbody scale-rows 15; ventrals 156-158; anal divided;
subcaudals 126-128; labials 7, the fourth and fifth entering the orbit.

Coloration. In life. cf. Above, leaf green, each scale heavily edged
with black, the black interstitial skin also conspicuous; upper labials
pale green anteriorly and yellow posteriorly in their upper portion, blue
below; five black lines on tail. Below, ventrals greenish flecked with
white, a brown line along each lateral angle, outer ends of ventrals pale
olive green; tail with a median dusky line flanked by irregular dark
flecks.

Thrasops jacksonii jacksonii Giinther

Thrasops jacksonii Giinther, 1895, Ann. Mag. Nat. Hist. (6), 15, p. 528:
Kavirondo, Kenya Colony.

9 (M. C. Z. 48341) Bundibugyo, U. 22.xii.38.
c? 9 (M. C. Z. 48342) Idjwi Id., B. C. 22.ii.39.

Native name. IVahimbiri (Lutoro and Luamba).

Synonymy. The snake from Mozambique described as Thrasops j.
mossambieus by Mertens (1937b, Abh. Senckenberg. Naturf. Ges., No.
435, p. 13) is a Dispholidus typus, a correction with which Dr. Mertens
concurs after reexamination of the dentition. Parker (1940a, p. 271),
by describing occiden talis, has elucidated the confusion arising from
western records of jacksonii. The following key, based on all records
in the literature, and a detailed study of M. C. Z. material, deals with
the recognizable forms in the genus.

1. Three labials in contact with the lowest postocular; midbody scale-

rows 15-19; ventrals 175-187; subcaudals 120-140; range: French

Guinea east to Togo occidentalis

Two, very rarely 3, labials in contact with the lowest postocular;
range : Nigeria east to Kenya 2

2. Midbody scale-rows 13-15; the dorsals much longer than the ven-

trals; range: Nigeria south to Cabinda and western Belgian

Congo flavigularis

Midbody scale-rows 17-21, the dorsals not or but slightly longer
than the ventrals; range: central Belgian Congo east to Kenya
Colony 3

3. Midbody scale-rows 19, rarely 17 or 21; ventrals 187-211; range:

central Belgian Congo east to western Tanganyika Territory

and western Kenya Colony j. jacksoni

Midbody scale-rows 17; ventrals 170-178; range: Mount Kenya to
Nairobi in southcentral Kenya Colony j. schmidti

278 . bulletin: museum of comparative zoology

Variation. Midbody scale-rows 19; ventrals 187-206; anal divided;
subcaudals 133-141; preoculars 1-2; postoculars 3. The young Idjwi
male would actually fall to occidentalis in the above key for it has a
third labial just reaching the lowest postocular.

Diet. A chameleon (C. b. ellioti) in the Bundibugyo snake; an agama
{A. atricollis) in each of the reptiles from Upper Mulinga, Idjwi Island,
Lake Kivu.

Meizodon semiornata (Peters)

Coronella semiornata Peters, 1854, Monatsb. Akad. Wiss. Berlin, p. 622: Tete,
Mozambique.

1 (M. C. Z. 48337) Kitaya, Ruvuma R., T. T. 29.iii.39.
3 (M. C. Z. 48338-40) Amboni Estate, T. T. 19.vi.39.

Native name. Namedi (Kimakonde).

Variation. Midbody scale-rows 21; ventrals 175-196; anal divided
subcaudals 83-85; labials 8, the fourth and fifth entering the orbit
4-5 lower labials in contact with an anterior sublingual; preocular 1
postoculars 1-2; temporals 2 + 2, rarely 2 + 3 or 3 + 3 (M. C. Z.
48338).

Corrigenda. Having examined the type of Coluber smithii Boulenger
from Somaliland, I must refer the Kenya snakes which I described as
Coronella semiornata fuscorosea to the synonymy of smithii, entirely
concurring with the remarks of Bogert (1940, p. 48) who independently
reached the same conclusion, and whom I follow in recognizing Meizodon
for the tropical African 'Coronella , so different from the European
genotype, austriaca.

Measurements. Largest <? (M. C. Z. 48338) measures 695+ (550 +
145+) mm., tail tip missing; only 9 (M. C. Z. 48339) measures 682
(552 + 130) mm.

Diet. A gecko (Hemidactylus gardineri) in one Amboni snake.

Habitat. I caught all beneath rotting vegetation, the Kitaya snake
at the base of a banana plant, the others in a sisal plantation as de-
scribed under Amboni in the account of the itinerary.

Meizodon coronata (Schlegel)

Calamaria coronata Schlegel, 1834, Phys. Serp., 2, p. 46" Gold Coast.

<? (M. C. Z. 47829) Gulu, Acholi, U. ix-x.38. (C.R.S. Pitman).

Variation. Midbody scale-rows 19; ventrals 185; anal divided; sub-
caudals 66+ ; labials 8, the fourth and fifth entering orbit; preocular 1;
postoculars 2; temporals 1 + 2.

loveridge: African reptiles 279

In its labial-sublingual arrangement of 4-5 in contact, this individual
combines the characters of both coronata and its synonym regularis
and possesses the blackish belly of the latter. For discussion on this
subject see Loveridge (1936j, p. 253).

Measurements, cf measures 597+ (482 4- 115+) mm., tail tip missing.

Grayia smythii (Leach)

Coluber Smythii Leach, 1818, in Tuckey, Explor. River Zaire, App., p. 409:
Embomma, i.e. Boma, Belgian Congo.

c? 9 (M. C. Z. 47813-4) Mlanji, L. Victoria, U. 14-20.vii.38.
(C.R.S. Pitman).

Variation. Midbody scale-rows 17; ventrals 155-164; anal 2; sub-
caudals 92-99; labials 7-8, the fourth entering the orbit; temporals
2 + 2 and 2 4- 3.

Measurements, c? measures 990 (675 4- 315) mm.; the 9 measures
1260 (900 + 360) mm.

Duberria lutrix abyssinica (Boulenger)

Homalosoma abyssinicum Boulenger, 1894, Cat. Snakes Brit. Mus., 2, p. 276,
pi. xiii, fig. 2: Lake Ashangi, Ethiopia.

2 (M. C. Z. 48345) S. Kinangop, K. C. 27.X.38.
11 (M. C. Z. 48346-53) Nyakabande, U. 27-31.L39.
1 (M. C. Z. 48354) Mushongero, U. l.ii.39.

Native name. Bulifu (Lukiga).

Synonymy — a correction. In reporting on a collection of D. I.
shirana, and with inadequate northern material, I (1933h, p. 241)
assumed that the holotype of abyssinica without a loreal, and holotype
of atriventris with a loreal, were simply aberrations of the highly
variable shirana.

Uthmoller (1937, Temminckia, 2, p. 112), reporting on three speci-
mens from the Kilimanjaro region, remarked that all three were like-
wise intermediate in character between shirana and lutrix.

Pitman (1936, p. 62) commented on Kigezi snakes being much nearer
abyssinica than to shirana in the matter of colouration.

More recently, Bogert (1940, p. 40) revived the name abyssinica in
a racial sense for an Ethiopian specimen, pointing out that its ventral
and subcaudal counts were lower than that of females in my series of
shirana from southern Tanganyika Territory.

280 bulletin: museum of comparative zoology

After going through the entire Duberria literature, I find that I was
in error in not recognizing abyssinica, to whose synonymy atriventris
(Sternfeld) should be transferred as all attempts to separate them failed.
The following key, based on counts of 53, 50 and 56 snakes respectively,
though not all the characters have been available in every specimen,
reflect the position as it stands at present. Slight changes in the per-
centages are due to the effect of additional data. I regard D. 1. abys-
sinica of the Northern highlands as being the original stock giving rise
to both shirana and tutrix between which it occupies an intermediate
position.

1 postocular (100%) ; a loreal (absent in 10%) ; belly usually very dark'
rarely yellowish in middle ; range : highlands of western Belgian Congo
northeast through Uganda, Ethiopia and south to the Usambara Moun-
tains in northeastern Tanganyika Territory /. abyssinica

1 postocular (85%, 2 in 15%); no loreal (100%); belly yellowish in
middle, rarely dark; range: highlands of southern Tanganyika and
Xyasaland around L. Xyasa /. shirana

2 postoculars (87%, 1 in 13%); a loreal (absent in 7%); belly yellow-
ish in middle; range: highlands and lowlands of Africa south of the
Zambesi I. lutrix

The lepidosis of the three forms, based on all available records and
scale-counts of the new material listed above, is as follows:

D. 1. abyssinica. Range of ventrals 118-149, subcaudals 17-39.

D. 1. shirana " 126-151, " 24^6.

D. 1. tutrix " 120-1341, " 225-51.

Omitted from abyssinica is Pitman's (1938b, p. 117) record of 151 ven-
trals and 46 subcaudals, as the latter is so much higher than any other
Kigezi specimens and may be based on a native's count.

Measurements. Largest C? (M. C. Z. 4S354) measures 332 (281 +
51) mm.; largest 9 (M. C. Z. 48346) measures 369 (332 4- 37) mm.

Pitman (1936, p. 62) is mistaken in thinking that Sternfeld's meas-
urements refer to the Bukoba snake. Throughout his papers on the
Fauna der deutschen Kolonien, Sternfeld furnishes condensed transla-
tions of Boulenger's descriptions, renders millimetres into centimetres,
then adds such localities as Bukoba from which he has material. The
390 mm. snake, whose large size surprised Pitman, was a typical
lutrix lutrix from South Africa.

'144 fide Boulenger.
= 21 fide Boulenger.

loveridge: African reptiles 281

Breeding. The following records of developing eggs were made.
Nyakabande:
January 27, 1939, a 9 held uncounted eggs with large embryos.

9 "11 eggs measuring 14 x 7 mm.

9 " 10 " " 8 x 5 mm.

9 " 7 " " 12x9 mm.

Diet. Slugs were present in the stomachs of four Nyakabande snakes.

Duberria lutrix shiraxa (Boulenger)

Homalosoma shiranum Boulenger, 1894, Cat. Snakes Brit. Mus., 2, p. 276.
pi. xiii, fig. 1 : Shire Highlands, Nyasaland.

7 (M. C. Z. 44114 & Vienna Mus.) Ugano, T. T. 1935-6. (H. Zerny).

Distribution. Ugano is in the Matengo highlands just east of Lake
Nyasa at about 1400-1600 metres. As these snakes constitute the
third known series of the race, their data is furnished below. All other
references, other than mine (1933h, p. 241) and part of Bogert's (1940,
p. 39) should be transferred to abyssinica, except of course those based
on the type.

Variation. Midbody scale-rows 15; ventrals 132-148; anal entire;
subcaudals 29-41; labials 6, the third and fourth (9 sides) or second,
third and fourth (4 sides) entering orbit, or 7, the third, fourth and
fifth entering (1 side); loreal absent; preocular 1; postocular 1; tem-
porals 1 + 2.

Measurements. Largest, a 9 (M. C. Z. 44114), measures 379 (326 +
53) mm.

Prosymna ambigua stuhlmanni (Pfeffer)

Liginirostra stuhlmanni Pfeffer, 1893, Jahrb. Hamburg. Wiss. Anst., 10. p. 78,
pi. i, figs. 8-10: Usambara, Tanganyika Territory.

d* c? (M. C. Z. 48355-6) Amboni Estate, T. T. 19.vi.39.

Variation. Midbody scale-rows 15; ventrals 127-129; anal entire;
subcaudals 29-28; labials 6, the third and fourth entering the orbit;
preocular 1; postoculars 2; temporals 1 + 2.

Coloration. One is white beneath, the other black.

Measurements. Both cTcT measuring the same, viz. 189 (161 + 28)
mm.

Habitat. Ploughed up by tractor in sisal plantation.

282 bulletin: museum of comparative zoology

Dasypeltis scaber medici (Bianconi)

Dipsas medici Bianconi, 1859, Mem. Accad. Sci. Bologna, 10, p. 501, pi. xxvi:

Mozambique.
Dasypeltis scaber var. fasciolata Peters, 1868, Monatsb. Akad. Wiss. Berlin,

p. 451: Zanzibar.
Dasypeltis lineolatus Peters, 1878, Monatsb. Akad. Wiss. Berlin, p. 206:

Kitui, Ukamba, Kenya Colony.
Dasypeltis elongata Mocquard, 1888, Mem. Cent. Soc. Philom. Paris, p. 131,

pi. xii, fig. 2: Zanzibar.
Dasypeltis scabra var. bianconii Med. (sic) Boettger, 1893, Zool. Anz., p. 387:

(lapsus for var. medici Bianconi).

9 (M. C. Z. 48384) Mikindani, T. T. 21.iv.39.
cf. 4 9 (M. C. Z. 48385-9) Nchingidi, T. T. 18.V.39.

Native name. Kwararu (Kisambara).

Variation. Midbody scale-rows 23-25; ventrals 242-249; anal en-
tire; subeaudals 73-86; preoculars 1-2; postoculars 2.

Measurements. 9 (M. C. Z 48384) measures 750 (625 + 125) mm.

Breeding. At Mikindani, April 21, above 9 held 6 eggs measuring
24 x 8 mm. All the Nchingidi specimens were young, four were under
268 mm. in total length and appeared to be from the same litter, two
being taken in my tent and two under a nearby log at the forest-edge.

Remarks. Some remarks appear called for by my action in reviving
the name medici. I had already been puzzled by, and commented
upon, the high ventral and subcaudal counts of certain East African
'scaber,' but the position was masked in part by non-separation of
the sexes, and further confused by the occurrence in West Africa of
snakes with a similarly high range of scale-counts, though those occu-
pying the centre of the continent have the low counts usually asso-
ciated with typical scaber. A somewhat analogous position occurs
with the forms of the genus Thrasops in the sense that the eastern
and western forms are more closely related than those now occupying
the central part of the range.

Bogert (1940, p. 86) has recently clarified the position regarding
the status of the western virgin-forest form macrops, now made a
synonym of fasciatus, by advancing sound reasons for treating it as a
race of scaber. Its not altogether unexpected discovery in western
Uganda has caused me to attempt an elucidation of the status of the
various races occurring in East Africa. I had been long aware that in

loveridge: African reptiles 283

the montane forests of this region only a uniformly black, brown, or
pinkish brown form (palmarum) occurred, which form I now find
occupies the highland plateau of the Central Lake region also. In
general throughout this region the rhomboid form appears to be
found along rivers and in the steppe and savanna. Whether it can
be recognized as an ecological form separable from palmarum only by
colour — for the squamation is the same — remains to be seen.

D. s. medici is an eastern coastal-belt, or coastal-plain, race asso-
ciated with red laterite soils which are characteristic of this region.
It ascends up to 3,000 feet (Amani, Usambara mountains, and Nchin-
gidi, Rondo Plateau) where the red soils are to be found. The form
itself, is reddish, uniform, or more usually cross-striped with narrow
black striae.

At some not too distant date I hope to study the entire literature
of DasypcUis, but in the meantime offer the following tentative key
based on nearly a hundred specimens in the collection of the Museum
of Comparative Zoology. It is not to be expected that every indi-
vidual reptile will confine its movements to its allocated sphere; in
general, however, the arrangement appears to reflect the position very
closely, and such apparent contradictions as have been investigated
proved to have erroneous data or to be incorrectly sexed. I regard
the South African inornatus (inc. unicolor) as a recognizable south-
eastern race distinguished from scaber scaber by its high subcaudal
count and colour, the latter leading to its confusion with palmarum.

Ecological Races of Dasypeltis occurring in East and Central Africa

4. Subcaudals of males more than 77 (78-94), of females more than 62

(63-87) 2.

Subcaudals of males less than 75 (51-74), of females less than 66
(44-65) 3.

2. Above pinkish brown, uniform in northeast, usually with numer-
ous, narrow, dark, sometimes light-edged, crossbands, n-shaped
anteriorly, posteriorly transverse and coalescing on dorsum with
a more or less distinct vertebral band (On reddish laterite soil
of coastal plain up to 3,000 feet, from Somaliland to Mozam-
bique) s. medici

Above yellowish olive to pinkish brown, with numerous, broad,
dark, vertical, stripes on flanks which may alternate, or coalesce,
with blotches on dorsum (Virgin forests of West Africa from
Sierra Leone to extreme western border of Uganda) . . s. fasciatus

284 bulletin: museum of comparative zoology

3. Above pinkish brown, dark brown, or black, uniform, except near
range of fasciatus when there may be some transverse dark lines
anteriorly (Highlands of Central African Lake Region, in East
in montane forests of Elgon, Kenya and Kilimanjaro, possibly
Ethiopian form distinct; transAfrica in equatorial region)

s. palmarum

Above pale sandy or olive brown, with numerous, dark, stripes or
blotches on flanks alternating with a dorsal series of large rhom-
boidal or squarish spots which sometimes coalesce to form a
zigzag vertebral band (Savanna areas and river banks in Uganda,
Kenya and Tanganyika; ranging from Sudan to Cape) .s. scaber

Dasypeltis scaber fasciatus Smith

Dasypeltis fasciatus A. Smith, 1842, Illus. Zool. S. Africa, 3, footnote to pi.

lxxiii: Sierra Leone.
Dipsas carinatus Hallowell, 1845, Proc. Acad. Nat. Sci. Philadelphia, p. 119:

Africa, later given as Liberia.
Rachiodon scaber var. snbfasciatus Jan (nomen nudum), 1863, Elenco Sist.

Degli Ofidi, p. 106: Gold Coast.
Dasypeltis macrops Boulenger, 1907, Ann. Mag. Nat. Hist. (7), 19, p. 324:

Efulen, French Cameroon.

4$ 9 (M. C. Z. 48357-8) Bundibugyo, U. 21.xii.38.

Distribution. Bundibugyo lies northwest of the Ruwenzori Moun-
tains, i.e. is in the Ituri Forest region and probably is the only area
of LTganda where this western forest form (of which macrops is a syn.)
occurs. These specimens constitute the first LTganda records.

Native name. Bankei (Luamba).

Variation. Midbody scale-rows 21-25; ventrals 239-248; anal en-
tire; subcaudals 63-69; preoculars 1-2; postoculars 2.

This form is sharply distinguished from palmarum and scaber
by its higher ventral and subcaudal count, but I cannot find the larger
eye a character constant enough to be of value. It will be recalled that
brevicejis Peters from southeast Africa was separated from scaber on
the same basis. Large eyed individuals, or groups of such, appear to
crop up throughout the range of this widespread species.

Measurements. Largest 9 (M. C. Z 48358) measures 862 (730
+ 132) mm.

Breeding. One 9 held small ova, another 9 eggs measuring 16 x 6
mm., a third 5 eggs, each measuring about 43 x 10 mm.

loveridge: African reptiles 285

Dasypeltis scaber palmarum (Leach)

Coluber Palmarum Leach, 1818, in Tuckey, Explor. River Zaire, App. p. 408:

Embomma, i.e. Boma, Belgian Congo.
fRachiodon Abyssinus Dumenl & Bibron, 1854, Erpct. Gen., 7, p. 496:

Ethiopia.
Rachiodon scaber var. unicolor Jan, part (nomen nudum), 1863, Elenco Sist-

Degli Ofidi, p. 106: Gold Coast.
Dasypeltis scabra var. atra Sternfeld, 1912, Wiss. Ergeb. Deut. Zentral-

Afrika-Exped. 1907-1908, 4, p. 272: Virgin forest behind boundary

mountains northwest shore of Lake Tanganyika, Belgian Congo.

9 (M. C. Z. 48359) Mihunga, U. 29.xii.38.
<?, 3 9 9 (M. C. Z. 48360-1) Nyakabande, U. 28.L39.
c? 9 (M. C. Z. 48362-3) Mushongero, U. l.ii.39.
9 (M. C. Z. 48364) Kisenyi, B. R. 10.ii.39.
16 cf cf , 9 9 9 (M. C. Z. 48365-83) Idjwi Id., B. C. ii.39.

Native names. Utugu (Lulega, for black examples), kubajoka (Lu-
lega, pinkish brown specimens).

Variation. Midbody scale-rows 21-25; ventrals 203-236; anal en-
tire; subcaudals 51-73; preoculars 1-2; postoculars 1-2.

Coloration. Both uniformly black, brown, and pinkish red forms
occur at upper Mulinga on Idjwi Island, Lake Kivu. Two cf1 &
(M. C. Z. 48374-5) tend towrards the colouring of fasciata and it was
interesting to note that these two snakes were the only ones with a
subcaudal count above 69. The coloration in life of one unusual cf
was noted as follows :

Above, dull copper, red, inverted, chevron-shaped markings on
occiput, a chain of spots along vertebral line corresponding to a series
of vertical stripes on flanks; flanks light coppery hue. Below, chin and
throat white, rest of undersurface uniformly pink.

Measurements. Largest cf (M. C. Z. 48374) measures 623 (513 -f-
110) mm., largest 9 (M. C. Z. 48360) measures 775 (677 + 98) mm.

Breeding. Not one was gravid.

Enemies. Three egg-eaters were recovered from the stomachs of
cobras (Naja melanoleuca) on Idjwi Island.

286 bulletin: museum of comparative zoology

Dasypeltis scaber scaber (Linnaeus)

Coluber scaber Linnaeus, 1758, Syst. Nat., ed. 10, 1, p. 223: Indiis.

Anodon typus A. Smith, 1829, Zool. Journ., 4, p. 443: Near Cape Town, Cape

Province, Union of South Africa.
Dasypeltis scaber var. capensis Peters, 1864, Monatsb. Akad. Wiss. Berlin,

p. 644, footnote: Cape of Good Hope.
Dasypeltis scaber var. mossambicus Peters, 1864, Monatsb. Akad. Wiss. Berlin,

p. 644, footnote: Boror and Tete, Mozambique.
Dasypeltis scaber var. breviceps Peters, 1864, Monatsb. Akad. Wiss. Berlin,

p. 645, footnote: Kaffirland.

c? (M. C. Z. 47815) Busingiro, U. vii.38. (C.R.S.Pitman).

Variation. Midbody scale-rows 27; ventrals 196; anal entire; sub-
caudals 54; preocular 1; postoculars 2.

Geodipsas vauerocegae Tornier

Geodipsas vauerocegae Tornier, 1902, Zool. Anz., 25, p. 703: Usambara Moun-
tains, Tanganyika Territory.

J 9 (M.C.Z. 48390-1) Magrotto Mtn., T.T. 8.vii.39.

Affinities. Bogert's (1940, p. 38) action in placing Geodipsas in
juxtaposition to the aglyphous Xeustcrophis appears thoroughly
sound. In external appearance, in the forest, I have often had a cer-
tain amount of difficulty in distinguishing vauerocegae from X. olivaceus
uluguruensis. Until the relationships of all the aglyphous and opis-
thoglyphous genera of colubrines have been settled, however, I prefer
to adhere to the older grouping for convenience of reference.

Variation. Midbody scale-rows 17; ventrals 128-131; anal entire;
subcaudals 38 pairs in d\ tail of 9 lacks tip; labials 7, the third and
fourth entering the orbit; preoculars 1; postoculars 2; temporals
1 + 2.

Measurements. The a", a juvenile evidently just born or hatched,
measures only 125 (105 + 20) mm.

Habitat and defense. The adult was found among drift leaves
between the buttress roots of a giant tree in the dark forest. Her eyes
were opaque as if about to slough. On being picked up she emitted
cloacal secretions which smelt just like those of Xatri.v n. natrix. The
young male was in leaf mould beneath a nearby log. As I took it up
by the middle it flattened out to the thinness of stout paper and held
itself thus distended and rigidly immobile.

loveridge: African reptiles 287

Boiga blandingii (Hallowell)

Dipsas Blandingii Hallowell, 1844, Proc. Acad. Nat. Sci. Philadelphia, p. 170:
Liberia.

9 (M. C. Z. 47812) Kome Id., Lake Victoria, U. 24.vi.38.
(C.R.S. Pitman).

Distribution. Kome Island, near Entebbe, at the north end of the
lake, should not be confused with the island of the same name near
Mwanza, at the south end. The precise locality where the snake was
killed is Kibanga, on Kome.

Variation. Midbody scale-rows 23; ventrals 255; anal entire; sub-
caudals 123; labials 9, the fourth, fifth and sixth entering the orbit;
preoculars 1-2; postoculars 2; temporals 2 + 2.

Measurements. 9 measures 1880 (1440 + 440) mm.

Boiga pulverulenta (Fischer)

Dipsas pulverulenta Fischer, 1856, Abhand. Natur. Ver. Hamburg, 3, p. 81, pi.
iii, figs, la-lc: Edina, Grand Bassa County, Liberia.

9 (M. C. Z. 48392) Bundibugyo, U. 24.xii.38.

Distribution. The only other Uganda record for this species is that
of Pitman (1938, p. 134) who took one in 1933 in the Mabira Forest.

Variation. Midbody scale-rows 19; ventrals 256; anal entire; sub-
caudals 118; labials 8, the third, fourth and fifth entering the orbit;
preocular 1; postocular 2; temporals 2 + 2.

Measurements. 9 measures 912 (720 + 192) mm.

Dipsadoboa unicolor Giinther

Dipsadoboa unicolor Giinther, 1858, Cat. Snakes Brit. Mus., p. 183: West
Africa.

<? (M. C. Z. 48394) Mabira Forest, U. 8.xi.38.
& (M. C. Z. 48395) Mihunga Swamp, U. 17.L39.

Distribution. These snakes constitute the first records of this west-
ern forest form in Uganda.

Variation. Midbody scale-rows 17; ventrals 190-197; anal entire;
subcaudals 57-66, single ; labials 8-9, the fourth and fifth, or fourth-
fifth and sixth entering the orbit; preocular 1; postoculars 2; temp-
orals 1 + 2.

288 bulletin: museum of comparative zoology

Coloration. In life. Above, bright olive green shading to plumbeous
on tail. Below bright yellow on throat and ventrals, black beneath
tail (cf adult).

Above, soft olive green, becoming rather abruptly plumbeous on
tail. Below, hinder part of throat white, ten anterior ventrals tinged
with yellow, rest of under surface, including outer scale-row, uni-
formly blue. Vertical pupil is black, the iris a gray-greenish white,
(cf juvenile, its three umbilical ventrals still showing suture).

Measurements. Adult c? (M. C. Z. 48394) measures 853 (700 4-
153) mm., the juvenile d> only 356 (303 4- 53) mm.

Diet. A frog (Phrynobatrachus graueri) in stomach of juvenile.

Habitat. The young snake was taken on a tangle of vines growing
on a wild banana in the swamp below Mihunga ridge, it did not make
use of its tail when placed upon a twig.

Crotaphopeltis hotamboeia hotamboeia (Laurenti)

Plate 3, fig. 1.

Coronella hotamboeia Laurenti, 1768, Syn. Rept., p. 85: "India orientali," i.e.
Africa.

1 (M. C. Z. 47802) Lira, Lango, U. iv-viii.38 (C.R.S.Pitman).

1 (M. C. Z. 47811) Busiro, Kome Id., U. 17.vi.38 (C.R.S.Pitman).

1 (M. C. Z. 47816) Busingiro, Budongo F., U. vi.38 (C.R.S.Pitman).

1 (M. C. Z. 47831) Gulu, Acholi, U. iv-viii.38 (C.R.S.Pitman).

1 (M. C. Z. 48395) Butiaba, L. Albert, U. 5.xii.38.

3 (M. C. Z. 48396-7) Ujiji, L. Tanganyika, T. T. 13.iii.39.

5 (M. C. Z. 48398-401) Kitaya, Ruvuma River, T. T. 28.iii.39.
1 (M. C. Z. 48402) Mikindani, s.e. coast, T. T. 18.iv.39.

3 (M. C. Z. 48403-5) Mbanja, near Lindi, T. T. 27.iv.39.
1 (M. C. Z. 48406) Lake Rutamba, near Lindi, T. T. 8.V.39.

6 (M. C. Z. 48407-9) Nchingidi, Rondo Plateau, T. T. 18.V.39.

1 (M. C. Z. 48410) Siga Caves, near Tanga, T. T. 15.vi.39.

2 (M. C. Z. 48411-2) Amboni Estate, Tanga, T. T. 19.vi.39.

7 (M. C. Z. 44115 & Vienna Mus.) Ugano, T. T. 1935-6 (H. Zerny).

Native names. Chijamitela (Kimwera); nowlendi (Kimakonde at
Kitaya, but thought by both Konde and Yao at Kitaya to be the
young of Psammophis s. sibilans).

Variation. Midbody scale-rows 19; ventrals 147-178; anal entire;
subcaudals 33-48; labials 7-9, the third and fourth, or third, fourth
and fifth, or fourth and fifth, or fourth, fifth and sixth entering the
orbit; preocular 1, or 2 in three snakes; postoculars 2, or 1 on one

loveridge: African reptiles 289

side of one snake ; temporals 1 + 1 or 1 + 2 ; loreal noticeably longer
than high in the Butiaba and one Ujiji snakes which, however, are
not like degeni in any other respect.

Measurements. Largest d" (M. C. Z. 48412) measures 608 (513
+ 95) mm., largest 9 (M. C. Z. 47831) measures 533 (460 + 73) mm.

Sexual dimorphism. Though in each locality, when taken separ-
ately, males have a higher subcaudal count, in the series as a whole
no such separation is possible, despite a careful rechecking of the
sexual diagnosis. Ventrals in c?c? are 150-172, in 9 9 147-178;
subcaudals in c? <? are 35-48, in 9 9 33^5.

Breeding. At Kitaya, March 28, one 9 held very small ova; on
March 31, 5 eggs measuring 22 x 12 mm., thought to be those of this
species, were found beneath a heap of vegetable rubbish. At Siga
Caves, June 15, a 9 held 6 eggs measuring 27 x 11 mm.
Diet. Stomach contents at Ujiji consisted of a gecko (H. mabouia)
and frog (Hypcrolius argentovittis) in one snake, a toad {B. r. regu-
lar is) in another; at Kitaya, a green shield-bug of a species which
emits a pungent odour; at Ugano, Mbanja and Nchingidi, four frogs
(Arthroleptis s. stenodactylus and s. lonnbergi) in three snakes.

Trimerorhinus tritaeniatus multisquamis Loveridge

Trimerorhinus tritaeniatus multisquamis Loveridge, 1932, Proc. Biol. Soc.
Washington, 45, p. 84: Nairobi, Kenya Colony.

<? 9 9 (M. C. Z. 48413-5) S. Kinangop Plateau, K. C. 27.X.38.

Taxonomy. As pointed out by Mertens (1937c, Copeia, p. 70),
the laudable attempt by Stejneger (1936b, Copeia, p. 139) to avoid
complications regarding the genus Cerastes by designating a synonym
as genotype, was rendered nugatory by the prior action of Fejervary
(1923, Zool. Anz., p. 172) in designating rhombcatiis. Trimerorhinus
therefore becomes a synonym of Cerastes but the employment of a
name, so long associated with the horned desert vipers of North
Africa, which occurs so frequently both in medical and popular litera-
ture appears to me to call for action by the International Committee
on Nomenclature on the grounds that 'no name shall be employed
if the result will be greater confusion than uniformity' (c.f Crossland,
1939, Nature, p. 942).

Variation. Midbody scale-rows 17; ventrals 163-166; anal divided;
subcaudals 50-55; labials 8, the fourth and fifth entering the orbit;
preoculars 1; postoculars 2; temporals 2 + 3.

290 bulletin: museum of comparative zoology

Breeding. Both 9 9 held eggs, apparently 7, measuring 16x5
mm. That such small snakes should be breeding appears to indicate
that the form is dwarfed in the alpine zone at 10,000 feet, where the
uncongenial climatic conditions restrict the hours of feeding and
basking.

Measurements. The c? (M. C. Z. 48413) measures 490 (405 + 85)
mm., the larger 9 (M. C. Z. 48414) measures 516 (426 4- 90) mm.

Rhamphiophis rubropunctatus (Fischer)

Dipsina rubropunctatus Fischer. 1884, Jahrb. Hamburg. Wiss. Anst., 1, p. 7,
pi. i, fig. 3: Near Arusha, Tanganyika Territory.

d" c? (M. C. Z. 48416-7) Amboni Estate, nr. Tanga, T. T. 20.iv.39.

Variation. Midbody scale-rows 19; ventrals 220; anal divided;
subcaudals 140 & M; labials 8, the fourth and fifth entering the orbit;
preoculars 2; postoculars 2; temporals 2 + 3 and 2 + 5.

Measurements. Larger c? measures 1492 (1000 + 492) mm., the
other, though 15 mm. longer in body, lacks the tip of its tail.

Habitat. Disturbed by a tractor distributing piles of rotting vege-
tation in a cleared sisal plantation.

Rhamphiophis oxyrhynchus rostratus Peters

Rhamphiophis rostratus Peters, 1854, Monatsb. Akad. Wiss. Berlin, p. 624:
Tete; Mesuril and Quitangonha, Mozambique.

1 (M. C. Z. 48418) Mikindani, T. T. 21.iv.39.
1 (M. C. Z. 48419) Mbanja, Lindi, T. T. 4.V.39.
1 (M. C. Z. 48420) Siga Caves, T. T. 15.vi.39.

Native name. Ninhyongolihanga (Kimakonde at Mbanja).

Taxonomy. In view of Pitman's finding both the western oxyrhyn-
chus and the eastern rostratus in Uganda, a subspecific recognition of
their close relationship appears advisable.

Variation. Midbody scale-rows 17; ventrals 164-176; anal divided;
subcaudals 99 & M: labials 8-9, the fifth entering the orbit; preoculars
2-3; postoculars 2-3; temporals 2 + 3 and 3 + 4.

Measurements. Size of all three moderate, the two larger snakes lack
tips to their tails.

Diet. Foot of a small mammal in Mbanja snake.

Habitat. I captured the Siga snake beneath the debris of a collapsed
hut at the edge of a large crocodile-inhabited swamp; the Mbanja
reptile was taken on the landing field.

loveridge: African reptiles 291

Dromophis lineatus (Dumeril & Bibron)

Dryophylax lineatus Dumeril & Bibron, 1854, Erpet. Gen., 7, p. 1124: White
Nile, Africa.

1 (M. C. Z. 47806) Lira, Lango, U. iv-viii.38 (C.R.S.Pitman).

2 (M. C. Z. 47832) Gulu, Acholi, U. iv-viii.38 (C.R.S.Pitman).

Variation. Midbody scale-rows 17; ventrals 145-152; anal divided;
subeaudals 91 & M; labials 8, the fourth and fifth entering the orbit;
preocular 1; postoculars 2; temporals 1 + 2.

The data from these specimens were utilized in the recently published
(1940c, Bull. Mus. Comp. Zool., 87, pp. 1-69) revision of this genus.

Psammophis sibilans sibilans (Linnaeus)
Coluber sibilans Linnaeus (part), 1758, Syst. Nat., ed. 10, 1, p. 222: "Asia."

1 (M. C. Z. 47857) Butiaba, U. 29.xi.38.

5 (M. C. Z. 47858-62) Bundibugyo, U. 21.xii.38.

2 (M. C. Z. 47863-4) Kitaya, T. T. 25.iii.39.

3 (M. C. Z. 47865-7) Mikindani, T. T. ll.iv.39.
1 (M. C. Z. 47868) Nchingidi, T. T. 25.V.39.

Distribution. Also seen at LTjiji, Mbanja, and Amboni Estate,
Tanganyika; and eight donated by Captain Pitman from Lira, Lango;
Serere, Teso; and Gulu, Acholi, LTganda.

Native names. Sebusaru (Lutoro); nachungu (Kimakonde and
Kiyao).

Coloratimi. Usually Uganda snakes have well-developed lateral
stripes on their dusky gray bellies as is commonly the case with speci-
mens from the Central Lake Region. See revision (Loveridge, 1940c,
p. 38) for further comments.
Breeding.

Butiaba, November 29, 9 held ? eggs measuring 17 x 11 mm.

Bundibugyo, December 21, 9 " 9 " " 15 x 9 mm.

Diet, etc. At Mbanja I was sitting on a low veranda, my feet upon
the ground, when there was a rushing sound in the foot-high grass be-
side me, and a s.kink (Mabuya v. varia) dashed out and into another
patch on my left. As my eyes swung back from following its flight I was
aware of, rather than saw, the arrival of a four-foot hissing sand snake
which halted the pursuit, its head poised nine inches from the ground,
within a foot of me. Before I could take action, it turned, and, like
a flash, disappeared into the grass from which it had emerged: I
searched for it immediately but in vain.

292 bulletin: museum of comparative zoology

At Nchingidi I passed close by a large (1336 mm.) snake as it lay
basking beside a narrow path. Returning, I picked it up by the back of
the neck without it offering any serious resistance: no stick was em-
ployed. The reptile was, in fact, almost moribund, yet I found no
nematodes in its stomach — only the tail of a Sundevall's skink Riopa
sundevallii) and a little fur of some mammal (? Crocidura); nor had it
suffered any injury.

Parasites. Numerous nematodes (Ophidascaris sp. probably 0. mom-
basica) in the Butiaba snake.

Enemies. A large sand-snake, as well as two P. s. sudanensis, recov-
ered from the stomach of an eagle (Cireaetus einerens) at Amboni,
near Tanga.

Psammophis subtaexiatus sudaxexsis Werner

Psammophis subtaeniatus var. sudanensis Werner, 1940, Denks. Akad. Wiss.
"\Yien. 96, p. 504: Kadugali, Anglo-Egyptian Sudan.

1 (M. C. Z. 47869) Kitaya, Ruvuma R., T. T. 3.iv.39.
7 (M. C. Z. 47870-6) Mikindani, T. T. 10-ll.iv.39.

1 (M. C. Z. 47877) Nchingidi, T. T. 13.iv.39.

2 (M. C. Z. 47878-9) Amboni Estate, T. T. 20.vi.39.

Distribution. Also occurs at Mbanja, near Lindi, southeastern T. T.

Native name. Nam (Kimakonde and Kimwera).

Variation. Embodied in the recently published revision of the
genus Psammophis (1940c, Bull. Mus. Comp. Zool., 87, pp. 1-69).

Diet. A frog (Arthroleptis s. stenodactylus) in each of two Mikindani
snakes.

Enemies. Three adults recovered from the stomach of one eagle
(Cireaetus cinereus) shot at Mikindani, and two from another eagle of
the same species killed on Amboni Estate, near Tanga.

Thelotorxis kirtlaxdii kirtlaxdii (Hallowell)

Leptophis Kirtlandii Hallowell, 1844, Proc. Acad. Nat. Sci. Philadelphia, p. 62:
Liberia.

c? (M. C. Z. 48421) Bundibugyo, U. 21.xii.38.

Native name. Mbeya (Luamba).

Remarks. Boulenger (1896d, p. 186) separated his material into
two groups on the basis of certain colour characters. Boettger in 1913,
and Mertens in 1937, rightly treated the southeastern form as a variety
or race under the name of T. k. capensis Smith. More recently Bogert

loveridge: African reptiles 293

(1940, p. 70) raised the latter to specific rank. This disposition, how-
ever, cannot be accepted in view of the intermediate nature of most
East African specimens which renders their allocation to one race or
the other almost arbitrary.

If, after examination of a topotype Liberian kirtlandii and topotype
Natal capensis, one applies the differential characters cited by Bogert
to a series of eleven snakes from Amani, Usambara Mountains, one
finds that all have immaculate crowns but only three, of which two
are young, have almost immaculate labials. Half the series have the
rostral strongly recurved (though not so strongly as in Liberia), the
rest agree with capensis. Half the series possess nasals which show
well from above, the rest are indistinguishable from capensis. Only
four of them have less than 161 ventrals, viz. 158, 158, 159 and 160.
The same variability holds good for material from the Uluguru Moun-
tains, all of which I refer to kirtlandii on the basis of the immaculate
upper surface of the head. On the other hand material from the
savanna areas and southern Tanganyika Territory are definitely much
nearer to capensis.

After examination of the 150 titles in the literature and tabulating
the available data for ventral and subcaudal counts of 93 snakes, I
have eliminated the few which are so extreme as to appear erroneous
and made appropriate allocation of the remainder to their geographical
form. The resultant ranges overlap so heavily that I reject them for
diagnostic regions; the numbers in parenthesis give the extent of
variation.

T. k. kirtlandii has ventrals 155-189 (35); subcaudals 137-175 (39).

T.k. capensis " " 147-170(24); " 130-166(37).

After omitting also the hemipeneal character advanced by Bogert, I
cannot improve on his diagnostic characters but amend the range. The
two forms may be separated as follows, though it should be borne in

MIND THAT TANGANYIKA AND ANGOLA ARE AREAS OF INTERMEDIATES.

Rostral and anterior ends of nasals broadly visible from above; crown
of head immaculate, labials more or less immaculate, neck cross-banded.
Inhabits forested areas of tropical Africa from Liberia to northern
Angola, east to southern Somaliland and1 central Tanganyika Terri-
tory k. kirtlandii

Rostral and anterior ends of nasals narrowly visible from above ; crown
of head speckled with black, labials heavily speckled with black,

1 An occasional lowland specimen in the Voi region, southeast Kenya may preponderate in
capensis attributes.

294 bulletin: museum of comparative zoology

neck not cross-banded though black lateral blotches usually present;
range: savanna areas of semitropical Africa from central Angola and
central Tanganyika Territory south to Natal k. capensis

Diet. An arboreal lizard {Agama atricollis) in its stomach, a 9
snake from Buta, Bas Uele, Belgian Congo, had swallowed a skink
(Mabuya m. maculilabris) and then two large nestling weavers (Sper-
mophaga r. ruficapilla) which Dr. J. P. Chapin kindly identified for
me.

Thelotornis kirtlandii capensis Smith

Thelotornis capensis A. Smith, 1849, Illus. Zool. S. Africa, 3, App., p. 19:
Kaffirland and the country towards Port Natal.

<? 9 (M. C. Z. 48422-3) Kitaya, T. T. 28.iii.39.

9 (M. C. Z. 48424) Mikindani, T. T. 13.iv.39.
d" 9 (M. C. Z. 48425-6) Nchingidi, T. T. 18.V.39.

Distribution. For distribution of this form, see remarks above.

Native names. Lukukuti (Kiyao); lukukutu (Kimakonde) ; likukutu
(Kimawiha). These small variations were carefully checked and dis-
cussed with the elders of these tribes.

Variation. Midbody scale-rows 19; ventrals 154-159; anal divided;
subcaudals 137-163; labials 8, the fourth and fifth, or third, fourth and
fifth, entering the orbit; preocular 1; postoculars 3; loreals 2.

Measurements. The largest, a cT (M. C. Z. 48425), measures 1453
(875 + 578) mm.

Diet. Feathers of a finch or weaver in a Kitaya snake, toads (Brevi-
ceps mossambicus) in both Nchingidi specimens, the smaller having
eaten a very young toad. The presence of burrowing toads in an
arboreal snake bears out the observation that the bird snake frequently
descends to the ground.

Dispholidus typus (Smith)

Bucephalus typus A. Smith, 1829, Zool. Journ., 4, p. 441: Old Latakoo, South
Africa.

d1 juv. (M. C. Z. 48427) Mabira Forest, U. 12.xi.38.

Native name. Kalwekalwe (Luganda).

Variation. Midbody scale-rows 19; ventrals 188; anal divided; sub-
caudals 108; labials 7, third and fourth entering the orbit.

Habitat. This pinkish brown juvenile was entwined among some
coffee berries on a tree growing at the forest edge, Mubango.

loveridge: African reptiles 295

Calamelaps unicolor warreni Boulenger

Calamelaps warreni Boulenger, 1908b, Ann. Natal Mus., 1, pp. 230, 234, fig. 3:

Kosi Bay, Zululand.
Calamelaps mellandi Boulenger, 1915a, Proc. Zool. Soc. London, p. 214:

Chirini Island, Lake Bangweulu, Northern Rhodesia.

9 (M. C. Z. 48428) Mbanja nr. Lindi, T. T. 5.V.39.
9 (M. C. Z. 48429) Nchingidi, Rondo, T. T. 18.V.39.
9 (M. C. Z. 48430) Amboni Estate, T. T. 21.vi.39.

Native name. Mbitu (Kimakonde at Mbanja, but applied to amphis-
baenids also).

Variation. Midbody scale-rows 19; ventrals 193-203; anal divided;
subcaudals 17-19; labials 6, the third and fourth entering the orbit.

Remarks. Misled by the confusing records of Tanganyika Territory
— where three races occur — I (1933h, p. 260) united several forms
under the oldest name, unicolor (Reinhardt). Like Thelotornis k. kirt-
landii, however, the western unicolor occurs on the forested or recently
deforested, areas (Teita; L/sambara and Uluguru Mtns.) in the east,
but also on the coast (early German records of Peccetoni, Mombasa
and Bagamoyo, which should be reexamined). Otherwise the East
Coast form is warreni, while the Angolan race polylepis just pene-
trates the southwestern corner of Tanganyika Territory with the
single example from Tukuyu (Langenburg) recorded by Tornier
(1901a, Zool. Jahrb. Syst., 14, p. 86).

After elimination of Calamelaps pellegrini Angel, 1921 — which I
regard as a synonym of Rhinocalamus ventrimaculatus Roux, 1907—1
have revised the genus, utilizing all records and the concolor and uni-
color material in the Museum of Comparative Zoology. The following
key is offered as a result.

Key to the Species

1. Frontal as long as, or shorter than, its distance from the rostral;
temporals 1 + 1; upper labials 7; fifth lower labial largest; pos-
terior sublinguals as long as, or almost as long as, the anterior;

ventrals 133-148; range: Natal concolor (inc. mironi)

Frontal as long as, or longer than, its distance from the end of the
snout; temporal 1 only; upper labials 6 or 5; fourth lower labial
largest; posterior sublinguals often scarcely differentiated, if
distinct then much shorter than the anterior; ventrals 161-194;
unknown in Natal 2.

296 bulletin: museum of comparative zoology

2. Midbody scales in 21 rows; range: Angola and Transvaal north

to extreme southwestern Tanganyika Territory (at Tukuyu nr.

Lake Nyasa) u. polylepis (inc. miolepis)

Midbody scales in 19 to 15 rows 3.

3. Midbody scales in 19 rows; range: Southern Rhodesia (at Em-

pandeni, wdiere it meets with polylepis) and Zululand, north to
Kenya Colony (at Ngatana, Tana River)

u. warreni (inc. mcllandi)
Midbody scales in 17 to 15 rows 4.

4. Midbody scales in 17 rows; range: Tanganyika Territory (Uluguru

Mtns.) north to Kenya Colony (at Peccatoni, fide Boettger)
west to Sierra Leone and "Guinea"

u. unicolor (inc. hildebrandtii and niangarae)

Midbody scales in 15 rows; range Portuguese Guinea (Rio Cassine,

known only from the type) u. feae

Sexual dimorphism. The marked dimorphism in the number of
subcaudals in the races of unicolor is best shown in tabular form. It
is important to note, however, that the se.ves have had to be assumed
in the ease of polylepis as authors have not furnished the sex, more-
over, though Boulenger stated that the type of feae was a cf, from
the scale-counts it would appear to be 9 .

C? c?. 9 9. cf cf 99.

C. u. polylepis 163-194, 200-212 ventrals; 27-27, 16-20 subcaudals.
C. u. warreni 161-177, 179-203 ventrals; 26-30, 17-21 subcaudals.
C. u. unicolor 164-182, 201-208 ventrals; 28-38, 21-27 subcaudals.
Cxi. feae 196. ventrals; 23 subcaudals.

Habitat. The Mbanja snake was dug up and brought in alive, the
very fat Nchingidi female was wandering on a path near camp shortly
after daybreak, the Amboni specimen was ploughed up by tractor.

Miodon and its forms

Since the turn of the century, the most valuable contribution to an
understanding of this genus is that contributed by Bogert (1940,
p. 45) who has painstakingly investigated the vexed question of its
dentition in relation to that of its synonym Cynodontophis Werner.

When, however, he refers eollaris to the synonymy of gabonensis
(Dumeril), it is largely in the sense of gabonensis Boulenger, which
appears to me to be a composite. Apparently Boulenger had no ex-

loveridge: African reptiles z\)<

amples of the striped form and the material which he referred to
gaboncnsis was, in reality, collaris.

It seems to me that within the genus we have a transAfrican (west
to east) series of species or races beginning with the Liberian 5-striped
acanthias, passing to the 3-striped g. gaboncnsis (inc. neuwiedi) and
the 2-row spotted notatus (inc. aemulans) to g. collaris (inc. fulvicollis,
caeutiens and werneri) of the Cameroon and Congo, passing into often
entirely and uniformly black christyi (inc. unicolor) of Uganda and
graueri.

When better known it may be that graueri and notatus will have
to be regarded as full species on the grounds of their lower number
of subcaudals in both sexes, particularly noticeable in the case of
graueri. Owing to the non-sexing of much of the material referred to
in the literature, the following ranges must be regarded as largely
hypothetical, but they are founded on a basis of sexed snakes.
Suggested range of scale-counts by sex, the latter sometimes assumed.

Species cfcf ventrals 9 9 ventrals 9 9 caudals dV caudals

acanthias

183-195

207-216

16-18

-22

g. gaboncnsis

219-238

-246

11-

-21

g. notatus

178-228

201-

14-

17-19

g. collaris

200-232

195-252

15-19

19-25

g. christyi

202-217

221-241

15-18

19-24

g. graueri

237-238

254-258

13

16-18

Key to the Species

1. Anal entire; dorsum with pattern of 5 parallel black lines; ventrals

183-216; subcaudals 16-22; range: Sierra Leone to Togoland

(or Nigeria, fide Angel) acanthias

Anal divided1 2.

2. Dorsum with pattern of 3 parallel black lines; ventrals 219-246;

subcaudals 11- 21; range: Gold Coast to Dahomey (with type

allegedly from Gabon, i.e. French Congo) g. gabonensis

Dorsum without pattern of parallel black lines 3.

3. Dorsum with pattern of 2 parallel series of black spots; ventrals

178-228; subcaudals 14-19; range: French Cameroon south to

French Congo g. notatus

Dorsum uniformly dark 4.

1 Allegedly entire in a Congo specimen of M. g. collaris fide Bocage (1895a, p. 126).

298 bulletin: museum of comparative zoology

4. Nape and crown pale fawn with dark mottlings ; throat white ; ven-

trals 181-252; subcaudals 15-25; range: (Togo! fide Werner)
southeastern Nigeria south to Angola east to Uganda (west of

Ruwenzori) g. collaris

Nape and crown of head entirely black, or with a sharply distinct
broad white band across the parietal region 5.

5. Top of head iridescent black like nape and dorsum; throat white

or black; ventrals 202-241; subcaudals 15-21; range: western
Congo (Poko) east to central Uganda (Mabira Forest) . g. christyi
Top of head from snout to behind eyes, black, posterior half of
head and nape pure white; ventrals 237-258; subcaudals 13-18;
range: eastern Congo (Idjwi Island, Lake Kivu) to central
Uganda (Entebbe) g. graueri

Miodon gabonensis collaris (Peters)

Microsoma collare Peters, 1881, Sitz. Ges. Naturf. Freunde. Berlin, p. 148;
Macange, Cuango = Kwango, French Equatorial Africa.

c? (M. C. Z. 48431) Bundibugyo, U. 22.xii.38.

Distribution. This constitutes the first record for collaris in Uganda,
but Bundibugyo — in the Bwamba country northwest of the Ruwen-
zori Mountains — is really a part of the Ituri Forest region.

Variation. Midbody scale-rows 15; ventrals 211; anal divided;
subcaudals 23.

Coloration. Quite typical. Head and nape light fulvous with a
few dark markings on the frontal region, dorsum irridescent plumbeous
with the edges of the scales darker. Below, pure white except for the
dorsal coloration impinging on the outer lateral edges of the ventrals.

Measurements. Total length of 9 , 237 (315 -f- 22) mm.

Miodon gabonensis graueri Sternfeld

Miodon Graueri Sternfeld, 1908, Sitz. Ges. Naturf. Freunde Berlin, p. 94:
Entebbe, Uganda. t

J> 9 9 (M. C. Z. 48432-3) Idjwi Id., B. C. 21-22.ii.39.

Distribution. These constitute the first records of this form in the
Belgian Congo; graueri being known until now only from the type.

Variation. Midbody scale-rows 15; ventrals 237-258; anal divided;
subcaudals 13-16. The c? has 237 (type had 238) ventrals, and 16
(type had 18) subcaudals. The 9 9 have 254-258 ventrals and only
13 subcaudals, indicating a slight sexual dimorphism.

Measurements. Total length of d" (to go to Mus. Congo Beige) 325
(310 + 15) mm.; of larger 9 (M. C. Z. 48432) 370 (358 + 12) mm.

loveridge: African reptiles 299

Diet. Remains of a blind snake (Typhlops b. Icstradei) in male, an
egg, possibly a lizard's, in one of the females.

Habitat. The two females were taken near the lake shore, where,
according to the native captor, the species is usually encountered.
The male, however, was killed when crossing a path about a mile
below our camp, circa 6000 feet? As the type locality, Entebbe, is also
on the lake (Victoria) shore, perhaps this race favours such an envir-
onment.

Aparallactus modestus (Giinther)

Elapops modestus Giinther, 1859, Ann. Mag. Nat. Hist. (3), 4, p. 161, pi. iv,

fig. C: West Africa.
Periaspis plumbeatra Cope, Proc. Acad. Nat. Sci. Philadelphia, p. 252:

Liberia.
Elapops {Calamaria) Petersi Jan, 1862, Arch. Zool. Anat. Fisiol., 2, p. 32:

Gold Coast.
Aparallactus boulengeri Werner, 1896, Verh. Zool. Bot. Ges. Wien, 46, p. 363,

pi. vi, figs. 6-6b: Cameroon.
Aparallactus peraffinis Werner, 1897, Verh. Zool. Bot. Ges. Wien, 47, p. 404,

pi. ii, fig. 3: Interior of Cameroon.
* Aparallactus ubangensis Boulenger, 1897, Ann. Mag. Nat. Hist. (6), 19, fig.:

Zongo, Ubangi Rapids, Belgian Congo.
Aparallactus flavitor que s Boulenger, 1901, Ann. Musee Congo (1), 2, p. 11,

pi. iv, fig. 3: Lubue, Kasai, Belgian Congo.

* Aparallactus dolloi Werner, 1902, Verh. Zool. Bot, Ges. Wien, 52, p. 346

Banzyville, Ubangi River, Belgian Congo.
*Aparallactus congicus Werner, 1902, Verh. Zool. Bot. Ges. Wien, 52, p. 346

Lingunda, Belgian Congo.
* Aparallactus Batesii Boulenger, 1907, Ann. Mag. Nat. Hist. (7), 19, p. 325

5 miles inland from Kribi, French Cameroon.
Aparallactus christyi Boulenger, 1910, Ann. Mag. Nat. Hist. (8), 5, p. 512

Mabira Forest, Chagwe, Uganda.
*Aparallactus nigrocollaris Chabanaud, 1917 (1916), Bull. Mus. Nat. Hist.

Paris, 22, p. 377, figs. 18-19: French Congo.
*Aparallactus nigrocollaris Roucheti Chabanaud, 1917 (1916), Bull. Mus. Nat.

Hist. Paris, 22, p. 378, figs. 20-21 : French Congo.
Guyomarchia unicolor Angel, 1923, Bull. Mus. Nat. Hist. Paris, 29, p. 348,

figs. 1-4: French Congo (probably from near Sangha).

* Aparallactus Grdueri Werner, 1924, Sitz. Akad. Wiss. Wien, 133, p. 42: Beni,

Belgian Congo.

c? 9 (M. C. Z. 48435-6) Mabira Forest, U. 13.xi.38.
9 (M. C. Z. 48437) Budongo Forest, U. 29.xi.38.
d" (M. C. Z. 48438) Bundibugyo, U. 22.xii.38.

* Here referred to the synonymy for the first time as a result of a revision of the genus now
in MS.

300 bulletin: museum of comparative zoology

Native name. Kileba (Luamba*).

Synonymy. I was particularly anxious to obtain topotypes of
christyi, a species which Parker (in Pitman, 1937, p. 338) recently
detected as synonymous with modest us, for I thought it might be pos-
sible to recognize it as an eastern form of this sylvicoline snake.
On the contrary, however, the generic revision which resulted showed
it necessary to add seven additional alleged species (as indicated above
with an asterisk) to the already lengthy synonymy. This 'lumping'
has not been done in any spirit of desperation, but after careful consid-
eration of the claims of each species in conjunction with a study of the
extensive series of modestus in the Museum of Comparative Zoology.
The reasons for this synonymizing will be furnished later upon publi-
cation of the revision. The only one of which there could be the slight-
est doubt is nigrocollaris (inc. roucheti) which I regard as an aberration;
it has twice been taken in Uganda by Pitman.

As long ago as 1923, Schmidt pointed out that this species has teeth
which may, or may not, be grooved, but it is only recently that Bogert
(1940, p. 43) presented the argument for merging Elapops with
Aparallactus.

Variation. Midbody scale-rows 15; ventrals 135-157; anal entire;
subcaudals 36-43 ; labials 7, the third and fourth entering the orbit.

Measurements. The larger d> (M. C. Z. 48438) measures 348 (288
+ 60) mm. and smaller 9 (M. < \ Z. 48436) 503 (435 + 68) mm.

Breeding. The latter held 7 eggs, the average size of which is about
25 x 8 mm.

Aparallactus jacksoxii (Gunther)

Uriechis Jacksonii Gunther, 1888, Ann. Mag. Nat. Hist. (6), 1, p. 325, pi. xix,
fig. E: Foot of Mt. Kilimanjaro, Tanganyika Territory.

9 (M. C. Z. 48442) Nchingidi, Rondo Plateau, T. T. 17.V.39.

Variation. Midbody scale-rows 15; ventrals 156; anal entire; sub-
caudals 44; labials 7, the third and fourth entering the orbit.

Coloration. In life. Above, head black, a six-scale wide black, trans-
verse band, edged before and behind by scale-wide bands of bright
yellow, on nape; back and tail a delicate pinkish brown. Below, bright
yellow, the lower ends of the black collar not extending on to the ven-
trals.

Measurements. Total length of this, the largest known 9 , 259
(213 + 46) mm.

loveridge: African reptiles 301

Aparallactus werxeri Boulenger

Aparallactus werneri Boulenger, 1895, Ann. Mag. Nat. Hist. (6), 16, p. 172:
Usambara Mountains, Tanganyika Territory.

<? cf (M. C. Z. 48445-6) Magrotto Mtn., T. T. l.vii.39.

Variation. Midbody scale-rows 15; ventrals 145-148; anal entire;
subeaudals 41 & M; labials 6, the second and third entering the orbit.

Coloration. In life. Above, head black, a six-scale wide black, trans-
verse band, separated from the head by a scale-wide olive band, on
nape ; back and tail olive. Below, bright lemon yellow except for throat
which is tinged with whitish.

Measurements. Total length of larger cf, 308 (255 + 53) mm.

Aparallactus capexsis uluguruexsis Barbour & Loveridge

Aparallactus uluguruensis Barbour & Loveridge, 1928, Mem. Mus. Comp.
Zool., 50, p. 132: Xyange, Uluguru Mountains, Tanganyika Territory.

d* cf 9 (M. C. Z. 48439-41) Magrotto Mountain, T. T. l.vii.39.

Remarks. A. uluguruensis was based on eleven uniformly plumbeous
specimens from the Uluguru and Usambara Mountains, a revisionary
study of the genus reveals that it differs in no essential from the pale
fawn eapensis of the savanna except in its much larger size and incolour,
moreover, small snakes intermediate in colour are to be found on the
recently deforested western Usambara, Kilimanjaro and on Mount
Mbololo. I propose to regard uluguruensis as a race of eapensis though
the latter is undoubtedly descended from the virgin forest form.

I was in error in referring .4. coneolor boulengeri to the synonymy
of uluguruensis as has been pointed out by Scortecci. At the time I
attributed less importance to the first lower labials being in contact
than I should have done.

Variation. Midbody scale-rows 15; ventrals 140-158; anal entire;
subeaudals 43-44; labials 7, the third and fourth entering the orbit.

Diet. A centipede in the stomach of the larger male.

Aparallactus capexsis capexsis Smith

Aparallactus eapensis A. Smith, 1849, 111. Zool. S. Africa, Rept., App., p. 16:
Kaffirland to the eastward of Cape Colony.

9 (M. C. Z. 48443) Ujiji, T. T. 12.iii.39.
9 (M. C. Z. 48444) Mbanja, T. T. 27.iv.39.

302 bulletin: museum of comparative zoology

Native name. Yamitera (Kimakonde).

Remarks. I now confirm my (1936J, p. 269) previous suggestion
that punctatolineatus is a synonym of capensis. Peter's nigriceps is a
good species with only 108-123 ventrals and 20-35 subcaudals, but
nigriceps of Boulenger (1896d, p. 260) is a composite of Peter's original
description and a specimen of capensis, for the latter may have 6 upper
labials, second and third entering the orbit (left side of head of Mbanja
snake) or 7 upper labials, third and fourth entering the orbit (right side
of head of Mbanja snake), the latter being the normal condition for
capensis though the former crops up throughout its range as in the
Ujiji snake and in the 9 of a pair collected on Mount Mbololo, Kenya
Colony.

In studying the genus, it was interesting to note that a block of six
western and northern species invariably possess seven upper labials
of which the third and fourth enter the orbit. On the other hand, three
eastern species (werneri, turneri and nigriceps) have six labials, the
second and third entering the orbit. Three other forms, (c. uluguru-
ensis, c. capensis and bocagii), for bocagii is scarcely more than a
western form of capensis, occupy an intermediate position normally
having seven labials but not infrequently six, resulting in confusion
with nigriceps sensu strictu.

Variation. Midbody scale-rows 15; ventrals 162-162; anal entire;
subcaudals 41-44; labials 6-7, see remarks above.

Breeding. At Mbanja, on April 27, a 9 held 2 eggs, each measuring
31 x 4 mm.

Elapsoidea guntherii Bocage

Elapsoidea guntherii Bocage, 1866, Jorn. Sci. Lisboa, 1, p. 70, pi. i, figs. 3-3b:
Cabinda, Portuguese Congo; Bissao, Portuguese Guinea.

cT (M. C. Z. 48447) Mabira Forest, U. 14.xi.38.

12 (M. C. Z. 48448-56) Magrotto Mtn., T. T. l-19.vii.39.

Native navxc. Kifutu (Kisambara).

Variation. Midbody scale-rows 13; ventrals 150-157; anal entire;
subcaudals 15-25; labials 7, the third and fourth entering the orbit
except on left side of M. C. Z. 48453, where it is 8, the fourth and fifth
entering.

Measurements. Largest c? (M. C. Z. 48448) measures 585 (548 +
37) mm.; largest 9 (M. C. Z. 48453) only 401 (377 + 24) mm.

Sexual dimorphism. 10 d71 d1 have 151-157 ventrals and 19-25 sub-
caudals while the 3 9 9 have 150-157 ventrals and 15-16 subcaudals.

loveridge: African reptiles 303

Diet. The only food present was a caecilian (Boulcngerula boulen-
geri) in the stomach of a Magrotto snake.

Enemies. Two of the males had lost the ends of their tails, perhaps
through fighting, one had a very truncated stump.

Habitat. With the exception of the Mabira snake, which I caught in
rank grass between the kitchen and back door, the other four that I
captured personally were wandering on paths towards evening, their
empty stomachs suggesting the cause for their being abroad.

Naja haje haje (Linnaeus)

Coluber haje Linnaeus, 1762, in Hasselquist, Reise Palestine, p. 386: Lower
Egypt.

d* <? (M. C. Z. 47808) Lira, Lango, U. iv-viii.38.
9 (M. C. Z. 47809) Gulu, Acholi, U. ix-x.38.

Distribution. Lt. Col. C. R. S. Pitman, to whom we are indebted
for these juvenile Egyptian cobras, tells me that the species is plentiful
on the flats along the eastern shores of Lake Albert where specimens
eight feet in length are not uncommon. The largest he has measured
was eight and a half feet.

Variation. Midbody scale-rows 21; ventrals 54-60; anal entire
(cf cf) or divided (9 ); upper labials 7, separated from the orbit by
suboculars.

Naja nigricollis nigricollis Reinhardt

Naja nigricollis Reinhardt, 1843, Dansk. Vidensk. Selsk. Skrift., 10, p. 269,
pi. iii, figs. 5 and 7: Guinea.

9 9 (M. C. Z. 48457-8) Kitaya, T. T. 29.iii & 2.iv.39.

c? 9 (M. C. Z. 48459-60) Mikindani, T. T. 20.iv.39.

c? (M. C. Z. 48461) Mbanja near Lindi, T. T. 29.iv.39.

9 (M. C. Z. 48462) Magrotto Mountain, T. T. 8.vii.39.

Native names. Liteweo (Kiyao) ; lilekela (Kimawiha) ; lipatera (Kima-
konde at Kitaya) ; lipeta (Kimakonde at Mbanja) ; siveela (Kisambara).

Variation. Midbody scale-rows 21; ventrals 183-204; anal entire;
subcaudals 48-58; labials 6, the third entering the orbit.

Coloration. The five southeastern snakes are all of the form mos-
sambica, the juvenile from forested Magrotto was uniformly black
except for white crossbars on throat.

Measurements. Largest <? (M. C. Z. 48460) measures 1228 (1007 +
221) mm.; largest 9 (M. C. Z. 48459) measures 1061 (880 + 181) mm.

304 bulletin: museum of comparative zoology

Diet. Four young rats (Rattus r. kijabius) in a Kitaya cobra, a
large shrew (Crocidura h. hirta) in the young Mbanja snake.

Parasites. Ticks (Aponomma falsolaeve) and worms at Kitaya.

Enemies. The large Kitaya 9 has lost practically its entire tail, the
stump, long since healed, terminates close behind the anus.

Defence. Mr. H. Tanner of Amboni Estate, near Tanga, tells me
that he was called upon to kill one of these cobras on the driveway
in front of his house. He shot it at close range with the result that
the charge severed head and neck from body, and the blast carried
the head and neck to where his son was standing some distance away.
The head, rearing on its stump, opened its mouth and discharged two
jets of venom at the boy.

This interesting case of reflex action may, I consider, be entirely
relied upon, as Mr. Tanner, who formerly kept cobras and other
snakes, is deeply interested in natural history. During my brief stay
on his estate a native killed a black cobra near my tent, it was not
preserved as he had chopped it in half with his bush knife (panga).

Venom. Sir Charles Belcher, the ornithologist, with whom I hunted
reptiles earlier in the trip, writes (1941) me that, when collecting at
Soysambu, near Lake Elmenteita, he encountered four bat-eared fox
(Otocyon m. nirgatus) cubs playing about a big black cobra. The head
of the latter was upraised three feet from the ground, and when Sir
Charles struck at the reptile with an ashplant he was carrying, the
snake discharged its venom full in his face from a distance of about
three feet. Fortunately his eyes were protected by the glasses he was
wearing, but he received a spray of the bitter tasting venom full in
the lips. Later, when washing, he got a fresh taste of the poison.

Naja melaxoleuca Hallowell

Naia haie var. melanoleuca Hallowell, 1857, Proc. Acad. Nat. Sci. Philadel-
phia, p. 61: Gaboon.

<? <? (M. C. Z. 48463-4) Mabira Forest, U. 15.xi.38.
& (M. C. Z. 48465) Budongo Forest, U. 22.xi.38.
c? (M. C. Z. 48466) Bundibugyo, U. 23.xii.38.
5 cf &, 2 9 9 (M. C. Z. 48467-71) Idjwi Id., B. C. 19-28.ii.39.
c? (M. C. Z. 48472) Mikindani, T. T. 20.iv.39.

Distribution. This Mikindani snake involves a mainland south-
eastward extension in range of 500 miles. Pakenham has recently
collected several on Zanzibar, where, however, they may have been

loveridge: African reptiles 305

introduced by escapees from the Wayeye 'snake-charmers' who visit
the island from time to time.

Lt. Col. Pitman has recently presented us with examples of this
cobra from Lira, Lango Province, and Gulu, Acholi Province, Uganda,
where they occur together with Naja h. haje.

Native names. Nsweela (Luganda); nchewera (Lutoro); bata (Lu-
amba); irizi (Lulega). Not distinguished from X. n. nigricollis by the
Makonde and Mawiha at Mikindani.

Variation. Midbody scale-rows 19; ventrals 197-220; anal entire;
subcaudals 59-71; upper labials 7, the third and fourth entering the
orbit, sixth largest and in contact with postoculars.

Coloration. This usually jet black forest cobra is very variable on
Idjwi Island; one cf was brown anteriorly, and black posteriorly;
some were all black, others brown mottled with black. The latter type
of coloration reaches its climax in the Mikindani snake which is light
brown peppered with darker in conformity with its savanna habitat
in this locality.

Measurements. Largest cf (Mikindani) measures 2364 (1960 + 404)
mm.; largest 9 (Idjwi Island) measures 1625 (1355 + 270) mm.

Diet. A lizard (Algiroides africanus) was present in the third-
grown cobra from Mabira; half-grown egg-eating snakes (Dasypeltis s.
palmarum) in each of three young or half-grown cobras on Idjwi
Island; two new-born rats (Lophuromys a. laticeps) in another from
the same locality!

At Budongo Forest camp I was summoned at 3:15 one afternoon to
catch a cobra which had been disturbed while swallowing a rat on a
garden path. The reptile, with rat in mouth, had darted into a long
pile of rubbish, mostly matete grass, piled against a fallen and rotting
tree trunk. Peering in, I caught a glimpse of its disappearing tail. We
cleared the rubbish, turned the log, but failed to find the cobra in the
maze of rat holes down which it had vanished.

Parasites. Ticks (Aponomma falsolacve) and a heavy infestation of
nematodes (Kalicephalus sp.; Opliidascaris naiae) was present in one of
the Idjwi Island cobras.

Habits. The six-pound cobra measuring about seven and a half feet;
was killed in the Mikindani bush by a native. The man surprised it
sunning, crept up and struck it a blow. The snake made off and climbed
a tree. The native then cut two long poles, split one and inserted
the other some little way down the cleft; with this weapon he reached
up into the tree and caught the snake in the cleft by withdrawing the
second pole. After pulling the snake down he was then in a position to

306 bulletin: museum of comparative zoology

belabour it with the free pole while holding it at a safe distance in the
fork of the other. The man told me that the snake made no attempt
to 'spit' under conditions where it certainly would have done so if
able. He brought the reptile to me under the impression that it was a
python and I had some difficulty in convincing him to the contrary.
In this connection it is worth recording that Mr. H. Tanner of Am-
boni Estate, told me that he had once run over a strange-looking cobra,
as large as a python, while motoring not far from my camp at Siga
Caves. Looking back he saw the snake in the road, rear up and spread
its hood.

Pseudohaje goldii (Boulenger)

Naia goldii Boulenger, 1895, Ann. Mag. Nat. Hist. (6) 16, p. 34: Asaba, Niger
River, Nigeria.

<? (M. C. Z. 48473) Bundibugyo, U. 22.xii.38.

Distribution. This three-pound snake, is not only the second record
for its occurrence in Uganda, but, being only 7 feet, 1 inch in length,
is surpassed by several Congo snakes recorded by Bogert (1942a, p. 6)
in his important paper re-establishing Giinther's genus Pseudohaje.

Native names. Neither Bwamba nor Batoro distinguish it from
N. melanoleuca to which they apply the same names, viz. bata and
ncheiccra.

Variation. Midbody scale-rows 15; ventrals 199; anal entire; sub-
caudals 82; labials 7, the third and fourth entering the orbit.

Coloration. This is as described by Boulenger except that the an-
terior ventrals are pink in alcohol, not white.

Measurements. Total length of d\ 2175 (1730 + 445) mm.

Dendroaspis angusticeps (Smith)

Naia angusticeps A. Smith, 1849, Illus. Zool. S. Africa, Rept., pi. lxx: Natal.

5 9 9 (M. C. Z. 48474-6) Kitaya, Ruvuma R., T. T. 25-30.iii.39.

d" 9 (M. C. Z. 48477-8) Nchingidi, Rondo Plat., T. T. ll.v.39.
2 9 9 (M. C. Z. 48479-80) Magrotto Mountain, T. T. 8-12.vii.39.

Native names. Namasambi (Kiyao) ; namahamba (Kimakonde) ; ngoe
(Kisambara, but it is thought to be the adult of Philothamnus s. semi-
variegatus).

Variation. Midbody scale-rows 19-23; ventrals 111-126; anal di-
vided; upper labials 7-9, the fourth entering the orbit.

loveridge: African reptiles 307

Coloration. The smaller Magrotto 9 was a rich green, very sparsely
necked with yellow above and below; the larger Magrotto 9 was olive,
though about 300 mm. smaller than the very rich green Xchingidi cf .

Measurements. The only d1' (M. C. Z. 48477) measures 7 feet (1640
mm. from snout to anus, end of tail missing) ; the largest 9 (M. C. Z.
4S474) measures 1981 (1515 + 466) mm.

Diet. A short-tailed nestling rodent in one Kitaya snake; three
young rats in the smaller Magrotto mamba.

At Xchingidi my attention was attracted by the vociferous calling
of a bulbul (Phyllastrcphus flawstriatus tenuirostris) in thick forest.
I shot the bird and my gunbearer ran in to recover the corpse; as he
did so the flickering of a leaf in the dense verdure from whence the
bulbul had fallen, directed my gaze to a large and beautiful green
mamba which was almost 'swimming' over the foliage. I halted it
with a charge of No. 12 from the .410, but as it seemed to be recovering
I gave it No. 8 from the second barrel in the heart while it was still
twenty feet above ground. Its stomach was empty.

Dendroaspis jamesoni kaimosae (Loveridge)

Dendraspis jamesoni kaimosae Loveridge, 1936, Proc. Biol. Soc. Washington,
49, p. 64: Kaimosi, Kakamega, Kenya Colony.

tf> (M. C. Z. 48481) Mabira Forest, U. 18.xi.38.
4c?c?8? 9 (M. C. Z. 48482-91) Idjwi Id., B. C. 18-28.ii.39.

Native names. Temankima (Luganda) ; mkubwe (Lulega).

Variation. Midbody scale-rows 15-17; ventrals 208-226; anal di-
vided; subcaudals 96-111; labials 8, rarely 7, the fourth entering the
orbit.

The tails, being uniformly black, are characteristic of the eastern
race (kaimosae), but the Lake Kivu series, coming as they do from a
geographically intermediate area, increase the overlap between the
subcaudal counts of the two races, that of the western form (jamesoni)
being 103-122.

Measurements. Largest <? (M. C. Z. 48482) measures 2087 (1592 +
495) mm.; the largest 9 (M. C. Z. 48484) only 1977 (1520 + 457)
mm., though when freshly killed this same snake measured 2052
(1580 + 472) mm. Eight of the series are over six feet in length.

Breeding. On February 22, small ova were present in the largest of
the Idjwi females; evidently it was not their breeding season.

Diet. In addition to unidentifiable rodent remains, the following

308 bulletin: museum of comparative zoology

rats and mice were recovered from the stomachs of six of the Idjwi
snakes: (1) one Rattus r. kijabius; (2) six half-grown Praomys j.
montis; (3) two Leggada a. grata; (4) four new-born Lophuromys a.
laticcps; (5) two Lophuromys a laticeps; (6) an adult Dasymys b.
medius.

Parasites. Nematodes (Kalicephalus sp.) in an Idjwi snake.

Enemies. The Mabira cf was killed on the road by a native. The
six-and-three-quarter-foot 9 by a woman who said that it chased a
rat from the forest into her hut; a third very large example was also
killed by a woman, but in the bush.

VIPERIDAE

Causus rhombeatus (Lichtenstein)

Sepedon rhombeatus Lichtenstein, 1823, Verz. Doubl. Mus. Berlin, p. 106:
No locality.

3 & & 1 9 (M. C. Z. 47837) Gulu, Acholi, U. iv-viii.38.(C.R.S.P.).

2 9 9 (M. C. Z. 47838) Lira, Lango, U. iv-viii.38. (C.R.S.P.).
2 cf cf (M. C. Z. 44117) Ugano, Matengo, T. T. 1935-6. (H.
Zerny).

Distribution. I am indebted to Lt. Col. Pitman for the Uganda
specimens; it seems strange that I did not encounter this common
night adder during the eight months of my safari, good testimony to
the fact that this species is not a forest-dweller.

Variation. Midbody scale-rows 17-19; ventrals 147-155; anal en-
tire; subcaudals 19-29 pairs; labials 6.

Causus resimus (Peters)

Heterophis resimus Peters, 1862, Monatsb. Akad. Wiss. Berlin, p. 277, pi. — ,
fig. 4: Gebel Ghule, Sennar, Anglo-Egyptian Sudan.

c? 9 (M. C. Z. 48492-3) Nyakabande, U. 27.L39.

Variation. Midbody scale-rows 21; ventrals 134-145; anal entire;
subcaudals 20-22 pairs; labials 6.

Measurements. The d" measures 491 (450 + 41) mm.; 9 meas-
ures 220 (200 + 20) mm.

Breeding. On January 27, 6 eggs measuring 36 x 16 mm. in the 9 .

Diet. A young toad (Bufo r. regularis) in her stomach.

loveridge: African reptiles 309

Causus defilippii (Jan)

Heterodon De Filippii Jan, 1862, Arch. Zool. Anat. Fisiol., 2, p. 225: Africa.

2 cf cf 2 9 9 (M. C. Z. 48494-6) Mikindani, T. T. 12.iv.39.
9 (M.C.Z. 48497) Mbanja. Lindi, T. T. l.v.39.

2o,tf29 9 (M. C. Z. 48498-501) Nchingidi, T. T. 13.V.39.

Distribution. Add Amboni Estate, near Tanga, T. T. 24. vi. 39.

Native names. Chipili (Kimawiha); lipili (Kimakonde). Neither
distinguishing it from the Puff Adder (Bitis arietans).

Variation. Midbody scale-rows 16 (M. C. Z. 48494 only) or 17;
ventrals 112-122; anal entire; subcaudals 13-15 pairs; labials 6.

Measurements. Largest, a 9 (Field Museum), measures 412 (382
+ 30) mm.; all the others were under 270 mm. in total length.

Enemies. One recovered from stomach of eagle (Circaetus cinereus)
at Amboni.

Causus lichtensteinii (Jan)

Heterodon Lichtensteinii Jan, 1859, Revue Mag. Zool., p. 511: Gold Coast.

c? 9 (M. C. Z. 48502-3) Mabira Forest, U. 19.xi.38.
2 d" & 2 9 9 (M. C. Z. 48504-7) Budongo Forest, U. 22-30.xi.38.
9 (M. C. Z. 48508) Kibale Forest, U. 9.xii.38.

Variation. Midbody scale-rows 15; ventrals 136-147; anal entire;
subcaudals 16-21 pairs; labials 6.

Measurements. Largest d71 (M. C. Z. 48504) measures 598 (540
+ 58) mm.; largest 9 (M. C. Z. 48503) measures 549 (507 + 42) mm.

Breeding. At Mubango, on November 19, 6 eggs measuring 15
x 9 mm. in 9 ; at Bisu, Budongo, on November 30, 8 eggs measuring
10 x 5 mm. in 9 ; at Kibale Forest, on December 12, 4 eggs measur-
ing 17x6 mm. in 9 •

Parasites. Nematodes (Ophidascaris sp. and Proteocephalus sp.) in
a Budongo snake.

Temperament. Our tent boy disturbed two of these velvety-green
night adders on the path between our fire and tent. When I reached
the spot the two snakes were returning from the rank vegetation bor-
dering the path. The anterior third of each was raised high, and so
flattened that for a second I thought that they were young cobras.
They were almost intertwined, their lower portions so close together
that I pinned both down with my T-ended stick. At this, one reptile
began to bite savagely at the other's neck, enabling me to seize both
together by their necks with the forceps. I supposed it was a court-

310 bulletin: museum of comparative zoology

ship performance that I had interrupted, but on examination, after
anaesthetization, found both were males (M. C. Z 48504-5) which
were presumably fighting.

Bitis arietans (Merrem)

Vipera arietans Merrem, 1820, Vers. Sj'st. Amphib., p. 152: Cape of Good
Hope.

d" (M. C. Z. 48509) Budongo Forest, U. 6.xii.38.
9 (M. C. Z. 48510) Idjwi Island, B. C. 3.iii.39.
9 (M. C. Z. 48511) Ujiji, T. T. 13.iii.39.
c? (M. C. Z. 48512) Kitaya, T. T. 31.iii.39.
c? (M. C. Z. 48513) Mikindani, T. T. 16.iv.39.
cf (M. C. Z. 48514) Siga Caves, T. T. 8.vi.39.

Native names. Viserosero (Lulega); lipili (Yao and Kimakonde);
chipili (Kimawiha). The last two names are applied to Causus also.

Variation. Midbody scale-rows 31-34; ventrals 133-141; anal en-
tire; subcaudals 13-33; labials 12-14.

Measurements. The entire series are juvenile, the smallest (M. C. Z
48514) measures only 209 (185 + 24) mm.

Habitat. At Siga I shot a wood hoopoe in a tree overhanging the
path leading to my tent; as my gunbearer stooped to pick up the
dead bird, which had fallen on the leaf-strewn path, he paused, for
right beside it was the young puff adder.

Bitis gabonica (Dumeril & Bibron)
Plate 3, fig. 2.

Echidna Gabonica Dumeril & Bibron, 1854, Erpet. Gen., 7, p. 1428, pi. lxxx b:
Gaboon.

9 (M. C. Z. 48415) Budongo Forest, U. 26.xi.38.
<? 9 (M. C. Z. 48416) Nchingidi, T. T. 12.V.39.
1 d* 2 9 9 (M. C. Z. 48417) Magrotto Mtn., T. T. l-9.vii.39.

Distribution. An even more southeasterly record in Tanganyika
Territory than Nchingidi, Rondo Plateau, is that of Lulindi, where,
according to Dr. L. Stirling of the U. M. C. A., it occurs in a patch of
forest at an altitude not much over a thousand feet.

Native name. Moma (Kisambara).

Variation. Midbody scale-rows 38-43; ventrals 128-140; anal en-
tire; subcaudals 20-26; labials 14-16.

loveridge: African reptiles 311

Measurements. One cf measures 1155 (1010 + 145) mm.; largest 9
measures 1311 (1225 + 86) mm.; both from Magrotto. This 4 foot
5 inch 9 is, however, easily surpassed by one of 5 feet 8^ inches which
Lt. Col. Pitman informs me (17.xi.34) he obtained in the Mabira
Forest.

It is interesting to note the difference between the measurements of
freshly killed Magrotto snakes and that of their dried skins though
care was taken not to stretch the latter unduly.

cf measured 1155 mm. in the flesh, its dried skin 1300 mm.

9 " 1162 mm. " " 1210 mm.

9 " 1311 mm. " " 1560 mm.

Weights. A 3 foot 9M inch c? weighed 4 lbs. ; its stomach being empty.

3 foot 9% inch 9 weighed 5 lbs. ; stomach held 4 rats.

4 foot 3 inch 9 weighed 8 lbs.; stomach held 2 rats.

4 foot 5 inch 9 weighed 11 lb.; stomach held 1 rat, and
oviducts eggs as listed below.

Breeding. At Magrotto, on July 1, 43 eggs, of which the largest
measured 35 x 24 mm., in a 9 • No other 9 9 taken were gravid.

Diet. Four house rats (Rattus r. kijabius) in Budongo and Magrotto
snakes, three field rats (Mastomys c. durumae) in a Magrotto snake.

Parasites. Nematodes (Ophidascaris sp. and Kalieephalus sp.) num-
erous in Budongo snakes, a rodent thread worm (Mastophorus m.
muris) and porocephalid larva in a Magrotto snake.

Defence. Mr. W. E. Hartmann, formerly engaged in making zoologi-
cal collections for a Swiss museum, told me how in the eastern Usam-
baras he had once found himself standing beside one of these huge
vipers which he had not noticed, though he had been standing some
little time. Having a spike-ended stick in his hand, he drove the point
down through the snake's neck close behind the head, transfixing it to
the ground. Thereupon the snake opened its mouth three times and
each time two jets of venom issued from the fangs and were projected
to a distance of a foot or two. They were not directed at Mr. Hart-
mann, however, for he was standing beside, and somewhat behind, the
head.

This species produces so much venom that the cloudy white liquid
dripped from one's fangs as I pushed back the vagina dentis to show
the teeth to Fimbu, a headman on Magrotto Estate, who had lost an
arm through the bite of one of these snakes. Fimbu told me that he
had retired to his hut for several days after being bitten, and was
treated with native medicines. His arm, however, 'went bad' and
was amputated by a mission doctor.

312 bulletin: museum of comparative zoology

Elsewhere I (1940a, p. 502) have described the sluggishness of one
of these big snakes on a hot afternoon. Secure in its camouflage, it
remained motionless while search was being conducted in close prox-
imity, nor even when first picked up did it attempt to struggle or strike.

Bitis nasicornis (Shaw)

Coluber Nasicornis Shaw, 1802, Nat. Miscell., 3, pi. xciv: Interior of Africa
(from the master of a Guinea vessel) .

& 9 (M. C. Z. 48518-9) Mabira Forest, U. 8.xi.38.
3 c? <? (M. C. Z. 48520-1) Bundibugyo, U. 21.xii.38.
4 & cf 3 9 9 (M. C. Z. 48522-3) Idjwi Id., B. C. 17-28.ii.39.

Native names. Salambwa (Luganda) ; vipoma (Lutoro); heli (Lu-
amba) ; mpili or mpiri (Lulega).

Variation. Midbody scale-rows 29-40; ventrals 118-131; anal en-
tire; subcaudals 19-30; labials 15-19.

Coloration. Males, irrespective of size, have bellies which are
marbled and mottled like those of the females.

Measurements. Largest d71 measures 824 (710 + 114) mm.; largest 9
only 875 (805 + 70) mm., both from Idjwi Island.

Breeding. Females non-gravid in both November and February.

Diet. Three house rats (Rattus r. kijabius) in Mabira and Idjwi
snakes, a rat (Lophuromys a. laticcps) in other Mabira viper; rodent
fur in two other Idjwi reptiles.

Atheris squamigera squamigera (Hallowell)

Echis squamigera Hallowell, 1854, Proc. Acad. Nat. Sci. Philadelphia, p. 193:
Near the Gaboon River, Guinea, i.e. French Congo.

4 (M. C. Z. 48524-6) Mabira Forest, U. 12.xi.38.

Variation. Midbody scale-rows 19-22; ventrals 154-161; anal en-
tire; labials 9-12.

Measurement. Largest cf measures 537 (445 + 92) mm., but the
other two are only one or two mm. shorter; 9 measures 655 (555 -f-
100) mm.

Diet. A pigmy mouse (Leggada sp.) in one, unidentifiable rodent fur
in another.

Parasites. Ticks (Aponomma falsolaeve) were present on one of these
vipers.

loveridge: African reptiles 313

«

Atheris xitschei nitschei Tornier
Plate 3, fig. 3.

Atheris nitschei Tornier, 1902, Zool. Jahrb. Syst., 15, p. 589, fig.: Mpororo

Swamp, southwest Uganda.
Atheris woosnami Boulenger, 1906, Ann. Mag. Nat. Hist. (7), 18, p. 37:

Mubuku Valley, Mount Ruwenzori, Uganda.

5 (M. C. Z. 48527-30) Mihunga swamp, U. 13-18.xi.38.
1 (M. C. Z. 48531) Nyakabande, Ruanda, U. 28.i.39.
4 (M. C. Z. 48532-5) Mushongero. Ruanda, U. l.ii.39.
22 (M. C. Z. 48536-50) Idjwi Id., B. C. 17-28.ii.39.

Native names. Xchia (Lukonja); wahimberi (Lutoro); chirazi or
kirazi (Lulega).

Taxonomy. The first five specimens are topotypes of woosnami as
the swamp is in the Mubuku Valley immediately below "Woosnam's
camp on the Mihunga ridge. The next five snakes must be near topo-
types of nitschei according to the location of Mpororo district in
Steiler's atlas, though I have failed to locate any particular swamp to
which the name is especially applied.

The series from Lake Kivu are somewhat intermediate between the
Uganda material and the snake from Tanganyika Territory recently
described by Bogert (1940, p. 104) as A. n. rungweensis, and which
may be distinguished by the strongly keeled gulars. This additional
material reveals that the other characters cited by Bogert are not
constant enough to be dependable, though the fact that there is a
definite trend for increased head scales as between the Uganda snakes
and those from the Congo, makes it appear possible that when more
material of the southern form is available an average difference will
be demonstrable.

With the object of emphasizing this trend in lepidosis, the scale-
counts of the Uganda (U) snakes are separated from those of the
Congo (C) as follows.

Variation. Midbody scale-rows 27-29 (U), 25-33 (C); ventrals
142-162 (U), 150-162 (C); anal entire; subcaudals 38-51 (U), 44-51
(C); labials 9-12 (U), 10-13 (C); interorbital scales across crown 8-12
(U), 9-13 (C); circumorbital scales 12-16 (U), 13-17 (C); scales
between eye and nasal 2-4 (U), 3-5 (C); scales between mental and
first ventral 5-7 (U. & C).

The type of rungweensis had 12 supralabials, a condition found on

314 bulletin: museum of comparative zoology

one or both sides of the head in seven of the above series; rungweensis
had 13 interorbital scales, as have six of these snakes; rungweensis
had 4 scales between eye and nasal, as have three snakes but eighteen
more have 3 scales in the upper row and 4 in the lower; rungweensis
had 7-8 scales between mental and first ventral, three snakes in my
series have 7.

Coloration. In life. 9 (M. C. Z. 48528) Above, pale yellowish
green, a conspicuous black arrow-head marking on crown, an obso-
lescent vertebral line, becoming more definite on posterior third of
body and on tail where it is almost zigzag, is flanked along its entire
length by black diamond-shaped or triangular spots; a black line
from end of snout through orbit. Below, uniform greenish yellow
except on tail which has an ill-defined, dark olivaceous, median line
that broadens on tip to cover the entire under surface.

c? like the 9 but with dusky streak along entire length of belly.
Measurements. Largest d* (M. C. Z. 48537) measures 672 (555
+ 117) mm.; largest 9 (Congo Mus.) measures 646 (555 + 91) mm.;
smallest snake, a c? (M. C. Z. 48535), measures 204 (172 + 32) mm.
It should be noted, however, that there is a considerable shrinkage
between specimens measured in the field and later as alcoholics.
Thus the beautiful pair captured at Mihunga, as described below,
measured respectively cf 623 (520 + 103) mm. in the field, 605
(505 + 100) mm. in alcohol, while the 9 was 620 (532 + 88) mm.
in the field, and 619 (537 + 82) mm. in alcohol.

Breeding. On January 18 this 9 held 9 embryos, of which a cf
measured 60 (50 + 10) mm. No other females were gravid.

Diet. Pigmy mice (Leggada g. grata) in stomachs of snakes from
Mihunga swamp and Idjwi Island; a shrew (Crocidura ? h. hilde-
gardae) in another of the island vipers.

Parasites. The Mushongero adult was heavily infested with tape-
worms (Ophioiaenia sp.) and nematodes (Ophidscartts sp.), while the
stomachs of many Idjwi snakes were similarly parasitized as well as
with Capillaria sp. and Kaliccphalus sp.

Habitat. In the late afternoon, (one was taken sunning at 4 p.m.,
the other at 5 p.m. or three quarters of an hour after the sun had
dropped behind the upper Mihunga ridge,) I caught a pair of fine
adults as they lay coiled like plates on the top of dense tangles of
creepers which smothered the elephant grass growing along the banks
of the little stream which ^meanders through the swamp. Both rep-
tiles were at a height of six feet or more from the ground and I was
able to seize each of them by the neck with forceps without trouble.

LOVERIDGE: AFRICAN REPTILES 315

When I transferred them from the forceps to my fingers they gaped
widely and raised their long fangs.

At Mushongero, where there is little, if any, elephant grass, these
vipers live in enormous beds of papyrus.

Though 'tree viper' is something of a misnomer, and I do not regard
them as sylvicoline, I was following a path through heavy forest on
Idjwi Island when my gunbearer remarked in the vernacular ''You
are not interested in collecting snakes then?" Turning at this sar-
castic shaft, I saw him regarding one of these vipers. The snake, at
a height of three feet from the ground, was ensconced in a shrub that
I had just passed, in fact I must almost have brushed against it; a
bar of sunlight striking down through a rift in the forest canopy,
illuminated the spot.

Apparently then, the requirements of these vipers — apart from
small mice and frogs — is a moist or humid habitat with vegetation
in which they can climb. Such conditions may be found on the edges
of lakes, in swamps, or on the outskirts of virgin forest.

Atractaspis irregularis (Reinhardt)

Elaps irregularis Reinhardt, 1843, Dansk. Vidensk. Selsk. Skrift., 10, p. 264,

pi. iii, figs. 1-3: Gaboon.
Atractaspis schoutedeni Witte. 1930, Revue Zool. Bot. Africaine, 19, p. 224,

figs. 1-3: X'Goma. north of Lake Kivu, Belgian Congo.

d1 (M. C. Z. 48551) Mabira Forest. U. 12.xi.38.
9 (M. C. Z. 48552) Budongo Forest, U. l.xii.38.
d" 9 (M. C. Z. 48553-4) Bundibugyo, U. 24.xii.38.
d" (M. C. Z. 48555) Goma, B. C. 13.ii.39.

Synonymy. A. schoutedeni was based on a single specimen which
its author stated differed only from irregularis in (1) Its frontal
being longer than broad, instead of as long as broad. Actually both
conditions are common in irregularis, the frontal being longer than
broad in the Budongo, one Bundibugyo, and the Goma snake listed
above. (2) The first lower labials are just separated, instead of being
broadly in contact, behind the mental.

I consider this to be an individual aberration in the type of schoute-
deni, and, as irregularis occurs in the same locality, cannot recognize
a species based on a single character. See also remarks on katangae
below.

Variation. Midbody scale-rows 23-25; ventrals 219-252; anal di-
vided; subcaudals 22-27, paired; labials 5, the third and fourth enter-

316 bulletin: museum of comparative zoology

ing the orbit; first lower labial in contact with its fellow behind the
mental.

Measurements. Largest cf (M. C. Z. 48551) measures 457 (425 +
32) mm.; larger 9 (M. C. Z. 48552) measures 656 (621 + 35) mm.

Atractaspis bibronii Smith

Atractaspis bibronii A. Smith, 1849, Illus. Zool. S. Africa, Rept., pi. Ixxi:

Eastern districts of Cape Colony.
Atractaspis katangae Boulenger, 1901, Ann. Mus. Congo, Zool., 2, p. 13, pi. v,

figs. 2-2c: Lofoi, Katanga, Belgian Congo.

2 c? cT (M. C. Z. 48557-8) Ujiji, T. T. 13.iii.39.
3 <? c? 4 9 9 (M. C. Z. 48559-65) Mbanja, T. T. 27-30.iv.39.
9 (M. C. Z. 48560) Nchingidi, T. T. 13.V.39.
9 var. (M. C. Z. 48556) Amboni Estate, T. T. 21.vi.39.

Native name. Mbitu (Kimakonde at Mbanja, but applied to any
limbless burrowing reptile.

Synonymy. A. katangae was based on a single juvenile specimen
which its author compared with aterrima, considering its snout
rounded, instead of cuneiform, a point often difficult to decide, par-
ticularly in the young. During the forty years since its description,
four further examples have been recorded, viz. Katanga (Witte,
1930i); Msamwia, Ufipa, T. T. (Sternfeld, 1910a); and Elisabethville,
Katanga (Witte, 1933m).

One of the latter (now M. C. Z. 42978), a juvenile 9 , has been
carefully studied and its snout, like that of the Amboni snake listed
above, is cuneiform, differing quite noticeably from that of the Ulu-
guru aterrima in the collection of the M. C. Z. (23466).

The only other ways in which katangae allegedly differs from bibronii,
within whose range it occurs, is (1) Midbody scale-rows 25, instead of
21-23, but our Elisabethville snake has 23 which is the normal number
in tropical Africa. As most other well-known species of the genus
Atractaspis have a range of two more or two less than the normal num-
ber, this minor difference should carry little weight. (2) The posterior
point of the mental is in contact with the anterior sublinguals, thus
separating the first labials. Now that a snake (M. C. Z. 48556) with
this condition has turned up at Amboni, near Tanga on the East Coast,
it does not seem unreasonable to assume that it is an individual aber-
ration (cf. irregularis + schoutedeni above) of which there is a marked
strain in the Katanga region, but insufficiently numerous to be regarded
as a race, surrounded as they are by typical bibronii.

loveridge: African reptiles 317

For the convenience of those who may entertain doubts, however, I
have separated the data below.

Variation, bibronii type (M. C. Z. 48557-60). Midbody scale-rows
21-23; ventrals 224-252; anal entire; subcaudals 17-26, single; labials
5, the third and fourth entering the orbit ; first lower labial broadly in
contact with its fellow behind the mental.

A. katangae type (M. C. Z. 48556). Midbody scale-rows 23; ventrals
259; anal entire; subcaudals 23, single; labials 5, the third and fourth
entering the orbit; first lower labial well separated from its fellow by
the mental which is in contact with the sublinguals.

Measurements. Largest cT (M. C. Z. 48557) measures 429 (397 +
32) mm.; largest 9 (M. C. Z. 48556) measures 504 (476 + 28) mm.

Diet. At Mbanja a naked nestling shrew (Crocidura h. hirta) was
recovered from one, three large and one small nestling mice (Leggada
b. vicina) in another.

Habitat. I caught the largest male under rotting vegetation in a
banana plantation at Ujiji. Kizamba unearthed the Xchingidi snake
in forest close to camp, and, under the supposition that it was harm-
less, grabbed its head in his hand and so carried it back to camp. I
captured the largest female when she was turned up by the tractor
engaged in spreading piles of rotting vegetation in a cleared sisal plan-
tation at Amboni.

GEKKONIDAE

Cnemaspis quattuorseriatus (Sternfeld)

Gonatodes quattuorseriatus Sternfeld, 1912, Wiss. Ergeb. Deut. Zentral-Africa-
Exped. 1907-1908, 4, p. 202, pi. vi. fig. 1: Rugege Forest and Kisenyi,
Lake Kivu, Belgian Ruanda-Urundi ; TJvira, Lake Tanganyika, Belgian
Congo.

c? (M. C. Z. 47305) Kibale Forest. U. 17.xii.38.
Eggs (M. C. Z. 47306) Idjwi Island, B. C. 3.iii.39.

Remarks. To some extent this Kibale gecko bridges the narrow gap
which separates quattuorseriatus from dickcrsoni (Schmidt) of the
northeastern Belgian Congo, which Bogert tells me cannot be regarded
as even subspecifically distinct.

Variation. Within the known range of variation except for having
(1) 7 lower labials on right side of head, (2) the upper, or dorso-lateral,
row of tubercles is almost non-existent, being represented only by one
or two scattered tubercles, (3) 6 preanal pores.

318 bulletin: museum of comparative zoology

Measurements, cf measures 66 (31 4- 35) mm. A hatchling only
27 (14 4- 13) mm.

Breeding. On March 3, on Idjwi Island, a dozen eggs measuring
7x6 mm., were found in crevices of bark and among debris at the foot
of two large trees. The eggs contained embryos, one egg which hatched
resulted in the little gecko whose measurements are given above.

Habitat. The male was taken just before sunset as it was running up
a tree trunk. Intensive search at Kibale and on the Upper Mulinga
failed to discover any more of these elusive geckos.

Cnemaspis africanus elgonensis Loveridge

Cnemaspis africanus elgonensis Loveridge, 1936 (1935), Proc. Zool. Soc.
London, p. 820: Above Sipi, w. slopes of Mount Elgon, Uganda.

9 & eggs (M. C. Z. 47304) Mubuku Valley, U. 7.L39.

Distribution. This constitutes the first record of the occurrence of
this species on the Ruwenzori Mountains and, at the same time, forms
a westerly extension of its range.

Correction. In the diagnosis of this species (loc. cit.) I made the
stupid mistake of transposing the number of preanal pores of africanus
to elgonensis and vice versa. The present individual, being a poreless 9 ,
cannot be determined on this important characteristic of the race.

Variation. It agrees with the typical form, a. africanus, in having
(1) a single granule between the supranasals, (2) in having five granules
in addition to the rostral surrounding the nostril, (3) 8 upper labials.
None of these characters is of importance except the first, which was
found to vary even in the type series.

Measurements. 9 measures 120 (56 + 64) mm.

Breeding. Between January 1 and 7, we unearthed more than a
dozen eggs, each measuring about 11 x 9.5 mm.

Habitat. This gecko is exceedingly scarce, or difficult to find, in the
Ruwenzori forests. As I devoted so much time to its acquisition, it
might be of interest to publish the details. Shortly after our arrival in
the Mubuku Valley we started digging about the bases of the larger
trees and unearthed three pairs of Cnemaspis eggs around the bole of
the largest, a gigantic old tree with some holes at its base and hollow
branches visible at a height of twenty feet from the ground. On sub-
sequent days we found four more pairs of eggs at the base of as many
trees of which we thoroughly excavated round about a hundred. A
special reward of ten cents was offered for the first specimen which

loveridge: African reptiles 319

should be captured, but during the whole week only one was sighted
and it contrived to slip into a hole and escape.

Night after night I visited the big tree with a flashlight and scanned
its trunk as well as those of many others. Morning after morning, dur-
ing the few hours of sunshine vouchsafed to us, we examined trunks in
case geckos had been tempted to descend to sun themselves. We con-
structed a twelve-foot ladder with four-foot subsidiary, but failed to
reach the two hollow branches of the big tree, then we felled a tree
against the giant and Kizamba, our most accomplished climber, as-
cended and chopped open the hollow branch, but all to no purpose.

Our last morning for collecting dawned after a torrential downpour
the previous evening. Kizamba, on his own initiative, went to inspect
the big tree upon whose trunk a shaft of sunlight was playing, illumi-
nating a fissure which was five feet from the ground. On the edge of
this fissure he glimpsed a gravid female basking; she promptly vanished
into the crevice, but he was successful in extracting her!

Cnemaspis africanus africanus (Werner)

Gymnodactylus africanus Werner, 1895, Verh. Zool. Bot. Ges. Wien, 45, p. 190,
pi. v, fig. 5: Usambara Mountains, Tanganyika Territory.

2 c? cf 2 9 9 & eggs (M. C. Z. 47301-3) Magrotto Mtn., T. T.

5.vii.39.

Distribution. Also a tail of one of these geckos from the Siga Caves.
These caves, on the banks of the Mkulumuzi River, are surrounded by
gallery forest which has suffered considerably. The presence of this
sylvicoline gecko in the lowlands helps to harmonize the earlier German
records of its occurrence at Tanga (M. C. Z. 21918) Kuettner leg.

Correction. In my revision of this genus, I (1936a, p. 818) made the
foolish mistake in the diagnostic key of switching the number of
preanal pores of C. a. africanus with those of its subspecies elgonensis.

Variation. These fall within the range given in the redescription of
africanus cited above.

Coloration. In life. cf. Above, olive green mottled with brown and
black. Below, throat white, belly, base of tail, thighs, groin, and an-
terior aspect of tibia, chrome.

Breeding. On July 5, a 9 , together with one egg, measuring 9.5 x
7.5 mm., and evidently just laid, were found beneath a log. Two fresh
eggs, measuring 9 x 7.5 mm. were dug from a hole at the base of a tree.
Unfortunately one hatched on the journey to the States and the gecko
is too dried to measure. In all a score of eggs were collected.

320 bulletin: museum of comparative zoology

Habitat. I observed a cf basking at 9 a.m. where a beam of sunlight
was striking through the canopy into the interior of a partially decayed
tree in the forest beside the river. The gecko immediately descended,
disappearing into a hole among the roots, from this I dug it together
with the two eggs mentioned above.

At Siga Caves I saw a gecko basking at edge of a crevice in the cliff
face. I fired at it from afar so as not to damage it. The lizard, how-
ever, discarding its tail, disappeared into the fissure, from which it
could not be retrieved.

Hemidactylus tropidolepis barbouri subspec. no v.

Hemidactylus tropidolepis Andersson (not Mocquard), 1912, Jahrb. Nassau.

Ver. Naturk. Wiesbaden, 65, p. 227, figs. 1-4 (Tanga). Barbour &

Loveridge, 1928, Mem. Mus. Comp. Zool., 50, p. 142 (Kilindini).
Hemidactylus tropidolepis squamulatus Loveridge (not Tornier), 1933, Bull.

Mus. Comp. Zool., 74, p. 284 (Kilindini), and idem 79, p. 287 (Chan-

gamwe) .

5 (M. C. Z. 47320-2) Siga Caves, Tanga, T. T. 13.vi.39.
7 (M. C. Z. 47323-8) Opposite Kilindini, K. C. 25.vii.39.

Type. Museum of Comparative Zoology, No. 40907, an adult 9
from Changamwe, near Mombasa, Kenya Colony, collected by Arthur
Loveridge, July 4, 1934

Paratypes. Museum of Comparative Zoology, Nos. 24649-50 and
30447, from Likoni on mainland opposite Kilindini, Mombasa, and the
material listed above.

History. Andersson (1912) with scanty material, decided to merge
tropidolepis (Somaliland), floweri (Sudan), squamulatus (Kenya and
Tanganyika) together with a Tanga gecko which he regarded as inter-
mediate in certain respects. Barbour & Loveridge (1928) with two
geckos from Likoni opposite Kilindini and no tropidolepis material,
accepted Andersson's conclusions. Later, realizing the distinctness of
the southern gecko, I (1933, 1936) erroneously applied Tornier's name
squamulatus to two further specimens.

One of the principal objectives of my recent visit to Tanga was to
obtain an adequate series of Andersson's gecko. Now, with Kenj'a
examples of all three — tropidolepis, squamulatus and the Tanga form,
it is necessary to describe the latter and I take pleasure in naming it
for Dr. Thomas Barbour, who has so generously furthered these studies
of African herpetofauna.

LiOVERIDGE: AFRICAN REPTILES

321

Diagnosis. It belongs to that section of the genus characterized by
the possession of imbricate scales, instead of granules or tubercles on
the dorsum. From its nearest allies it may be differentiated as follows:

Character
Dorsal scales

Median sub-
caudal scales

Male pores
Snout to anus

H. t. tropidolepis

[ Considerable disparity

in size.

Both large and

small keeled.

Slightly

enlarged

transversely.

6-8

37 mm.

H. t. squamulatus

Great disparity

in size.

Only largest

strongly keeled.

Moderately

enlarged

transversely.

10-20

48 mm.

H. t. barbouri

Little disparity

in size.

Only largest

feebly keeled.

Strongly

enlarged

transversely.

16-23

44 mm.

The keeling, quite apart from other characters, distinguishes bar-
bouri from homoeolepis, ophiolepis, isolcpis and the so-called Bunoc-
nemis modestus.

H. t. tropidolepis

Dorsal lepidosis of

Type from Somaliland

(after Angel, 1925)

H. t. squamulatus
Dorsal lepidosis of

Type of alluaudi
from Kenya Colony
(after Angel, 1925)

' H. t. barbouri

Dorsal lepidosis of

Type from Kenya

Colony

(M. C. Z. 40907— 9 )

Description. (Figures in parenthesis are those of the sixteen para-
types, though in none of them does barbouri differ from tropidolepis or
squamulatus, the latter including werneri, tomieri and alluaudi).

Nostril bordered by the rostral, first labial, and (3-) 4 small nasals,
the uppermost in contact with its fellow (or separated by a single
granule in M. C. Z. 47320 only) ; upper labials 7-8 (6 in three paratypes,
see fig. 1 in Andersson) ; lower labials 6-7 (5 in two paratypes) ; digits
moderately dilated, inferiorly with oblique lamellae, 5 (4-6) under the
first toe, 7 (6-8) under the fourth toe (see fig. 2 in Andersson; owing to

322 bulletin: museum of comparative zoology

the gradual diminution in size of enlarged lamellar-like shields, count-
ing becomes somewhat arbitrary and on some individuals might be
reckoned up to 10 or 11).

Back covered with heterogeneous, imbricate scales, the largest of
which are feebly keeled; tail covered above and below by smooth,
imbricate scales, of which the median subcaudal series are transversely
enlarged, resembling the ventrals of ophidia. (Males with 16-23
preano-f emoral pores) .

Measurements. Total length of type 9 , 84 (44 + 40) mm., largest
paratype cf , 70+ (38 + 32+) mm., the tail being regenerated.

Habitat. All were taken beneath the great piles of palm fronds
assembled in the coconut plantations, the Siga series within a hundred
yards of my tent, those from opposite Kilindini within a mile of the
ferry landing at Likoni.

Hemidactylus mabouia (Jonnes)

Gecko mabouia Moreau de Jonnes, 1818, Bull. Soc. Philom. Paris, p. 138:

Antilles and adjacent mainland.
Hemidactylus tasmani Hewitt, 1932, Ann. Natal Mus., 7, p. 120: Gwelo,

Southern Rhodesia.

1 (M. C. Z. 47318) Uvira, B. C. 7.iii.39.

2 (M. C. Z. 47307) Ujiji, T. T. 10.iii.39.
Eggs & 4 (M. C. Z. 47308) Kitaya, T. T. 25.iii.39.

2 (M. C. Z. 47309) Mikindani, T. T. ll.iv.39.
2 (M. C. Z. 47311) Nchingidi, T. T. 9.V.39.
1 (M. C. Z. 47312) Lindi, T. T. l.vi.39.
1 (M. C. Z. 47313) Siga Caves, T. T. 25.vii.39.

Distribution. Common on the screening of Jinja Hotel, Uganda;
seen also at Magrotto, T. T.

Native names. Nangwagwa (Kimawiha); the common Swahili name
of mjusi (lizard) is used for this gecko by Wamakonde and Wayao.

Synonymy. H. mabouia has usually been described as having
"round" or "conical" tubercles, and this has led Dr. Hewitt to conclude
that his pair of Gwelo geckos with distinctly keeled tubercles are dis-
tinct. As a matter of fact the tubercles of our extensive Antillean
material frequently exhibit well defined keels and striae and neither
Mr. Shreve nor I can distinguish between this topotypic material and
our Southern Rhodesian (Birchenough Bridge, Sabi River) specimens
in this character which shows considerable variation throughout its
range. The fact that tasmani has 8 subdigital lamellae beneath its

loveridge: African reptiles 323

median toe, together with its large size of 142 (69 + 73) mm., preclude
the possibility of its being identified with the so closely allied gardineri.

Measurements. Largest c? (M. C. Z. 47311) measures S3 mm. from
snout to anus; tail regenerated.

Breeding. On March 3, at Kitaya, three pairs of eggs, measuring
11 mm. in diameter, were taken in grass and under a collapsed hut;
others at Mikindani on March 23.

Enemies. Adult, with tail intact, recovered from the stomach of a
half-grown house snake (Boacdon I. lineatus) at Kitaya; one from a
spotted wood snake (Phihthamnus s. semivariegatus) at Kitaya, and
one from a white-lipped snake (Crotaphopeltis h. hotamboeia) at Ujiji.

Habitat. The huge male was sunning on a forest tree in an over-
grown clearing of the Rondo Plateau forest, far from the haunts of
man! These clearings, however, were made by natives who fled up to
the plateau during the war of 1914-1918. They were removed in the
interests of forest conservation about 1920. The geckos were doubtless
introduced in bundles of thatching material. Ujiji specimens were
taken in abandoned tannery or refinery vats.

Hemidactylus gardineri Boulenger

Hemidactylus gardineri Boulenger, 1909, Trans. Linn. Soc. London (2), 12,
p. 296, pi. xl, fig. 4: Farquhar Island, Seychelle Islands.

Hemidactylus persimilis Barbour & Loveridge, 1928, Mem. Mus. Cornp. Zool.,
50, p. 140, pi. iv, figs. 1 and 3: Dar es Salaam, Tanganyika Territory.

Hemidactylus mandanns Loveridge, 1936, Proc. Biol. Soc. Washington, 49,
p. 60: Kitau, Manda Island, Kenya Colony.

1 (M. C. Z. 47310) Mbanja near Lindi, T. T. l.v.39.
1 (M. C. Z. 47319) Nchingidi, Rondo, T. T. 9.V.39.
. 5 (M. C. Z. 47314) Siga Caves, Tanga, T. T. 13.vi.39.
6 (M. C. Z. 47315) Amboni Estate, T. T. 19.vi.39.
4 (M. C. Z. 47316) Tanga, T. T. 22.vii.39.
1 (M. C. Z. 47317) Opp. Kilindini, K. C. 25.vii.39.

Synonymy. H. gardineri has never before been recorded from the
African mainland; indeed, except for Boulenger (191 Id) noting its
occurrence in five localities of the Aldabra Islands, it has not appeared
in the literature, being omitted by Parker (1936b) from his revised
list of the Seychellois herpetofauna. When Barbour and I described
persimilis we discussed the possibility of its identity with gardineri,
but with no specimen of the latter available, we ruled it out as im-
probable.

324 bulletin: museum of comparative zoology

Since then we obtained by exchange a 9 cotype (M. C. Z. 28652),
and from the Berlin Museum one of the Seychelle geckos referred by
Peters (1869a) to maculatus. This specimen is specifically identical
with gardineri and is not maculatus Dumeril & Bibron as now re-
stricted to their Indian material. The types of persimilis and man-
danus have also been compared with the cotype of gardineri
and found to be indistinguishable, for the minor differences displayed
by mandanus are within the range of variation displayed by fifty
specimens from nineteen localities which have been individually ex-
amined for the purpose of a revision of the family.

Breeding. On March 23, at Mikindani, several eggs measuring
8x7 mm., presumably of this species, were found under palm
fronds.

Habitat. The series (M. C. Z. 47314-7), like the young geckos,
with tails conspicuously barred black and white on a pinkish ground,
from Lamu and Mombasa Islands captured in 1934, were taken,
without exception, from beneath piles of fronds of the coconut palm.
When such piles were in the vicinity of a palm, each gecko, as soon
as disturbed, would make a dash for the nearest trunk, dart round to
its further side and so upwards to safety.

Enemies. One was recovered from the stomach of a spotted wood
snake (Philothamniis s. semivariegatus) at Mbanja, another from a
smooth snake (Meizodon semiornata) at Amboni, near Tanga.

Lygodactylus grotei grotei Sternfeld

Lygodactylus grotei Sternfeld, 1911, Sitz. Ges. Naturf. Freunde Berlin, p. 245:
Mikindani, Lindi Province, Tanganyika Territory.

2 (M. C. Z. 47329) Ujiji, T. T. ll.iii.39.
11 (M. C. Z. 47330) Kitaya, T. T. 15.iii.39.

3 (M. C. Z. 47331) Mbanja, T. T. 29.iv.39.
22 (M. C. Z. 47332-3) Mikindani, T. T. iv.39.

Native names. Those applied to L. p. picturatus with which it
occurs in the last three localities, as well as at Lindi.

Variation. Upper labials 6-9; lower labials 6-8; 4 pairs of lamellae
under fourth toe; median subcaudals transversely enlarged except
in one (M. C. Z. 47329) of the Ujiji geckos where a few pairs of scales
occur in the transversely enlarged series, i.e. it is approaching the
condition of capensis.

Coloration in life. d\ Ujiji. Above, dark brown, a pale, black-
edged light streak from nostril passes over upper part of eye and

loveridge: africax reptiles 325

*

along flank (where it becomes pinkish brown) to base of tail; crown
of head pale, vermiculated with black; back with a scarcely distin-
guishable light vertebral line; limbs and tail with light mottling.
Below, throat pure white; chest and belly cream color; tail pinkish.
In young the tails are distinctly reddish.

Measurements. Largest cT (M. C. Z. 47329) measures 71 (31 + 40)
mm., largest 9 (M. C. Z. 47333) measures 61 (29 + 32) mm.

Breeding. On March 23, at Mikindani, a pair of eggs measuring
about 6.5 x 5 mm., were preserved. On March 25, at Kitaya, several
pairs measuring about 5.5 x 5 mm. were found beneath (1) fallen
thatch of collapsed hut, (2) bundles of grass assembled for thatching,
(3) piles of weeds.

Parasites. Bright red acarines present on Ujiji geckos.

Enemies. Eleven were recovered from stomachs of five spotted
wood snakes (Philotlunmius s. semivariegatus) at Ujiji, Mbanja, and
Mikindani, of these six were in one snake from the last locality.

Habits. Adults occasionally squeak when captured. At Lindi one
was seen on the same tree with a L. p. picturatus.

Habitat. The topotypic series were mostly taken when basking on
the trunks of coconut palms about my tent. At Kitaya pawpaw trees
were favoured. Towards evening these geckos were frequently dis-
turbed among rubbish piled about the base of trees, up whose trunks
they would dart; it would appear as if they passed the night in the
rubbish if the trees offered no refuge such as holes or deep fissures.

Lygodactylus picturatus gutturalis (Bocage)

Hemidactylus gutturalis Bocage, 1873, Jorn. Sci. Lisboa, 4, p. 211: Bissao,
Portuguese Guinea.

1 (M. C. Z. 47340) Budongo Forest edge, U. l.xii.38.

1 (M. C. Z. 47341) Bundibugyo, Uamba, U. 23.xii.38.

2 (M. C. Z. 47342) Ujiji, T. T. 12.iii.39.

Native name. Abagwakulu (Luamba).

Variation. Upper labials 6-7; lower labials 6-6; 5 pairs of lamellae-
under fourth toe; median subcaudals transversely enlarged.

Measurements. Largest d" (Ujiji) measures 80 (41 + 39) mm.

Enemies. Recovered from stomachs of spotted wood snakes (Philo-
thamnus s. semivariegatus) at Katwe and Budongo.

326 bulletin: museum of comparative zoology

Lygodactylus picturatus mombasicus Loveridge

Lygodactylus picturatus mombasicus Loveridge, 1935, Proc. Biol. Soc. Wash-
ington, 48, p. 198: Kilindini, Mombasa Id., Kenya Colony.

2 (M. C. Z. 47339) Tanga, T. T. 22.vii.39.

Variation. Upper labials 7-8; lower labials 7-8; 6 pairs of lamellae
under fourth toe; median subcaudals transversely enlarged.

Measurements. The adult c? measures 76 (38 + 38) mm., and is
an undoubted example of this race which occurs together with the
typical form in the Voi-Mombasa-Tanga triangle.

Lygodactylus picturatus picturatus (Peters)

Hemidactylus picturatus Peters, 1870, Monatsb. Akad. Wiss. Berlin, p. 115,
n.n. for variegatus Peters, preoccupied: Zanzibar.

1 (M. C. Z. 47334) Kitaya, T. T. 25.iii.39.

4 (M. C. Z. 47335-6) Mikindani, T. T. 23.iii.39.
Embryo & 1 (M. C. Z. 47337) Siga Caves, T. T. 10.vi.39.

3 (M. C. Z. 47338) Amboni Estate, T. T. 19.vi.39.
Seen also at Lindi on the same tree as a Lygodactylus g. grotei.

Native names. Nankwakwa (Kiyao); nangwagwa (Kimahiwa); kihe-
tupetu (Kimakonde). But all three applied also to L. g. grotei.

Variation. Upper labials 6-8; lower labials 6-8; 6 pairs of lamellae
under fourth toe; median subcaudals transversely enlarged.

Measurements. Largest cfcf (Mikindani) each measured 43 mm.
from snout to anus; tails reproduced. The large size attained in this
region being reflected in the egg dimensions.

Breeding. On June 10, at Siga, an egg, containing embryo, preserved.

Enemies. Three recovered from the stomach of a hawk {Accipiter
b. polyzonoides) shot in palm at sunset, one in another {Accipiter m.
tropicalis) shot in palm at noon, both at Mikindani.

AGAMIDAE

Agama mossambica mossambica Peters

Agama mossambica Peters, 1854, Monatsb. Akad. Wiss. Berlin, p. 616: Coastal
Province, Mozambique.

12 (M. C. Z. 47358-9) Kitaya, T. T. 25-31.iii.39.
6 (M. C. Z. 47360-1) Mikindani, T. T. ll.iv.39.

2 (M. C. Z. 47362) Mbanja, T. T. 27.iv.39.

3 (M. C. Z. 47363) Nchingidi, T. T. 13.V.39.
1 (M. C. Z. 47364) Lindi, T. T. 2.vi.39.

LOVERIDGE: AFRICAN REPTILES 327

Native names. Nampopo (Kiyao); nankandindumba (Kimakonde at
Kitaya); nangwangula (Kimakonde at Mikindani and Mbanja); li-
wangula (Kimawiha).

Variation. Midbody scale-rows 74-80; preanal pores 23-26.

Coloration. In life. cf. Above, crown (sometimes also occiput and
sides of head) olive; occiput and sides of head mottled with pale blue
and white; back, limbs, and tail, pale olive tending to very pale blue
along the vertebral line. Below, chin and throat anteriorly blue, dew-
lap black; remaining undersurfaces pale bronzy olive to white.

9 . Above, as male, but with a chain of broad, blood-red markings
along the dorso-lateral region sometimes uniting on the vertebral line
with the corresponding series on the other side.

Young 9 . Above, crown black, occipital scale pale blue, parietal
region and sides of head bright ultramarine blue, circumorbital region
pale buff as is also the vertebral line and a dorso-lateral line which
unites with the vertebral line and its fellow by a series of saddle-like
sepia markings; these are continued on to the tail as a series of dark
blotches or bands which become fainter posteriorly ; upper parts other-
wise plumbeus. Below, throat faintly white, almost obscured by a
network of dark and light blue, base of throat with a black patch;
chest, belly, limbs, and tail cream colored.

The adults described above came from Kitaya, the young from
Mikindani.

Measurements. Largest cf (Kitaya) measures 310 (120 + 190)
mm., largest 9 9 (Kitaya) measure 259 (99 + 160) mm., smallest
(Nchingidi) measures 84 (34 + 50) mm.

Breeding. Between March 25-31, at Kitaya, all four females were
gravid, the 6 (?) or 8 eggs measuring about IS x 10 mm. in three exam-
ined. At this same time the largest male was shot while chasing
another from his tree. At Mbanja and Nchingidi only very young
agamas were seen, so that subspecific determination is only assumed.

Diet. Stomachs held: (1) Ants; (2) ants; (3) ants; (4) ants and big
beetle; (5) ants; grasshopper; (6) ants; beetles; grasshoppers; chrysid
bee; millipede; (7) ants; beetle larva; termite; millipede.

Parasites. At Kitaya and Mikindani almost all adults were heavily
infested by acarines beneath the ventral scales; at Kitaya a single
small 9 nematode (Abbreviate sp.) was present in a stomach full of
big black ants.

328 bulletin: museum of comparative zoology

Agama mossambica Montana Barbour & Loveridge

Agama mossambica montana Barbour & Loveridge, 1928, Mem. Mus. Comp.
Zool., 50, p. 147: Below Bagilo, Uluguru Mountains, Tanganyika Terri-
tory.

7 (M. C. Z. 47365-6) Magrotto Mtn., T. T. l.vii.39.

Native name. Kokolwe (Kisambara) .

Variation. It is interesting to come straight from collecting the
larger coastal form to the mountains where this differently colored
dwarf race occurs, the males with only 12-14 preanal pores. Six of
the series are adult, or subadult, males.

Measurements. Largest cf (M. C. Z. 47365) measures 259 (86 +
173) mm., only 9 measures 237 (87 + 150) mm., and is gravid.

Agama atricollis Smith

Agama atricollis A. Smith, 1849, Illus. Zool. S. Africa, Rept., App., p. 14: Natal,
South Africa.

8 (M. C. Z. 47343-4) Mabira Forest, U. 7.xi.38.
11 (M. C. Z. 47345-6) Budongo Forest, U. 23.xi.38.
4 (M. C. Z. 47347-8) Kibale Forest, U. 9.xii.38.
3 (M. C. Z. 47349-50) Bundibugyo, U. 21.xii.38.
6 (M. C. Z. 47351-2) Bugoye, U. 27.xii.38.
1 (M. C. Z. 47353) Nyakabande, U. 26L39.

1 (M. C. Z. 47354) Mushongero, U. l.ii.39.

2 (M. C. Z. 47355) Kisenyi, B. R. 12.ii.39.

13 (M. C. Z. 47356-7) Idjwi Id., B. C. 22-28.ii.39.

Native names. Konkome (Luganda, Lutoro, Luwamba); chihangara
(Lukiga); burubttru (Lulega).

Variation. Twenty males possess two, occasionally three, rows of
preanal pores, apparently increasing with age, ranging from 15-30,
average 22.

Color in life. Considerable difference as between Uganda and
Congo males was noted, and recorded as follows: (Budongo). Above,
head, anteriorly and on cheeks, turquoise blue; posteriorly, whole of
back, sides, limbs, and middle third of tail, cobalt blue; anterior third
of tail, olive, posterior third dark brown or black. Below, throat a
mixture of turquoise and cobalt blue; yellow spots on cheeks and
nape; a black patch in front of forelimbs which are blackish; chest

loveridge: African reptiles 329

cobalt blue; belly buff mottled with black; anterior third of tail bufhsh
white, median third tinged with cobalt, posterior third dark.

cf (Idjwi). Above, head and cheeks turquoise blue; back and limbs
yellow green; centre of flanks with a large patch of cobalt blue which
reaches upwards towards vertebral line; base of tail slightly tinged
with orange, followed by a band (40 mm. in width) of grayish brown,
followed by a band (40 mm. in width) of pale blue, the remaining 40
mm. of tail grayish brown like the upper surface of hands and feet.
Below, chin bright cobalt blue, centre of throat greenish necked with
yellow, base of throat black; chest, abdomen, and lower side of thighs,
metallic old gold with patches of cobalt anteriorly and on sides and a
black patch posteriorly; base of tail buff, followed by pale gray-brown
and yellowish gray.

9 (Idjwi). Above, head bluish, dark on crown, clearer on sides,
circumorbital region and supraorbital ridge yellowish; nape blackish,
the darker pigment extending midway along the back where it merges
into pale bronze, a dorso4ateral chain of orange-red blotches; tail
pale bronze with ill-defined black cross bands. Below, metallic pale
bronze somewhat greener on chest.

Measurements. Largest cf (M. C. Z. 47343) measures 330 (120
+ 210) mm., largest 9 (M. C. Z. 47344) measures 263 (107 + 156)
mm., smallest agama (Idjwi) measures 92 (40 + 52) mm.

Breeding. On November 7, at Mabira, embryos were discernible
in the spherical, 8 mm. diameter eggs in a female! On February 23,
on Idjwi, all three adult females were gravid, each holding 6 eggs
measuring 10x17, 12x20, and 13x24 mm. respectively, while the
last mentioned had a second series developing of about one-third the
size.

Parasites. Nematodes (Oochoristica sp.) were present in several
stomachs, and one, together with an immature 9 trematode (Strongij-
luris sp.), was found in the peritoneum of a 9 agama from Mabira;
Abbreviata britaniea recovered from a Budongo lizard.

Enemies. Recovered from the stomachs of two tree snakes (Rham-
nophis j. jacksoni and Thelotornis k. kirtlandii) at Idjwi and Bundi-
bugyo respectively.

Habitat. At Kisenyi, though one was taken on a palm trunk in
town, many were seen on the lava road to the lava flow, as often on
the ground as on shrubs!

330 bulletin: museum of comparative zoology

ZOXURIDAE

ZONURUS TROPIDOSTERNUM Cope

Zonurus tropidosternum Cope, 1869, Proc. Amer. Philos. Soc, 11, p. 169:
"Madagascar." (error).

<? (M. C. Z. 47367) Nchingidi, T. T. 9.V.39.

Native name. Chigologolo (Kimwera).

Variation. Nasals form a suture, thus differing from the type
(M. C. Z. 5742). Incidentally Cordylus Laurenti should probably be
used in preference to Zonurus.

Measurements. Total length 170 (90 -f- 80) mm.

Habitat. Shot while basking outside its hole in a tree at the forest
edge ; a glowworm was in its mouth as it came tumbling down.

Chamaesaura tenuior Giinther

Chamaesaura tenuior Giinther, 1895, Ann. Mag. Nat. Hist. (6), 15, p. 524;
pi. xxi, fig. B: Kampala, Uganda.

9 (M. C. Z. 47368) Fort Portal, U. 19.xii.38.

Native name. Nyarunyansi (Lutoro and Luamba).
Breeding. On December 12, this 138 mm. (snout to anus) 9 held
embryos.

VARAXIDAE

Varanus ocellatus Rtippell

Varanus ocellatus Riippell, 1827, Atlas Reise nord. Afrika, p. 21, pi. vi: Kordo-
fan, Anglo-Egyptian Sudan.

Skull & 5 (M. C. Z. 47369-74) Mikindani, T. T. 15-21.iv.39.
Skull & skin (M. C. Z. 47375) Siga Caves, T. T. 12.vi.39.

Native names. Liongondo (Kimakonde); ngondo (Kimawiha).

Color in life, cf (Mikindani). Above, crown black, back and flanks
blue-gray, three broad (i.e. 3-4 scales wide) black lines on nape, the
median forked anteriorly and posteriorly so as to surround two large
light ocelli, each of which has three black scales in its centre, these
form the first of seven transverse rows of similar ocelli, the interme-
diate rows being composed of small, less well-defined, light spots,
with or without black scales, the intervening area is covered by a

loveridge: African reptiles 331

heavy black network which almost obscures the ground color; limbs
black, spotted with yellow, such spots being usually formed of groups
of three yellow scales; tail black with thirteen alternating broad and
narrow, light (yellowish to bluish) crossbands speckled with black.
Below, lower jaw blue-gray ; throat olivaceous with obsolescent dark
spots; soles of feet brown, the centre of each scale often black; rest
of undersurfaces yellowish or white with a blue-gray network.

Young cf (Mikindani). Circumnasal area, limbs anteriorly, and
throat, greenish, otherwise like adult, but the blue-gray ground color
much more distinct so rendering the black network more conspicuous;
spots clear white; bars on tail less well-defined.

Measurements. Largest cf (M. C. Z. 47375) measures 1500 + (700
+ 800 + tip of tail) mm., i.e. 4' 9^"; weight 13 lbs.; skull length
135 mm. Another rf1 (M. C. Z. 47370) measures 1440 + (620 + 820
4- tip of tail) mm., i.e. 4' 8" (2' + 2' 8").

Breeding. On April 21, at Mikindani, a 9 held numerous small
ova 14 mm. in diameter.

Diet. Her stomach held a large slug and five snails (Achatina sp.),
that of a Mikindani male, two giant crickets (Brachi/trypetes mem-
branaceus).

Parasites. Ticks and nematodes on, and in, last monitor.

Enemies. At Mikindani some Yao showed me a big male (skull is
M. C. Z. 47369) which they were about to eat. Fortunately the poor
beast, lashed to a pole, was dead, for they had severed each foot at the
base, being afraid, so they said, of its claws.

Habitat. These big lizards are distressingly tough. One, which I
shot as it was basking high in a coconut palm, though mortally
wounded, made a dash for the burlap-like fibre about the stem and hid
there till fetched down by my gunbearer.

The Siga male was basking on the branch of a wild fig at a height of
sixty feet from the ground. Except that its head sank slowly to the
branch, it never moved when I shot it with No. 3 from a twelve gauge.
Two hours later, having in the meantime returned to camp for a 120 ft.
rope, I climbed a cliff and, after the fifth attempt, succeeded in throw-
ing a stone, attached to the rope, over the limb and eventually beneath
the monitor's chin. Each time that the rope was pulled the reptile's
head waggled to and fro lifelessly. Suddenly, to our amazement, the
tail wriggled, the head was raised, the tongue flickered, and the almost
five-foot lizard turned about on the broad limb. Quickly I gave it a
charge of No. 8 from the .410, which brought it crashing down sixty
feet. Even then it started to crawl away until I seized it by the tail.

332 BULLETIN": MUSEUM OF COMPARATIVE ZOOLOGY

Yaraxus niloticus (Linnaeus)

Lacerta nilotica Linnaeus, 1766, Syst. Nat., ed. 12, 1, p. 369: Egypt.

3 (M. C. Z. 47376) Mbanja, T. T. 28.iv.39.
1 (M. C. Z. 47377) Siga Caves, T. T. 14.vi.39.
1 (M. C. Z. 47378) Magrotto Mtn., T. T. l.vii.39.

Native names. Xc/ondo (Kimakonde) ; mbulu (Kisambara) .

Variation. In a young Mbanja monitor the nostril is only 1 mm.
nearer the orbit than it is to the end of the snout.

■Measurements. Largest (M. C. Z. 47377) measures 1170 (430 -f-
740) mm., i.e. 3' S3^'', and weighed 4^/9 lbs.

Habitat. The three young Mbanja lizards were shot in shrubs bor-
dering a stream, the Magrotto monitor in a tree at edge of rain forest,
it dropped from the tree and plunged into a stream, then swam under
water to a hole in the bank which it entered.

AMPHISBAENIDAE

As all pertinent information regarding the amphisbaenids collected
during the course of the expedition has been incorporated in my recent
revision of the African members of this family (Bull. Mus. Comp.
Zool., 87, pp. 353-451, figs. 1-53), they need be only listed here.

Amphisbaexa phylofixiexs Tornier

Amphisbaena phylofiniens Tornier. 1899, Zool. Anz., 22, p. 260: Ujiji, Tan-
ganyika Territory.

4 (M. C. Z. 47901-4 ) Ruanda, Ujiji, T. T. ll-15.iv.39.

Amphisbaexa oriextalis (Sternfeld)

Amphisbaenula orientalis Sternfeld, 1911, Sitz. Ges. Naturf. Freunde Berlin,
p. 246: Mikindani, Tanganyika Territory.

1 (M. C. Z. 47905) Mikindani. T. T. 24.iv.39.

Amphisbaexa ewerbecki ("Werner)

Chirindia ewerbecki Werner, 1910 (1909), Mitt. Zool. Xat. Mus. Hamburg,
27, p. 37: Mbanja (Banja), Tanganyika Territory.

70 (M. C. Z. 47906-49) Mbanja, T. T. 26-30.iv.39.
1 (M. C. Z. 47950) Lindi, s.e. T. T. l.vi.39.

loveridge: African reptiles 333

Amphisbaexa roxdoensis Loveridge

Amphisbaena rondoensis Loveridge, 1940, Bull. Mus. Comp. Zool., 87, p. 394
fig. 23: Nchingidi, Rondo Plateau, Tanganyika Territory.

49 (M. C. Z. 47951-99) Nchingidi, T. T. 9-19.V.39.

These are the Type and Para types of a form which would be referred
to the genus Chirindia if grounds for its revival should be forthcoming.
The series showed no overlapping with the seventy topotypes of its
nearest ally, ewerbecki, occurring in low-lying country fifty miles to
the north.

LACERTIDAE

Nucras boulexgeri kilosae Loveridge

Nucras kilosae Loveridge, 1922, Proc. Zool. Soc. London, p. 314: Kilosa, Usa-
gara, Tanganyika Territory.

3 (M. C. Z. 47379-80) Nchingidi, T. T. 13.V.39.

Variation. It is interesting to note that the keels were not notice-
able on these very young specimens unless dry ; their scale-counts fall
within the range of this race though the record involves a southeast-
ward extension of the range nearly 300 miles.

Color in life. Above and on sides, black, a cream stripe from rostral
to frontal, latter, as well as some other head scales, edged with creamy
white ; upper lip and sides handsomely spotted with white ; a vertebral
and two slightly narrower, light, dorso-lateral lines tinged with salmon;
tail translucent pink with a broad, black, median, longitudinal line on
the base and two ill-defined ones on the sides. Below, chin, throat,
breast, abdomen, and fore limbs, white; hind limbs and tail pink.

Habitat. Taken on sandy paths in eroded areas of the forest.

Lacerta versus Algiroides

Though the genotypes of these two groups are distinct enough, some
of the intermediate forms are not so readily distinguishable. Lacerta
jacksoni, as stated by Tornier, appears closely related to vauereselli
in the moderate size and keeling of the dorsal scales which are juxta-
posed or subimbricate in jacksoni, slightly imbricate in vauereselli; in
fact, the quickest way to distinguish the two, apart from the lower
femoral pore counts of the latter, is in the presence in preserved speci-
mens of jacksoni of a white background to the femoral pores and cir-

334 bulletin: museum of comparative zoology

cumanal areas which are uniform in vauereselli. The transference of the
latter to Algiroides, makes necessary the following amendment to the
key to those genera as presented by Boulenger (1920, Monogr. Lacer-
tidae, 1, p. 2), viz.
Dorsal scales small or moderate, smooth or strongly keeled, juxtaposed

or subimbricate Lacerta

Dorsal scales moderate or large, strongly keeled, imbricate. . Algiroides

Lacerta jacksoni Boulenger

Lacerta jacksoni Boulenger, 1899, Proc. Zool. Soc. London, p. 96, pi. x: Ravine
Station, 7500 feet, Mau Mountains, Kenya Colony.

1 (M. C. Z. 47381) Nyakabande, U. 29.L39.

1 (M. C. Z. 47382) Mushongero, U. 30.i.39.

2 (M. C. Z. 47383-4) Kisenyi, B. R. 10.ii.39.

85 (M. C. Z. 47385-94) Idjwi Id., B. C. 13-28.ii.39.

Native names. Ngondochero (Lukiga); kavunduguru (Lulega).

Variation. Femoral pores 14-19, the former number on 9 sides only,
the latter only on 1 side (M. C. Z. 47383), 15-17 being normal with an
average of 16.6 pores for 159 sides (the 160th side being damaged).

Coloration, cf (Kisenyi). Above, brown, marbled with black on
dorsum and tail and with more or less longitudinal series of cream-
colored spots, each consisting of one or two scales; sides spotted with
cream and pale blue. Below, lower labials, like the upper, spotted with
black; throat to collar white; chest and belly yellowish white; limbs
bright yellow; tail white.

Measurements. Largest tf1 (M. C, Z. 47385) measures 231 (83 +
148) mm., but half-a-dozen others are 82 mm. from snout to anus;
largest 9 (M. C, Z. 47390) measures 187 (76 +111) mm.

Breeding. On January 29, at Nyakabande, a 9 held 3 eggs measur-
ing 14 x 8 to 15 x 7 mm. Between February 13 and 28, on Idjwi Island,
many 9 9 were gravid, five (of 65-76 mm. head and body length)
examined held from 3-5 eggs ranging from 11 x 9 to 15 x 9 mm., i.e.
about ready for laying.

Diet. Stomachs held (1) two large crickets, (2) limbs of big spider.

Enemies. One recovered from stomach of a green snake (Chlorophis
irregularis) on Idjwi.

Habitat. 'Where, as at Mushongero and Kisenyi, deforestation has
occurred, it is interesting to observe the adaptability of this arboreal
species which was living in holes in the earth bank of the terraced

loveridge: African reptiles 335

roadway leading to Lake Mutanda; while at Kisenyi I captured two
that were basking on the road but retreated into crevices between the
blocks of lava which formed the foundation for the road down to
Lake Kivu.

Algiroides vauereselli (Tornier)

Lacerta vauereselli Tornier, 1902, Zool. Anz., 25, p. 701: Kagera, west of Lake

Victoria, Tanganyika Territory.
Algiroides boulengeri Peracca, 1917, Atti Ace. Torin, 52, p. 351: Fort Portal,

Toro, Uganda.

16 (M. C. Z. 47395-404) Idjwi Id., B. C. 20-28.ii.39.

Native name. Kavunduguru (Lulega).

Synonymy. Though I have not seen types of either of the two species
listed above. I am confident that the synonymizing of boulengeri is
correct. Its type locality is only 120 miles north of that of vauereselli,
taking the Kagera River at its nearest point. L. vauereselli, already
recorded from Idjwi Island, has been considered a great rarity until
now, so that it seems advisable to furnish the statistical data derived
from this fine series in some detail.

Variation. Nasals separated (M. C. Z. 47395) or in contact (M.C. Z.
47396); upper labials anterior to subocular 3-5, normally 4; dorsal
scales across midbody 35-43; gular scales from mental to collar 20-27;
plates in collar 6-8; transverse rows of ventrals from collar to crotch
17-22; longitudinal rows of ventrals 6-8; femoral pores 8-12, average 9;
lamellae beneath fourth toe 17-22; adpressed hind limb reaches axilla
or shoulder in females, shoulder or collar in males.

Coloration. In life. Adult cf • Above, head dark bronze ; tip of snout
and a broad dorsal band, light bronze; from nostril through eye and
along flanks a broad, brown, lateral band, darker above; upper lip
light-colored the coloring continued as a light line to insertion of fore
limb; between fore and hind limbs are two rows of black-edged, cream-
centered ocelli; on vertebral line, at least posteriorly, a chain of dark
spots; tail flecked with darker brown. Below, metallic whitish gray
(black in formalin).

Young — 67 mm. Above, as adult, but end of snout greenish; dorsal
band edged on either side by a row of cream-colored spots, those on
the sides brighter, some tinged with green. Below, chin and throat,
extending to upper lip, and tail, from a short distance behind anus,
bright blue; rest of lower surface of body and limbs, bright grass-
green.

336 BULLETIN': MUSEUM OF COMPARATIVE ZOOLOGY

Measurements. Largest d1 (M. C. Z. 47400) measures 197 (62 +
135) mm., largest 9 (M. C. Z. 47399) measures 171 (57 + 114) mm.

Breeding. Several 9 9 gravid, one examined held 3 eggs measuring
8x6 mm.

Diet. A grasshopper in each of five stomachs examined.

Enemies. One recovered from the stomach of a wolf snake (Lycophi-
dion c. ornatum).

Algiroides africaxus Boulenger

Algiroides africanus Bbulenger, 1906. Proc. Zool. Soc. London, 2, p. 570, fig. 96:
Entebbe, Uganda.

2 (M. C. Z. 47405-6) Mabira Forest, U. 10-16.xi.38.
8 (M. C. Z. 47407-12) Budongo Forest, U. 29-30.xi.38.

1 (M. C. Z. 47413) Kibale Forest, U. 9.xii.38.

2 (M. C. Z. 47414-5) Idjwi Id., B. C. 6.iii.39.

Native name. Kalumbalunda (Luganda).

Variation. Dorsal scales across midbody 20-24; plates in collar 6-9;
transverse rows of ventrals from collar to crotch 18-22; longitudinal
rows of ventrals 6; femoral pores 12-16; lamellae beneath fourth toe
17-22; adpressed hind limb reaches shoulder or collar in females, collar
or beyond in males.

Coloration. In life. Budongo. Above, dark rufous brown with an
irregular vertebral series of black spots ; flanks and limbs sepia brown
spotted with black; tail transversely barred with darker and lighter.
Below, throat grayish with a metallic lustre; chest from collar, abdo-
men to basal portion of tail, bright green, rest of tail brown.

Measurements. Largest d71 (M. C. Z. 47414) measures 167 (60 -f-
107) mm., largest 9 measures 67 mm. in head and body length, tail
missing.

Breeding. On November 16, at Alubango, ova small, about 6x5
mm.

Diet. Stomachs held: (1) Large spider, (2) spiders and egg packet,
(3) orthopteran and very small insects, (4) black cricket, (5) cricket,
(6) cricket.

Parasites. Two larval nematodes (Physalopteridae) in stomach of
Mubango lizard.

Enemies. Recovered from stomachs of green snake (Chlorophis hop-
logaster) and cobra (Xaja melanoleuca) at Mubango, Mabira Forest.

Habitat. Most usually to be seen basking at whatever height the

loveridge: africax reptiles 337

sunlight happened to be striking the trunks of trees fringing the forest
edge or clearing: frequently beside some knot hole or piece of loose
bark, in, or beneath, which they could seek refuge when disturbed.
Judging by their diet, much of their hunting is done on the ground,
and a Mabira lizard was discovered beneath a pile of mulch in a banana
plantation adjacent to the forest.

Ichxotropis squamulosa Peters

Ichnotropis squamulosa Peters, 1854, Monatsb. Akad. Wiss. Berlin, p. 617:
Tete, Mozambique.

c? (M. C. Z. 47416) Kitaya, T. T. 2.iv.39.

Coloration. Differs from all our comparative material in that the
entire under surface is black, perhaps this is correlated with the
specimen being an adult breeding male, whose testes are much en-
larged.

Diet. Two spiders and two young crickets.

Habitat. Shot on a sandy path through dry scrub, the two or three
seen were exceedingly wary and the species unknown to the local
natives — Konde and Yao, though it has been recorded already from
the not far distant Konde Plateau.

Eremias spekii spekii Giinther

Eremias spekii Giinther, 1872, Ann. Mag. Nat. Hist. (4), 9, p. 381 : Unyamwezi
Tanganyika Territory.

3 (M. C. Z. 47417-9) Siga Caves, T. T. 8.vi.39.
Also seen on Amboni Estate, both localities being near Tanga.

Coloration. In life. Young — 80 (26 + 54) mm. Above, head
brown, dorsum black, three slightly pinkish lines commencing behind
parietals converge to form a single vertebral line which becomes
pink on tail, a cream-colored dorso-lateral line commencing at orbit
broadens and becomes pink on tail, a cream-colored lateral line from
rostral, over labials, through tympanum, after which it breaks up
into a series of dots and dashes on flank, but is continued (as dots)
along the anterior surface of hind limb to the foot; both fore and
hind limbs spotted. Below, white, except for tail, which is red.

338 bulletin: museum of comparative zoology

Holaspis guentheri Gray

Holaspis guentheri (A. Smith) Gray, 1863, Proc. Zool. Soc. London, p. 153,
pi. xx, fig. 1 : No type locality.

d" (M. C. Z. 47420) Budongo Forest, U. 29.xi.38.
cf c? (M. C. Z. 47421-2) Magrotto Mtn., T. T. 7.vii.39.

Native name. Chungura (Kisambara).

Color in life, cf (Budongo). Above, pale creamy green with black
vertebral line bifurcating on nape and head; flanks with three black
longitudinal lines; median line of tail pale blue edged with black
and flanked with yellow. Below, throat and fore limbs grayish white
tinged with pink; chest and belly bright orange extending on to the
hind limbs which otherwise are yellow; anterior two-thirds of tail
yellow transversely barred with black, posterior third pale blue but
similarly barred with black.

The tails of the Magrotto males differed somewhat, that of the
larger being yellow below, and of the smaller bright blue bordered
by bright yellow, each segment with a pair of black spots.

Measurements. Largest (M. C. Z. 47420), apparently a record,
measures 123 (53 + 70) mm

Diet. A moth in its stomach.

Parasites. Red acarine parasites on its flanks anteriorly, added
to the appearance of this beautiful species.

Preserved in amber. Five years ago Dr. Thomas Barbour showed
me a specimen of Holaspis guentheri, minus the posterior portion of
its hind limbs and tail, which was embedded in a small block of
amber-like substance. So well preserved was the specimen that one
could count the 24 rows of dorsals at midbody, and the 31 or 32
gular scales between mental and medium collar plate. The femoral
pores, however, were uncountable, nor was the precise position of
the nostril clear. The lizard appears indistinguishable from the spe-
cies which today ranges right across tropical Africa, in the west from
Sierra Leone to Angola, in the east from the Lsambara Mountains
to Nyasaland.

The specimen had been found in a desk drawer of the late Director
of the Museum of Comparative Zoology, Mr. Samuel Henshaw, and
unfortunately bore no label. Mr. Henshaw, being approached regard-
ing its history, stated that it was one of "two lizards in amber" which
formed part of the collection of the late Dr. Hagen of Konigsberg.

On hearing of the Konigsberg association, my interest was aroused

loveridge: African reptiles 339

because of the implications regarding the lizard allegedly embedded
in Baltic amber in the Konigsberg Museum. Fifty years ago this
reptile was thought to be a Cnemidophorus by Boettger, but sub-
sequently named Xucras succinea by Boulenger (1917, Ann. S. African
Mus., 13, p. 195, footnote) on the basis of the description given by
Klebs (1910, Schrift. Physik. Ges. Konigsberg, 51, p. 217) who had
shown him the lizard in 1891.

Dr. Frank M. Carpenter kindly identified two diptera which were
embedded with the Holaspis, as Phlcbotomus and Trigona, wide-rang-
ing genera which shed no light on the place of origin. Then, becom-
ing interested, Dr. Carpenter, together with Mr. F. J. Gettens, staff
chemist of the Fogg Art Museum at Harvard, investigated the com-
position of the material and definitely demonstrated that it was not
Baltic amber. This amber has a melting point of about 50° Cent,
higher than that of any of the copals. They reached the conclusion
that the melting point of this specimen was nearer to that of Congo
copal than to the Zanzibar product.

In this connection it may be recalled that Peters (1865, Monatsb.
Akad. "\Yiss. Berlin, pp. 455-457) discussed the presence of a gecko
embedded in Zanzibar copal. It had formed one item in a collection
of embedded animals and plants brought back by F. O'Swald after
several years sojourn on the island. Peters referred the gecko to
Lygodactylus capcusis (A. Smith) which, with its synonym strigatus
Gray, was the only member of the genus known at that time. It is
probable, however, that it should be identified as the scarcely dis-
tinguishable grotei Sternfeld (1911) which is the common species of
the adjacent coast and neighbouring islands.

This raises the question as to whether Xucras succinea is really
embedded in Baltic Oligocene amber, and therefore "the oldest known
Lacertid" as stated by Boulenger (1920, Monogr. Lacertidae, 1, p. 7).
When he examined it in 1891, the genus was known only from south of
the Zambesi, since then its range has been extended to equatorial East
Africa, though not the Congo. It has not been recorded from Zanzibar
but does occur on the adjacent mainland 150 miles due west of Zanzi-
bar, where the form which I named A. boulengcri kilosae is found, a
race which ranges from Kajiado, Kenya Colony, south to the Rondo
Plateau, some fifty miles southwest of Lindi Bay, Tanganyika Terri-
tory.

Very little, if any, of what is known as "Zanzibar copal" came from
the island itself; Capt. J. F. Elton, writing from Kisiju, Kwale District,
a dozen miles south of Dar es Salaam, says: (1879, Travels and Re-

340 bulletin: museum of comparative zoology

searches among the Lakes and Mountains of Eastern and Central
Africa, p. 78) "Everywhere signs of copal diggings were visible. In fact
we were passing through the main fields from which the Zanzibar
market was once almost entirely supplied, and which still produce the
valuable gum in considerable quantities."

Should it ever be proved that the embedding medium of succinea
is Zanzibar copal, the name succinea might have to take precedence
over that of kilosae. Unfortunately, according to Klebs, the specimen
suffered so severely at the time of its examination that few of the
characters once observable, can now be seen.

GERRHOSAURIDAE

Gerrhosaurus nigrolineatus nigrolineatus Hallowell

Gerrhosawrus nigro-lineatus Hallowell, 1857, Proc. Acad. Nat. Sci. Philadelphia,

p. 49: Gaboon.
Gerrhosaurus flavigularis forma intermedia Lonnberg, 1907, in Sjostedt, Wiss.

Ergebn. Schwed. Zool. Exped. Kilimandjaro, Meru . . . , No. 4, p. 7,

pi. i, figs, la-b: Steppe near Lake Natron, Tanganyika Territory.
Gerrhosaurus nigrolineatus auritus FitzSimons, 1939a, Ann. Transvaal Mus.,

20, p. 10: Kaapmuiden, eastern Transvaal.

3 (M. C. Z. 47423-5) Kitaya, T. T. 25.iii.39.

1 (M. C. Z. 47426) Mikindani, T. T. 18.iv.39.

2 (M. C. Z. 47427-8) Mbanja, T. T. 28.iv.39.

1 (M. C. Z. 47429) Nchingidi, T. T. 9.V.39.

2 (M. C. Z. 47430) Lindi, T. T. 2.vi.39.

Native names. Ligondo (Kiyao); liwalawahi (Kimakonde at Kitaya;)
nangkwakata (Kimakonde at Mbanja).

Synonymy. Reasons for reverting to the opinion that nigrolineatus
of East Africa is inseparable from the typical form of West Africa, and
also for referring australis to the synonymy, were recently furnished
in a revision of the family (1942, Bull. Mus. Comp. Zool, 89, p. 511).

Variation. Dorsal scales transversely 22-26, average 24; supracili-
aries 4, except for No. 47428 with 5; length from snout to hind edge of
ear contained in that from snout to anus 3.25-4.33 times. The lower
figures are those of emergent young.

Breeding. On March 25, at Kitaya, six eggs were found beneath a
pile of rotting vegetation ; on opening one egg I found an embryo so

loveridge: African reptiles 341

small that I placed the remaining five in a tin of damp sand and grass.
During the succeeding months the eggs were examined periodically,
one egg which had dried up was discarded. On opening the tin on
June 8 I found three lizards so recently emerged that the albumen
upon them was still moist. I measured one of them and found it
(M. C. Z. 47425) was 181 (56 + 125) mm., but after two years in
preservative only 170 (53 +117) mm. The remaining egg seemed to
have swollen somewhat since first collected, it now measured 30 x
22 mm.

In anticipation of a long wait, I sat down with the egg lying on my
palm. Almost immediately, however, there was a convulsive movement
and a long slit appeared at one end of the parchmentdike shell, about
one minute later the little lizard's snout (up to and including the eyes)
was thrust out, another minute elapsed and then the rest of the head
appeared. For seven minutes thereafter nothing more occurred, then
the forepart up to the hind legs crawled out, to be followed, a couple
of minutes later, by the long tail. The little creature lay breathing
heavily on my hand and offered no objection to being picked up and
placed upon the table. Suddenly, with the unexpectedness which
characterized most of its actions, and as if a fall of two and a half feet
was of no account, it leaped to the ground and raced away with a fine
turn of speed. I let it go. The whole emergence had taken place be-
tween 10.45 and 11.10 a.m.

Apparently there is a definite season for such hatching along the
East Coast, for hatchlings, either 6 mm. smaller or larger than the
dimensions given above, were taken on April 8, May 5, and June 6
(loc. cit. supra), while recently emerged young were seen on paths at
Kiponda, Rondo Plateau, and in Amboni Estate, near Tanga,on May 8,
and June 20, respectively. Moreover, ovules were small in an adult 9
taken on June 2 at Lindi.

Diet. Stomach of this 9 held grasshoppers, termites and an Ennea
shell; that of a young lizard, grasshoppers.

Enemies. Both Mbanja hatchlings, and another (not preserved)
from Amboni, were recovered from the stomach of two hawks (Kaupi-
falco monogrammiea); one had been swallowed by a house snake
(Boaedon I. lineatus) at Nchingidi, and two by wolf snakes (Lyco-
phidion c. capense) at Mikindani and Amboni. At Kitaya a big adult
was taken in a rat trap baited with meat which probably attracted
insects.

3-12 bulletin: museum of comparative zoology

SCINCIDAE

Mabuya maculilabris maculilabris (Gray)
Plate 4, fig. 2.

Euprepis maculilabris Gray, 1845, Cat. Lizards Brit. Mus., p. 114: West

Africa.
Mabuia maculilabris var. kwidjwiensis Sternfeld, 1912, Wiss. Ergebn. Deut.

Zentral-Afrika-Exped. 1907-1908, 4, p. 233: Kwidjwi Island, Lake Kivu,

Belgian Congo.

2 (M. C. Z. 47431-2) Mabira Forest, U. 5.xi.38.

3 (M. C. Z. 47433-5) Budongo Forest, U. 20.xi.38.
1 (M. C. Z. 47436) Butiaba swamp, U. 5.xii.38.

1 (M. C. Z. 47437) Kibale Forest, U. 16.xii.38.

1 (M. C. Z. 47438) Bundibugyo, U. 26.xii.38.

64 (M. C. Z. 47451-500) Idjwi Id., B. C. 16-28.ii.39.

2 (M. C. Z. 47439-40) Ujiji, T. T. 13.iii.39.

2 (M. C. Z. 47405-6) Mikindani, T. T. 10.iv.39.
2 (M. C. Z. 47445) Amboni Estate, T. T. 19.vi.39.

4 (M. C. Z. 47601-4) Likoni, K. C. 25.vii.39.
Seen also at Kisenyi, B. R., and near Siga Caves, T. T.

Native names. Munya (Luganda); kifumbatanjoka (Lutoro); hila-
merembi (Luamba); sazi (Lulega).

Variation. The disposition of skinks of this difficult group must
remain somewhat unsatisfactory until such time as a thorough revision
of the genus can be undertaken. The long series of topo types of
kwidjwiensis were secured in the hope that it might be possible to
recognize the race; however, after taking a series of one thousand
observations on the fifty catalogued specimens, no grounds were
found for doing so. This Idjwi series furnished the following data:

Centre of nostril above first labial, rarely above the suture between
rostral and first labial; postnasal not, or but rarely, in contact with the
second labial; frontonasal in contact with the rostral (12 ex.) though
more usually separated (38 ex.) ; frontonasal in contact with frontal
(17 ex.), usually separated (33 ex.); interparietal subequal to, or
larger than, the frontoparietals (40 ex.) sometimes definitely smaller
(10 ex.) ; frontal separated from (40 ex.) or in contact with (10 ex.)
first supraocular; supraoculars 4 (constant on both right and left);
supraciliaries (one side only examined) 4 (1 ex.), 5 (46 ex.), or 6
(3 ex.) ; lobules on anterior border of ear-opening normally 3, rarely 1
or 2 or indistinguishable; nuchals obtusely .carinate; dorsals with 5
(young) to 7 (adults) or very rarely 8, keels; midbody scale-rows 30

loveridge: African reptiles 343

(11 ex.), 31 (5 ex.), 32 (28 ex.), 33 (3 ex.), or 34 (3 ex.); from 2 to 4
preanals slightly, sometimes scarcely appreciably, enlarged; the
adpressed hind limb fails to meet or meets the fingers of the backward
pressed fore limb in 14 females and 1 male (M. C. Z. 47496), the wrist
or elbow in 30 males (all of which have the hemipenes extruded).

The data derived from the Uganda, Tanganyika, and Kenya material
falls within this range except that some young have tricarinate
dorsals, that midbody scale-rows number 32-34 in about equal pro-
portions, and that the sexual dimorphism of limb length noted in
Idjwi skinks is not so reliable in coastal material.

Coloration. On Idjwi one gained the impression that the skinks
were consistently more handsome than elsewhere, particularly at the
coast. However, the following color notes were made in the field.

Budongo Forest, cf. Above, olive, with ill-defined longitudinal
rows of black spots on dorsum with finer green flecks interspersed be-
tween them; lips yellow, the latter color continuing on towards ear
where it changes to orange, then fades out on flank about midbody;
tail spotted with black. Below, pale yellow, spotted with black on
throat and tail; chest and belly slightly brighter yellow, immaculate.

Idjwi Id., cf . Above, black flecked with pale green, or olive flecked
with black and pale green. Below, throat and flanks blood-orange;
chest, abdomen, and underside of limbs very pale yellow; soles of feet
dusky; tail whitish, uniform or spotted with black and tinged with
pink towards the tip. X.B. Young cf c/1 have throats which are
white, spotted with black, like those of the females.

Idjwi Id., 9 . Above, dark olive or brown, uniform or flecked with
black and usually also with pale green. Below, throat lemon yellow;
rest of underparts pale greenish white, or else this coloring reversed,
viz. throat white spotted with black, chest and abdomen bright
yellow; soles of feet and underside of tail as in c? cf.

Idjwi Id., young. Above, dark brown, a bar of sepia brown from
eye along flank. Below, white, immaculate.

Ujiji. cf. Above, substantially similar to those from Idjwi Id. as
described above. Below, throat white, spotted with black; otherwise
in agreement with the Idjwi skink.

Measurements. Largest Idjwi cf (M. C. Z. 47451) measures 282
(92 + 190) mm., largest 9 (M. C. Z. 47480) 222 (94 + 12S) mm., but
tail regenerated; females appear to average slightly smaller than
males. From elsewhere largest d1 (M. ('. Z. 47434) measures 250
(88 + 162) mm., and 9 (M. C. Z. 47601) 212 (86 + 126) mm., but
tail of latter, possibly that of former also, regenerated.

344 bulletin: museum of comparative zoology

Breeding. In February, on Idjwi, and July, at Likoni, females were
found with both small ova and large eggs; 6 of the latter from a
Likoni skink, measured 10 x 15 mm.

Diet. Stomachs held: large spider; several ants; real wasps; para-
sitic wasp; plant bugs; small weevil; plant beetles; fly; moths; grass-
hoppers and a cockroach.

Parasites. Nematodes were present in one of the Likoni skinks.

Enemies. One recovered from the stomach of a Lycophidion c.
ornatvm on Idjwi, another from a L. e. capense at Amboni, near Tanga.

Habitat. On wall of house at Mubango, Mabira; beneath vegeta-
tional debris in a banana plantation at Ujiji; under piles of palm
fronds and coconut husks at Likoni.

Mabuya maculilabkis comorensis (Peters)

Euprepes comorensis Peters, 1854, Monatsb. Akad. Wiss. Berlin, p. 619:
Anjuan, i.e. Johanna Island, Comoro Islands.

7 (M. C. Z. 47446-50) Magrotto Mtn., T. T. 3.vii.39.

Native name. Ghondo (Kisambara), but generic.

Variation. These skinks are subspecifically identical with the
extensive series from the F/sambara Mountains discussed by Barbour
and Loveridge (1928, p. 156). They differ from typical maculilabris
in having 34-36 midbody scale-rows, together with a more robust
build and shorter tail.

Measurements. Largest cf (M. C. Z. 47447) measures 243 (90 +
153) mm.

Habitat. Not uncommon at forest-edge where they may be seen
basking on tree-trunks, both vertical and fallen.

Mabuya maculilabris boulengeri Sternfeld

Mabuia boulengeri Sternfeld, 1911, Sitz. Ges. Naturf. Freunde Berlin, p. 248:
Makonde Plateau, Lindi Province, Tanganyika Territory.

1 (M. C. Z. 47441) Kitaya, T. T. 2.iii.39.

3 (M, C. Z. 47442-3) Mikindani, T. T. 21.iv.39.

1 (M'. C. Z. 47444) Nchingidi, T. T. 18.V.39.

Native names. Chengamawta (Kiyao); ehipakamaicta (Kimakonde);
liny era nen da (Kimawiha) .

Variation. In 1928 I relegated boulengeri to the synonymy of macu-
lilabris, now, after collecting the above series of both sexes and ages

LOVERIDGE: AFRICAN" REPTILES 345

from points around the type locality, I am prepared, though with some
misgivings to recognize it. It appears to differ from typical maculila-
bris in having 4 supraciliaries as a constant feature, whereas in the
typical form 4 occurs only as a very exceptional deviation from 5-6;
it has a longer tail and much more uniform coloring. Midbody scale-
rows 30 (3 ex.) or 32 (2 ex.), whereas the typical form on the East
Coast has 32-34.

Coloration. 9 • Kitaya. Above, uniform brown; sutures of upper
and lower labials black; a series of black dots from eye to ear; flanks
paler than dorsum. Below, throat pure white; belly tinged with yellow;
tail buffy white, the lateral borders of the scales edged with greenish
black resulting in four, irregular, longitudinal lines (as in mcgalura).

Measurements. Largest cf (M. C. Z. 47442) measures 268 (76 +
192) mm., only 9 (M. C. Z. 47441) 232 (84 + 148) mm., tail regen-
erated.

Sexual dimorphism. In males the toes of the adpressed hind limb
overlap the fingers of the backward pressed fore limb, in the female
they fail to meet.

Diet. One stomach held a large spider, two others a large cricket
each.

Habitat. The female I found asleep on the frond of a doom palm at
the edge of a swamp, she was at a height of about five feet from the
ground. The three males from Mikindani were shot while basking on
the trunks of three adjacent coconut palms, it is interesting to note
that two typical maculilabris were obtained at Mikindani, also upon
the stems of coconuts, eleven days before.

Mabuta planifrons (Peters)

Euprepes (Euprepis) planifrons Peters, 1878, Monatsb. Akad. Wiss. Berlin,
p. 203, pi. ii, fig. 2: Teita, Kenya Colony.

4 (M. C. Z. 46707-9) Amboni Estate, T. T. 19.vi.39.

Variation. Midbody scale-rows 28; dorsals with 3-7 keels; supra -
nasals in contact; prefrontals separated; supraoculars 4-4; supracili-
aries 5-6.

Measurements. Larger cf (M. C. Z. 47609) measures 261 (111 +
150) mm., and larger 9 (M. C. Z. 47607) 328 (116 + 212) mm.

Diet. One disgorged a very large grasshopper when caught.

Habitat. All were dislodged by a tractor from piled and rotting
vegetation.

346 bulletin: museum of comparative zoology

Mabuya megalura (Peters)

Euprepes (Mabuia) megalura Peters, 1878, Monatsb. Akad. Wiss. Berlin,
p. 204, pi. ii, fig. 4: Teita, Kenya Colony.

33 (M. C. Z. 47610-9) Idjwi Id., B. C. 18-28.ii.39.

Native name. Kahirira (Lulega).

Variation. Midbody scale-rows 24-26; dorsals very obtusely tri-
carinate, or (in old females) smooth; supranasals in contact -(28 ex.),
or separated (5 ex.); prefrontals in contact (5 ex.), or separated (28
ex.); supraoculars 4-4; supraciliaries 3-5.

In skinks of this species the exceptionally slender tail is more often
regenerated than otherwise, nor are such regenerated tails readily
recognized without magnification. Of the present series only six pos-
sess their original tails intact and in all of these the length of the tail
is more than twice that of the head and body. Recently de Witte
(1933m, p. 76) revived the name massaianus (Fischer, 1884) for some
Congo skinks on the ground that their tails were not twice the length
of the respective head-and-body length.

Measurements. Largest cf (M. C. Z. 47612) measures 199 (56 +
143) mm.; largest 9 (M. C. Z. 47616) 215 (69 + 146) mm., but both
are surpassed in head-and-body length by others of (cf ) 61 mm. and
( 9 ) 75 mm. respectively which have regenerated tails.

Breeding. In February all adult females appeared to be gravid, five
examined held 4, 4, 6, 6, and 8 ova respectively. In all, except the first
lot, were white, but well-formed, embryos.

Diet. Examination of ten stomachs revealed the following: 6 grass-
hoppers, 1 pigmy locust, 1 mantid, 2 crickets, 1 dragonfly (with an
estimated wing spread of three inches), 1 leafhopper (Homoptera),
1 moth, 2 caterpillars, and 2 spiders. Mr. F. G. Werner, to whom I
am indebted for the identifications, expressed surprise that so small a
species of skink should be capable of overpowering prey so large as
many of the orthoptera proved to be.

Parasites. Nematodes (Cosmocercidae) were present and preserved,
also larval Spiruroidea which Dr. Lucker suggests are not parasitic
on this skink but were introduced with arthropod prey.

Habitat. In the grass fringing the footpaths.

loveridge: African reptiles 347

Mabuya varia varia (Peters)

Euprepes (Euprepis) varius Peters, 1867, Monatsb. Akad. Wiss. Berlin, p. 20:
Tete, Mozambique.

9 (M. C. Z. 47620) S. Kinangop, K. C. 27.X.38.
d> (M. C. Z. 47621) Ujiji, T. T. 13.iii.39.
9 (M. C. Z. 47622) Kitaya, T. T. 3.iv.39.
c? 9 (M. C. Z. 47623) Mikindani, T. T. 18.iv.39.
9 (M. C. Z. 47624) Mbanja, T. T. 27.iv.39.
Seen also at Lindi and on Nchingidi Plateau.

Native names. Jagasi (Kikuyu); namkwakwa (Kiyao); liwalawahi
namahonta (Kimakonde at Kitaya); mjusi islam (Kimakonde at
Mbanja). The Konde and Yao names are not specific, however.

Variation. Midbody scale-rows 32-34; dorsals tricarinate; supra-
nasals in contact; prefrontals separated; supraoculars 4-4; supra-
ciliaries 4-5.

Measurements. Not exceptional.

Breeding. All four females are gravid, and in only the Kinangop
skink do the ova not contain embryos.

Enemies. One recovered from the stomach of a hawk (Kaupifalco
monogrammica) at Mbanja, two others from wolf snakes (Lycophidion
c. capense) at Mbanja and Lindi, while a third was chased past my feet
by a sand snake (Psammophis s. sibilans) as recorded under that species.

Mabuya striata (Peters)

Tropidolepisma striatum Peters, 1844, Monatsb. Akad. Wiss. Berlin, p. 36:
Mozambique.

c? (M. C. Z. 47625) Mabira Forest, U. o.xi.38.
9 (M. C. Z. 47626) Budongo Forest, U. 5.xii.38.
9 (M. C. Z. 47627) Bundibugyo, U. 21.xii.38.
c? (M. C. Z. 47628) Mihunga Ridge, U. 17.L39.
o* (M. C. Z. 47629) Nyakabande, U. 27.L39.
<? (M. C. Z. 47630) Bugoie, B. R. 8.U.39.
3 (M. C. Z. 47631) Kisenyi. B. R. 10.ii.39.
8 (M. C. Z. 47632-5) Idjwi Id., B. C. 16.ii.39.
<? (M. C. Z. 47636) Kitaya, T. T. 28.iii.39.
c* 9 (M. C. Z. 47637-8) Mikindani, T. T. lS.iv.39.
tf1 9 (M. C. Z. 47639-40) Nchingidi, T. T. lO.v.39.
See also at Ujiji, Lindi, Siga, Amboni, and on Magrotto Mtn.

Native names. Munya (Luganda and Lutoro) ; kistumbu (Luamba);
machercrcra (Lulega) ; liwalawahi namahonta (Kimakonde); namkwa-
kwa (Kiyao); nangwagwa (Kimawiha); ghondo (Kisambara).

348 bulletin: museum of comparative zoology

Variation. Midbody scale-rows 34-38; dorsals with 3-5-7 keels of
which the median 3 only are prominent; supranasals in contact and
prefrontals separated except in two (Idjwi and Xchingidi) skinks;
supraoculars 4-4; supraciliaries 2-6.

Measurements. Largest d71 (M. C. Z. 47636) measures 252 (97 +
155) mm., largest 9 (M. C. Z. 47638) 200 (97 + 103) mm., but tail
of latter regenerated.

Diet. Lygaeus bug, golden chrysomelid beetle, and grasshopper.

Enemies. One recovered from the stomach of a wolf snake (Lycophi-
dion c. eapense) at Ujiji.

Parasites. On Idjwi Island even very young skinks harboured large
nematodes (Cosmocencidae) in stomach and intestines.

Habitat. On house in Mabira; rare at Budongo; common only in
vicinity of native huts on Mihunga; in garden rubbish and on tree at
Kisenyi; one shot twelve feet up a hardwood tree at Kitaya, while
at Nchingidi two were shot when basking on tree trunks at a height
of twenty feet from the ground. These trees were in forest clearings
which had been made by refugee natives during the World War and
subsequently abandoned on orders from the Forestry Department.

It was interesting to note the adaptability of this essentially
savanna species which is accompanying man in his incursions into
primary forest.

Riopa feknandi (Burton)

Tiliqua fernandi Burton, 1836, Proc. Zool. Soc. London, p. 62: Fernando Po.

9 9 (M. C. Z. 47641-2) Budongo Forest, U. 24.xi.3S.
cf (M. C. Z. 47643) Bundibugyo, U. 21.xii.38.

Native names. Ngurukisi (Lutoro) ; korukutendi (Luamba).

Variation. Midbody scale-rows 34-35, and in other respects en-
tirely typical.

Measurements. The d1 measures 226 (116+ HO) mm., the larger
9 , 261 (138 + 123) mm.

Breeding. Ova small in both these large females, testes large in the
male.

Diet. Orthoptera in one, stomachs of the others empty.

Parasites. Female nematodes (Ascaridoidea, probably Cosmo-
cercidae) present in Bundibugyo skink.

Defence. One, which I caught beneath a log, threatened to bite but
made no sound either then or later in camp when it was provoked.

loveridge: African reptiles 349

Riopa sundevallii (Smith)

Eumices (Riopa) sunderallii (sic) A. Smith, 1849, 111. Zool. S. Rept., App.,
p. 11: Natal, South Africa.

1 (M. C. Z. 47644) Butiaba swamp, U. 5.xii.38.

1 (M. C. Z. 47645) Mikindani, T. T. 14.iv.39.

2 (M. C. Z. 47646) Mbanja, T. T. l.v.39.

6 (M. C. Z. 47647) Nchingidi, T. T. 13.V.39.

1 (M. C. Z. 47648) Siga Caves, T. T. 10.vi.39.

2 (M. C. Z. 47649) Amboni Estate, T. T. 19.vi.39.
4 (M. C. Z. 47650) Magrotto Est., T. T. l.vii.39.

N alive 7iames. Chiyoweja mikiongo (Kimakonde at Mikindani) ;
liwalawahi (Kimakonde at Mbanja); kitoiva ivaimi (Kisambara).

Variation. Midbody scale-rows 26-28; supranasals in contact;
prefrontals separated; supraoculars 4-4; supraciliaries 6-8.

Measurements. Largest & (M. C. Z. 47650) measures 196 (110 +
86) mm., largest 9 (M. C. Z. 47649) 220 (140 + 80) mm., youngest
(Butiaba) 79 (45 + 34) mm.

Defence. From the stomach of a sand snake (Psammophis s.
sibilans) I recovered the tail of a Sundevall's skink, tangible evidence
of the advantage derived from a readily detachable tail. This was at
Nchingidi, where these lizards were found both without and within
the rather dry forest.

Riopa pembanum (Boettger)

Lygosoma (Riopa) pembanum Boettger, 1913, in Voeltzkow, Reise in Ostafrika,
3, p. 350, pi. xxiv, figs. 4-5: Pemba Island.

3 (M. C. Z. 47651-3) Likoni, K. C. 25.vii.39.

Variation. Midbody scale-rows 24-26, and in other respects en-
tirely in agreement with seven cotypes recently received from Pemba.

Habitat. These sandy-yellow young were taken beneath piled palm
fronds in a coconut plantation half-a-mile from the ferry landing
opposite Kilindini. They constitute the second record of the occur-
rence of this species on the mainland.

350 bulletin: museum of comparative zoology

Lygosoma kilimense Stejneger

Lygosoma kilimensis Stejneger, 1891, Proc. U. S. Nat. Mus, 14, p. 405: Mt.

Kilimanjaro, Tanganyika Territory.
Lygosoma gromieri Angel, 1925, Bull. Mus. Hist. Nat. Paris, 31, p. 419: Tsavo,

Kenya Colony.
f Lygosoma (Siaphos) compressicauda Witte, 1933, Revue Zool. Bot, Afri-

caine, 23, p. 175, figs. 1-4: Sandoa, Belgian Congo.
f Siaphos dewittei Loveridge, 1934, Copeia, p. 184: n.n. for compressicauda

de Witte (not of Werner, 1897) preoccupied.

7 & eggs (M. C. Z. 47654-6) Magrotto Mtn., T. T. l-15.vii.39.

Synonymy. L. gromieri was described as having a pair of supra-
nasals but no frontonasal and was consequently referred to Riopa,
then regarded as a subdivision of Lygosoma. Actually I believe it to
represent an aberration in which the frontonasal is divided so as to
simulate a pair of supranasals. It will be noted that a similar con-
dition occurs in 2 of the 128 topotype L. bhchmanni (vide infra),
and if it occurs as a rare variant in one member of the group it is
not unreasonable to assume that it may occur in another, as gromieri
differs from kilimense in no other respect I refer it to the synonymy.
In reply to a query regarding the type locality, Mons. Angel, with
customary kindness, replied in a letter dated November 13, 1937,
that Gromier had informed him that the skink was taken on a termi-
tarium near Tsavo station. As the station is beside the Tsavo River
which drains the Kyulu and Ongolea Rivers which descend from
Kilimanjaro, its geographical location in the arid Tsavo district does
not seem unreasonable.

The description of L. compressicauda Witte, which I renamed de-
wittei differs in no way from our kilimense material, some of the
series possessing regenerated tails apparently just as compressed as
that of the holotype of compressicauda. If my conclusion is correct
then we are confronted with an amazing extension of range, circa
1500 miles, for Sandoa is not far from the border of Angola, A simi-
lar case of discontinuous distribution is provided by this skink's
chief enemy, Lycophidion meleagris, which is also known only from
the Congo-Angola region in the west and the Usambara-Uluguru
Mountains in the east.

Variation. Midbody scale-rows 22-24; supraciliaries 7-8; fingers 5;
toes 5; lamellae beneath fourth toe 14-15.

Measurements. Largest c? (M. C. Z. 47655) measures 156 (67 + 89)

loveridge: African reptiles 351

mm., larger 9 (M. C. Z. 47656) 129 (60 + 69) mm., a hatchling
(M. C. Z. 47654) 54 (27 + 27) mm.

Breeding. On July 1 a 9 held small ova, another 4 eggs almost
ready for deposition. On July 10, 4 eggs, measuring 14x11 mm.,
and holding embryos on the point of hatching, were found in leaf
mold between the buttress roots of a giant tree. On July 12, 4 eggs,
measuring 13.5 x 11 mm., and two clutches so recently hatched that
the shells were still soft and fresh, were found in a similar situation.

Enemies. Two eggs and the end of a tail of one of these skinks
were recovered from the stomach of a wolf snake (Lycophidion
meleagris).

Defence. On being picked up, only one skink of the entire series
attempted to bite, and that but feebly.

Habited. All were taken in deep forest between the buttress roots
of trees, in which situation their eggs and enemy were also unearthed.

Lygosoma graueri graueri Sternfeld

Lygosama graueri Sternfeld (part), 1912, Ergebn. Deutschen Zentral-Afrika-
Exped., 4, p. 240, fig. 3, and quinquedigitata, p. 241, pi. vi, fig. 5: Karisimbi
and vicinity, Belgian Ruanda-Urundi.

19 (M. C. Z. 27657-66) Mubuku Valley, U. 29.xii.38-7.i.39.

Variation. Midbody scale-rows 22-24; supraciliaries 7-8; fingers 5;
toes 5; lamellae beneath fourth toe 8-10. The fingers in all are well
developed and show no tendency towards the race quattuordigitata
Sternfeld, 1912, of the Rugege Forest.

Coloration. Above, much as in meleagris, but very different below
and without such marked sexual dichromatism, viz. Below, whitish,
each scale with a central black spot, these spots forming longitudinal
rows; preanal region and hind limbs of adults salmon pink and less
spotted ; spotting in adult males like that of females but denser though
the throat is not black as in meleagris.

Measurements. Largest cf (M. C. Z. 47659) measures 160 (60 -f
100) mm., largest 9 (M. C. Z. 47660) 166 (65 + 101) mm., but both
surpassed in length from snout to anus by c? c? of 67-68 mm., and
9 9 of 72-73 mm. with regenerated tails. Young hatchling (M. C. Z.
47657) 51 (25 + 26) mm.

Breeding. Between January 1-7 five females examined were in all
stages of pregnancy, pairs of eggs in the most advanced measured
10x5, and 11 x 5.5 mm. respectively. Doubtless many of the eggs

352 bulletin: museum of comparative zoology

listed under L. meleagris are those of graueri a point which can be
determined only by dissection of those containing embryos with
digits developed.

Diet. Ten stomachs revealed the following: 1 spider; 2 small
mites, 1 rose beetle, 1 tenebrionid beetle, 2 weevils, 1 beetle larva —
probably a wood borer, 2 grasshoppers, 3 moths, 1 large worker ant,
1 small bee, 2 fern sporangia — presumably ingested with prey.

Habited. These skinks were found about the roots of ferns in the
wet forest, their favorite habitat being on the east side of large trees,
preferably those with buttress roots. Here they enjoyed the maxi-
mum benefit from the few hours of sunshine which filtered through
the forest canopy, and the powdery loam tended to be drier for the
frequent rainstorms came mostly from the west. To find these very
secretive skinks it is first necessary to cut away the ferns which are
massed about the base of almost every tree, this done the mat of
fern roots which carpets the forest floor may be rolled back so that
the skink, or skinks, are exposed. They burrow quickly into the
powdery subsoil, though not so active in escaping as Riopa sunde-
vallii would be under similar circumstances. This is probably to be
attributed to the lower temperature prevailing in the Mubuku Valley
of Ruwenzori at 7000 feet.

Lygosoma meleagris Boulenger
Plate 4, fig. 3.

Lygosoma meleagris Boulenger, 1907, Ann. Mag. Nat. Hist. (7), 19, p. 488:

Mubuku Valley, 7000 feet, Ruwenzori Mountains, Uganda.
Siaphos meleagris hellcri Loveridge, 1932, Proc. Biol. Soc. Washington, 46,

p. 113: Bugongo Ridge, 9500 feet, Ruwenzori Mountains, Belgian Congo

(not Uganda).
Lygosoma (Siaphos) Burgeoni de Witte, 1933, Revue Zool. Bot. Africaine,

24, p. 116, figs. 1-2: Kalonge, 6725 feet, Ruwenzori Mountains, Belgian

Congo.

40 & eggs (M. C. Z. 47667-700) Mubuku Valley, U. 29.xii.38-7.i.39.

Synonymy. L. meleagris was described from a single unsexed indi-
vidual which, on the basis of its color description, we can confidently
assume to be a male. For twenty-five years no second example was
recorded. In 1932 I described a female from the western (Congo)
side of the range which differed from Boulenger's description in five
distinct ways. Eight years later Parker (1940a, p. 267), reporting on
a juvenile from 8000 feet, pointed out that the original description of

loveridge: African reptiles 353

meleaaris was inaccurate in 2, if not 3, of my points, leaving helleri
as distinguished only by the number of subdigital lamellae (partly
bridged by his skink), and coloration.

In the hope that an adequate series of topotypes would shed light
on the range of variation to be expected, I visited the Mubuku Valley
and there, at 7000 feet, personally captured the series listed above.
This material disposed of the remaining differences, embracing the high
number of subdigital lamellae of melcagris and the low number dis-
played by the type of heMeri, and revealing that the striking color differ-
ences were only sexual {vide infra). It enables me not only to corrobo-
rate Mr. Parker's doubts as to the validity of helleri but to definitely
refer it to the synonymy.

L. burgeoni was differentiated solely on the grounds of coloration
and a semidivided nasal. Actually the description of the former cor-
responds well with some of our topotypes. As to the nasal character,
though in our series probably entire, the nasal not infrequently presents
the appearance of being divided or semidivided, usually on one side of
the head only. It will be recorded that Boulenger described meleagris
as having the "nostril pierced between two nasals" but on reexamina-
tion Parker found that "the nasal is undivided, though an artifact on
one side simulates a suture." The recognition of burgeoni on this single
character, therefore, would appear unjustified.

Variation. Midbody scale-rows 22-24; nostril in an entire nasal
though not infrequently (M. C. Z. 47669, 47680, etc.) having the ap-
pearance of being divided or semidivided; supraciliaries 5-8; 7 being
normal; fingers 4; toes 4; lamellae beneath fourth (i. e. third, as hallux
missing) toe 9-12. (Based on examination of 33 skinks only).

Coloration. In life, cf • Above, dark or black, each dorsal scale
with a central or two lateral light flecks; flanks white, usually sharply
distinct from back but sometimes sparsely flecked with black. Below,
throat mottled black and bluish white; chest, belly, hind limbs, and
base of tail salmon red (white in alcohol) ; rest of tail bluish white
heavily mottled with black.

Young c? cf resemble 9 9 in being yellowish white or uniform pink
from chin to base of tail, this tint turning to salmon red as they
reach maturity. In alcohol 6 sexed males, ranging from 43-63 mm.
from snout to anus, had throats which were white or heavily spotted
(progressively so with age), 9 others, ranging from 61-69 mm., had
black throats. On the other hand, 14 sexed females, ranging from 52-
75 mm., had throats which were pure white, only 3 exhibiting any
spotting and then but very few spots.

354 bulletin: museum of comparative zoology

9 . Above, pale brown, each dorsal scale with a blackish central
shaft and the dorsal coloring shading off to translucent pink on base of
tail; a dark brown streak from nostril, through eye, to flank a short
distance behind the fore limb; flanks spotted. Below, throat yellowish
white; sometimes one or two lines of black spots along the sides of
chest and belly; rest of under surface, including limbs, to base of tail
salmon red, remainder of tail bluish white heavily mottled with black.

Measurements. Largest c? (M. C. Z. 47668) measures 196 (68 +
128) mm., largest 9 (M. C, Z. 47667) 193 (69 + 124) mm., but the
latter is surpassed in length from snout to anus by others of from
70-75 mm. with regenerated tails. A hatchling measures 53 (24 + 29)

mm.

Breeding. Fourteen females examined were in all stages of preg-
nancy, a pair of eggs in each of the four most advanced measured
8 x 5, 9 x 5, 10 x 5, and 10 x 6 mm. respectively. Some skinks had al-
ready laid for numerous pairs of very fresh eggs, measuring 12 x 8.5
mm., were unearthed (for habitat see graueri above). Other eggs were
in various stages of development and we uncovered one pair from which
a skink had just hatched and its fellow emerged as we watched. Later,
in camp, I observed the hatching of two more pairs. To escape from
the egg the little skink makes a cut two-thirds the length of one side
and involving an end of the shell also. Many such hatched-out shells
were collected, usually they were in pairs but sometimes several skinks
had resorted to the same spot to lay. The greatest number found in
any one spot was 22 in an area of about two inches in diameter. Half
of these were fresh and about half consisted of shells from some
previous season.

Diet. Ten stomachs revealed the following: 1 centipede, 1 scorpion
or pseudoscorpion, 1 spider, 2 mites, 1 woodlouse or sow bug (Isopod),
2 beetles, 2 weevils, 2 grasshopper nymphs, 2 leaf hoppers (Homop-
tera), 2 small flies, 1 midge, 3 caterpillars.

Defence. Only one of the entire series bit on being captured, their
surest defence consists in discarding the exceptionally long tail con-
sequently few specimens with unregenerated tails are found.

Habitat. The habitat of this species is fully described under L.
graueri, for both 5- and 4-toed species occur together. The first three
were taken on the day of our arrival during the process of clearing ferns
for a camp site. Though usually only one or two are to be found at the
base of any one tree, half-a-dozen were encountered among the buttress
roots of one giant.

loveridge: African reptiles 355

Lygosoma blochmanni Tornier

Lygosoma blochmanni Tornier, 1903, Zool. Jahrb. Syst., 19, p. 173: Lake
Kivu, Belgian Congo.

128 (M. C. Z. 47701-50) Idjwi Id., Lake Kivu, B. C. 16-28.ii.39.

Native name. Kashondanegu (Lulega).

Variation. Except for six examples from Idjwi Island recorded by
Stemfeld in 1912, this species is known only from the c? and 9 cotypes.
I therefore took the opportunity of securing the above series of topo-
types from which, after the tridactyle character of the entire series
had been checked, I selected fifty specimens for examination in con-
siderable detail and noted the following results.

Midbody scale-rows 20-24, only 2% with the former and 8% with
the latter; supraciliaries 5-7, 6 being normal; fingers 3; toes 3; lamellae
beneath fourth (i. e. median, as hallux and first toe are absent) toe
8-10. Two males (M. C. Z. 47702, 47728) have the frontonasal divided
or with an artifact simulating a division and thus presenting the ap-
pearance of possessing a pair of supranasals.

Coloration. In life. cf. Above, dark brown or black, many scales
flecked laterally with yellowish white ; tail suffused with pink. Below,
throat uniform black, body and fore limbs greenish white, each scale
with a dark brown spot ; hind limbs and tail brownish pink, each scale
spotted with black and white ; soles of all four feet black.

Of 27 sexed males it was observed that 14, ranging from 40-48 mm.,
and 1, of 51 mm. snout to anus, had throats which were white or heavily
spotted (progressively so with age), 12 others, ranging from 46-54
mm., had the throats so heavily spotted as to appear black. There is
considerable variation in the belly coloring, which may be white, each
flank scale, or all scales, with a black central spot.

Females differ from the males in substantially the same way as has
been described for L. meleagris.

Measurements. Largest <? (M. C. Z. 47701) measures 139 (50 +
89) mm., but is surpassed in length from snout to anus by d" cf
(M. C. Z. 47721, 47734) of 54 mm., and 9 9 (M. C. Z. 47727, 47737)
of 55 mm., with regenerated tails. A hatchling measures 44 (20 + 24)
mm., and slightly older skinks, 60 (25 + 35) and 62 (25 + 37) mm.

Breeding. Each of four females examined held either 2 or 3 eggs,
measuring 8x4 mm. and 9x4 mm. respectively.

Diet. Fragments of five centipedes, a scorpion, spider, beetle, and
grasshopper.

356 BULLETIN : MUSEUM OF COMPARATIVE ZOOLOGY

Enemies. No fewer than twenty-two of these skinks were recovered
from stomachs of wolf snakes (Lycophidion c. ornatum).

Habitat. More in evidence than either of the two species taken on
the Ruwenzori Mountains, for on sunny mornings they might be heard
rustling, and occasionally observed basking, among the fallen leaves
and short grass fringing the paths in the montane forest.

Ablepharus boutonii africanus Sternfeld

Ablepharus boutonii africanus Sternfeld, 1918, Abh. Senckenberg Nat. Ges.,
36, p. 423: Manda Island and Malindi, Kenya Colony; Pemba Island.

4 (M. C. Z. 47751) Mbanja, T. T. 2.V.39.

Said to occur at Mikindani also, but I failed to find it.

Variation. Midbody scale-rows 22-23.

Measurements. Larger d71 measures 105 (45 + 60) mm., and 9 , 98
(47 + 51) mm.

Breeding. On May 2, the larger 9 held 2 eggs measuring 11x5 mm.

Diet. Stomachs of two examined by Dr. F. A. Chace held at least
six crabs, three being referable to the genus Sesarma and three to Uca,
in^addition were unidentifiable remains of some other creature, possibly
an arthropod.

Ablepharus wahlbergii (Smith)

Cryptoblepharus wahlbergii A. Smith, 1849, 111. Zool. S. Africa, Rept., App.,
p. 10: Natal, South Africa.

21 (M. C. Z. 47752-4) Mikindani, T. T. 24.iii & 12.iv.39.
10 (M. C. Z. 47755-6) Kitaya, T. T. 28.iii-6.iv.39.
5 (M. C. Z. 47757-8) Mbanja, T. T. 26.iv.39.

1 (M. C. Z. 47759) Lindi, T. T. l.vi.39.

2 (M. C. Z. 47760-1) Mitunga, T. T. 8.V.39.

8 (M. C. Z. 47762-9) Nchingidi, T. T. lO.v.39.

Native names. Not unnaturally, this diminutive skink is thought to
be the young of Mabuya m. boulengeri and Riopa sundevallii by the
Yao and Konde at Mikindani and Kitaya.

Variation. Under certain conditions this species appears to produce
individuals which are markedly larger than the average. I (1933h, p.
324) had occasion to comment on such a skink from Nyamkolo, North-
ern Rhodesia. Usually such specimens occur spasmodically with those
of average size but on the Rondo Plateau (Mitunga and Nchingidi)

loveridge: African reptiles 357

the majority appear to be larger than those on the surrounding coastal
plain. Such an increase in size is accompanied by an increase in lamel-
lae beneath the toes and in the number of scales about midbody
(which, throughout East Africa are normally 24-26 with 22 (and
possibly 20) and 2S occurring as rare variants). To clarify the position
I furnish below the data derived from 26 skinks of the coastal plain
(1) that they may be contrasted with that derived from 10 examples
taken on the Rondo Plateau (2).

(1). Midbody scale-rows 24-26-28, only one with latter number,
average 25.2; lamellae beneath fourth toe 13-16, average 14.3; length
from snout to anus of six largest 40-51 mm., average 44 mm.

(2) Midbody scale-rows 26-28, only two with lower number, aver-
age 27.6; lamellae beneath fourth toe 15-18, average 16.6; length
from snout to anus of six largest 33-41 mm., average 36.8 mm.

Diet. A surprisingly large harvestman (Arachnida) in one.

Enemies. Recovered from the stomachs of wolf snakes (Lycophidion
c. capense and capense acutirostre) at Mbanja, from a house snake
(Boaedon I. lineatns) at Xchingidi.

Habitat. At Mikindani under piles of palm fronds, coconut husks,
and other rubbish; at Kitaya beneath heaps of wet weeds taken from,
and piled at edge of, a rice swamp. At Mbanja and Mitunga in clear-
ing camp sites beneath mango trees. At Xchingidi these little skinks
were commonly seen along the edges of the paths.

Scelotes tetradactylus tetradactylus (Peters)

Sepsina (Rhinoscincus) tetradactyla Peters, 1874, Monatsb. Akad. Wiss. Berlin,
p. 374: Zanzibar Coast.

4 (M. C. Z. 47776-9) Xchingidi, T. T. 12.V.34.

Variation. Midbody scale-rows 24; supraciliaries 4 (or 5?); fingers
4; toes 4; lamellae beneath fourth toe 3.

Remarks. The structural characters used by Hewitt to differentiate
the southwestern albcrti (n.n. for hessei Hewitt, 1927, not of Boettger,
1887) from tetradactylus, which he had not seen, are invalid. Even in
our small series of five (including adult from Uluguru Mountains)
the frontal may be as broad as (young) or broader than (adult) long,
while the relative lengths of the digits appear to be the same as those
of alberti, which I have not seen. Doubtless albcrti is a good race of
tetradactylus, though I do not know how to distinguish them except
by the former being said to have the scales margined with black,

358 bulletin: museum of comparative zoology

i

whereas in tetradactylus of the same size there is a dark central shaft
while the margins are pale brown. In adults these dark shafts ap-
parently disappear, being only faintly discernible with a lens.

Coloration. In life. 9 . Above, pale brown, head darker, each
dorsal and lateral scale with a median, longitudinal, black shaft,
which, by coalescing, form sixteen conspicuous dorsal and dorsolateral
lines, those on the flanks being narrower and fainter; tail with similar
dark lines on a background of rich ultramarine blue. Below, body and
limbs white; tail dull ultramarine with series of dusky lines resulting
from a dark streak down the centre of each. See also remarks above.

Measurements. All young or subadult, the largest (M. C. Z. 47776)
being only 114 (78 + 36) mm.

Habitat. All taken in damp sandy soil beneath logs lying at the
forest edge.

SCELOTES TETRADACTYLUS HEMPTINNEI ("Witte)

Plate 4, fig. 1.

Sepsina Hemptinnei de Witte, 1933, Revue Zool. Bot. Africaine, 23, p. 188:
Lukafu, Kundelungu, Katanga, Belgian Congo.

15 (M. C. Z. 47770-5) Ruanda near Ujiji, T. T. 10-15.iii.39.

Variation. Midbody scale-rows 22-24; supraciliaries 4 (or 5?);
fingers 4; toes 4; lamellae beneath fourth toe 1-2.

Remarks. This race, which is new for Tanganyika, apparently
occupies an intermediate position structurally as well as geographically
between the eastern tetradactylus and the western alberti. De Witte
distinguished hemptinnei from the former by its supposedly less de-
pressed snout, no such difference exists when direct comparison is
made between our types of hemptinnei and topotypes of tetradactylus.

Adult hemptinnei do differ, however, in their slightly more at-
tenuated bodies, their slightly more rudimentary digits as shown also
in lamellae reduction, and their slightly longer tails which are bronze,
instead of bright ultramarine blue.

These slight differences can best be appreciated by direct comparison,
and under these circumstances it would seem advisable to regard
hemptinnei as a race which has become rather more specialized for its
fossorial life than is the typical form.

Coloration. In life. 9 (M. C. Z. 47770). Above, iridescent bronze,
each dorsal scale, excepting the two median series, with a median,
longitudinal, black shaft, which go to form six, interrupted, longi-

loveridge: African reptiles 359

tudinal lines. Below, throat pinkish (as a result of blood vessels
showing through) except on buccal border which is white; rest of
undersurface white, each scale with a small, dusky, apical spot, such
spots being larger on the tail.

Measurements. Largest cf1 (M. C. Z. 47774) measures 136 (76 +
60) mm., largest 9 (M. C. Z. 47770) 153 (87 + 66) mm., but both
surpassed by others with a length from snout to anus of 85 and 91 mm.
respectively, with regenerated tails.

Breeding. In mid-March the testes of the males were large but the
ova in five females small, in a sixth were 2 eggs, measuring 10 x 4 mm.

Diet. Two stomachs held seven termites, a cockroach, and a spider.

Habitat. In the slightly damp sandy soil beneath piles of dry or
rotting vegetation in the gardens adjacent to the rice fields of Ruanda,
from four to six miles east of Ujiji on Lake Tanganyika.

Melanoseps — Key to the Species1

For more than a decade I have been hoping for an opportunity to
secure a series of Melanoseps which might enable one to get an idea
of the probable extent of variation and sexual dimorphism. The 24
specimens which we have obtained on the Rondo Plateau make it
necessary for me to correct my (1933h, p. 326) mistake of synonymiz-
ing var. lonaicauda Tornier with the typical form. As the short-tailed
type of ater was apparently a 9 , it seemed in accordance with the
other data then available to postulate that the long-tailed type of
longieauda was a male.

It now appears from the Rondo series that there is no appreciable
sexual dimorphism for the largest c? measures 93 + 27 mm., and the
largest 9 92 + 29 mm. It may be postulated that the genus ori-
ginated in Tanganyika Territory and has undergone tail reduction in
its process of fossorial specialization. If this be accorded priority as
the most important character, then increase in midbody scale-rows,
which is correlated with increase in size in favourable habitats, can
take its place as a secondary development. It is on this suggestion that
the following synoptical key is based.

1. Midbody scale-rows 18-20; size small, the maximum length from

snout to anus being 93 mm 2.

Midbody scale-rows 22-28; size large, the maximum length from
snout to anus being 160-210 mm 3.

1 Eastern forms only. M. occidentalis (Peters) of Cameroon omitted. Melanoseps acontias
Werner, 1913 (1912) is referred to Scoleeoseps.

360 bulletin: museum of comparative zoology

2. Length of tail contained 1.7 times in length from snout to anus;

range : northeastern Tanganyika a. longicauda

Length of tail contained 2.6 to 4.6 times in length from snout to
anus; range eastcentral (Mpwapwa) to southeast (Nchingidi)
Tanganyika a. rondoensis subsp. nov.

3. Midbody scale-rows 22-24; length of tail 2.6 to 4.1 times in length

from snout to anus 4.

Midbody scale-rows 26-28; length of tail 3.8 times in length from
snout to anus 5.

4. Midbody scale-rows 22; color below from chin to anus whitish

with blackish-brown lines resulting from the fusion of a blackish-
brown spot on the centre of each scale; maximum length from
snout to anus 160 mm.; range: (? Uluguru Mountains, Tan-
ganyika south to) Zambesi a. ater

Midbody scale-rows 22-24 (only 1 of the 12 types with the lower
number); color below from chin to anus pure white in sharp
contrast to the plumbeous tail, only occasional scales with a
dusky brown spot showing a tendency to form lines; maximum
length from snout to anus 166 mm.; range: Matengo Highlands,
southwestern Tanganyika a. matengoensis subsp. nov.

5. Color below substantially that of the typical form, viz. from chin

to anus pinkish white with brown lines resulting from the fusion
of a brown spot on the centre of each scale; range: Uzungwe
Mountains, southcentral Tanganyika a. uzungwensis subsp. nov.

Melanoseps ater rondoensis subspec. nov.
24 (M. C. Z. 47780-800) Nchingidi, T. T. 11-21.V.39.

One found dead on path between Kiponga and Mitunga, Rondo Plateau,
8.V.39.

Type. Museum of Comparative Zoology, No. 47780, an adult cf
from Nchingidi, 3000 feet, Rondo Plateau, Lindi Province, south-
eastern Tanganyika Territory. Collected by Arthur Loveridge, May
11-21, 1939.

Paratypes. Museum of Comparative Zoology, Nos. 47781-800, be-
ing 23 specimens with same data as type.

Diagnosis. As given in preceding key.

Description. Midbody scale-rows 18 (18-20 in paratypes); length
of unregenerated tail included in length from snout to anus 3.4

loveridge: African reptiles 361

(2.6-4.6) times. Otherwise similar to the typical form though colora-
tion probably paler.

Measurements. Largest c?1 (T}rpe) measures 120 (93 + 27) mm.,
largest 9 (M. C. Z. 47781) 121 (92 + 29) mm.

Breeding. In mid-May this female held two spherical eggs measur-
ing about 4 mm. in diameter.

Diet. Pupa of a midge and a click beetle larva in one stomach.

Habitat. In damp sandy soil beneath logs lying at the forest edge.
The majority were taken around the periphery of the forest-clearing
in which my tent was pitched, and it was across the sandy floor of
the latter on the cool day of our arrival that I found the first skink
wriggling. The entire series were taken by my two gunbearers and
self.

Range. I refer to this race, at least provisionally, two skinks taken
under similar habitat conditions at Mpwapwa and the Mkata River
in central Tanganyika Territory.

Melanoseps ater matengoensis subspec. now

Cotypes. Museum of Comparative Zoology, No. 44119, and eleven
others in Vienna Museum, from Ugano, Matengo Highlands, west of
Songea, Tanganyika Territory. Collected by H. Zerny, 1935-6.

Diagnosis. As given in preceding key.

Description. Midbody scale-rows 22-24 (22 in only one of the
twelve cotypes) ; length of unregenerated tail included in length from
snout to anus (2.6 to 4.1 times; coloration (see key) strikingly differ-
ent from that of the typical form.

Measurements. One cotype 9 (M. C. Z. 44119) measures 154 (115
+ 39) mm., but is surpassed by nine others (Vienna Museum) whose
length from snout to anus ranges from 121-166, the unregenerated
tails from 30-49 mm.

Melanoseps ater uzungwensis subspec. no v.

Melanoseps ater Loveridge (part). 1933h, Bull. Mus. Comp. Zool., 74, p. 326
(Kigogo, Uzungwe Mountains).

Type. Museum of Comparative Zoology, No. 51076, an adult 9
from Kigogo, Uzungwe Mountains, southern Tanganyika Territory.
Collected by Arthur Loveridge, November 22, 1929.

362 bulletin: museum of comparative zoology

Paratype. Museum of Comparative Zoology, No. 51077, a larger 9
with same data as the type.

Diagnosis. As given in preceding key.

Description. Midbody scale-rows 28 (26 in paratype); length of
unregenerated tail included in length from snout to anus 3.8 times.
Otherwise like the typical form.

Measurements. Type 9 (M. C. Z. 51076) measures 136 (108 +28)
mm., but is surpassed in length from snout to anus (210 mm.) by
the paratype with injured tail.

CHAMAELEOXTIDAE

As an offshoot of the Agamidae this specialized family might be
placed more appropriately in juxtaposition to the agamas; in the pres-
ent paper it is left in its old place for convenience of reference and to
preserve uniformity with the five previous reports dealing with the
herpetological collections which I have made in East Africa.

Chamaeleo senegalensis senegalensis Daudin

Chamaeleo senegalensis Daudin, 1802, Hist. Nat. Rept., 4, p. 203: Region
watered by the Senegal and Niger Rivers, Gambia and Guinea.

2 cf cf 3 9 9 (M. C. Z. 47151-2) Ujiji, T. T. 13.iii.39.

Coloration. In life. d". Above, pale pinkish brown, crown of head
darker; on vertebral line a series of eight rusty-orange blotches; flanks
with larger and smaller spots of the same color and a black line from
axilla to midbody where it breaks up into a few spots. Below, throat
purplish gray, the interstitial skin orange; gular-ventral crest pure
white; soles of feet and tail fawn to purplish brown.

In life. 9 . Above, yellow and grass green; on flanks a lateral series
of six black-edged white spots, below which is a more or less black-
edged white band from axilla to groin. Below, throat green, the inter-
stitial skin orange; gular-ventral crest pure white continued on as a
narrow line beneath tail; limbs and anal area pure white.

Measurements. Larger c? measures 190 (108 + 82) mm., largest 9 ,
240 (135 + 105) mm.

Breeding. On March 13 all three females were gravid, the largest
held 60 eggs, each measuring about 7 mm. in diameter.

loveridge: African reptiles 363

Chamaeleo dilepis quilensis Bocage

Chamaeleo dilepis var. Quilensis Bocage, 1866, Jorn. Sei. Lisboa, 1, p. 59:
Rio Quilo, north of Cabinda, Portuguese Congo.

3 (M. C. Z. 47153-4) Ujiji, T. T. 14.iii.39.

Variation. All are young under 150 mm. in length so that the occipi-
tal lobes are of little use in diagnosis, both males, however, have well
developed tarsal spurs.

Habitat. One was on a mhoga plant beside the path.

Chamaeleo dilepis dilepis Leach

Chamaeleo dilepis Leach, 1819, in Bowdich, Miss. Ashantee, App., p. 493:
Gaboon, i.e. French Congo.

3 c? 2 9 (M. C. Z. 47166-7) Kitaya. T. T. 25.iii.39.

3 d" 3 9 (M. C. Z. 47168-9) Mikindani, T. T. 18.iv.39.

1 9 (M. C. Z. 47170) Lindi, T. T. 26.iv.39.

5 9 (M. C. Z. 47171-4) Mbanja, T. T. 27-30.iv.39.

5 9 (M. C. Z. 47175-6) Nchingidi, T. T. 12.V.39.

1 9 (M. C. Z. 47177) Magrotto Mtn., T. T. l.vii.39.

Native names. Naluii (Kiyao); naluiku (Kimakonde); nalwiu (Ki-
mawiha); Ivvi (Kisambara).

Measurements. Largest cf (M. C. Z. 47168) measures 229 (110 +
119) mm., largest 9 (M. C. Z. 47171) 390 (195 + 195) mm. This and
another female from Mbanja are really gigantic, surpassing by 60 mm.
the largest (Morogoro) females I have ever taken elsewhere. It is
difficult to believe that they are subspecifieally the same as the gravid
females of 190 mm. from localities immediately north and south of
Mbanja and in no way dissimilar climatically or topographically.

Breeding. Though females were taken in all localities they were
noticeably gravid only at Kitaya, Nchingidi, and Magrotto, on the
dates given above.

Chamaeleo dilepis idjwiensis subspec. now
Plate 5, fig. 2.
7 cf 11 9 4 yng. (M. C. Z. 47155-65) Idjwi Id., B. C. 24-28.ii.38.

Type. Museum of Comparative Zoology, No. 47155, an adult c?
from Upper Mulinga River, Idjwi Island, Lake Kivu, Belgian Congo.
Collected by Arthur Loveridge, February 24-28, 1939.

364 bulletin: museum of comparative zoology

Paratypes. The rest of the series listed above and having the same
data as the type, many of them coming from the coffee plantations
further down the mountain.

Native name. Lumvu (Lulega).

Diagnosis. Intermediate between dilepis and roperi, agreeing with
d. dilepis in possessing large occipital lobes but differing in the absence
of a tarsal process (spur), in this latter respect agreeing with the small-
lobed d. roperi. Both sexes have the spines of the gular crest more
strongly developed than is the case with d. dilepis.

On first seeing females of this form I was struck by their handsome
coloring (see below, and pi. v, fig. 2) which seemed strikingly dif-
ferent from any that I had previously encountered in East Africa.

Coloration. In life, cf Type. Above, light or dark green mottled
and spotted with black; dorsum with seven, green -centered, black,
saddle-like markings, ten or more on tail; from axilla to groin a white
lateral band which, anteriorly, is bordered above and below by a few
scattered orange-red spots; occipital flaps posteriorly pure white.
Below, yellowish, throat with dark longitudinal lines and spotted with
orange-red; gular crest dark brown; ventral crest white.

In life. 9 Paratype. Above, light or dark green conspicuously
marked with numerous large orange spots; occipital flaps posteriorly
yellowish white. Below, dark green, interstitial skin of throat bright
yellow spotted with orange-red; gular crest brown; ventral crest white.
(Noticeably no white lateral band.)

Measurements. Total length of type cf, 250 (130+ 120) mm., of
paratype 9 (M. C. Z. 47156) 270 (140 + 130) mm., of young (M. C. Z.
47157) 66 (35 4- 31) mm.

Breeding. Four young, with dimensions approximating those
given above, were brought in. Ten of the females are conspicuously
gravid, the smaller eggs measuring 8 mm. in diameter.

Diet. One stomach held six green caterpillars and a Mylabris beetle.

Defence. Though, in protest at being photographed, the large female
bit a forefinger and hung on persistently, the teeth failed to break the
skin.

Chamaeleo bitaeniatus bitaeniatus Fischer

Chamaeleo bitaeniatus Fischer, 1884, Jahrb. Hamburg. Wiss. Anst., 1, p. 23,
pi. ii, figs. 7a-b: Lake Naivasha, Kenya Colony.

17 (M. C. Z. 47178-90) Mabira Forest, U. 7-19.xi.38.
16 (M. C. Z. 47191-9) Kibale Forest, U. 9-19.xii.38.
1 (M. C. Z. 47200) Bundibugyo, U. 26.xii.38.

loveridge: African reptiles 365

Native names. Nawolovu (Luganda); waruju (Lutoro); ameuli
(Luamba).

Coloration. In life. cf. Mabira. Above, pale blue green; centre of
casque pale blue; side of head with bright yellow horizontal streaks;
dorsal crest orange; flank with a light lateral line partially obscured by
three irregular vertical bars of brick red. Below, throat yellow mar-
gined with white, a very distinct black gular spot; ventral crest white;
inner side of limbs and tail lemon yellow.

In life. 9 . Mabira. Above, slightly olivaceous dirty white; dorsal
crest olive; flanks with a white lateral line immediately below which
is an interrupted olive line which commences at the orbit. Below, a
distinct black gular patch; throat, ventral crest, inner side of limbs,
and tail, more or less white.

Measurements. Largest cf (M. C. Z. 47191) measures 152 (80 +
72) mm., largest 9 (M. C. Z. 47178) 140 (80 + 60) mm., smallest
(Kibale) 64 (34 + 30) mm.

Breeding. In November small embryos present, in December
larger ones.

Chamaeleo bitaeniatus ellioti Gunther

Chamaehon Ellioti Gunther, 1895, Ann. Mag. Nat. Hist. (6), 15, p. 524, pi.
xxi, fig. A: Bugoye, eastern foot of Ruwenzori Mtns., Uganda.

25 (M. C. Z. 47201-9) Bugoye, U. 26-28.xii.38.

3 (M. C. Z. 47210-1) Mihunga, U. 9-19.1.39.

3 (M. C. Z. 47212-3) Nyakabande, U. 26.L39.

6 (M. C. Z. 47214-5) Mushongero, U. l.ii.39.
67 (M. C. Z. 47216-50) Kiraga, B. R. 8-12.ii.39.

Native names. Waruju (Lutoro); nyarungu (Lukonjo); lurnvu
(Lukiga).

Variation. Though formerly, with inadequate material, I regarded
this form as indistinguishable from typical bitaeniatus, now, with a
good series of topotypes one can separate ellioti by its longer gular-
ventral and dorsal crests, the latter being brick red or dried-blood red
(orange in typical bitaeniatus). C. b. ellioti increases in size as one
proceeds southwards and probably presents an average difference, a
matter which I hope to deal with at some future time.

Coloration. This is likely to play so important a part in any future
revision of the races that I took pains to record it in detail.

In life, cf . Bugoye. Above, pale blue green; centre of casque pale
blue or vellow; side of head with, or without, a few ill-defined horizon-

366 bulletin: museum of comparative zoology

tal streaks; dorsal crest brick red; flank with a distinct light lateral
line bordered below by red while both above and below irregular
mottlings of brick red may be present or absent. Below, throat
yellow, or blue with a central shaft of yellow, one or tivo black, or blue,
gular spots; ventral crest white; inner side of limbs and tail lemon
yellow.

In life. 9 . Bugoye. Above, light green or yellow; dorsal crest
reddish; flank with a white lateral line bordered below by rufous,
sometimes a second, but narrower, white line extending from axilla
to groin. Below, throat and under surface, including limbs, light
green or yellow, one, though normally two, black gular spots; ventral
crest and soles of feet white.

In life. d\ Kiraga. Above, a very bright green; centre of casque
grass green; sides of mouth showing a little yellow; dorsal crest dried-
blood red; flank with a light lateral line partly obscured by a very
broad dried-blood-red band which tends to throw off six, indistinct,
vertical bars to unite with the dorsal crest. Below, throat uniformly
green without black gular spots; ventral crest bright yellow anteriorly,
only white posteriorly from midbody; inner side of limbs bright
yellow; underside of tail green, browner posteriorly.

In life. 9 • Kiraga. Above, light green; supraoccipital and dorsal
crests rusty brown; flank with a white lateral line bordered below
by an obsolescent, interrupted, rusty brown band anteriorly extending
to the orbit. Below, throat pale green without black gular spots; ventral
crest pure white; inner side of limbs white; tail white anteriorly, brown
posteriorly.

Measurements. Largest Uganda cT (M. C. Z. 47201) measures 149
(81 + 68) mm., largest Uganda 9 (M. C. Z. 47215) measures 171
(90 + 81) mm. Ruanda series not measured.

Breeding. On January 18, a very small young measuring 50 (26 +
24) mm. was found on sedges in swamp at Mihunga. On January 26
and February 1 all eight females were gravid, the ova being small jn
only one.

Enemies. One recovered from the stomach of a kestrel (Falco t.
rufescens) shot at Mihunga.

Chamaeleo bitaeniatus ? bergeri Sternfeld

Chamaeleon bitaeniatus bergeri Sternfeld, 1912, Wiss. Ergebn. Deut. Zent.-
Afrika-Exped. 1907-1908, 4, pp. 250-253, pis. viii, fig. 4, and ix, fig. 5:
"Sirgoi" i.e. Sergoit, Kenya Colony.

d* 9 (M. C. Z. 47251-2) S. Kinangop Plateau, K. C. 27.X.38.

loveridge: africax reptiles 307

Native name. Kimbu (Kikuyu).

Variation. These chameleons, taken at 10,000 feet, though adult
are much too small to be referred to C. b. hohnelii which they closely
resemble except for two conspicuous rows of large, flat plates on either
flank. A female chameleon from Thomson's Falls, western Aberdare
Mountains, which one might expect to be subspecifically similar to the
Kinangop reptiles, was recently referred by Parker (1940a, p. 269) to
C. b. leikipiensis Steindachner, a name used to describe a chameleon
taken with hohnelii at the same type locality and altitude — 6000 feet,
and therefore invalid as a race of bitaeniatus.

C. b. bergeri Sternfeld was based on a single male from Sergoit, a
locality given as 6000-8000 feet. Heretofore I have considered it a
synonym of hohnelii but prefer to apply it to these chameleons rather
than propose another name at this time. I might add that the possi-
bility of identifying the Kinangop chameleons with C. b. schubotzi
(Mt. Kenya), C. b. altacclgonis (Mt. Elgon, 10,500 feet), or with
C. b. rudis (Mt. Ruwenzori, 10,000 feet) has been rejected after careful
consideration and comparison with topotypes of all three. Two
races are present on each of these mountains and on Kenya one (pre-
sumably from high altitude) appears subspecifically identical with the
Kinangop chameleons.

Measurements. Total length of d\ 121 (62 4- 59) mm., of 9, 146
(78 4- 68) mm. The holotype d1 of bergeri is 128 (64 + 64) mm.
according to Sternfeld who assumed that it was not fully adult.

Habitat. I found these in long grass among rank vegetation on the
banks of the Chania River.

Chamaeleo fischeri matschiei Werner

Chamaeleon matschiei Werner, 1895, Verh. Zool.-Bot. Ges. Wien, 45, p. 192:
Usambara Mountains, Tanganyika Territory.

c? (M. C. Z. 47253) Magrotto Mtn., T. T. 5.vii.39.

Native name. Luvi (Kisambara, but generic only).

Synonymy. In the absence of typical /. fischeri from Nguru Mtn.,
I continue to use the name matschiei though it may well be a synonym
of fischeri.

Measurements. Total length 332 (17 4- 125 4- 190) mm., the first
measurement in parenthesis being the projecting horns.

Coloration. In life. cf. Above, a rich, dark, velvety green, edge of

368 bulletin: museum of comparative zoology

casque and vertebral ridge black; horns horn-color; tail barred with
darker Below, inner aspect of limbs and underside of tail grayish;
soles of feet creamy yellow. Turns dark black all over when molested.

Chamaeleo xenorhinus Boulenger
Plate 6, fig. 2.

Chamaeleon xenorhinus Boulenger. 1901, Proc. Zool. Soc. London, 2, p. 135,
pi. xii: Eastern Ruwenzori Mountains, 6000 feet, Uganda.
cf (M. C. Z. 47254) Mubuku Valley, U. 2.xii.38.
9 (M. C. Z. 47255) Mihunga Ridge, U. 18.L39.

These specimens are topotvpes from 7000 and 6000 feet respectively.
In addition to the original pair only four others have been recorded
in the literature over a period of forty years.

Coloration. In life. cf. Above, head and flanks pale bluish green,
each scale standing out against the purplish brown venations of the
skin- vertebral ridge and upper side of tail (except for its green tip)
purplish brown with four darker saddle-like markings on dorsum and
eight on tail; fore limbs pale yellow green; hind limbs dark moss green
interspersed with purplish brown. Below, chin and jaws tinged with
green, otherwise dirty white; tail mottled with greenish white and

purplish brown.

In life. 9 . Above, head and saddle-like markings a soft, olive,

moss green. . ,

Measurements. Total length of c? 209 (8 + So + 116) mm., the
first measurement in parenthesis being the projecting horn; largest 9

210 (95 +H5) mm.

Breeding. On January 18 the female held 6 eggs, each measuring

about 15 x 10 mm.

Diet. The female's stomach held two flies, two plant beetles, a
metallic wood-boring beetle, a caterpillar, and a stink bug.

Parasites Two nematodes (Entomelas chamaeleoms) present in
mesenterv, apparently lungs, of male, and several in rectum of female

Habitat Though strenuous efforts were made to obtain a series ol
these chameleons they met with no success, the female being the only
one brought in. It seems probable that they are dwellers in the forest
canopv The male was found crawling over vegetation piled at the
periphery of our camp clearing in the heart of the forest. It was
obviously ailing and died within half-an-hour, it might possibly have
been injured by the staves of porters slashing down the ferns when

loveridge: African reptiles 369

making the clearing a few days previously though there were no signs
of external injury to support such an idea, more probably parasitiza-
tion was responsible.

Chamaeleo melleri (Gray)

Ensirostris melleri Gray, 1864, Proc. Zool. Soc. London, p. -478, pi. xxxii.
fig. 1: Mountains in the interior of East Africa.

4 (M. C. Z. 47256-9) Mikindani, T. T. 9 & 21.i.39.

Native names. Litagamulu (Kiyao); natendalechi (Kimakonde at
Mikindani); nandendereji (Kimakonde at Mbanja); naliumondo
Kimawiha).

Measurements. Larger cf measures 555 (7 + 228 + 320) mm., and
9 500 (6 + 244 + 250) mm., their horns obviously worn down as in
the much younger specimens they project 10 and 12 mm. beyond the
snout. -.

Habits. Mr. Kent of the Public Works Department at Lindi, told
me that one of these giant chameleons lived for a long time in a tree
above his quarters. He said that each day about noon, with surprising
regularity, it went to the end of a certain branch and defaecated. It
was the sight of the accumulated droppings in one spot which at-
tracted his attention to the reptile's routine.

Habitat. Though occurring on mountains this is not a montane
species for my camp at Mikindani was almost at sea level and I cap-
tured the large (non-breeding) male in an acacia just behind my tent.
The three others were brought in by local Wamawiha.

Chamaeleo johnstoni johnstoni Boulenger

Plate 6, fig. 1.

Chamaeleon johnstoni Boulenger, 1901, Proe. Zool. Soc. London, 2, p. 136
pi. xiii: Eastern Ruwenzori Mountains at 6000 feet, LTganda.

1 d* 1 9 (M. C. Z. 47260-1) Mubuku Valley, U. l.xii.38.

3o"3 9 (M. C. Z. 47262-7) Mihunga Ridge, U. 9-21. i.39.

3 cf 3 9 (M. C. Z. 47268-73) Mushongero, U. l.ii.39.

7 cf 6 9 (M. C. Z. 47274-83) Nyondo, nr. Kisenyi, B. R. 8.ii.39.

Native names. Xyampimpina (Lutoro) ; nyarungu ya pembi (Lukonjo
for male) ; nyaruju (Lukiga for male, the hornless female being known
by the generic name of Lumvu).

Coloration. In life. d*. Mihunga (topo type). Above, light to dark
moss green, head a lighter, more yellowish green with dark spots on

370 bulletin: museum of comparative zoology

casque and five sepia brown bands on side radiating outwards from
the eyeball; vertebral line with fourteen pairs of saddle-like black
blotches of which the first and second are continued on flanks as an
ill-defined marbled band with bright burnt-orange centre, the third
and fourth similar but without the orange, these four are all on the
dorsum, the remaining ten on the tail; limbs spotted with black, each
spot consisting of a dusky centre and black outer ring. Below, paler
with a faintly indicated white ventral line; inner side of limbs dirty

white.

Measurements. Largest d* (M. C. Z. 47268) measures 267 (25 +
120 + 122) mm. the first measurement in parenthesis being the pro-
jecting horn; largest 9 (M. C. Z. 47269) measures 265 (135 4- 130)

mm.

Diet. One stomach held a grasshopper, striped coccinelid, banana
snail, and a seccnd snail 10 mm. in diameter, others held snails.

Habitat. I captured one 9 , which was at a height of twelve feet
from the ground, in a small tree beside the path. The Xyondo series
were brought down to my camp at Kiraga, where they are said not to
occur, by pupils of the Catholic Mission. It appears probable that
other collectors who labeled their John stent material as from "Kisenyi"
obtained them from natives who captured them up the mountain.

Brookesia spectrum boulexgeri (Steindachner)

Plate 5, fig. 1.

Rhampholeon boulengeri Steindachner, 1911, Anz. Akad. Wiss. "Wien, p. 178:
Rainforest behind sandhills, northwest of Lake Tanganyika, Belgian
Congo.

1 (M. C. Z. 47296) Mihunga, Ruwenzori Mtns., U. 184.39.

3 (M. C. Z. 47297-9) Upper Mulinga, Idjwi Id., B. C. 20.ii.39.

Native names. Nyampimpina (Lutoro); lumvu (Lulega). Both
Batoro and Balega regard these pigmy chameleons as the young of
the larger Chamaeleo, and therefore do not differentiate them by

name.

Distribution. The Mihunga record is the first for this species from
Uganda. I take this opportunity of pointing out that these represent
boulengeri of Steindachner and Sternfeld, not of Schmidt and Witte,
for though Schmidt (1919, p. 595) pointed out that there were prob-
ably two forms in Central Africa, he applied the wrong name to his
sixty-three females from the Ituri, these represent B. s. affinis (Stein-
dachner), type locality Beni, Ituri region.

LOVERIDGE: AFRICAN REPTILES 371

Variation. In making both affinis and bovlengeri races of spectrum
I am influenced by the fact that both, except for their much shorter
tails, differ only from the typical form in a number of average char-
acters, viz. the even shorter rostral process, the less developed or
even absent supraciliary crest's flexible process, the flatter parietal
region, the casque being even less elevated posteriorly, and the shorter
isolated spines at the base of the bicuspid claws, which Steindachner
apparently overlooked, stating that they were absent. The two forms
are geographically and structurally but stages in the evolutionary
development from the more simple platyceps of Xyasaland to the
better equipped spectrum of the French Congo and Cameroon.

Measurements. Larger cf measures 6-4 (4S + 16) mm., and 9
measures 61 (47 + 14) mm. Both from Idjwi Island, Lake Kivu.

Breeding. Though so small, 49 and 61 mm. respectively, both
Mihunga and Mulinga females were gravid, the latter holding 3
spherical eggs measuring about 6 mm. in diameter. The eggs in the
former were damaged.

Diet. Two stomachs held many small flies, a small cockroach,
spider and woodlouse.

Habitat. The L'ganda female was struck by a hoe as we were dig-
ging out a very dry and rotten clump of sedge roots at the base of a
giant wild banana growing on the banks of the small stream which
meanders through the swamp at the base of Mihunga Ridge. Some
yolky masses were ruptured by the hoe and it occurred to me that if
the species is oviparous, the female may have burrowed into the
powdery roots to lay, alternately it is feasible to suppose that it had
fallen in unnoticed.

Brookesia brevicaudata (Matschie)

Chamaeleon (Brookesia) brevicaudata Matschie, 1892, Sitz. Ges. Naturf.
Berlin, p. 107: Derema, Usambara Mountains, Tanganyika Territory.

1 (M. C. Z. 47300) Nchingidi, Rondo Plateau, T. T. 12.V.39.

Distribution. This interesting record helps to bridge the gap in
the distribution of this species, heretofore known from the Shire
Highlands in Xyasaland (brachyurus), otherwise only north of the
Rufigi River in Tanganyika Territory.

Variation. The only African Brookesia with a beard-like 'tuft' of
scales forming a flexible process on the chin.

Enemies. Recovered from the stomach of the Usambara green
snake (Chlorophis macrops).

372 bulletin: museum of comparative zoology

BIBLIOGRAPHY

of papers referred to in the text more than once.

Barbour, T., and Loveridge, A.

1928. "A Comparative Study of the Herpetologieal Fauna of the
Uluguru and Usambara Mountains, Tanganyika Territory, with
Descriptions of new Species." Mem. Mus. Comp. Zool., 60,
pp. 87-265, pis. i-iv.

Bogert, C. M.

1940. "Herpetologieal Results of the Vernay Angola Expedition."

Bull. Am. Mus. Nat. Hist., 77, Art. 1, pp. 1-107, figs. 1-18, pi. i.
1942a. "Pseudohaje Gunther, a valid Genus for two West African arboreal
Cobras." Am. Mus. Novit,, No. 1174, pp. 1-9, figs. 1-8.

BOULENGER, G. A.

1889a. "Catalogue of the Chelonians, Rhynchocephalians, and Croco-
diles in the British Museum (Natural History)." pp. i-ix + 1-311,
figs. 1-73, pis. i-vi. (London).

1896d. "Catalogue of the Snakes in the British Museum (Natural His-
tory)." 3, pp. i-xiv + 1-727, figs. 1-37, pis. i-xxx. (London).

Loveridge, A.

1923g. "Notes on East African Tortoises collected 1921-1923, with the
Description of a new Species of Soft Land Tortoise." Proc. Zool.
Soc. London, pp. 923-933, pis. i-ii.

1928d. "Field Notes on Vertebrates collected by the Smithsonian-
Chrysler East African Expedition." Proc. U. S. Nat. Mus., 73,
Art. 17, pp. 1-69, pis. i-iv.

1933h. "Reports on the Scientific Results of an Expedition to the South-
western Highlands of Tanganyika Territory. VII. Herpetology."
Bull. Mus. Comp. Zool., 74, pp. 197-416, pis. i-iii.

1936j. "Scientific Results of an Expedition to Rain Forest Regions in
Eastern Africa. V. Reptiles." Bull. Mus. Comp. Zool., 79, pp.
209-337, pis. i-ix.

1939c. "Revision of the African Snakes of the Genera Mehelya and
Gonionotophis." Bull. Mus. Comp. Zool., 86, pp. 131-162, figs. 1-2.

1940a. "A Serpent-seeking Safari in Equatoria. I. Uganda." Scientific
Monthly, 50, pp. 498-512, photos.

1941d. "Fevision of the African Terrapin of the Family Pe!omedusidae."
Bull. Mus. Comp. Zool., 88, pp. 467-524.

Parker, H. W.

1936c. "Dr. Karl Jordan's Expedition to South-West Africa and Angola:
Herpetologieal Collections." Novit. Zool., 40, pp. 115-146,
figs. 1-2.

loveridge: African reptiles 373

1936e. "Reptiles and Amphibians collected by the Lake Rudolf Rift

Valley Expedition." Ann. Mag. Nat. Hist. (10), 18, pp. 594-609,

figs. 1-13.
1940a. "Undescribed Anatomical Structures and new Species of Reptiles

and Amphibians." Ann. Mag. Nat. Hist. (11), 5, pp. 257-274,

figs. 1-3.

Pitman, C. R. S.

1936. "A Guide to the Snakes of Uganda." Uganda Journal (Kampala),
3, pp. 211-229, 259-281, pis. iii-v, col. pis. A-D. 4, pp. 41-61,
126-150, pis. vi-x, col. pis. E-G.

1937. hoc. cit. supra. 4, pp. 220-246, 319-349, pis. xi-xiii, col. pis. H-M.

1938. Loc. cit. supra. 5, pp. 160-244, pis. xv-xviii, col. pis. R-W.

Schmidt, K. P.

1919. "Contributions to the Herpetology of the Belgian Congo based
on the Collection of the American Museum Congo Expedition
1909-1915. Part I. Turtles, Crocodiles, Lizards and Chame-
leons." Bull. Am. Mus. Nat. Hist., 39, pp. 385-624, figs. 1-27,
pis. vii-xxxii.

Sternfeld, R.

1910a. "Die Schlangen Deutsch-Ostafrikas" in "Die Fauna der deuts-

chen Kolonien," 3, part 2, pp. i-iv + 1-47, figs. 1-54, map.

(Berlin).
1910e. "Neue Schlangen aus Kamerun, Abessynien, u. Deutsch-Ost-

afrika." Mitt. Zool. Mus. Berlin. 5, pp. 67-70.
1912c. "Reptilia" in "Wiss. Ergebn. Deut. Zent.-Afrika-Exped. 1907-

1908, 4, pp. 197-279, figs. 1-4, pis. vi-ix. (Leipzig, 1913).

TORNIER, G.

1896. "Die Kriechthiere Deutsch-Ost-Afrikas. Beitrage zur Systematik
und Descendenzlehre." pp. i-xiii + 1-164, figs. 1-11, pis. i-v.
(Berlin, republished 1897).

EXPLANATION OF PLATES

PLATE 1

Loveridge — African Reptiles

PLATE 1

Map showing Principal Collecting Localities

1938

Landing at Mombasa (25.x), except for a stopover at Naivasha and Kinangop
(26-31. x), Loveridge proceeded by rail direct to Jinja (1-5. xi). Thence to
Mabira Forest (5-21. xi), Budongo Forest (22.xi-7.xii), Kibale Forest (8-19.
xii), Bundibugyo near Bwamba Forest (19-26. xii), Bugoye, foot of Ruwenzori
Mountains (26-28.xii) and Mubuku Valley at 7000 ft. (29.xii-).

1939

On leaving Mubuku (-9.i) Loveridge descended down the valley to Mihunga,
circa 6000 ft. (9-21.i), then back to Bugoye (21-24.i), Nyakabande (25-30.i),
Mushongero (30.i-4.ii), returned to Nyakabande (4-8. ii.); Kisenye (8-13. ii),
Goma (13-14.ii), Mamvu on Idjwi Island (14-16. ii), Upper Mulinga River
on Idjwi (16.ii-6.iii), Uvira (7-8.iii), Ujiji (9-16. iii), Dares Salaam (18-19.iii),
Mikindani (22-24.iii), Kitaya (24.iii-7.iv), Mikindani (7-24.iv), Mbanja
(25.iv-6.v), Lake Rutamba (6-8.v), Nchingidi (9-21.v), Lindi (22.v-4.vi),
Siga Caves (7-17.vi), Amboni Estate (17-27.vi), Magrotto Mountain (27.vi-
21.vii), Tanga (21-23. vii), Likoni opposite Kilindini (24-26. vii).

BULL. MUS. COMP. ZOOL.

Loveridge. African Reptiles. Plate 1

^k\e£y forest
J$¥:.Kibal< ftCSj
MuouKu Ktlley-Mr&f JJNJ5

■?',Cte> Mabira

LaX< MataXiJS----

Late KivtMd%.

UGANDA,.-

Cw(*r«o

Uvira

BELGIAN
CONGO

KENYA
COLONY

Ami.*; Sstli^^^

TAN G AN Y^kXA
TER RI TORY

is

(f>

V

<

Tut

1 — ,

( f-V

<o

( o

Ui

; b

Q

C pJ

O

• o \

*

) 1

*

/ T

K. *

^

.^

^

V

H'

Zanzibar

iar ES SALAAM

'ik.inda.ni

Scale in Miles

0 100 200

1 .1 I I I I 1

PLATE 2

Loveridge — African Reptiles

PLATE 2

Fig. 1. Emin's Worm Snake (Leptotyphlops emini emini)

During his retreat with H. M. Stanley from the southern Sudan, Emin
Pasha discovered an example of this diminutive species. The uniformly
silvery-black emini ranges from the Belgian Congo to the East Coast, but is
replaced on Pemba Island by a form (pembae: 1941) which consistently differs
in coloration and proportions.

Fig. 2. Lestrade's Blind Snake (Typhlops blanfordii lestradei)

This beautifully glossy species varies considerably in coloration, rich coppery
brown or black examples are those most frequently encountered. Many were
found beneath lava boulders on the lava-strewn plains near Mount Mikeno,
others in termitaria upon whose owners, as well as ants, they were found to
be feeding. On Idjwi Island a Lestrade's blind snake was found in the stomach
of another rare species — Miodon gabonensis graueri.

Fig. 3. African Keeled Green Snake (Hapsidophrys lineata)

The figured female, a rich velvety green, its color accentuated by ten
longitudinal black lines, was captured while moving slowly along a branch in
deep forest. It was only five feet from the ground and as I took it by the neck
made no attempt to bite, nor later when subjected to considerable provocation
during a quarter-of-an-hour's posing for its photograph.

BULL. MUS. COMP. ZOOL.

Loveridge. African Reptiles. Plate 2

f

fWz*-

r4*

' • '

* C

'

■

.1 ••

I ** ■

'

^ ; *

PLATE 3

Loveridoe — African Reptiles

PLATE 3

Fig. 1. White-lipped Snake (Crotaphopeltis hoiamboeia hotamboeia) .

This opisthoglyphous colubrine was encountered in the roadway at Ujiji;
it immediately assumed the threatening attitude depicted and, flattening its
head, struck savagely at anything approaching. So effective is the display
that one instinctively accords such a snake the deference due to a more
dangerous species.

Fig. 2. Gaboon Viper (Bitis gabonica) from Budongo Forest, Uganda.

At the same time one of the most beautiful, deadly, and sluggish of African
snakes, this fifty-one inch reptile was found concealed in a patch of grass in
the middle of a path from which it was removed to camp to be photographed.
Her weight, eight pounds, was less than that of smaller examples though her
stomach held two rats and many parasites of several species.

Fig. 3. Sedge Viper (Athens nitschei nitschei) from Uganda.

The female shown, was one of a pair which I captured as they were basking
in the late afternoon sun in a swamp at the base of the Ruwenzori Mountains.
Both reptiles were at a height of sLx feet from the ground as they lay, tightly
coiled but flat as plates, on top of dense tangles of creepers which smothered
the elephant grass growing along the banks of a little stream meandering
through a swamp at Mihunga, type locality of the synonym A. woosnami.

BULL. MUS. COMP. ZOOL.

Loveridge. African Reptiles. Plate 3

/■ *.:5

PLATE 4

Lovebidge — African Reptiles

PLATE 4
Fig. 1. Hemptinne's Skink (Scelotes tetradactylus hemptinnei).

The series of these skinks found in slightly damp sandy soil beneath piles
of dry or rotting vegetation in the gardens adjacent to the rice fields of Ru-
anda, from four to six miles east of Ujiji on the eastern shore of Lake Tan-
ganyika, constitute the first record of the occurrence of this western race in
the Territory.

Fig. 2. Spotted-lip Skink (Mabuya maculilabris maculilabris) .

An examination of stomachs of this handsome species, revealed real wasps
as well as a parasitic wasp, ants, plant bugs, a weevil and other beetles, moths,
grasshoppers, a cockroach, fly and large spider. The figured specimen was
captured on Idjwi Island, Lake Kivu, and is therefore topotypic of the alleged
variety kividjwie7isis.

Fig. 3. Black-and-Gray Skink (Lygosoma meleagris) Mt. Ruwenzori.

On visiting the type locality of this little species, we found it plentiful
though laborious to collect on account of its secretive habits. Many eggs, in
various stages of development, were found, one pair was uncovered from
which a skink had just hatched and even as we watched its fellow emerged.
To escape from the egg the little. creature makes a cut two-thirds the length
of one side and involving the end of the parchment-like shell.

BULL. MUS. COMP. ZOOL.

Loveridge. African Reptiles. Plate 4

PLATE 5

Loveridge — African Reptiles

PLATE 5

Fig. 1. Pigmy Chameleon (Brookesia spectrum boulengeri).

Two-and-a-half inches in length, and lacking the prehensile tail of the
typical genus, these small creatures never assume the bright colors of the
larger forms, but are restricted to the shades usually associated with the dead
leaves which they so closely resemble. The photograph is of a male on Idjwi
Island.

Fig. 2. Kivu Chameleon (Chamaeleo dilepis idjwiensis).

The orange-spotted, green female figured above, is more handsome than her
black-spotted mate. In protest at being photographed she bit a forefinger
and hung on persistently, despite her size of over ten inches, however, the
teeth failed to break the skin. A stomach held sLx green caterpillars and a
Mylabris beetle so that these chameleons perform a useful service in the coffee
plantations which they frequent.

BULL. MUS. COMP. ZOOL

Loveridge. African Reptiles. Plate 5

■

PLATE 6

Loveridge — African Reptiles

PLATE 6

Fig. 1. Johnston's Chameleon (Chamaeleo johnstoni johnstoni).

Topotypes of this moss-green and burnt-orange species were taken on the
Ruwenzori Mountains where snails form an item in their diet. The males
alone bear horns, a character which distinguishes the typical form from the
eastern Congo race — itiiriensis — in which both sexes are hornless.

Fig. 2. One-horned Chameleon (Chamaeleo xenorhinus) of Uganda.

Among the choicest acquisitions resulting from the expedition were a
topotypic pair of this rare species, being difficult to obtain on account of its
habitat in the forest canopy. With this species again it is the male alone that
is horned, though, as may be seen from the photograph, the female displays
an indication of one on the tip of the snout.

BULL. MUS. COMP. ZOOL

Loveridge. African Reptiles. Plate 6

IlvNIlvW

19

» »

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE

Vol. XCI, No. 5

SCIENTIFIC RESULTS OF A FOURTH EXPEDITION
TO FORESTED AREAS IN EAST & CENTRAL AFRICA

V

AMPHIBIANS

By Arthur Loveridge

The Library
Museum of Comparative Zoo

With Four Plates HaPVard UnlVeri.

CAMBRIDGE, MASS., U.S.A..

PRINTED FOR THE MUSEUM

December, 1942

PUBLICATIONS
OF THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

The Bulletin and Memoirs are devoted to the publication of
investigations by the Staff of the Museum or of reports by spec-
ialists upon the Museum collections or explorations.

Of the Bulletin, Vols. I to XC, Vol. XCI, Nos. 1, 2, 3 and
4, and Vol. XCII, No. 1 have appeared and of the Memoirs,
Vol. I to LVL

These publications are issued in numbers at irregular intervals.
Each number of the Bulletin and of the Memoirs is sold separately.
A price list of the publications of the Museum will be sent upon
application to the Director of the Museum of Comparative
Zoology, Cambridge, Massachusetts.

After 1941 no more Memoirs are to be published.

'M.C.-L-L-

Bulletin of the Museum of Comparative Zoology

AT HARVARD COLLEGE
Vol. XCI, No. 5

SCIENTIFIC RESULTS OF A FOURTH EXPEDITION
TO FORESTED AREAS IN EAST & CENTRAL AFRICA

V

AMPHIBIANS

By Arthur Loveridge

With Four Plates

CAMBRIDGE, MASS., U.S.A.
PRINTED FOR THE MUSEUM
December,/1942J

»* Zootomy ***\

U£C 23 1942)

No. 5. — Scientific Results of a Fourth
Expedition to Forested Areas in East and Central Africa

V

Amphibians

By Arthur Loveridge

CONTENTS

Page

Introduction 377

Acknowledgements 378

Summary of Taxonomic Alterations 379

Index to List of Species collected 380

Systematic Discussion 382

Bibliography -436

INTRODUCTION

The collection on which the following report is based, was made by
the author while investigating the fauna of certain forested regions of
East and Central Africa. The enquiry was carried out on behalf of
the Museum of Comparative Zoology with a fellowship granted by
the John Simon Guggenheim Memorial Foundation of New York.

A synopsis of the itinerary is given in the caption accompanying
Plate 1 — a map showing the principal position of the collecting locali-
ties. Altitudes and detailed information regarding the various camps
will be furnished in the final report of this series which will deal with
the general conclusions arrived at.

The period of collecting amphibians was from October 27, 1938, to
July 25, 1939, during which time 1,681 amphibians, representing 77
forms, were secured. Ten of these were new to the Museum of Com-
parative Zoology, exclusive of certain others not previously recognized,
but now considered valid as a result of this study of additional ma-
terial.

378 bulletin: museum of comparative zoology

Three forms are here described as new, one (in parenthesis) is based
on material taken during an earlier expedition. The new forms are:

Bufo micranotis rondoensis Nchingidi, Rondo Plateau, T.T.

Hyperolius parkeri rovumae Kitaya, Rovuma River, T.T.

(Arthroleptis xenodactyloides nkukae) Nkuka Forest, Rungwe Mtn., T.T.

In addition to these new forms, the undermentioned races or species
are recorded for the first time for certain countries.

New for Uganda

Bufo funereus Bocage
Chiromantis rujescens (Gunther)
Hyperolius alticola Ahl
Rana fuscigula angolensis Bocage
Rana christyi Boulenger
Phrynobatrachus plicatus (Gunther)
Phrynobatrachus versicolor Ahl

Four of these, taken in the Budongo Forest, were already known
from the Ituri Forest of the eastern Belgian Congo.

New for Tanganyika Territory

Leptopelis concolor Ahl

Hyperolius kivuensis Ahl

Rana mascareniensis venusta Werner

New for Belgian Congo
Xenopus laevis bunyouiensis Loveridge
Arthroleptis xenochirus Boulenger

New for Belgian Ruanda
Rana mascareniensis venusta Werner

Several species, most of which may be regarded as rareties, might
be singled out for special mention, among them: Leptopelis n. christyi,
Megalixalus uluguruensis, Hyperolius sehubotzi, H. flawmaeulatus,
Arthroleptis xenochirus, Spelaeophryne methneri, and Hoplophryne
rogersi.

ACKNOWLEDGEMENTS

The opportunity is taken of thanking Dr. Thomas Barbour, Director
of the Museum of Comparative Zoology, for the support and encour-
agement which he has given to the prosecution of this undertaking,

loveridge: African amphibians 379

and to the John Simon Guggenheim Memorial Foundation without
whose generous aid this expedition would not have been possible.

In appreciation of the action of His Excellency the Governor of the
Congo Beige in granting permission to collect on Idjwi Island, a se-
lection of duplicates of such species as were collected in Belgian terri-
tory are being set aside for dispatch to the Congo Museum, Tervueren,
after the German evacuation of Belgium.

The photographs illustrating this report were taken by my son,
Brian A. Loveridge, and for permission to use the blocks for plate 4
we are indebted to the Editor of the Scientific Monthly, in which
journal (June and July, 1940) they appeared as illustrations to a
popular account of the safari.

The colored plates 2 and 3 showing age and sex dichromatism in
the RHACOPHORIDAE are furnished through the generosity of
Dr. Barbour. They represent the skillful work of Mr. Eugene N.
Fischer, being based on color notes made in the field.

SUMMARY OF TAXONOMIC ALTERATIONS

The following forms are accorded subspecific rank :

Hylambates christyi Boulenger as Leptopelis notatus christyi (Boulenger).
Rappia sansibarica Pfeffer as Hyperolius aitrinus sansibaricus (Pfeffer).

Topotypes of numerous doubtful species were collected and result
in my considering the following as synonyms:

*Bufo r. kisoloensis Loveridge = *B. regularis regularis Reuss

Leptopelis budduensis AM = *? Leptopelis n. christyi Boulenger

* Hyperolius macrodactylus Ahl = Hyperolius schubotzi Ahl
Hyperolius ornatus Laurent = Hyperolius schubotzi Ahl

"Hyperolius kwidjwiensis Ahl = "Hyperolius kivuensis Ahl

*Hyperolius kandti Ahl = "Hyperolius kivuensis Ahl

* Hyperolius koehli Ahl = * Hyperolius kivuensis Ahl
"Hyperolius substriatus Ahl = "Hyperolius puncticulatus (Pfeffer)
"Hyperolius microps Gunther = * Hyperolius pusillus (Cope)
"Hyperolius milnei Loveridge = "Hyperolius pusillus (Cope)

Arthroleptis schubotzi Nieden = "Arthroleptis xenodactylus Boulenger

"Arthroleptis affinis Ahl = Arthroleptis adolfifriederici Nieden

* Arthroleptis schoenebecki Ahl = Arthroleptis adolfifriederici Nieden
"Arthroleptis vagus Ahl = "Arthroleptis s. lonnbergi Nieden
"Arthroleptis ukamiensis Ahl = "Arthroleptis s. lonnbergi Nieden

♦Topotypes or paratypes compared.

380 bulletin: museum of comparative zoology

LIST OF SPECIES COLLECTED

Page

CAECILIIDAE

Boulengerula boulengeri Tornier

Scolecomorphus vittatus (Boulenger)

PIPIDAE

Xenopus laevis victorianus Ahl . . . '

Xenopus laevis bunyoniensis Loveridge ^

Xenopus muelleri (Peters)

BUFONIDAE

384
Bufo brauni Nieden

Bufo regularis regularis Reuss **

Bufo funereus Bocage

Bufo lonnbcrgi nairobiensis Loveridge &"

Bufo micranotis rondoensis subsp. nov ^°'

Neetophrynoides tornieri (Roux)

RHACOPHORIDAE

Chiromantis ? rufescens (Giinther) 389

Chiromantis xerampelina Peters ^

Leptopelis eoncolor Ahl

Leptopelis johnstoni (Boulenger) «j*

Leptopelis notatus christyi (Boulenger) *»-

Leptopelis vermiculatus (Boulenger) ^

Hylambates maadatus Dumeril ^

Kassina senegalensis (Dumeril & Bibron) . _ ^°

Megalixalus fornasinii fomasinii (Bianconi) «*95

Megalixalus leptosomus (Peters) _ &™

Megalixalus ulugurucnsis Barbour & Loveridge o9<

Megalixalus brachycnemis Boulenger 39b

Hyperolius schubotzi Ahl

Hyperolius kivuensis Ahl

Hyperolius sp

Hyperolius alticola Ahl

401

Hyperolius montanus (Angel) *j

Hyperolius undulatus (Bou'
Hyperolius udjijiensis Ahl

Hyperolius undulatus (Boulenger) ^0-

loveridge: African amphibians 381

Page

Hyperolius flavomaculatus Giinther 403

Hyperolius argentouttis Ahl 404

Hyperolius ahli Loveridge 404

Hyperolius puncticulatus (Pfeffer) 405

Hyperolius mariae Barbour & Loveridge 406

Hyperolius citrinus sansibaricus (Pfeffer) 407

Hyperolius citrinus citrinus Giinther 407

Hyperolius parkeri parkeri Loveridge 409

Hyperolius parkeri rovumae subsp. nov 410

Hyperolius nasutus Giinther 411

Hyperolius pusillus (Cope) 412

RANIDAE

Arthroleptides martiensseni Nieden 413

Rana albolabris albolabris Hallowell 413

Rana galamensis bravana (Peters) 413

Rana fuscigula fuscigula Dumeril & Bibron 414

Rana fuscigula chapini Noble 414

Rana fuscigula angolensis Bocage 415

Rana christyi Boulenger 415

Rana oxyrhynchus oxyrhynchus Smith 416

Rana oxyrhynchus gribinguiensis Angel 417

Rana mascareniensis mascareniensis Dumeril & Bibron 417

Rana mascareniensis venusta Werner 418

Rana occipitalis Giinther 419

Rana edulis (Peters) 419

Phrynobatrachvs spp 420

Phrynobatrachus natalensis (Smith) 421

Phrynobatrachus perpalmatus Boulenger 421

Phrynobatrachus acridoides (Cope) 421

Phrynobatrachus kinangopensis Angel 422

Phrynobatrachus plicatus (Giinther) 422

Phrynobatrachus krefftii Boulenger 422

Phrynobatrachus dendrobates (Boulenger) 423

Phrynobatrachus versicolor Ahl 423

Phrynobatrachus graueri (Nieden) 424

Arthroleptis minutus Boulenger 425

Arthroleptis xenodactylus Boulenger 426

Arthroleptis xenodactyloides nkukae subsp. nov 427

382 bulletin: museum of comparative zoology

Page

Arthroleptis xenochirus Boulenger 428

Arthrolcptis near poecilonotus Peters 429

Arthroleptis adolfifrirdrrici Xieden 429

Arthrolcptis stenodactylus lonnbergi Xieden 430

Arthroleptis stenodactylus stenodactylus Pfeffer 431

Cacosternum boettgeri boettgeri (Boulenger) 432

Hcmisus marmoratum marmoratum (Peters) 432

BREVICIPITIDAE

Callulina kreffti Xieden 432

Spelaeophryne methneri Ahl 434

Probreviceps macrodactylus macrodactylus (Xieden) 434

Breviccps mossambicus Peters 434

Hoplophrync rogersi Barbour & Loveridge 435

PHRYXOMERIDAE

Phrynomerus bifaseiatiis (Smith) 435

CAECILIIDAE
Boulexgerula boulexgeri Tornier

Boulengerula boulengeri Tornier, 1896, Kriechthiere Deutsch-Ost-Afrikas, p.
164: Usambara Mountains, Tanganyika Territory.

10 (M. C. Z. 25001-9) Magrotto Mtn., T. T. 29. vi-10. vii. 39.

Distribution. These constitute the first record of its occurrence on
this mountain which, however, is only twenty miles from the type
locality. S. vittatus has already been recorded from Magrotto by
Nieden (1912b).

Native name. Mango (Kisambara for caecilians).

Variation. Annuli 129-133, i.e. well within the range.

Size. Total lengths 112-250 mm.

Enemies. The largest specimen was recovered from the stomach of
an elapine snake (Elapsoidea gilntherii).

Habitat. Two juveniles were taken beneath logs; a 230 mm. adult
in soggy wood of a rotten stump near clump of bamboos growing by
stream.

loveridge: African amphibians 383

Scolecomorphl's vittatus (Boulenger)

BdeUophis vittatus Boulenger, 1895, Proc. Zool. Soc. London, p. 412, pi. xxiv,
fig. 4: Usambara Mountains, Tanganyika Territory.

4 (M. C. Z. 25010-3) Magrotto Mtn., T. T. 1-10. vii. 39.

Native name. As last.
Variation. Annuli 132-149.

Size. Total lengths 132-145 mm.; diameter into length 29.5 to 36.6
times.

PIPIDAE

Xenopus laevis victorianus Ahl

Xenopus victorianus Ahl, 1924, Zool. Anz., 60, p. 370: Busisi, Lake Victoria,
Tanganyika Territory.

5 (M. C. Z. 25014-7) Mabira Forest, U. 8. xi. 38.
50 (M. C. Z. 25018-27) Fort Portal, U. 12. xii. 38.

Native name. Kikere (Luganda).

Color. All are typical in having the bellies heavily spotted, those
from Fort Portal against a background of rich yellow.

Size. Lengths from 43-64 mm.

Habitat. The Mabira frogs were taken from a cement tank at the
Mubango coffee factory, the Toro series were brought in by two men.

Xenopus laevis bunyoniensis Loveridge

Xenopus laeiis bunyoniensis Loveridge, 1932, Proc. Biol. Soc. Washington, 45,
p. 114: Bufundi, Lake Bunyonyi, Uganda.

1 (M. C. Z. 25029) Nyakabande, U. 28. i. 39.
81 (M. C. Z. 25030-9) Mushongero, U. 1. ii. 39.

2 (M. C. Z. 25040-1) Mamvu, B. C. 22. ii. 39.

Native names. Magulumberi (Lukiga); derima (Lulega).

Color. Bright yellow below, at least posteriorly, only one light yel-
low one noted in the entire Mushongero series. My earlier assertion,
based on an even larger series, that the belly is usually spotted, does
not hold for the Mushongero series in which it is unspotted in 43
(lengths 25-48 mm.), and spotted only in 38 (lengths 28-68 mm.).

Size. This dwarf form from the cold waters of the Central African
Highlands, commonly breeding at 35 mm., furnished six examples
over 45 mm. of which the giant 68 mm. was a female, thus actually ex-

384 bulletin: museum of comparative zoology

ceeding victorianus! These half-dozen larger frogs, however, possess
the slender habitus of bunyoniensis, their greatest body width being
but slightly broader than the head.

Enemies. The feet of one frog were recovered from the stomach of
an otter (Lutra m. tenuis).

Habitat. These little frogs were extremely abundant in the waters
of Lake Mutanda, on whose shore Mushongero is situated. At times
we would see individuals which were unable to submerge, swimming
desperately on the surface with something of the appearance of half-
drowned shrews. My boatmen said that they were ill so I captured
one and found that it was inflated almost to the roundness of a ping-
pong ball. As already pointed out by Parker, the majority are afflicted
with helminths.

Xenopus muelleri (Peters)

Dactylcthra Muelleri Peters, 1844, Ber. Akad. Wiss. Berlin, p. 37: Mozambique.

14 (M. C. Z. 25042-50) Ujiji, T. T. 11. iii. 39.

Color. Bellies uniform, or only slightly spotted.
Size. Lengths 37-61 mm.

BUFONIDAE

Bufo brauni Nieden

Bufo brauni Nieden, 1910, Sitz. Ges. naturf. Freunde Berlin, p. 450: Amani,
Usambara Mountains, Tanganyika Territory.

2 (M. C. Z. 25051-2) Magrotto Mtn., T. T. 11. vii. 39.

Native name. Jula (Kisambara corruption of chura and applied to
most toads and frogs).

Color. Above, rich reddish leaf -brown paling towards the periphery ;
from snout to anus a hair-like, yellow, vertebral line; an interorbital
marking which is rich black finely edged with yellow; pairs of similar
markings on the scapular, lumbar, and supra-anal regions, between
these larger blotches there are several pairs of smaller spots; sides of
face, and flanks, black, sharply demarcated from the dorsal coloring;
limbs grayish or cream-colored, distinctly barred and spotted with
black. Below, whitish, the limbs showing some black spotting; palms
and soles black, but some of the tubercles white.

loveridge: afkican amphibians 385

bltfo regularis regularis reuss

Bufo regularis Reuss, 1834, Mus. Senckenberg., 1, p. 80: Egypt.
Bufo regularis kisoloensis Loveridge, 1932, Occ. Papers Boston Soc. Nat. Hist.,
8, p. 52: Kisolo, Kigezi District, Uganda.

8 (M. C. Z. 25053-8) S. Kinangop Plateau, K. C. 29. x. 38.
1 (M. C. Z. 25059) Budongo Forest, U. 29. xi. 38.

7 (M. C. Z. 25060-1) Kibale Forest, U. 9. xii. 38.

1 (M. C. Z. 25062) Bundibugyo, U. 20. xii. 38.

2 (M. C. Z. 25063) Mubuku Valley, U. 31. xii. 38.

3 (M. C. Z. 25064) Mihunga Ridge, U. 9. i. 39.

1 (M. C. Z. 25065) Nyakabande, U. 28. i. 39.
16 (M. C. Z. 25066-9) Mushongero, U. 1. ii. 39.

9 (M. C. Z. 25070-1) Kisenyi, B. R. 9. ii. 39.

2 (M. C. Z. 25072) Idjwi Id., B. C. 17. ii. 39.
1 (M. C. Z. 25073) Ujiji, T. T. 11. iii. 39.

• 1 (M. C. Z. 25074) Kitaya, T. T. 30. iii. 39.

1 (M. C. Z. 25075) Mikindani, T. T. 14. iv. 39.

2 (M. C. Z. 25076-7) Lindi, T. T. 1. vi. 39.

Native names. Kihere (Luganda; Lutoro; Lukonjo; Lulega); mtuvu
cf and ichcheri 9 (Lukiga); kuvifata (Luruanda); ligumi (Kiyao);
liumi (Kimakonde; Kimawiha).

Variation. It is with some misgivings that I refer all these toads to
the typical form: those from the Kinangop Plateau at 10,000 feet
certainly appeared very different in life from those taken at the coast
at altitudes of about 50 feet.

The numerous spinose warts on the interorbital space of the former
may, and probably are, only conspicuous during the breeding season,
which had just begun; nevertheless, breeding toads from elsewhere
exhibit noticeably smoother crowns.

One might reasonably have supposed, as I did in 1932 when de-
scribing kisoloensis, that toads from the wet highlands (circa 6000-8000
feet) of the Kigezi District in the vicinity of the Kivu Volcanoes,
would differ from typical Egyptian specimens, but though the type
series from Kisolo (or Kisoro) exhibited a greater degree of webbing
on their hind feet, the series now obtained from Nyakabande and
Mushongero (about six and twenty miles from Kisolo respectively) do
not differ sufficiently to justify the retention of this race.

Color. In life. The following notes were made in the field. Kinan-
gop. Above, adults (3 c^c?, 3 9 9) bright green, their dorsal spots
exceptionally elongate, thus presenting a different appearance from
those seen at lower altitudes. The youngest differed from the adults
only in possessing unusually rufous parotids.

386 bulletin: museum of comparative zoology

Kibale Forest. Above, backs flecked with cream color, giving them
an unfamilar appearance.

Mushongero. Above, much cream color; a creamy white interorbi-
tal band and a pair of interorbital markings which are velvety black
finely edged with cream ; pairs of similar markings on the scapular and
lumbar regions; a fine, cream, vertebral line and a creamy patch in
centre of dorsum; a broad, cream, dorso-lateral band; from parotids
to groin a black lateral band; limbs barred with black. Below, dirty
white covered with a blackish network which laterally spreads up on
to flanks; soles grayish black.

Kisenyi. Here the cf d71 are bright yellow-ochre or yellowish green,
but one, with a more rounded snout, has the markings of a 9 •

Lindi. The larger of two very young toads exhibits red on groin and
on back of thighs, the younger on back of thighs only.

Size. Length of largest cTcf, 87 (Kisenyi), 75 (Mushongero), 71
(Ujiji), 67 (Mubuku), 65 (Kinangop), 59 (Kibale); length of largest
9 9 , all of which, with the exception of Mihunga, are gravid, 100
(Budongo), 100 (Bundibugyo), 95 (Kisenyi), 90 (Mushongero), 80
(Kinangop), 80 (Nyakabande), 73 (Mihunga), 65 (Kibale), 60 (Kitaya).
Measurements are in millimetres.

Breeding. As indicated above, females were gravid in most localities
on the dates shown. Our arrival at Kinangop synchronized with that
of the rains so that amphibians were assembling in every pool, the
three males captured were all attempting to clasp the largest female
with the result that they kept rolling over and over in the water to the
accompaniment of curious little noises. Apparently assembling at
Kisenyi, for males predominated in the proportion of 3 to 1 of those
brought in by natives. Calling at Kitaya.

Parasites. Acarine parasites on Kinangop toads. See Arthroleptis
minvius also.

Enemies. Two young recovered from the stomach of a green snake
(Chlorophis irregularis) on Idjwi, one from a white-lipped snake
(C rota pho pelt is h. hotamboeia) at Ujiji.

Bufo funereus Bocage

Bufo funereus Bocage, 1866, Jorn. Sci. Lisboa, 1, p. 77: Duque de Braganca,
Angola.

2 (M. C. Z. 25078-9) Mabira Forest, U. 9. xi. 38.
2 (M. C. Z. 25080-1) Budongo Forest, U. 30. xi. 38.
2 (M. C. Z. 25082-3) Kibale Forest, U. 9. xii. 38.

loveridge: African amphibians 387

Distribution. Though widespread, and frequently recorded from
Angola and the Congo, this species is new for Uganda, doubtless having
been confused at times with /•. regularis alongside which it occurs in
at least two of the foregoing localities, and which it so closely re-
sembles.

Variation. Differs from B. r. regularis in the complete absence of a
tarsal fold, and the usually more extensive webbing of the toes.

Size. Length of only d\ 57 mm., of 9 9 , 60-66 mm.

Breeding. Apparently about to spawn in all three localities.

Bufo lonnbergi nairobiensis Loveridge

Bufo lonnbergi nairobiensis Loveridge, 1932, Occ. Papers Boston Soc. Nat.
Hist., 8, p. 48: Nairobi, Kenya Colony.

16 (M. C. Z. 25084-9) S. Kinangop Plateau, K. C. 29. x. 38.

Variation. Entire series measured with following results. Third
finger included 7.2-8.7 times in the length from snout to anus; fourth
toe 4.6-5.5 times; tibia 2.5-3 times. Thus they are quite clearly refer-
able to this form rather than to the typical one from Alt. Kenya as
defined in my key on p. 50 of the above citation.

Size. Lengths of 13 &&, 28-38 mm., of 3 9 9, 33-39 mm.

Breeding. The males were calling from large pools formed in cart
tracks crossing the plateau at 10,000 feet. In such pools as had grassy
bottoms these little toads might be seen walking about on the grass
as if on land.

Bufo micranotis rondoensis subspec. now
11 (M. C. Z. 25090-100) Nchingidi, T. T. 11-19. v. 39.

Type. Museum of Comparative Zoology, No. 25090, a non-breeding
adult from Nchingidi, 3000 feet, Rondo Plateau, Lindi Province,
southeastern Tanganyika Territory. Collected by Arthur Loveridge,
May 11-19, 1939.

Paratypes. Museum of Comparative Zoology, Nos. 35091-35100,
being 10 specimens with same data as type.

Diagnosis. Differs from micranotis Loveridge (1925, Proc. Zool.
Soc. London, p. 770, pi. i, fig. 1) of Kilosa and the Uluguru Mountains,
only in the throat being almost entirely white in the entire series,
whereas in both sexes of micranotis (taken in amplexus) the throat is
so heavily overlaid with black as to appear black.

388 bulletin: museum of comparative zoology

Color. In life. Above, brown marbled with black; from nape to anus
a hair-like, light, vertebral line (absent in some paratypes); warts
slightly rufous encircled with black; sides of head white, streaked and
spotted with black. Below, throat white very sparsely spotted or
vermiculated with brown; chest, belly, and limbs faintly bluish white
heavily overlaid with black.

The color often harmonized remarkably with the sandy soil on which
this tiny species is found.

Size. Length of type from snout to anus, 19 mm., of paratypes 19
(2 ex.), 18 (2 ex.), 17 (2 ex.), 14 (2 ex.), 12 (2 ex.) mm.

Remarks. In new of the fact that none of the above attain to the
size of typical micranotis (cfcf 20-22 mm., 9 23 mm.), the possibility
of the white throats being a subadult character should not be over-
looked. Xchingidi is about 300 miles southeast of Kilosa, type lo-
cality of micranotis.

Diet. Stomachs held: chrysomelid larva; phalacrid beetle; fly
(Cecidomyiidae) ; springtail (CoUembola); a true bug; five termite
workers; two land snails whose shells had been dissolved away.

Habitat. Following a night and morning of very heavy rain, the
first of these toads was seen at 11 a.m. as it sought shelter under the
awning of my tent which was pitched on sandy ground in a clearing
on the outskirts of the forest. The rest were taken beneath logs lying
at the forest-edge.

Xectophryxoides torxieri (Roux)

Nectophryne tornkri Roux, 1906, Proc. Zool. Soc. London, 1, p. 63, pi. ii, fig. 4:
Ukami, Tanganyika Territory.

68 (M. C. Z. 25101-5) Magrotto Mtn., T. T. 1-6. vii. 39.

Color. In life. Adults. Above, ranging from pale gray through
shades of olive and dusky brown to reddish brown with, or without,
mottlings, scarcely two alike; bright yellow spots — one in front of each
eye and three on either flank — present only in one female, which was
gravid with tadpoles ; a light vertebral line present on seven specimens
only. Young. For the most part, very young toads presented a rather
Arthroleptis-\\ke appearance, being pale gray with dark sepia marblings.

Size. Lengths from 10-34 mm., the largest a gravid female.

Breeding. Most females from 26-34 mm. long, were gravid with
eggs, a few with tadpoles.

Habitat. Apparently rare in the deforested area now occupied by
the oil-palm plantation, though I took one on an oil palm at a height

loveridge: African amphibians 389

of four feet and another on the floor of an adjacent patch of forest.
Hearing that the sole surviving stand of wild bananas on the mountain
were to be found on some rocky heights at Kitulwe, we made two
journeys to the spot and there obtained 56 frogs of which about half
were young.

RHACOPHORIDAE

Chiromantis ? rufescens (Giinther)

Polypcdates rufescens Giinther, 1868, Proc. Zool. Soc. London, p. 486: West
Africa.

2 Nests (M. C. Z. 25106) Budongo Forest, U. 5. xii. 38.

Distribution. The only record of the occurrence of a Chiromantis
in Uganda known to me, is the inclusion of C. xerampelina in Boulen-
ger's (1902a) list in Johnston's "Uganda Protectorate." As however,
C. xerampelina in East Africa is chiefly a species of the coastal belt,
though penetrating into the interior further south, it seems more
probable that the species occurring in northwest Uganda is rufescens;
unfortunately no frogs were collected.

Breeding. The eggs when fresh, were pale greenish, but turned
bright orange on being dried, possible through exposure to formalin
fumes. According to the native collector, the shrub, to whose leaves
the nests are attached, was overhanging a stagnant stream at the
forest edge.

Chiromantis xerampelina Peters

Chiromantis xerampelina Peters, 1854, Monatsb. Akad. Wiss. Berlin, p. 627:
Tete and Sena, Mozambique. •

Nest & 8 (M. C. Z. 25107-9) Kitaya, T. T. 25-31. iii. 39.
2 (M. C. Z. 25110-1) Mikindani, T. T. 12. iv. 39.

Native names. Chitowa (Kiyao); kitowa (Kimakonde), but applied
to Megalixalus, Hyperolius, and Rana edulis as well.

Color. In life. 9 . Hind legs and groin of breeding frog tinged with
yellow.

Size. Lengths of tftf 48-50 mm., of 9 9 67-72 (M. C Z. 25110).

Breeding. On March 25, nests, which appeared to be about two
weeks old, were found on a fence in a rice swamp. On the 30th a
freshly deposited nest was found in grass, no higher objects being in

390 bulletin: museum of comparative zoology

the vicinity, six inches above the surface of a recently-formed pool, a
female frog was found resting on the bank within a yard of the nest.
Habitat. Two other females and three males were taken one night
as they rested on grass stems projecting from a pool and palm fronds
overhanging it. The Mikindani females were taken at a height of nine
feet while resting in a bush and tree respectively.

Leptopelis concolor Ahl
Plate 2, figs. 5-6.

Leptopelis concolor Ahl, 1929, Sitz. Ges. naturf. Freunde Berlin, p. 192: Witu,
Kenya Colony.

2 juv. (M. C. Z. 25112, 25300) Mikindani, T. T. 12. iv. 39.
1 juv. (M. C. Z. 25113) Siga Caves, T. T. 14. vi. 39.
1 hgr. (M. C. Z. 25114) Tanga town, T. T. 22. vii. 39.

Distribution. The above constitute the first records of the occur-
rence of this species in Tanganyika Territory.

Variation. The subadult frog has been compared with a series of
topotypes from Witu. Identification of the juveniles, however, should
be accepted with reserve as the characters are not too well defined in
such very young frogs. The one from Siga Caves, near Tanga, appears
to have a slightly longer hind limb, a condition possibly resulting from
its subjection to the action of gastric juices and consequently slightly
macerated state.

Color. This Siga froglet should possibly be referred to johnstoni for
its back is green spotted with yellow as is the case with the young of
that species, its digital characters, however, seem to be those of con-
color. The Mikindani froglets, which are 2-3 mm. smaller, have uni-
formly green backs. The Tanga frog had assumed the brown livery
of the adult as depicted in the accompanying plate.

Size. Lengths (M. C. Z. 25300) 16+10 mm., (25112) 15 mm.,
(25113) 18 mm., (25114) 30 mm.

Breeding. On April 12, at Mikindani, one froglet with the stump of
a tail, another of same length but with tail already absorbed.

Enemies. Three juveniles, two partly digested and so discarded,
recovered from the stomach of a spotted wood snake (Philothamnus s.
semivariegatus) near Siga Caves.

Habitat. At Mikindani taken in rice swamp and on sandy soil in
camp; at Tanga beneath rubbish on the outskirts of the township.

loveridge: African amphibians 391

Leptopelis johnstoni (Boulenger)
Plate 2, figs. 3-4.

Hylambates johnstoni Boulenger, 1897, Proc. Zool. Soc. London, p. 803, pi.
xlvi, fig. 4: Kondowe; Karonga; and Nyika Plateau, Nyasaland.

13 (M. C. Z. 25125-35) Magrotto Mtn., T. T. 7. vii. 39.

Native navie. Xkewc (Kisambara).

Variation. The series consists of a typically colored brown 9 , adult
of 68 mm., and twelve frogs ranging from 18-48 mm. which I assume
to be the young as they agree in all physical characters such as outer
finger with one and a half joints free of web ; outer toe with a narrow
margin of web on the last joint; fourth toe with the last two joints
free of web.

In coloration, however, they are amazingly different from the adult.
In the field I thought they represented some distinct species but dis-
section of the 35 and 48 mm. frogs leads me to think that all are im-
mature. See also remarks (Loveridge, 1936k, p. 387) on johnstoni from
Ngatana.

Color. In life. 48 mm. Above, rich light green as if enameled,
flecked with yellow; a conspicuous cream spot on each elbow, knee,
and heel.

In life. 35 mm. As last, but lips, as well as spots white. In alcohol
there is now no difference in the appearance of the lips of these two
frogs, and the entire series, with two exceptions, exhibit a white line
on the outer aspect of forearm and finger, and another on the outer
aspect of foot and toe. No mention of such a mark was made in the
field, however, yet they are too conspicuous to have escaped notice.

In life. Tailed froglet 18 + 25 mm. Above, as larger with back
flecked with yellow, and white spots on elbow, knee, and heel, but no
white on lips.

Size. Lengths 18-68 mm.

Breeding. On July 7, the largest female was distended with numer-
ous white ova.

Diet. Orthoptera including some quite large grasshoppers.

Parasites. Large nematodes (Amplicaecum involution) in intest-
ines.

Habitat. I found the 48 mm. frog seated on my arm after I had
brushed past some ferns projecting from the stem of an oil palm grow-
ing at the edge of the swamp — in which the tadpole was taken — im-
mediately below my camp. The 35 mm. frog was in one of the fifty

392 bulletin: museum of comparative zoology

wild bananas searched at Kitulwe. Most of the other young were
taken one morning after heavy rain as they were squatting on broad-
leaved plants in a little dell or clearing in the forest.

Leptopelis notatus christyi (Boulenger)
Plate 2, figs. 5-6. Plate 4, fig. 1.

Hylambates christyi Boulenger, 1912, Ann. Mag. Nat. Hist. (7), 12, p. 556:

Mabira Forest, Kyagwe, Uganda.
? Leptopelis budduensis Ahl, 1929, Sitz. Ges. naturf. Freunde Berlin, p. 199:

Northwest Buddu Forest, Tanganyika Territory.

9 (M. C. Z. 25114) Mabira Forest, U. 7. xi. 38.
6 yng. (M. C. Z. 25115-9) Kibale Forest, U. 12. xii. 38.
2 9 9,2 yng. (M. C. Z. 25120-3) Mihunga Swamp, U. 18. i. 39.

Native name. Lulengaya miti (Luganda + Kiswahili = frog of the
tree) .

Variation. The series consists of three gravid females (43-50 mm.)?
and 8 young (18-26 mm.) which bear little resemblance to the adults!
These young, however, agree closely with the brief description of the
23 mm. Cameroon frog called notatus by Buchholz and Peters, 1875,
and later figured by Xieden. Subsequently two 33 mm. frogs from
Zima, Cameroon, formed the basis of the description of Hylambates
cubitoalbus Boulenger, 1906, which Parker considers synonymous with
notatus, but the two additional frogs from Bunyoro, Uganda, men-
tioned by Boulenger, Parker refers to christyi, which may, or may not,
be subspecifically distinct. Trinomials are employed to show this close
relationship and avoid further confusion.

The 50 mm. topotypic 9 agrees with the Ruwenzori 9 9 in differ-
ing from Boulenger's description of the 53 mm. holotype 9 in the
vomerine teeth being directly between (not behind the level of the
hinder edge) the choanae; the snout is longer than (not as long as) the
orbital diameter; the inner metatarsal tubercle is relatively small (not
large) in the Mabira frog, but is large in the two Mihunga 9 9 •

The young differ from the adults, not only in coloration as described
below, in their less extensive webbing which might fairly be described
as "a rudiment" for the fingers, while the fourth toe has two joints
free of web, and the fifth only one joint free. In the adults, on the other
hand, the outer finger has only 1 joint free of web; the fourth toe also
has only one joint free though there is a narrow, or moderately broad,

LOVERIDGE: AFRICAN AMPHIBIANS 393

margin of web on the penultimate joint ; the fifth toe is fully, though
somewhat narrowly, webbed.

Color. In life. Kibale Forest. Young 18-20 mm. Above, rich, light
green as if enameled; a conspicuous white spot on each elbow, knee, and
heel. Below, pure white.

In life. Mihunga Swamp. Young 23-26 mm. Both bearing perma-
nent spots like the foregoing, but those of the smaller are yellow.

In life. Mihunga Swamp. Gravid 9 46 mm., taken within 100 feet
of the last, had, on its elbows only, fugitive light spots that disappeared
on preservation. Above, bright green grass, on end of snout a cream-
colored vertical spot, two more on lip; a dark brown canthal streak
from nostril through eye to a short distance behind tympanum; an
interorbital bar and vertebral markings of genus in darker green more
or less edged with brown; a dark brown circum-anal spot edged with
cream and flanked on thighs by creamy flecks ; flanks with brown and
creamy vermiculations; fore and hind limbs with very indistinct trans-
verse dark bars. Below, white, except for brown infuscations on throat
and outer side of hind feet.

In life. Mihunga Swamp. Gravid 9 4S mm. Above, pale brown;
a broad, black, canthal streak from end of snout through eye to above
tympanum; dorsum devoid of markings; a dark circum-anal spot;
flanks and groin irregularly flecked with brown, the lower flanks with
a number of fine yellow spots; limbs and feet indistinctly barred with
broad brown bands. Below, creamy white, except for brown infusca-
tions on throat, breast and from elbow to wrist of fore limb, hind limb
and underside of feet, while the belly is tinged with yellow from the
spawn within.

In life. Mabira Forest. Gravid 9 50 mm. Above, pale brown with
darker flecks indicating vertebral markings of genus; flanks spotted
with brown; limbs and feet barred with broad brown bands; axilla and
groin slightly bluish; Below, throat a faintly greenish, dirty white;
breast and belly anteriorly sparsely spotted with brown, and belly
stained with rust color; limbs obscurely yellow-buff, unspotted; soles
slightly darker. Inside of mouth green, dilated pupil milky blue sur-
rounded by golden-brown iris.

The differences in coloration are reflected by habitats as recorded
below.

Size. See above.

Breeding. On November 7 and January 18 females were gravid,
while on December 12 very young frogs were found, slightly larger ones
on Januarv 18.

394 bulletin: museum of comparative zoology

Diet. Two stomachs held small grasshoppers in addition to inde-
terminate insect remains.

Habitat. The Kibale series were found on a shrub by a stream, a
couple on a felled tree, others on a forest trail after a heavy downpour
which had apparently dislodged them from the forest canopy. The
three green Ruwenzori frogs were taken on rich green sedges within
twenty feet of a stream, the brown one was clambering over felled
sedges that had been cut earlier in the day. The brown Mabira female
was beneath loose bark on the trunk of a tree at a height of only two
feet from the ground, a second brown frog was seen calling at midday,
as is their custom, from a yellow-brown shrub growing among straw-
like dry grass in the garden.

Leptopelis vermiculatus (Boulenger)

Hylambates vermiculatus Boulenger, 1909, Ann. Mag. Nat. Hist. (8), 4, p. 497:
Amani, Usambara Mountains, Tanganyika Territory.

juv. (M. C. Z. 25136) Magrotto Mtn., T. T. 8. vii. 39.

Variation. This emergent frog possesses all the key characters
except that the webbing on the hind feet is a little less full, it has been
compared with a slightly larger topotype.

Color. In life. Tailed froglet. Above, pale, slightly yellowish, green
finely vermiculated with black; a conspicuous white spot on each elbow,
knee, and heel; limbs white, heavily crossbarred with black; elbow
region of forearm and tibia of hind limb green and vermiculate like
back; digital disks of fingers white, of toes dusky grayish. Below,
whitish with dusky speckling on throat and sides of belly.

Size. Length 18 -f- 10 mm., the stump of tail being in process of
absorption.

Habitat. Taken in swamp immediately below my camp.

Hylambates maculatus Dumeril

Hylambates maculatus A. Dumeril, 1853, Ann. Sci. Nat. (3), 19, p. 165, pi. vii,
figs. 1— lb and 4: Zanzibar.

1 c? 3 9 9 (M. C. Z. 25137-9) Kitaya, T. T. 28. iii. 39.
13 c? & 1 9 (M. C. Z. 25140-9) Amboni, T. T. 17. vi. 39.

Native name. Nanhengo (Kimakonde, for Kassina also).
Size. Length of cTcf 63-68 mm., of 9 9 58-71 mm.
Breeding. Great numbers were calling in a pond close to Amboni
Village, near Tanga, though the rains were nearly over in this dis-

loveridge: African amphibians 395

trict. The cry is quek-quek, with an explosive pop in it, hence more
resonant than that of Rana m. mascareniensis.

Habitat. At Kitaya beneath damp weeds recently uprooted and
piled at the edge of a rice swamp, one on a grass stem a foot above the
water. At Amboni the majority were immersed in the pond with only
their heads showing.

Kassina sexegalexsis (Dumeril & Bibron)
Plate 4, fig. 2.

Cystiganthus Senegalensis Dumeril & Bibron, 1841, Erpet. Gen., 8, p. 418:
Galam, Senegal.

9 (M. C. Z. 25150) Idjwi Id., B. C. 23. ii. 39.
juv. & cf (M. C. Z. 25151-2) Kitaya, T. T. 25. iii. 39.

Native name. Kabunda (Lulega) ; nanhengo (Kimakonde, for Hylam-
bates also).

Remarks. The generic definition of Kassina appears badly in need
of revision for not only do young Kassina often lack vomerine teeth
but on occasion the adults also. It was the absence of such teeth that,
in part, led Hoffman (1939) to describe Semnodactylus thabanchuensis
from the Transvaal, and Laurent (1940) the genus Kassinula wittei
from the extreme southern Belgian Congo, the latter based on juveniles
of 12.5-14 mm. in length. That they are not the young of senegalensis
seems clear from the fact that our tailed froglet listed above measures
23 mm. from snout to anus and all the smallest of our completely
metamorphosed young are 20 mm. in length. Both thabanchuensis and
wittei bear a striking resemblance to tvealii with which they should be
compared.

Size. Length of c? 45 mm., of 9 44 mm., of tailed froglet 23 + 10
mm.

Breeding. On February 23 the 9 was gravid, and in that condition
led to the description of an gel i Witte.

Megalixalus fornasinii forxasixii (Bianconi)

Euchnemis Fornasinii Bianconi, 1848 (not 1850), Spec. Zool. Mosamb., Rept.,
pi. v, fig. 1 : Mozambique.

3 <? 2 9 (M. C. Z. 25153-4) Kitaya, T. T. i. iv. 39.

3 9 (M. C. Z. 25155-6) Siga Caves, T. T. 9. vi. 39.

4 cf 4 9 (M. C. Z. 25157-8) Amboni Est., T. T. 17. vi. 39.

1 9 (M. C. Z. 25159) Magrotto Mtn., T. T. 1. vii. 39.

396 bulletin: museum of comparative zoology

Native names. Chitowa (Kiyas); kitoica (Kimakonde), but applied
to Chiromantis and Hyperolius also.

Remarks. Trinomials are employed on account of the recently de-
scribed M. f. gongicus Laurent (1941) which, to judge by our four
Congo specimens may be differentiated in pattern from M. f. dorsalis
(Peters) of Cameroon. The differences are slight, however, and occa-
sionally complicated by the occurrence of uniformly colored examples
in both races. The typical form has physical distinctions.

Color. Uniform or with a vertebral streak, normally dark, but in
M. C. Z. 25155 the normal arrangement is reversed and the streak is
light, flanked by darker.

Size. Length of cfc? 31-37 mm., of 9 9 31-38 mm.

Megalixalus leptosomus (Peters)

Hyperolius leptosomus Peters, 1877, Monatsb. Akad. Wiss. Berlin, p. 619, pi.
— , fig. 5: Chinchoxo, Portuguese Congo.

Remarks. In a recent paper on the Megalixalus of the Congo Beige,
Laurent (1941c, p. 127) suggests that the Kabengere from which I
(1936h, p. 103) recorded leptosomus, is not on the Luapula River, but
lies farther north in the neighbourhood of Xyonga. In this he is un-
doubtedly correct for Mr. J. T. Zimmer, who collected the frogs in
1926, writes:

"my locality Kabengere is not on the Luapula River but is in
another part of the 'Haut Luapula' District, at least it was so
called when I was there although I believe they call it now Haut
Katanga. In this District my collections were all made at Buk-
ama, Katobwe and Kabengere. At Kabengere we were on the
edge of true forest and found some things that showed this
ecological association, but in the main the area was exactly the
same as that at Katobwe. We were only two days' walk southeast
from Katobwe and it was two days' walk southwest from Kaben-
gere to Bukama, so the exact locality is about 9° 15' x 26° 15'. If
you look up Reichenow's 'Yog. Afr. Atlas,' page 22, you will find
a Katapana which is our Kabengere. For some reason which I
have forgotten Katapena is an alternative name for Kabengere;
at least I have the alternative name in my journal without com-
ment."
I quote Mr. Zimmer at length as he collected a great many species of
frogs at both Kabengere and Katobwe, which are not on the Luapula
River.

loyeridge: African amphibians 397

Dr. Laurent then breaks up leptosomus into three forms and sug-
gests that the Kabengere frogs should probably be referred to one of
his new forms, viz. M. 1. upembae (p. 125). Of the original series of
19 frogs which I listed, 9 are now in the Museum of Comparative
Zoology. Using the key (p. 132) furnished by Dr. Laurent, I should
say that two of these might be referred with equal justification to
either I. leptosomus or /. upembae as they combine the key characters;
one (M. C. Z. 19176) has broad, the other (M. C. Z. 19177) narrow,
dorsal lines; in the former the tibio-tarsal articulation reaches the
shoulder, in the latter it fails to do so; in both the length of the tibia
is about 2}4: times in the length from snout to anus, in which respect
they do not appear to differ from our Cameroon and French Congo
material, so I fail to see how upembae can be recognized.

In the remaining 7 specimens (M. C. Z. 19178-84) the dorsal bands
unite in a point in the interocular region and are therefore referable to
M. wittei Laurent (p. 127), an apparently valid species differing only
from leptosomus in this marking and in averaging a trifle larger. In
answer to my enquiries, Mr. K. P. Schmidt has kindly informed me
that the rest of the series in the Field Museum is divided equally as
between leptosomus and wittei.

Megalixalus tjluguruexsis Barbour & Loveridge

Megalixalus uluguruensis Barbour & Loveridge, 1928, Mem. Mus. Comp.
Zool., 50, p. 231, pi. iii, fig. 2: Vituri, Uluguru Mountains, Tanganyika
Territory.

9 (M. C. Z. 25160) Magrotto Mtn., T. T. 12. vii. 39.

Color. In life. 9 . Above, enamel white; from nostril through eye
to halfway along flank a brownish band; back with a few irregularly
scattered brown flecks; upper arms and thighs translucent; forearm,
tibia, foot, and to a lesser extent the thigh also, with minute, black-
centered, enamel-white spots (under high magnification these black
centers are seen to be irregular, 5-pointed, black stars) ; hands and feet
translucent yellow; disks of digits and toes almost orange. Below,
chest pure white, remaining under surfaces transparent.

Size. Length 28 mm.

Breeding. On July 12, three large cream-colored eggs were (and
still are) visible through the transparent abdominal skin.

Habitat. I captured this frog among leaves between the mighty
buttress roots of a tree in a patch of forest remote from my camp. It

398 bulletin: museum of comparative zoology

was taken at 10 a.m. immediately following several hours of very
heavy rain and constitutes the first record of this very Hyperolius-like
species on Magrotto Mountain. Recognizing it, I immediately checked
its pupil which was plainly vertical.

Megalixalus brachycnemis Boulenger

Megalixalus brachycnemis Boulenger, 1896, Ann. Mag. Nat. Hist. (6), 18,
p. 403, pi. xvii, fig. 2: Chiradzulu, Nyasaland.

60 d" 17 9 (M. C. Z. 25161-2) Kitaya, T. T. 25-31. iii. 39.

4 cf (M. C. Z. 25163-4) Mikindani, T. T. 8. iv. 39.

Id1 1 9 (M. C. Z. 25165-6) Siga Caves, T. T. 16. vi. 39.

36 & 2 9 (M. C. Z. 25167-8) Amboni Est., T. T. 17. vi. 39.

2d1 3 9 (M. C. Z. 25169-70) Magrotto Mtn., T. T. 1. vii. 39.

Variation. Of this species, which has on four occasions been de-
cribed as a Hypcrolius by Ahl, I noted at Kitaya: "Pupil definitely
vertical," and on Magrotto: "After anaesthetization the pupil was
indistinctly vertical in a young frog and almost round in an adult, the
latter retaining only a slight v at base."

Color. In life. Kitaya. cf • Above, dusky yellow minutely punc-
tate with black; from nostril through eye to groin extends a broad,
dusky, irregular, lateral band which is punctate with white; back with
a dusky gray M-shaped marking. Below, sac chrome ; breast and belly
pure white; throat between sac and breast like all remaining under-
parts a semitransparent flesh color with the digits slightly yellow.

Thus below it is very similar to Hyperolius c. citrinus though above
quite dissimilar.

In life. Kitaya. 9 • Above, substantially like the male but with
the dorsal markings more frequently absent. Below, throat, breast,
and belly uniformly pure white.

In life. Magrotto. Adult and young of 17 mm. Above, both were
golden bronze with mottlings of greenish bronze on which were super-
imposed white specks or dots.

Size. Length of c?c? 18-23 mm., of 9 9 21-25 mm. A slight, but
definite increase in size is noted as one proceeds southwards, viz. cf d1
20-21-19-22-23 mm., 9 9 22-21-22-25 mm. About March 31, the
largest of six tailed froglets measured 11-4-14 mm.

Breeding. Our arrival at Kitaya almost synchronized with the break-
ing of the big rains. On March 25 males were calling from a rice swamp
at the edge of the Rovuma River and one clasping pair were taken on

loveridge: African amphibians 399

a horizontal blade of rice. On each of four succeeding days the number
of males taken decreased, the number of females increased. On June
17, though the rains were almost over, a few males might still be heard
calling at Amboni in a swamp near the village. Females, of course,
were gravid when taken.

Enemies. One recovered from the stomach of a spotted wood snake
(Philothamnus s. semivariegatus).

Habitat. The great majority, having assembled, were taken in
recently-formed rice and sedge swamps, but a couple were taken be-
neath rubbish a good way from water on March 29. A search of fifty
wild bananas at Kitulwe, on Magrotto, resulted in the capture of two
whose unusual coloring, noted above, may be attributed to habitat.

Hyperolius schubotzi Ahl

Hyperolius schubotzi Ahl, 1931, Das Tierreich, no. 55, p. 329, fig. 202: Kissenji,

i.e. Kisenyi, Lake Kivu, Belgian Ruanda-Urundi. ( 9 . 33 mm. Schubotz

coll.).
Hyperolius macrodactylus Ahl, 1931, Das Tierreich, no. 55, p. 368: Lake Kivu,

Belgian Congo. ( 9 . 37 mm. Kandt coll.).
Hyperolius ornatus Laurent, 1940, Revue Zool. Bot. Africaine, 34, p. 4, pi. ix,

figs. A. C. D.: Ruchuru, Belgian Congo. (Type 32 mm.).

11 c? 28 9 (M. C. Z. 25174-89) Mamvu Bay, B. C. 15. ii & 6. iii. 39.

Native names. Kashembero (Lulega); neosa was also used on Idjwi.

Synonymy. I personally captured the entire series listed above in
a small patch of sedges growing in Mamvu Bay, Idjwi Island (where
Kandt lived), Lake Kivu, and concluded that the two types — the
finely vermiculate 9 schubotzi and largely marmorate 9 ornatus were
one species for they appear to grade from one to the other. In size
there was no difference for the schubotzi ranged from 37-41 mm. and
the ornatus type from 36-41 mm. By its dimensions and pattern I
judge that macrodactylus is yet another synonym. One of Ahl's para-
types of variabilis, that from Idjwi Island, should doubtless be included
though the Bukoba cotypes of variabilis, sent me by Ahl, appear to
represent a smaller species.

Variation. Outer finger with 1 joint free, fourth toe with 1 joint
free, outer toe webbed to disk; tibio-tarsal articulation reaches eye
rather fully in all but one frog where it just falls short.

Color. In life. 9 . Above, black with well separated wavy yellow
lines and a few spots ; around anus silvery with black vermiculations ;
thighs whitish with blood red vermiculations; tibia and foot like back.

400 bulletin: museum of comparative zoology

Below, throat silvery blotched with black; chest and belly immaculate
white; lower side of thighs with red vermiculations. (This is the
or nat us phase).

9 . Similar to the last except that it is the ground color which is
yellowish green with very numerous black vermiculations. (This is
the schubotzi phase).

Twelve females (M. C. Z. 25174-7), 37-41 mm., are light, finely
vermiculate with darker as figured for schubotzi, which species they
undoubtedly represent. Fourteen females (M. C. Z. 25178-82), 36-41
mm., are so heavily overlaid as to appear black with lighter, and very
varied, vermiculations (ornatus) some even forming longitudinal lines
(macrodactylus). Two apparently young females (M. C. Z. 25183-4),
34 mm., resemble the males.

The eleven males (M. C. Z. 25185-9), 35-39 mm., now an almost
uniform putty color, were, as I recollect, yellowish in life; one or two
show faint traces of a dark can thai and/or postocular band.

Size. Length of- cfcf 35-39 mm., average 36.3 mm.; of 9 9 34-41
mm., average 34.6 mm.

Breeding. On February 15, four males were taken when calling.

Enemies. Three in the stomach of a green snake (Chlorophis irregu-
laris) up the mountain.

Habitat. On reeds and sedges in knee-deep water in Mamvu Bay.

Hyperolius kivuensis Ahl

Hyperolius kivuensis Ahl, 1931, Das Tierreich, no. 55, p. 280, fig. 151: Lake

Kivu, Belgian Congo. ( 9 . 34 mm. Kandt coll.).
Hyperolius kwidjwiensis Ahl, 1931, Das Tierreich, no. 55, p. 296, fig. 172:

Idjwi Island, Lake Kivu, Belgian Congo. (d\ 32 mm. Schubotz coll.).
Hyperolius kandti Ahl, 1931, Das Tierreich, no. 55, p. 327: Lake Kivu, Belgian

Congo. (d\ 33 mm. Kandt coll.).
Hyperolius koehli Ahl, 1931, Das Tierreich, no. 55, p. 405: Kisenyi, Lake Kivu,

Belgian Ruanda-Urundi. (o71. 33 mm. Koehl coll.).

lc^l9(M.C. Z. 25190-1) Mamvu Bay, Idjwi Id., B. C. 6. iii. 39.
2^19 (M. C. Z. 25192-4) Ujiji, L. Tanganyika, T. T. 11. iii. 39.

Synonymy. Idjwi Island, long the home of Kandt, is in Lake Kivu,
and I see no reason to regard Ahl's four 'species' as distinct. Laurent
has recently (1941a, p. 152, pi. ix, figs. B. E. F. G.) figured the striking
changes in pattern encountered in a developmental series of this spe-
cies. He remarks that kivuensis is an extremely common frog in Kivu,

loveridge: African amphibians 401

Urundi, Tanganyika, and the Katanga districts. The Ujiji record,
however, is actually the first for its occurrence on the Tanganyika
Territory side of the lake.

Variation. Outer finger with 1 joint free but webbing narrow on
second; fourth toe with 1 joint free, outer toe feebly webbed to disk;
tibio-tarsal articulation reaches eye.

Color. In life the Ujiji series were noted as being green, both series
are now discolored by formalin but the dark canthal streak and lateral
band may be more or less present in both sexes.

Size. Length of & c? 30 mm., of 9 9 30-33 mm.

Hyperolius sp.
cf (M. C. Z. 25171) Mushongero, U. 3. ii. 39.

Native name. Kitembi (Luganda).

Variation. Outer finger with 1 joint free except for a narrow seam,
fourth toe with 1 joint free except for a narrow seam, outer toe webbed
to disk; tibio-tarsal articulation reaches eye.

In the absence of females, this nondescript putty colored frog may,
or may not, be referable to the foregoing or following species.

Size. Length of cf 34 mm.

Habitat. I captured this frog on a papyrus stem six feet above the
surface of Lake Mutanda, the only one secured in two hours of paddling
a dugout and using a flashlight, yet, to judge by their calls, these frogs
were not uncommon.

Hyperolius alticola Ahl

Hyperolius alticola Ahl, 1931, Das Tierreich, no. 55, p. 379, fig. 255: Ruwenzori
Mountains, 1800 metres, Belgian Congo.

9 (M. C. Z. 25172) Mubuku Valley, Ruwenzori Mtns., U. 4. xii. 38.

Variation. Outer finger with 1 joint free except for a narrow seam,
fourth toe with 1 joint free except for a narrow seam, outer toe webbed
to disk; tibio-tarsal articulation reaches eye.

Color. In life. 9 • When caught, putty colored, but changed to :
Above, head, back, and limbs, a dead-leaf rufous brown sparsely
speckled with black; from nostril through eye a dark line which is con-
tinued on flank only as a series of dark spots; back anteriorly sparsely
flecked with yellow; exposed portions of fore and hind limbs similarly

402 bulletin: museum of comparative zoology

speckled. Below, throat and belly yellow, the latter on white; remain-
ing under surfaces flesh-colored; palms and soles pink.

Size. Length of 9 37 mm., as was 9 holotype.

Habitat. On leaf of a tall swamp plant at a height of seven feet.
The altitude was about 1000 feet higher than that at which Schubotz
took the type.

Hyperolius montanus (Angel)

Rappia montana Angel, 1924, Bull. Mus. Hist. Nat. Paris, 30, p. 269: Mount
Kinangop, Aberdare Mountains, Kenya Colony.

juv. (M. C. Z. 25173) S. Kinangop Plateau, K. C. 27. x. 38.

Variation. Outer finger with 1 joint free, fourth toe with 1 joint
free, outer toe webbed to disk; tibio-tarsal articulation barely reaches
eye. Compared with adult cotype and juvenile topotypes.

Size. Length 8 mm., tail recently absorbed.

Hyperolius undUlatus (Boulenger)

Rappia undulata Boulenger, 1901, Ann. Mus. Congo (1), 2, fasc. 1, p. 4, pi. ii,
fig. 2: Pweto & Lofoi, Lake Mweru, Belgian Congo.

d1 (M. C. Z. 25299) Kitaya, T. T. iv. 39.

Native names. Chitowa (Kiyao); kitoiva (Kimakonde), but applied
to all species and Chiromantis and Megalixalus also.

Variation. Outer finger with 1 joint free, fourth toe with 1 joint
free, outer toe webbed to disk; tibio-tarsal articulation reaches eye.
Compared with adult cotype.

Size. Length 25 mm.

Hyperolius udjijiensis Ahl

Hyperolius udjijiensis Ahl, 1931, Das Tierreich, no. 55, p. 370, fig. 246: Ujiji,
Tanganyika Territory (restricted).

1 c? 7 9 (M. C. Z. 25195-200) Ujiji, T. T. 11. iii. 39.

Remarks. Some years ago I (1933h, pp. 401-2) placed the cotype of
udjijiensis from Kibwezi, Kenya Colony, in the synonymy of striolatus
Peters, but pointed out that the Ujiji specimen, to which the name is
now restricted, might prove to be distinct. The topotypic series now

loveridge: African amphibians 403

obtained proves clearly that this is the case, in fact the relationship
to the callichromus form of argentovittis is so close that in the field I was
inclined to doubt their distinctness, thinking that udjijiensis might be
a subadult phase, at least one female, however, is gravid.

If not an actual synonym of H. vermiculatus Peters, 1882 (not vcr-
miculatus Pfeffer, 1893), of Mlange, Angola, the relationship would
appear to be subspecific, to judge by comparing a 36 mm. gravid 9
from Bella Vista, Angola, with a 34 mm. gravid 9 from the Ujiji
series.

Variation. Outer finger with 1 joint free, fourth toe with last joint
narrowly webbed, outer toe fully webbed to disk; tibio-tarsal articula-
tion reaches eye, or to between eye and nostril.

Size. Length of c? 32 mm., of 9 9 30-34 mm.

Hyperolius flavomaculatus Gunther

Hyperolius flavomaculatus Gunther, 1864, Proc. Zool. Soe. London, p. 310, pi.
xxvii, fig. 1 : Rovuma Bay, Tanganyika Territory.

■3oM 9 (M. C. Z. 25201-4) Kitaya, T. T. 25-31. ii. 39.

Remarks. The above series are practically topotypes for my camp
on the banks of the Rovuma River was scarcely twenty miles from the
Bay where Kirk collected the type. Boulenger (1882 b, p. 128) trans-
ferred the species to MegaUxalus and Parker wrote me that, after
examining the type which until now has remained the only known
specimen, he saw no reason to reverse Boulenger's action. The pig-
mentation, however, was so wholly like that of a Hyperolius that I
felt sure the pupil must be horizontal, and visited Kitaya with the ob-
ject of securing fresh material. The series obtained have horizontal
pupils, as I noted in the field, and so I restore the species to its original
allocation.

The Mt. Kenya and Mt. Mbololo frogs which I identified as H.
flavoguttatus AM, have some webbing on the terminal joint of the
fourth toe and represent a slightly larger (34 mm.) species.

Variation. Outer finger with 1 joint free, fourth toe with 1 joint
free, outer toe webbed to disk; tibio-tarsal articulation reaches the eye.

Color. In life. c? . Above, dusky yellow, heavily blotched or vermic-
ulated with black; thighs and concealed portions of feet dull reddish;
tibia dusky yellow with, or without, black vermiculations. Below,
throat cream-colored (not yellow) ; interspace between disk and breast
colorless like the lower flanks; median portion of breast and belly
creamy white; limbs dull opaque red.

404 bulletin: museum of comparative zoology

Size. Length of c7 cT 28-29 mm., of 9 29 mm.

Habitat. Apparently a rare species as compared with H. c. citrinus
with which it occurred in the same rice swamp, as also in another pond
(B) where a pair were taken on grasses (Setaria palmifolia) much be-
loved by all the species of Hyperolius taken at Kitaya. I am indebted
to the Gray Herbarium for identifying the specimens of this sedge
which I brought back.

Hyperolius argentovittis Ahl

Hyperolius argentovittis A 1, 1931, Das Tierreich, no. 55, p. 345, fig. 220: Ujiji,
Tanganyika Territory.

5 <? 15 9 (M. C. Z. 25205-9) Ujiji, T. T. 11. iii. 39.

Variation. Outer finger with 1 joint free, fourth toe with 1 joint
free, outer toe webbed to, or almost to, disk; tibio-tarsal articulation
reaches eye, or to between eye and nostril.

Color. In life. Above, maroon, a black-edged, pure white, vertebral
line (argentovittis) in some, irregular or incomplete or broken up into
spots (calliehromus Ahl, 1931c, fig. 248) in others, from snout to anus.

Size. Length of cf cf 21-32 mm., of 9 9 30-36 mm., average 32
mm.

Breeding. On March 11a male was taken calling from a tree in our
camp beneath the mangoes and far from any water. The larger females
were gravid, but the majority were not.

Enemies. One in stomach of a white-lipped snake (Crotaphopeltis h.
hotamboeia) at Ujiji.

Hyperolius ahli Loveridge
Plate 3, figs. 1-2.

Hyperolius ahli Loveridge, 1936, Bull. Mus. Comp. Zool., 79, p. 402: Lake
Peccatoni, northeast of Witu, Kenya Colony.

2 cf 7 9 (M. C. Z. 25210-3) Kitaya, T. T. 1-4. iv. 39.
1 9 (M. C. Z. 25214) Lake Rutamba, T. T. 8. v. 39.

Variation. Outer finger with 1 joint free, fourth toe with last joint
narrowly webbed, outer toe fully webbed to disk; tibio-tarsal articu-
lation reaches to between eye and nostril, or nostril.

Color. The sexual dichromatism of this species was noted in the
original description. In life these males rather resembled the straw-

LOVERIDGE: AFRICAN AMPHIBIANS 405

colored c. citrinus, but the lateral streaks were clearer and the size
larger. In the females the large spots were pale red broadly edged with
black; limb surfaces which would be exposed when at rest were pale
dusky brown; thighs and digits brick red.

Size. Length of tf& 30-32 mm., of 9 9 30-33 mm.

Habitat. Taken at night on sedges in three different ponds, all
Kitaya females were taken in one pond on April 4.

Hyperolius puncttculatus (Pfeffer)

Rappia puncticulata Pfeffer, 1893 (1892), Jahrb. Hamburg. Wiss. Anst., 10,

p. 31, pi. ii, fig. 2: Zanzibar.
Hyperolius substriatus AM, 1931, Das Tierreich, no. 55, p. 358, fig. 234:

Magrotto Mountain, near Tanga, Tanganyika Territory.

1 J (M. C. Z. 25215) Amboni Estate, T. T. 17. vi. 39.
27 & 8 9 20 yng. (M. C. Z. 25216-29) Magrotto Mtn., T. T. 1-17. vii. 39.

Synonymy. The Museum of Comparative Zoology possesses a co-
type of substriatus AM in addition to the long series of topotypes now
secured; there seems to have been no reason for its having been de-
scribed.

Variation. Outer finger with 1 joint free, fourth toe with 1 joint
free or narrowly margined with web, outer toe webbed to disk; tibio-
tarsal articulation reaches eye (at all ages).

Color. In life, cf (M. C. Z. 25216). Above, orange yellow or yel-
lowish green; from nostril over eye to above axilla a broad, mustard-
colored band which is broadly edged with black above and below, pos-
teriorly it breaks up into one (on right) or two (on left) black-edged,
mustard-colored spots (but such spots are not only variable in number
but may be entirely lacking or the dorso-lateral band may be con-
tinuous); an azygous, black-centred spot in lumbar region; thighs,
except for a narrow band of orange yellow or yellowish green, like
other concealed parts of hind limb when at rest, blood red. Below,
gular disk and belly lemon yellow; remaining underparts greenish,
more or less transparent and tinged with red.

Another Magrotto specimen, though taken with typical examples in
sedges bordering a stream where it passed through a swamp, appeared
so strikingly different in the field that I concluded it might be distinct.
Its color was noted as follows: c? (M. C. Z. 25229). Above, pinkish
brown, from nostril over eye almost to groin a broad, enamel-white
band narrowly edged with light yellow and broadly with jet black

406 bulletin: museum of comparative zoology

above and below; on either side of anus and on either heel are white
spots similarly ringed with yellow and black; forearm, tibia, and feet
speckled with black; thighs with dusky patches formed by minute
speckling. Below, bright lemon yellow, gular disk sparsely necked with
brown; limbs, fingers, and toes show some transparent red.

d71 (M. C. Z. 25215). This 29 mm. variant is of the spotted type of
which an example (M. C. Z. 9531) from Morogoro, where it occurs
with more typically colored specimens, has been figured by Proctor
(1920, p. 415, fig. 2). This condition comes about by persistence as
spots (see also 26 mm. cf M. C. Z. 2521) of some of the faint lines
which are to be observed in the young.

Five young, taken on July 6 during a search of fifty wild bananas
at Kitulwe, were so variable that it was difficult to believe that all
were of one species. Below, their throats were whitish, bluish, or rich
emerald green; bellies white or bright yellow; hind limbs and feet
showing much red.

Nine young, taken with five adults in sedges and on leaves of trees
in a forest clearing, all possessed the U-shaped yellow canthal band,
but the young had dusky lines on the dorsum; a dark-edged, light,
vertebral line from snout to anus was present in some, and there was
no red on the limbs of these juveniles.

Size. Length of cf cf 21-29 mm., average 26.7 mm., of 9 9 30-33
mm., average 32 mm., of young 11-24 mm.

Breeding. All females gravid.

Habitat. In sedges, domestic and wild bananas, and on trees in a
forest clearing.

Hyperolius mariae Barbour & Loveridge

Hyperolius maricu Barbour & Loveridge, 1928, Mem. Mus. Comp. Zool., 60,
p. 217, pi. iii, fig. 1 : Derema, Usambara Mountains, Tanganyika Territory.

12 <? 1 9 (M. C. Z. 25230-4) Siga Caves, T. T. 16. vi. 39.
20 cT 7 9 (M. C. Z. 25235-9) Amboni Estate, T. T. 17. vi. 39.

Variation. Outer finger with 1 joint free, fourth toe with 1 joint
free, outer toe webbed to disk; tibio-tarsal articulation reaches eye or
just beyond.

Color. In life. Below, hinder part of gular disk bright lemon yellow;
belly usually white though sometimes red in either sex; limbs red.

The characteristic subdermal, black, lateral streak of mariae is
present in all, but the dark canthal streak or circum-nasal spot is

loveridge: African amphibians 407

very rarely present; this, in conjunction with the fact that these frogs
average about 2 mm. shorter than the montane (3000 feet) mariae
might cause some to regard melanophthalmus Ahl, 1931, as a recog-
nizable lowland form. I doubt if Siga Caves and Amboni are much
more than thirty-five miles from Derema as the crow flies.

Size. Length of c?(? 22-25 mm., average 23.5 mm., of 9 9 22-25
mm., average 23.1 mm.

Breeding. Calling in swamps near my camp at Siga and in one close
to Amboni village; the females, taken June 16-17, are distended with
eggs.

Hyperolius citrinus sansibaricus (Pfeffer)

Rappia sansibarica Pfeffer, 1893 (1892), Jahrb. Hamburg. Wiss. Anst., 10, p.
97, pi. ii, fig. 4: Zanzibar.

30 cf 5 9 (M. C. Z. 25251-6) Siga Caves, T. T. 14-16. vi. 39.
28 d" (M. C. Z. 25257-9) Amboni Estate, T. T. 17. vi. 39.

Remarks. When Pfeffer described sansibaricus he remarked on its
similarity to citrinus, and as I can distinguish them only by size it
seems advisable to regard the former (1893) as a smaller northern
form of the latter (1864).

Variation. Outer finger, etc. as in typical form.

Color. In life. a*. Below, throat bright lemon yellow. 9. Below,
throat whitish but tinged with yellow like the belly. In both sexes the
thighs, back of knees, and such parts of the feet as would be concealed
when at rest, blood red (rubripes Ahl).

Size. Length of c?c? 22-29 mm., average 25.4 mm., of 9 9 26-32
mm., average 29 mm.

Breeding. Though the rains were drawing to a close, the swamps
still resounded with the sharp "snap-snap" call of th£ males. All five
females were distended with ova.

Enemies. One was entangled in the strong mesh of a spider's web
which had been woven between the sedges, another was recovered
from the stomach of a larger frog (Ran a cdulis).

Hyperolius citrinus citrinus Giinther

Hyperolius citrinus Giinther, 1864, Proc. Zool. Soc. London, p. 311, pi. xxvii,
fig. 2: Senegal (rejected) and Zambezi Expedition (restricted). ?Possibly
from Rovuma Bay, Tanganyika Territory (A. L.)

120 <? 13 9 (M. C. Z. 25240-5) Kitaya, T. T. 25-31. iii. 39.
lie? 2 9 (M. C. Z. 25246-50) Mikindani, T. T. 8-20. iv. 39.

408 bulletin: museum of comparative zoology

Remarks. In his report on the material received from Livingstone
and Kirk's "Zambezi Expedition", Giinther figured (1864, pi. xxvii)
H. flavomaculatus and microps from Rovuma Bay and a third species,
citrinus, based on two males, from Senegal and Zambezi Expedition.
This does not necessarily imply that citrinus came from the Zambezi,
and we know that Livingstone tramped over the low hills from Mikin-
dani to some point on the Rovuma near to the site where the little
village of Kitaya stands today.

Of the three frogs — flavomaculatus, microps, and citrinus — which I
collected at Kitaya, the latter was by far the commonest and the males
bore a striking resemblance to the male figured by Giinther. Perhaps
the original label of citrinus was illegible or lost, hence "Zambezi
Expedition" only. However that may be, I propose to restrict the
type to the Zambezi Expedition frog and reject the Senegal specimen
as being unlikely to be conspecific. Unfortunately, owing to war condi-
tions, it would be impossible to get further light on this subject from
the British Museum.

Variation. Outer finger with almost 2 joints free, for second only
narrowly margined with web, fourth toe with 1 joint free and second
only narrowly margined with web on one side, outer toe with last joint
half webbed or continued as a narrow seam to disk; tibio-tarsal articu-
lation reaches eye or nostril.

Color. In life, cf . Above, green, dusky yellow, or straw brown,
paler on flanks, which are demarcated from the back by a rather in-
distinct (disappears on preservation) cream-colored, dorsolateral line
commencing high above axilla and terminating high above groin;
thighs largely colorless but tinged with red by blood vessels one of
which is clearly visible as a fine red line. Below, gular disk chrome;
breast and belly pure white ; interspace between disk and breast semi-
transparent like all remaining underparts.

To ascertain the proportion of the two dominant color phases in
regard to sex, catches made in three separate ponds on three different
days were sorted and noted down as follows:

19 straw cf cf, 19 green cf cf , 1 green 9 .
36 " cfcf, 15 " cfcf, 1 " 9.
16 " cfcf, 1 " cfcf, 1 " 9.

Size. Length of cfcf 28-34 mm., of 9 9 34-38 mm.

Thus citrin us is a larger frog but otherwise does not differ from san-
sibaricus; unfortunately no measurements were given for either type,
and I failed to note that of the latter when examining the type some

LOVERIDGE: AFRICAN amphibians

409

years ago. I did, however, find it resembled our specimens from Baga-
moyo, which is on the mainland exactly opposite Zanzibar. The posi-
tion is best set forth by listing the measurements of all the material
in the Museum of Comparative Zoology. All places are in the coastal
belt — Kilosa being the furthest inland — and are arranged below from
north to south.

Localities in Kenya and Tanganyika. Length cfcf. Length 9 9-

Kililani (Berlin d" and 9 cotypes of rubripes).

Mkonumbi

Peccatoni

Witu

Xgatana

Karawa

Malindi

Amboni

Siga Caves

Bagamoyo

Dar es Salaam

6 (?<?

6 cfc?

15c?c?

4

1
1

9 9

9
o

26-30 mm.
26-29 mm.
23-28 mm.

32 mm.

30-33 mm.

29 mm.

30 mm.
30 mm.

1 cf 31 mm.

28 <? c? 22-28 mm.

30 c? <? 5 9 25-29 mm. 26-32 mm.

3 cT cf 26-28 mm.

Id" 30 mm.

Zanzibar (Hamburg type sansibaricus). ?

Summary of 90 cf c? 13 9 9 c. sansibari-
cus 22-31*mm. 26-33 mm.
Kilosa 1 9 38 mm.
Mikindani 11 & tf 2 99 31-34 mm. 35-38 mm.
Kitava 120 c? & 13 9 9 28-34 mm. 34-36 mm.
Summary of 131 tf1 c? 16 9 9 c. eitrinus**28-te mm. 34-38 mm.
Breeding. On [March 25, on all sides, males might be heard calling
vociferously; the call consisting of an explosive double click or snap.
[Most of the females taken were gravid, their scarcity, relative to the
males, apparently indicating that the breeding season was largely
over. One mass of eggs was found at the base of a rice plant just above
water level, the latter having recently subsided.

Hyperolius parkeri parkeri Loveridge
Plage 3, figs. 3-4.

Hyperolius parkeri Loveridge, 1933, Bull. Mus. Comp. Zool., 74, p. -410: Baga-
moyo, Tanganyika Territory.

1 & (M. C. Z. 25270) Siga Caves, T. T. 16. vi. 39.
13 cf 4 9 (M. C. Z. 25271-9) Amboni Estate, T. T. 17. vi. 39.

*Only 1 frog of the 90 attains 31 mm.
**Only 7 frogs of 131 under 30 mm., average 31 mm.

410 bulletin: museum of comparative zoology

Variation. Outer finger with almost 2 joints free for second only
narrowly margined with web, fourth toe with 1 joint free the second
being well webbed, outer toe webbed to, or almost to, disk; tibiotarsal
articulation reaches eye or nostril (in both sexes).

Size. Length of <?<? 21-24 mm., average 23 mm., of 9 9 19-21
mm., average 19.7 mm.

Breeding. On June 17 males were observed emitting their trilling
call; females, taken on the same date, were all gravid.

Hyperolius parkeri rovumae subspec. nov.

Plate 3, figs. 5-6.

9 & 5 9 (M. C. Z. 25260-9) Kitaya, T. T. 31. iii-4. iv. 39.

Type. Museum of Comparative Zoology, No, 25260, a gravid 9
from Kitaya, Rovuma River, Lindi Province, southeastern Tangan-
yika Territory. Collected by Arthur Loveridge, March 31, 1939.

Paratypes. Museum of Comparative Zoology, Nos. 25261-25269,
being 13 specimens as listed above.

Diagnosis. Principal differences from II. p. parkeri Loveridge are
as follows :

Breeding spinosites of males confined to hind feet; sexes dichromatic;
females without a well-defined, black-edged canthal and lateral band,
at most indicated by an upper and lower series of black spots; size
smaller, cfcf 21-24 mm., 9 9 19-21 mm.; range: From Witu, Kenya
Colony, south along coast to Dar es Salaam, Tanganyika Territory.

p. parkeri
Breeding spinosities of males on belly, thighs, and soles of feet;
females resemble males in possessing a well-defined, black-edged,
cantho-lateral band; size larger, cTc? 24-27 mm., 9 9 21-23 mm.;
range : Kitaya, Rovuma River, 275 miles south of Dar es Salaam.

p. rovumae

Description. Outer finger with 2 joints almost free for second only
narrowly margined with web; fourth toe with 1 joint free or narrowly
margined with web, outer toe webbed to disk; tibio-tarsal articulation
reaches between eye and nostril, or to nostril.

Color. In life, d71 Paratype. Above, dark brown, back minutely
punctate with brown; from nostril through and above orbit along
flanks to groin a broad white band, edged above and below by lines
formed by a concentration of dusky spots; limbs clear, transparent
brown stippled with black, such stippling forming large spots on fore-

loveridge: African amphibians 411

arm and tibia; thigh tinged with pink from subcutaneous blood vessels.
Below, pale bu fly-white; buccal borders finely punctate with black;
rest of undersurface translucent flesh color.

In life. 9 Type. Above, dark green, back minutely punctate with
black; from nostril through and above orbit along flanks to groin a
broad white band, edged above and below by black; limbs a clear, but
paler, green, immaculate on thigh and foot but stippled with brown
on forearm and tibia. Below, throat verdigris green; belly largely
transparent so that eggs are clearly discernible; limbs and feet trans-
parent green.

Size. Length of cf cf 24-27 mm., average 25.3 mm., of 9 9 21-23
mm., average 21.8 mm. Length of type 9 23 mm.

Hyperolius nasutus Giinther

Hyperolius nasutus Giinther, 1864, Proc. Zool. Soc. London, p. 482, pi. xxxiii,
fig. 3: Duque de Braganga, Angola.

9 d" (M. C. Z. 25280-4) Kitaya, T. T. 25-30. iii. 39.
16 a" 5 9 (M. C. Z. 25285-9) Amboni Estate, T. T. 17. vi. 39.

Remarks. It might be pointed out that breeding males can be dis-
tinguished from those of parkeri by the absence of black spinosities,
they are a brighter green and the dorso-lateral band is narrower and
not edged above and below with black.

Variation. Outer finger with almost 2 joints free for second only
narrowly margined with web, fourth toe with 1 joint free and second
moderately webbed, outer toe with last joint half-webbed ; tibio-tarsal
articulation reaches eye or nostril (in both sexes).

Color. In life. <? . Above, pale green, head alone minutely punctate
with brown; from above eye (only rarely from nostril) along flanks to
groin a white band; limbs a clear, pale green; thigh immaculate except
for a subcutaneous blood vessel showing as a red line; forearm, tibia,
and exposed surface of foot stippled with black. Below, gular disk
yellow; throat dark green; belly pure white; limbs colorless except for
a faint greenish tinge; feet colorless except for a slight yellowish tinge.

Size. Length of cfc? 20-24 mm., average 22.8 mm., of 9 9 21-23
mm., average 22 mm.; a tailed froglet measures 17 + 15 mm.

Breeding. The above-mentioned froglet was taken about March 30.
On June 17 all five females were distended with ova which are clearly
visible through the semitransparent abdominal wall.

412 bulletin: museum of comparative zoology

Hyperolius pusillus (Cope)
Plate 3, figs. 7-8.

Crumenifera pusilla Cope, 1862, Proc. Acad. Nat. Sci. Philadelphia, p. 343:

Umvoti, Natal. (9).
Hyperolius microps Gunther, 1864, Proc. Zool. Soc. London, p. 311, pi. xxvii,

fig. 3: Rovuma Bay, Tanganyika Territory, (cf).
Hyperolius milnei Loveridge, 1935, Bull. Mus. Comp. Zool., 79, p. 18: Witu,

Coast Province, Kenya Colony. ( 9 ).

12 9 9 (M. C. Z. 25290-8) Kitaya, Rovuma River, T. T. 4. iv. 39.

Synonymy. These are practically topotypes of microps Gunther,
with which milnei, from a locality 500 miles further north, appears con-
specific as I (1941h, p. 291) have recently suggested. They would ap-
pear to differ from pusillus (inc. translucens Power) of which we have
good topotypic series, in generally exhibiting a black spot on knee and
heel. This, however, is by no means constant in the big series of milnei
and is absent in a series from Golbanti, Tana River, as well as from the
types of usaramoae Loveridge, from Dar es Salaam. I prefer, therefore,
to regard all as representing one species which is very variable in the
size of the dorsal spots which may be present or absent.

Variation. Outer finger with 1 joint free, fourth toe with last joint
• narrowly webbed, outer toe fully webbed to disk; tibio-tarsal articu-
lation reaches eye or nostril

Color. In life. 9 . Above, head and anterior part of back rich
green; from nostril to orbit a line of relatively large black spots some-
times coalescing, others are scattered over head and anterior part of
back, in addition the entire upper surface is peppered by minute
black points (visible only with a lens); back posteriorly, as well as
limbs, a paler green; fingers and toes bright orange. Below, pale bluish
green, breast silvery white.

Size. Length of 9 9 20-23 mm., average 21 mm.

Habitat. Though especial search was made for this species in several
small swamps we failed to find it until, almost at the end of our stay,
I located them in the leech- and crocodile-infected lagoon some miles
below our camp in the direction of Rovuma Bay.

loveridge: African amphibians 413

RANIDAE

Arthroleptides marttenssexi Nieden

Arthroleptides martiensseni Nieden, 1910, Sitz. Ges. naturf. Freunde Berlin, p.
445: Amani, Usambara Mountains, Tanganjika Territory.

<? 9 & juv. (M. C. Z. 25380-2) Magrotto Mtn., T. T. 3. vii. 39.

Color. In life. cT. Above, dark olive with a brown interorbital bar;
from nostril to eye, and from eye over tympanum to axilla, a rather
indistinct dark band; lips dark brown flecked with white; limbs,
fingers and toes blotched with brown. Below, throat slightly brown
flecked with white; belly, and remaining underparts, whitish tinged
with yellow towards the sides; palms and soles dark brown.

Size. Length c? 57 mm., 9 67 mm., young 22 mm.

Habitat. The male was shot as he squatted on a sloping rock far in
under an overhanging slab at the side of a torrent, the others were
taken beneath vegetable debris or logs in the sodden rain forest.

Raxa albolabris albolabris Hallowell
Plate 4, fig. 3.

Rana albolabris Hallowell, 1856, Proc. Acad. Nat. Sci. Philadelphia, p. 153:
West Africa.

8 (M. C. Z. 25301-4) Idjwi Id., B. C. 20-28. ii. 39.

Native name. Mote (Lulega).

Variation. Snouts somewhat intermediate between topotypical
Liberian R. a. alboleibris and a paratype Angolan R. a. parkeriana
Loveridge, 1941 (n.n. for acutirostris Parker, 1936, preoccupied by
acutirostris Fatio, 1872).

Size. Length 44-61 mm., average 54.7 mm.

Breeding. Xonbreeding, possibly subadult.

Diet. Black crickets, ant, ? stone fly, spider, slug, snail.

Raxa galamexsis bravaxa (Peters)

Limnodytes bravanus Peters, 1882, Sitz. Ges. naturf. Freunde Berlin, p. 3:
Barawa, i.e. Brava, Italian Somaliland.

3c,39 (M. C. Z. 25305-6) Ujiji, T. T. 11. iii. 39.

414 bulletin: museum of comparative zoology

Size. Length cfcf 64-74 mm., 9 9 63-74 mm.

Breeding. Not breeding.

Diet. In one stomach a cricket, caterpillars, and a large (15 mm.
long) shield bug; nothing in stomachs of three emaciated frogs taken
from deep water in an old sugar vat from which there was no way of
escape.

Rana fuscigula fuscigula Dumeril & Bibron

Rana fuscigula Dumeril & Bibron, 1841, Erpet. Gen., 8, p. 386: South Africa.
1 d1 16 9 (M. C. Z. 25315-9) Mushongero, U. 1. ii. 39.

Native name. Senyamiganda (Lukiga for all ranids).

Variation. Characters of fuscigula, viz.: Length of tibia not more
than half the length from snout to anus; fifth toe webbed to the very
tip, but not so fully as in R. f. chapini.

Size. Length cT 55 mm., 9 9 60-75 mm., average 70 mm.

5 & 15 9 13 juv. 2 tad. (M. C. Z. 25320-5) Mushongero, U. 1. ii. 39.

Variation. Characters of fuscigula and chapini, i. e. intermediates
vjz.: Length of tibia more than half the length from snout to anus;
fifth toe webbed to the very tip (except for the two smallest young,
M. C. Z. 25322-3, whose condition shows an approach to that of/.
angolensis) but not so fully as in R. f. chapini; size that of fuscigula
and not of chapini.

Size. Lengths cfcf 49-54 mm., 9 9 50-70 mm., average 62 mm.,
young 35-45 mm., average 39 mm. tadpoles 55-76 mm.

Enemies. One frog recovered from stomach of a green snake
{Chlorophis irregularis).

Rana fuscigula chapini Noble

Rana chapini Noble, 1924, Bull. Am. Mus. Nat. Hist., 49, p. 214, fig. 6a:
Batama, Belgian Congo.

3 J 19 15 juv. (M. C. Z. 25326-9) Budongo Forest, U. 23. xi. 38.
2 9 1 juv. (M. C. Z. 25330-2) Kibale Forest, U. 9. xii. 38.
4 d" 5 9 4 juv. (M. C. Z. 25333-7) Magrotto Mtn., T. T. 5. vii. 39.

Variation. Length of tibia more than half the length from snout to
anus; fifth toe webbed to the very tip though in some of the smallest
Budongo and Kibale specimens rather less fully than in the adults,
i.e. showing an approach to the ( ? ancestral) condition found in R. f.
angolensis.

loveridge: African amphibians 415

Size. Length cT & 60-65 mm., 9 9 65-90 (75-76 at Kibale, 84
43udongo, 65-90 Magrotto) mm., young 27-59 mm., average 40 mm.,
tailed young 25 + 15, 26 + 12, and 27+16 mm.

Habitat. In or beside streams in the forest in all three localities, of
which Kibale appeared to be the driest.

Rana fuscigula angolensis Bocage

Rana angolensis Bocage, 1866, Jorn. Sci. Lisboa, 1, p. 73: Duque de Braganca,
Angola.

1 juv. (M. C. Z. 25307) Mabira Forest, U. 11. xi. 38.
6(^1 95 juv. (M. C. Z. 25308-12) Mubuku Valley, U. 31. xii. 38.
1 9 5 juv. (M. C. Z. 25313-4) Kisenyi, B. R. 9. ii. 39.

Native name. Lulenga (Luganda for all ranids).

Variation. Length of tibia very variable but always more than half
the length from snout to anus; fifth toe with last phalange free of web.

Color. Very variable, the most striking variant being one from the
Mubuku Valley which, in life, was: Above, rich reddish brown, a
broad, dull golden, vertebral stripe, edged with black anteriorly, from
snout to anus.

Size. Length d^cf 45-52 mm., 9 9 60-70 mm., young 25-47 mm.,
average 39.6 mm.

Habitat. Boulenger recorded and figured frogs from the Mubuku
Valley, Ruwenzori Mountains, under the name of Rana nutti, which
I regard as a synonym of angolensis. It might reasonably be expected
that frogs from this forested region would have been referable to the
sylvicoline race chapini, instead we must assume that the valley has
been populated by frogs ascending the Mubuku River from the sa-
vanna surrounding the eastern foot of the mountain. In Mabira the
hot savanna is everywhere encroaching on the forest though actually
I captured our single specimen in a rain-filled rut of the track which
passes through a stretch of forest.

Rana christyi Boulenger

Rana christyi Boulenger, 1919, Revue Zool. Africaine, 7, p. 5: Madie, i.e.
Medje, Ituri Forest, Belgian Congo.

9 (M. C. Z. 25369) Budongo Forest, U. 29. xi. 38.

Distribution. This constitutes the first record of the occurrence of
this species in Uganda. In the Congo it has been confused with aequip-

416 bulletin: museum of comparative zoology

licata Werner, (which is a synonym of longirostris Peters, fide Nieden
(1908) who compared the types of both), first by Noble (1924) for
two of his Boyulu "ckristyi" (now M. C. Z. 6612-3) are really longiros-
tris, and later by Witte (1934, p. 171) for six of his Dika series of
"aequiplicata (now M. C. Z. 17956-7, 21757-60) are really ckristyi,
though at least two of his Djamba frogs (now M. C. Z. 17958-9),
mentioned in the same paper, were correctly identified as christyi.

Variation. Toes of the hind limb, taken in order from first to fifth,
exhibit the following phlanges free of web: V/2, 1, 1, 2, 0.

Color. In life. 9 • Above, yellow leaf-brown, a short, transverse,
interorbital, black bar; round snout to behind tympanum a rich black
streak; back irregularly flecked with black, the dorso-lateral glandular
fold carrying a series of black dashes; flanks with a few flecks of black;
forelimbs and feet distinctly, hind limbs indistinctly, barred with
black; thighs posteriorly bright mustard yellow marbled with black;
hind feet yellowish faintly tinged with pink. Below, throat cream-
colored; belly bright lemon yellow; limbs yellowish; soles of forefeet
light with dusky markings, those of hind feet black.

Size. Length 9 51 mm.

Habitat. Captured while taking tremendous leaps over the sodden,
leaf-strewn, forest floor.

Rana oxyrhynchus oxyrhynchus Smith

Rana oxyrhynchus A. Smith, 1849, 111. Zool. S. Africa, Rept., pi. lxxvii, figs.
2-2c : Kaffirland & region of Port Natal, South Africa.

cT (M. C. Z. 25338) Budongo Forest, U. 30. xi. 38.
4 cf 1 9 (M. C. Z. 25339-40) Kisenyi, B. R. 9. ii. 39.

9 (M. C. Z. 25341) Mikindani, T. T. 8. iv. 39.
1 juv. (M. C. Z. 25342) Mbanja, T. T. 3. v. 39.
8 juv. (M. C. Z. 25343) Siga Caves, T. T. 9. vi. 39.

9 (M. C. Z. 25344) Magrotto Mtn., T. T. 1. vii. 39.

9 (M. C. Z. 25345) Tanga township, T. T. 23. vii. 39.

Native names. Nanhengo and nanihengo (Kimakonde at Kitaya and
Mbanja respectively); nanmiengo (Kimawiha). But all applied to
R. m. maseareniensis also.

Variation. Fourth toe with 1^-2 phlanges free of web, all other toes
webbed to the tip except in very young frogs where the webbing is a
little less extensive, the Mbanja froglet, for example, has the terminal
phlange of the third toe apparently free.

loveridge: African amphibians 417

Color. In life, of an unusually colored 9 from Magrotto. Above,
gray, but so heavily overlaid and mottled with black as to appear
black; from nostril a line passes below eye to base of forearm; the
dorsolateral glandular fold and a line on hinder side of thigh from anus
to back of knee, whitish or greenish. Below, lips flecked with white;
throat silvery white flecked with gray merging into the slightly yel-
lowish belly coloring, distinctly yellowish on its periphery and on hind
limbs.

Size. Length cfcT 41-45 mm., 9 9 41-54 mm., young 19-26 mm.

Breeding. From February 9 to June 20 all females were gravid,
their abdomens distended with ova; the female taken on July 23,
however, had finished spawning.

Habitat. It should be explained that the Magrotto frog was taken
in a swamp in open country far down the mountain ; had it been cap-
tured in the forest one would have expected the sylvicoline form which
follows.

Raxa oxyrhynchus gribinguiensis Angel

Rana (Ptychadena) Gribinguiensis Angel, 1922, Bull. Mus. Hist. Nat, Paris,
28, p. 399, fig. — : Fort Crampel, Lake Chad, French West Africa,

9 (M. C. Z. 25346) Amboni Estate, T. T. 20. vi. 39.

Variation. Fourth toe with 1 phlange free as is characteristic of this
large race.

Size. Length 9 5S mm.

Breeding. Gravid, distended with ova.

Habitat. Taken in the patch of forest, beside which I was camped,
or on the nearby land which was being cleared by tractors. ■ To find
this large and unmistakable montane-forest form — so far as the east
is concerned — almost at sea-level, was a considerable surprise.

Rana mascareniensis mascareniensis Dumeril & Bibron

Rana Mascareniensis Dumeril & Bibron, 1841, Erpet. Gen., 8, p. 350: Mada-
gascar; Mauritius; Seychelles.

2 c? 2 9 (M. C. Z. 25347-50) Kinangop Plateau, K. C. 29. x. 38.
7 juv. (M. C. Z. 25351-2) Kitaya, Rovuma R.. T. T. 28. iii. 39.
9 (M. C. Z. 25353) Siga Caves, Tanga, T. T. 9. vi. 39.
9 (M. C. Z. 25354) Tanga township, T. T. 22. vii. 39.

Native names. The same as for B. o. oxyrhynchus.
Variation. Toes of the hind limb, taken in order from first to fifth,
exhibit the following phlanges free of web: 2, 1}/?, 2, 3, 13^ for the

418 bulletin: museum of comparative zoology

series from Kinangop at 10,000 feet, and 1-1 J^, 1-lJ^, 1-lJ^, 23^-3,
\-\Yl on those from the coastal belt almost at sea-level. The alloca-
tion of the Kinangop series to the typical race, therefore, may be re-
garded as tentative. They are separated from the larger western, and
eastern montane, race apparently only by size!

Size. Length cTcf 41-42 mm., 9 9 34-48 mm., young 13-33 mm.

Breeding. On October 29 males were calling "quek-quek" and
females spawning in rain-filled pools along a cart track traversing
the plateau. On March 28 and June 9 newly emerged young were
numerous.

Enemies. Two young were recovered from the stomach of an egret
(Egretta g. dimorpha) and one from a house snake (Boaedon I. lineatus)
near the Siga Caves.

Rana mascareniensis venusta Werner

Rana venusta Werner, 1907, Sitz. Akad. Wiss. Wien, 116, 1, pp. 1889 and 1892,
pi. iv. fig. 11: Entebbe, Uganda; Mongalla, Belgian Congo; and Lagos,
Nigeria.

1 9 2 juv. (M. C. Z. 25355-6) Budongo Forest, U. 30. xi. 38.
1 juv. (M. C. Z. 25357) Kibale Forest, U. 16. xii. 38.
1 juv. (M. C. Z. 25358) Nyakabande, U. 28. i. 39.
3 cT (M. C. Z. 25359-60) Mushongero, U. 1. ii. 39.
6 c? 3 9 (M. C. Z. 25361-2) Kisenyi, B. R. 9. ii. 39.
1 & 7 9 6 juv. (M. C. Z. 25363-4) Idjwi Id., B. C. 25. ii. 39.
4 & 2 9 (M. C. Z. 25365-6) Ujiji, T. T. 11. iii. 39.
3^29 (M. C. Z. 25367-8) Magrotto Mtn., T. T. 1. vii. 39.

Native name. Marembera (Lulega).

Variation. Toes of the hind limb, taken in order from first to fifth,
exhibit the following phlanges free of web: 1-2, 1-lH, 1_1M> 2-3,
1-1J/2. The tibio-tarsal articulation of the adpressed hind limb, in both
this and the typical form, reaches the end of the snout or beyond, both
conditions being found in any locality from which an adequate series
is available. This apparently reduces the recognition of venusta as a
western, or forest, form to a matter of size.

Size. Length cfc? 40-53 mm., 9 9 52-65 mm., young 20-46 mm.

Breeding. On November 30, juvenile frogs measured 20-22 mm.
and increased more or less progressively to February 25, when those of
30^6 mm. were encountered.

Habitat. The Budongo frogs were sitting on a grassy bank beside
a ditch across which they leaped into sedges of a small swamp separat-

loyeridge: African amphibians 419

ing road from forest. The Kibale frog was in a swamp by palms with-
out the forest. The Idjwi series were on the lake shore. Ujiji in a
sugar vat. The Magrotto series in a swamp near, but outside, the
forest.

Rana occipitalis Giinther
Plate 4, fig. 4.

Rana occipitalis Giinther, 1858, Cat. Batr. Sal. Brit. Mus., p. 130, pi. xi: West
Africa; Africa; Gambia.

1^29 (M. C. Z. 25370-1) Ujiji, T. T. 11. iii. 39.

Size. Length <? 97 mm., 9 9 125-125 mm.

Diet. Bones of a small frog and a large Dytiscid beetle (Cybister
sp.) in one.

Breeding. Apparently not; spawn small in females.

Rana edulis (Peters)

Pyxicephalus edulis Peters, 1854, Monatsb. Akad. Wiss. Berlin, p. 626: Boror;
Mozambique; and Tete, Mozambique.

1 juv. (M. C. Z. 25372) Kitaya, T. T. 31. iii. 39.
3 9 3 juv. (M. C. Z. 25373-7) Mikindani, T. T. 10. iv. 39.
4 juv. (M. C. Z. 25378) Siga Caves, T. T. 15. vi. 39.
1 juv. (M. C. Z. 25379) Amboni Est., T. T. 17. vi. 39.

Native names. Liola (Kiyao); kitowa (Kimakonde, but also for
Chiromantist).

Size. Length 9 9 125-145 mm., young 29-115 mm.

Breeding. On June 15 the ground was literally a-hop with hundreds
of emergent frogs which had come from a large swamp and were seek-
ing shelter under bundles of grass and among the rubble of a collapsed
hut.

Diet. These bullfrogs seem impervious to stings, for example, the
stomach of one Mikindani frog held: 3 scorpions each measuring l%"
from head to end of sting, a centipede 4" long and 1/3" broad, a milli-
pede 2%" long, a scutigera, a carabid beetle \}/%' long of a species
that ejects formic acid, 3 black "stink" ants 11/16" long, remains of a
snail with shell 7/16" diameter.

From other stomachs a frog (Hyperolius e. sansibaricus), 2 scorpions,
4 millipedes, a tenebrionid beetle \}/%' long, 3 hard-shelled tenebrionids

420

bulletin: museum of comparative zoology

each about ^" long without their heads, a nymph, and a hard black
tip of a sisal (aloe) leaf, the thorn measuring 1" in length.

Enemies. In their turn these big frogs are eaten by natives of the
Mawiha tribe and two tailed tadpoles were recovered from the stomach
of an egret (Egretta g. dimorpha). At Kitaya a frog was brought in
which lacked a left hind leg, if it had ever had one, the place was com-
pletely healed over though the frog was 60 mm. long.

Phrynobatrachus spp.

Recently Dr. Laurent, (1941b, p. 192) has discussed the subgenera
of Phrynobatrachus, his conclusions being based on his own anatomical
investigations; the following list in which the species are arranged
according to disks, webbing, and habitat, conforms very nearly with
his views except that graueri is anatomically affiliated with the first
five species listed, i.e. Phrynobatrachus (sensu strictu) and not with the
last three, which he refers to Xatalobatrachus.

As the digital disk, when present, occupies much of the terminal
phlange, a toe that is webbed to the disk is here considered to have one
phlange free, the necessity for such a rating is apparent when, as is
not infrequently the case, the digital expansion dries or macerates, or
when one has to compare the extent of webbing with such a species as
natalensis.

The list refers only to the material in the present collection, every
individual of which — except in the case of the long series from Idjwi
Island — has been examined.

Species

Disks

oresent

Phlanges on toes free of web

Habitat

1st. | 2nd.

3rd. | 4th. 5th.

natalensis

No

1

i

2

3

2

savanna swamp

perpalmatus

Yes

1

i

1

1-134

1

(<

acridoides

Yes

1

i

1

•)

1

n

kinangopensis

•Just

1

i

1

•>

2

" i

plicatus

Yes

1

i

1

2

1

forest pool

krentii

Yes

1

i

1

2

1

a

dendrobates

Yes

1

i

2

3

•)

a

versicolor

Yes

1

i

2

3

2-23/2

" 2

graueri

Yes

i-iy2

lJ^-2

2-2^

3-3^

2-2V2

forest & vie.

■In alpine meadows at 10,000 feet.
:See remarks under Habitat.

loveridge: African amphibians 421

Phryxobatrachus xatalexsis (Smith)

Stenorhynchus ?iatalensis A. Smith, 1849, 111. Zool. S. Africa, Rept., App., p.
24: Natal, Union of South Africa.

2 c? (M. C. Z. 25432-3) Mabira Forest, U. 15. xi. 38.
8 hgr. (M. C. Z. 25434-5) Budongo Forest, U. 23. xi. 38.
1 cf 1 9 (M. C. Z. 25436-7) Ujiji. T. T. 11. iii. 39.

Native name. Lulenga (Luganda for all ranids).

Variation. Tibio-tarsal articulation reaches eye (cf 9 ) or nostril
( 9 )• See also under Phrynobatrachus.

Color. A narrow, light, vertebral line in one Budongo frog only.

Size. Length cTcf 31-32 mm., 9 31 mm.; halfgrown 25-29 mm.

Breeding. On November 11, males calling from pool.

Enemies. One recovered from the stomach of a one-streaked hawk
(Kaupifalco monogrammicus) at Kitaya.

Habitat. Both Mabira and Budongo series were taken outside the
forest in a pool and stream respectively; the Ujiji pair were in an
abandoned sugar vat from which there was no means of escape.

Phryxobatrachus perpalmatus Boulenger

Phrynobatrachus perpalmatus Boulenger, 1898, Proc. Zool. Soc. London, pi.
xxxviii, fig. 1: Lake Mweru, Northern Rhodesia.

10 ad. 7 juv. (M. C. Z. 25414-8) Ujiji, T. T. 11. iii. 39.

Variation. Tibio-tarsal articulation reaches eye, sometimes only
barely. See also under Phrynobatrachus.

Color. A broad, light, vertebral line in two frogs only; the longitudi-
nal dusky lines on hinder and lower aspect of thighs are characteristic
and an aid to identification.

Size. Length 19-24 mm., young 10-14 mm.*

Habitat. Calling from swamped grasslands near the lakeshore.

Phryxobatrachus acridoides (Cope)

Staurois acridoides Cope, 1867, Journ. Acad. Nat. Sci. Philadelphia, 6, p. 198:
Zanzibar.

2 (M. C. Z. 25426-7) Kitaya, T. T. 4. iv. 39.

12 (M. C. Z. 25428-9) Mikindani, T. T. 11. iv. 39.

3 (M. C. Z. 25430-1) Siga Caves, T. T. 8. vi. 39.

422 bulletin: museum of comparative zoology

Variation. Tibio-tarsal articulation reaches eye or nostril, in young
even to end of snout. See also under Phrynobatrachus.

Color. A broad, or narrow, vertebral line in 3 Mikindani males
and 2 Siga females only.

Size. Length crc? 25-27 mm., 9 9 26-30 mm., young 17-23 mm.

Breeding. On April 11 males were numerous (7 to 3 females) and
calling, a hind-limbed tadpole measuring 14 + 32 mm. is presumed
to be this species. Males are recognizable by their dark throats which
become roughened at this season by accentuation of the dermal
granulations.

Habitat. At Mikindani in swamped grasslands beneath the coco-
nuts, at Siga on a swampy path beside the Mkulumusi River.

Phrynobatrachus kinangopensis Angel

Phrytwbatrachus Kinangopensis Angel, 1924, Bull. Mus. Hist. Nat. Paris, 30,
p. 130: Mount Kinangop, Aberdare Mountains, Kenya Colon y.

2 (M. C. Z. 25424-5) S. Kinangop Plateau, K. C. 29. x. 38.

Variation. Tibio-tarsal articulation reaches tympanum. See also
under Phrynobatrachus.

Size. Length cTd71 14-19 mm.

Habitat. In a rain-filled pool beside road at 10,000 feet.

Phrynobatrachus plicatus (Giinther)

Hyperolius plicatus Giinther, 1858, Cat. Batr. Sal. Brit. Mus., p. 88, pi. vii,
fig. C: Coast of Guinea.

1 (M. C. Z. 25419) Budongo Forest, U. 29. xi. 38.

Distribution. New for Uganda, though long known from the Ituri
Forest of the Belgian Congo.

Variation. Tibio-tarsal articulation reaches far beyond end of
snout. See also under Phrynobatrachus.

Size. Length 25 mm.

Habitat. In ditch at forest edge.

Phrynobatrachus krefftii Boulenger

Phrynobatrachus krefftii Boulenger, 1909, Ann. Mag. Nat. Hist. (8), 4, p. 496:
Amani, Usambara Mountains, Tanganyika Territory.

5 (M. C. Z. 25420-3) Magrotto Mtn., T. T. 8. vii. 39.

loveridge: African amphibians 423

Variation. Tibio-tarsal articulation reaches eye or end of snout.
See also under Phrynobairackus.

Color. In life, cf . Above, olive mottled with black; limbs barred
with black. Below, throat rich yellow, posteriorly crossed by a broad
dusky band vermiculated with white; rest of undersurface whitish
faintly tinged with yellow and spotted with pale brown, principally
along outer part of limbs; palms and soles black.

9 . Above, as in male. Below, whitish with dusky markings as in
male but the post-gular band is much less clearly defined.

Size. Length c?c? 31-36 mm., 9 9 33-41 mm.

Habitat. Some were taken in stony puddles formed by a spring
issuing from the mountainside in deep forest, others were resting on the
wet banks of, or stones in, a turbulent little torrent where it cascaded
through the shady forest.

Phrynobatrachus dendrobates (Boulenger)

Arthroleptis dendrobates Boulenger, 1919, Revue Zool. Afr. 7, p. 8: Madie, i.e.
Medje, Belgian Congo.

& (M. C. Z. 25438) Kibale Forest, U. 12. xii. 38.
5 (M. C. Z. 25439-41) Idjwi Id., B. C. ii. 39.

Variation. Tibio-tarsal articulation reaches nostril or end of snout.
See also under Phrynobatrachus.

Size. Length d71 31 mm., and of Idjwi series both sexes 23-33 mm.

Enemies. In stomach of a tree snake (Hapsidophrys lineata) taken
in Budongo Forest.

Habitat. On damp leaves of forest floor in Kibale Forest.

Phrynobatrachus versicolor AM.

Phrynobatrachus versicolor AH, 1924, Zool. Anz., 61. p. 100: Rugege Forest,
Belgian Ruanda-Urundi.

20 (M. C. Z. 25451-5) Nyakabande, U. 28. i. 39.
7 (M. C. Z. 25442-5) Mushongero, U. 1. ii. 39.
25 (M. C. Z. 25446-50) Idjwi Id., B. C. ii. 39.

Native name. Miusi (Lulega, but not specific).

Correction. Formerly I (1936h, p. 97) regarded this species as in-
distinguishable from dendrobates, with which I synonymized it. Now,
after seeing them both in life, I believe them to be distinct though I

424 bulletin: museum of comparative zoology

am at a loss to define versicolor except by stouter habit and pigmenta-
tion of lower surface, for in size, limb length, webbing, etc., they ap-
pear to be indistinguishable.

Variation. Tibio-tarsal articulation just fails to reach the eye in one
frog, reaches eye or nostril in majority, end of snout in a very few
instances. Minute spines are present on soles of hind feet in both sexes
but are better developed in the males. See also under Phrynobatrachus.

Color. A narrow, light, vertebral band in two (Mushongero and
Idjwi) frogs only, but flanks of three large females in the latter series
are strikingly light, as if each side bore a broad, light band.

Size. Length of cf cf 25-28 mm., 9 9 25-34 mm., average for 52
frogs of both sexes 29.7 mm.

Habitat. This frog is doubtless associated with forest as is indicated
by the type locality. I must state, however, that most of mine came
from the deforested uplands in the Kigezi district.

Phrynobatrachus graueri (Nieden)

Arthroleptis graueri Nieden, 1910, Sitz. Ges. naturf. Freunde Berlin, p. 441:
Rugege Forest, Belgian Ruanda-Urundi.

4^ 297 juv. (M. C. Z. 25456-8) Budongo Forest, U. 23. xi. 38.

(M. C. Z. 25459) Kibale Forest, U. .13. xii. 38.

(M. C. Z. 25460-1) Mubuku Valley, U. 31. xii. 38.
2 juv. (M. C. Z. 25462-5) Mihunga Swamp, U. 18. i. 39.

(M. C. Z. 25466-9) Mushongero, U. 3. ii. 39.

(M. C. Z. 25470-3) Kisenyi, B. R. 11. ii. 39.
2 juv. <M.*C. Z. 25474-7) Idjwi Id., B. C. 16-28. ii. 39.

Native names. Senyamiganda (Lukiga for all ranids) ; ntoli (Lutoro) ;
etoli (Lukonjo); kashakara (Lulega, but for Arthroleptis also).

Remarks. In the field, more especially at Mihunga and on Idjwi,
I was under the impression that I was dealing with two species, in part
owing to the fact that gravid females from 25-31 mm. seemed a large
range. In the laboratory I have failed to find any means of separating
two forms or species, for they occurred together.

Variation. Tibio-tarsal articulation reaches eye or nostril, rarely to
end of snout; this range is covered by both sexes though in general
females average a relatively shorter hind limb. See also under Phry-
nobatraehits.

Color. In life, c? ■ Mihunga. Above, dark olive, from nostril to eye
an indistinct brown streak, a broad and conspicuous one from eye,
through tympanum, to forelimb; from snout to anus a very fine, yel-

2

9

1 cf

1

9

17 cf 15

9

5 &

11 cf

4

9

69 cf 25

9

LOVERIDGE: AFRICAN AMPHIBIANS 425

low, vertebral line; a black eireum-anal area is flanked by a pair of
light lines; limbs barred with brown, thighs with light bars also, tibia
with a fine light longitudinal line. Below, throat black, posteriorly
minutely flecked with white; breast and limbs infuscated with brown;
hind limbs, from the waist, often chrome colored.

In life. 9 . Mihunga. Above, brown or olive though sometimes
quite green, snout paler; a dark interorbital band; a dark, light-edged,
circum-anal area ; limbs indistinctly barred with brown ; digits barred
with white and brown. Below, white, the throat infuscated with
brown, the belly often bluish and almost spotted writh brown; soles
brown.

A broad, light, vertebral line in 1 cf from Mubuku, 1 c? from
Mushongero, 1 d71 and 1 9 from Mihunga, 16 <?<? and 8 9 9 from
Idjwi; 1 Mushongero and 2 Mihunga males have the dorsal area be-
tween the glandular cordons light brown edged with black along the
cordons, thus presenting a very distinctive appearance.

Size. Length of 108 cTcT 20-24 mm., average 21.8 mm., of 46 9 9
24-31 mm., average 26.3 mm., of 11 young 9-14 mm.

Breeding. On January 18, males, recognizable by their dark and
granular throats, were heard calling "tink-tink" in the sw^amp at the
foot of the \Yeria Ravine; on February 3 they were noted as "clicking"
in pools close to the papyrus fringing the edge of Lake Mutanda.
Females were gravid during the entire three months of our stay in the
Central Lake Region which coincided with heavy rains.

Enemies. In stomach of a tree snake (Dipsodoboa nnicolor) in
Mihunga Swamp.

Habitat. On damp leaves beside streams in the forests of Budongo,
Kibale, and the Mubuku Valley, while at Mihunga they were calling
from the water-filled footprints of elephants among the long grass of
the swamp. At Mushongero and Kisenyi there wTas no real forest in
the vicinity though much of this region has suffered from deforestation.

Arthroleptis minutus Boulenger

Arthroleptis minutus Boulenger, 1895, Proc. Zool. Soc. London, p. 539, pi. xxx,
fig. 4: Durro, Western Somaliland, i.e. Duro, Ethiopia.

25 (M. C. Z. 25408-9) S. Kinangop Plateau. K. C. 27. x. 38.
2 (M. C. Z. 25410-1) Mikindani, T. T. 12. iv. 39.
1 (M. C. Z. 25412) Lindi, T. T. 1. vi. 39.

Variation. Tips of fingers not dilated; metatarsal tubercles and a

426 bulletin: museum of comparative zoology

tarsal tubercle present, small; tibio-tarsal articulation reaches eye or
just beyond.

Color. The Mikindani males differ from the Kinangop breeding
males by the possession of very black throats.

Size. Length (sexing questionable as based on external appearance)
cfcT 18-20 mm., average 19 mm., 9 9 21-24 mm., average 21 mm.,
only one frog over 22 mm.

Breeding. On October 27 the Kinangop frogs were assembling for
breeding, the small rains having just commenced. Probably the two
Mikindani males also, the big rains being in full force. The only juve-
nile taken during the entire eight months was captured on June 1
under grass thatching, it measured 12.5 mm.

Parasites. The buttocks of the Mt. Kinangop series are infested
with chigger mites (Endotrombicula sp.) possibly of the same species
(penetrans) which I found in minutus from Mt. Sagalla.

Habikit. Many of those taken on the Plateau at 10,000 feet were
clambering about in the long dew-drenched grass flanking the Chania
River, it seems difficult to believe that they are specifically identical
with those taken in the coastal belt under 100 feet alt.

Arthroleptis xexodactylus Boulenger

Arthroleptis xetwdactylus Boulenger, 1909, Ann. Mag. Nat. Hist. (8), 4, p. 496:

Amani, Usambara Mountains, Tanganyika Territory.
Arthroleptis schubotzi Nieden, 1910, Sitz. Ges. naturf. Freunde Berlin, p. 440:

Usumbura, Belgian Ruanda-Urundi.

1 (M. C. Z. 25403) Mikindani, T. T. 12. iv. 39.

2 (M. C. Z. 25404-5) Nehingidi, T. T. 12. v. 39.

5 (M. C. Z. 25406-7) Magrotto Mtn., T. T. 8-12. vii. 39.

Synonymy. Apart from our big series of topotypes, we have this
species from right across Tanganyika Territory to the Lake itself, as
well as from Buta, Bas Uele, Belgian Congo, determined by Dr. G. F.
de Witte, with whose identification I concur. There seems, therefore,
no reason to regard schubotzi as distinct, the frogs referred by me
(1933h, p. 380) to that species are much smaller and therefore de-
scribed below as new.

Variation. Tips of fingers more or less swollen, of toes distinctly
dilated into little disks; metatarsal tubercle small, oval, no outer or
tarsal tubercles; tibio-tarsal articulation reaches tympanum or eye
(posteriorly in 9, anteriorly in cfd71).

loveridge: africax amphibians 427

Color. In life, c? . Magrotto. Above, pale fawn; a few light flecks
on buccal border and a larger, slightly pink, spot posterior to commis-
sure of mouth; a dark brown interocular patch continuous with hour-
glass pattern on back; from upper eyelid a light, dorsolateral, glandu-
lar line and immediately below it, on flanks, scattered dark brown
patches; groin transparent red; limbs and feet pale fawn barred or
mottled with dark brown. Below, whitish, chin and throat flecked
with gray; center of belly pale yellow.

Another cf. Magrotto. Above, back, upper arm and tibia vermilion
red, remaining upper parts grayish olive; flanks flecked with white.
Below, grayish, indistinctly flecked with white.

Size. Length cTcf 19-20 mm., 9 9 20-22 mm., young 14-15 mm.

Breeding. On April 12 female gravid, distended with ova; on May
12 two young.

Enemies. One in stomach of a snake (Xcustcrophis o. ulugunirnsis).

Habitat. The Nchingidi juveniles were taken on a forest path and
beneath a log, the Magrotto adults in swamp by camp and near sawpit
at forest edge, immediately after heavy rain.

Arthroleptis xenodactyloides nkukae subspec. nov.

Arthroleptis schubotzi Loveridge (not of Nieden), 1933, Bull. Mus. Comp. Zool.,
74, p. 380.

Type. Museum of Comparative Zoology, Xo. 17029, an adult 9
from Nkuka Forest, Rungwe Mountains, Tanganyika Territory,
collected by Arthur Loveridge, March 1930.

Paratypes. Museum of Comparative Zoology, Nos. 17026-8 and
17030-53, being the entire series of material listed in the above citation
taken in the Rungwe, Ukinga and Uzungwe Mountains of southwestern
Tanganyika Territory.

Remarks. In 1933, though with serious misgivings, I referred these
frogs to schubotzi Nieden rather than describe three new species of
Arthroleptis from the Rungwe-Ukinga region at the northern end of
Lake Nyasa. Subsequent information, however, has led me to con-
clude that schubotzi, whose type measured 21 mm., is a synonym of
xenodactylus Boulenger, whereas the Nkuka material is but a form of
the smaller xenodactyloides Hewitt, whose type was 19 mm.

Diagnosis. A. xenodactyloides Hewitt (1933) was described from
Chirinda Forest, Mount Belinda, Southern Rhodesia, a thousand
miles south of the Nkuka Forest. The new race differs from the typical

428 bulletin: museum of comparative zoology

form, of which we have extensive topotypical material, in that the
belly of adult nkvkae is darkly marmorate whereas in xenodactyloides
it is immaculate.

Color. In life. This, together with much information about breed-
ing, diet, parasites, habitat, etc. will be found in the citation given
above.

Size. Length of Type 9 from snout to anus, 17 mm., but the largest
of several hundred specimens from the Nkuka Forest, Kigogo, and
Madehani, all measure IS mm.

Arthroleptis xenochirus Boulenger

Arthroleptis xenochirus Boulenger, 1905, Ann. Mag. Nat. Hist. (7), 16, p. 108,
pi. iv, figs. 2-2a: Marimba, northern Angola.

c? (M. C. Z. 25413) Idjwi Id., B. C. ii. 1939.

Native name. Kashakara (Lulega). The Balega, when questioned,
appeared not to have any special name for this peculiar little frog.

Remarks. A. .venochirus is known only from the cf holotype and
Nieden's (1908) record of one from Jaunde, Cameroon, presuming that
this is the same frog as that listed by Deckert (1938).

Idjwi Island, in Lake Kivu, lies just under a thousand miles north-
east of Marimba, yet our frog differs in only one particular from the
description; being formalin preserved, however, the color pattern
description cannot be checked.

Boulenger states that the third finger of the type is three times the
length of the second, this tallies with the proportions shown in his fig.
2, but in the enlargement of the hand, fig. 2a, it is three and two-third
times. The length of the type was 19 mm.

In the 22 mm. holotype of A. procterae Witte (1921 ) from Beni, 200
miles northwest of Idjwi, the third finger is only 2 times as long as the
second. This frog was said to differ from xenoehirus in that the tibio-
tarsal articulation reached the eye (instead of tympanum), the first
finger was as long as (instead of shorter than) the second. The latter
alleged difference, judging by the figures, is based on a difference of in-
terpretation ; the former is trifling and well within the range of variation
displayed by most members of the genus.

In the Idjwi frog the third finger is four times longer than the second,
but in the absence of comparative material I refrain from describing
an eastern race based on a single specimen which is the same length
as the type of .venochirus.

Size. Length 19 mm., length of third finger 9 mm.

loveridge: africax amphibians 429

Arthroleptis near poecilonotus Peters

Arthroleptis poecilonotus (Schlegel) Peters, 1863, Monatsb. Akad. Wiss. Berlin,
p. 446: Boutry, Ashanti, Gold Coast.

1 (M. C. Z. 25402) Mabira Forest, U. 10. xi. 38.

Remarks. The state of preservation leaves much to be desired and
a series is badly needed, I scarcely think that it is referable to poecilono-
tus which has never been recorded from Uganda.

Variation. Tips of digits slightly swollen or dilated (but no longer
so), first finger slightly shorter than second; a metatarsal, but no tarsal,
tubercle, small, rounded, shorter than inner toe, appears rather less
prominent than in poecilonotus; toes with a rudiment of web; tibio-
tarsal articulation reaches nostril.

Color. In alcohol. Above, grayish mottled with brown rather like
minutus from which it differs in lacking a second tubercle, etc.

Size. Length 20 mm.

Habitat. At base of banana plant in a plantation across the ravine
from the Mubango rubber-coffee factory.

Arthroleptis adolfifriederici Xieden

Arthroleptis adolfi-friederici Nieden, 1910, Sitz. Ges. naturf. Freunde Berlin,

p. 440: Rugege Forest, Belgian Ruanda-Urundi.
Arthroleptis affinis Ahl, 1939, Sitz. Ges. naturf. Freunde Berlin, p. 303, fig. 1:

Amani, Tanganyika Territory (Type 9 39 mm.).
Arthroleptis schoenebecki Ahl, 1939, Sitz. Ges. naturf. Freunde Berlin, p. 305,

fig. 2: Amani, Tanganyika Territory (Type c? 22 mm.).

3 (M. C. Z. 25400-1) Magrotto Mtn., T. T. 3. vii. 39.

Synonymy. Magrotto Mountain is but twenty miles distant from
Amani, Usambara Mountains, type locality of affinis and schoenebecki.
The Museum of Comparative Zoology possesses 190 Amani topotypes
which were referred by Barbour & Loveridge (192S, p. 212) to adol-
fifriederici. Though I have never examined the type of that species,
I see no reason to question the identification for the M. C. Z. possesses
an example (M. C. Z. 14696) from just north of Lake Kivu — and so
nearly topotypical — as well as others from intermediate localities in
the Poroto and Uluguru Mountains, as well as from Mt. Mbololo in
Kenya .

Ahl compares affinis with methneri Ahl from near Kilwa, a species

430 bulletin: museum of comparative zoology

I regard as synonymous with A. s. stenodactylus though comparison
may show it to be a synonym of A. s. lonnbergi seeing that the latter
has turned up on the Rondo Plateau to the south of Kilwa where it
occurs alongside Spelaeophryne methneri Ahl.

Variation. Tips of toes, and of some fingers, distinctly dilated;
metatarsal tubercle only moderately large, somewhat oval ; tibiotarsal
articulation reaches nostril or between eye and nostril.

Sir.e. Lengths 34-39 mm.

Enemies. One in stomach of a young house snake (Boaedon I.
lineatus) .

Habitat. One beneath a log in the forest.

Arthroleptis stenodactylus lonnbergi Nieden

Arthroleptis lonnbergi Nieden, 1915, Mitt. Zool. Mus. Berlin, 7, p. 361 : Mombo,
foot of Usambara Mountains, Tanganyika Territory.

Arthroleptis vagus Ahl, 1939, Sitz. Ges. naturf. Freunde Berlin, p. 306: Usam-
bara Mountains, Tanganyika Territory (cf 31 mm., $ 38 mm.).

Arthroleptis ukamiensis Ahl, 1939, Sitz. Ges. naturf. Freunde Berlin, p. 308,
fig. 3: Ukami, Tanganyika Territory ( 9 33 mm.).

1 (M. C. Z. 25399) Nchingidi, T. T. 12. v. 39.

Synonymy. Nchingidi, on the Rondo Plateau, lies close to the
Makonde Plateau which, together with Tendaguru and Amani, are
among the localities of Ahl's paratypes of vagus. It is fortunate that
Ahl definitely designated certain specimens as cf and 9 co types in
case the paratype from Chifumbazi, Mozambique, might prove refer-
able to the typical form. The Museum of Comparative Zoology pos-
sesses a good series of lonnbergi from Amani, therefore topotypes of
vagus.

Ukami, i.e. country of the Kami tribe, is practically synonymous
with the Uluguru Mountains where they dwell. When I first noticed
the difference between the two forms of stenodactylus, I submitted
an Uluguru specimen to Ahl for favour of comparison with the type of
lonnbergi which I thought it to represent. On account of Ahl's state-
ment that they differed, I (1932) described A. s. uluguruensis which
later (1936), on obtaining a topotype of lonnbergi from the original
series, referred to the synonymy of the latter. To that synonymy I
now add both vagus and ukamiensis.

Variation. Tips of fingers slightly swollen, of second, third, and

loveridge: African amphibians 431

fourth toes dilated; metatarsal tubercle large, shovel-shaped; tibio-
tarsal articulation reaches tympanum.

Size. Length 30 mm.

Enemies. One recovered from stomach of a green snake (Chlorophis
macrops) and three from white-lipped snakes (Crotaphopcltis h. hotam-
boeia) at Nchingidi.

Arthroleptis stenodactylus stenodactylus Pfeffer

Arthroleptis stenodactylus Pfeffer, 1893 (1892), Jahrb. Hamburg. Wiss. Anst.,
10, p. 93, pi. i, fig. 11: Kihengo, Tanganyika Territory.

2 (M. C. Z. 25383-4) Kitaya, T. T. 30. iii. 39.
21 (M. C. Z. 25385-9) Mikindani, T. T. 11. iv. 39.
2 (M. C. Z. 25390-1) Mbanja, T. T. 27. iv. 39.

2 (M. C. Z. 25392-3) Lindi, T. T. 1. vi. 39.

3 (M. C. Z. 25394-5) Magrotto Mtn., T. T. 1. vii. 39.
3 (M. C. Z. 25396-8) Likoni, K. C. 25. vii. 39.

Variation. Tips of digits not dilated; metatarsal tubercle, large,
shovel-shaped; tibio-tarsal articulation reaches, or just fails to reach,
tympanum except in 3 Mikindani, 2 Magrotto, and 1 Likoni frog
where it attains the eye. The Magrotto frogs are somewhat inter-
mediate between s. stenodactylus and s. lonnbergi, one would have ex-
pected them to be the latter.

Color. In life. Kitaya. Above, buff, so thickly speckled with red
as to appear rufous ; a large blotch beneath orbit and an interorbital
mark on crown; from end of snout a dark band passes upwards over
nostril, through orbit, above tympanum to flank where it breaks up
into spots; groin and thigh gray, the latter indistinctly barred; tibia
with four dark cross-bars. Below, pure white; soles of feet dusky
gray.

Mikindani. A most unusual variant is a 25 mm. frog (M. C. Z.
25385) which has a broad, black-edged, light, vertebral streak from
snout to anus.

Size. Length 18-38 mm., average 28 mm.

Breeding. Apparently not; the only specimen under 20 mm. in
length was taken on July 25.

Diet. Ants and beetles.

Enemies. In stomachs of many snakes, viz. Boaedon I. lineatus,
Chlorophis neglectus (2), Crotaphopeltis h. hotamboeia, and Psammophis
s. sudanensis (2).

432 bulletin: museum of comparative zoology

Habitat. Under rubbish at Kitaya; beneath rotting palm trunks
and piles of coconut husks, also in swamped grasslands under palms,
at Mikindani ; under piles of grass at both Kitaya and Lindi ; beneath
vicuti at Likoni.

Cacosternum boettgeri boettgeri (Boulenger)

Arthroleptis boettgeri Boulenger, 1882, Cat. Batr. Sal. Brit. Mus., p. 118, pi. xi,
fig. 8: Vleis, Kaffraria, Bechuanaland.

7 (M. C. Z. 25478-9) S. Kinangop Plateau, K. C. 29. x. 38.

Remarks. It is with some hesitation that trinomials are employed
for the status of C. b. albirentris Hewitt, 1926, is not too clear. A pair
(M. C. Z. 15807, 22260) received from Power, the first to recognise it,
seem scarcely distinguishable. The only other writer to consider the
form as valid is Hoffman, 1940, Avho obtained a single example at
Broedershoek, near Greytown, Natal.

Variation. Tympanum concealed; tips of digits not dilated; a dis-
tinct inner, and a scarcely distinguishable outer, metatarsal tubercle;
no tarsal tubercle; tibio-tarsal articulation reaches shoulder.

Size. Length J* tf 18-20 mm., 9 22 mm.

Breeding. These males, standing vertically in rain-formed pools
with their forelimbs resting on grass blades beneath overhanging tus-
socks, were calling loudly, but difficult to locate on account of the con-
cealed positions which they had selected.

Hemisus marmoratum marmoratum (Peters)

Engystoma marmoratum Peters, 1854, Monatsb. Akad. Wiss. Berlin, p. 628:
Cabaceira, Mozambique.

4 (M. C. Z. 25480-2) Kitaya, T. T. 30. iii. 39.

9 (M. C. Z. 25483-5) Mikindani, T. T. 15. iv. 39.

5 (M. C. Z. 25486-7) Lindi, T. T. 1. vi. 39.

9 (M. C. Z. 25488-9) Siga Caves, T. T. 15. vi. 39.

Native names. Kisianene (Kiyao); chihenene (Kimakonde); chenene
(Kimwika ). These names being applied also to Breviceps mossambicus.

Color. In life. Kitaya. 9 . Above, gray, or greenish gray, vermicu-
lated with black; sides paler. Below, white, uniform.

Size. Length 22-33 mm., those of 22-23 mm. still bearing stumpy
rudiments of tails.

loveridge: African amphibians 433

Breeding. On June 15 only, were young examples found, the entire
Siga series ranging from 22-26 mm.

Enemies. Two tailed young recovered from the stomach of an egret
(Egretta g. dimorpha) at Siga, an adult from a house snake (Boaedon
I. lineatus) at Mikindani.

Habitat. Three beneath thatching of a collapsed hut at Kitaya;
under rubbish or in damp sand beneath mango trees at Mikindani;
beneath bundles of grass at Lindi; under palm fronds at Siga Caves'.
When uncovered these plump little frogs squat down, then run fast
rather like a mouse.

BREVICIPITIDAE

Calltjlina kreffti Nieden

Callulina kreffti Nieden, 1910, Sitz. Ges. naturf. Freunde Berlin, 10, p. 449:
Amani, Usambara Mountains, and Tanga, Tanganyika Territory.

13 (M. C. Z. 25490-4) Magrotto Mtn., T. T. 6. vii. 39.

Native name. Kikorowe (Kisambara, but applied to Probrevieeps
also).

Color. In life. Above, pale brown mottled with sepia in adults, the
young are often particolored the head and posterior half of back being
dark brown edged with black, the anterior half of back from shoulder
to about midbody, and a spot on the lumbar region, creamy white.
Below, adults white, or pale chrome, minutely flecked with chinese
white; young dark gray, almost black, minutely flecked with chinese
white. The pupil of a toad with a bright pale chrome throat was hori-
zontal, black with bright orange iris, in all the others it was pale
bronze.

Size. Length 22-45 mm., average 32 mm., for only four were really
young 22-24 mm. at this time.

Defence. On being removed from their retreats these toads im-
mediately inflate tremendously and, rising stiffly on their ridiculously
short little legs, present a very Breviccps-\ike appearance. Simul-
taneously their pores exude an extremely sticky substance which was
very difficult to remove from one's fingers except with blue-wattle
soap which has a high soda content.

To ascertain whether this exudation possessed the same poisonous
properties as that of Phrynomerus bifasciatus, I purposely rubbed my
sticky fingers together. Apart from a slight tingling or pulsating sen-
sation, however, no serious ill-effects were noticeable.

434 bulletin: museum of comparative zoology

Habitat. The entire series were taken at Kitulwe, nine resulted from
the examination of fifty wild bananas which were systematically
searched. It was observed that the toads were almost all in the basal
portion of the outermost leaf-stalks, which are so often in a semi-
withered condition.

Spelaeophryne methneri Ahl

Spelaeophryne methneri Ahl, 1924, Zool. Anz., 61, p. 99: Xangoma Cave, Ma-
tumbi near Kilwa, Tanganyika Territory.

19 (M. C. Z. 25495-500) Nchingidi, T. T. 11-19. v. 39.

Range. The finding of this rare and interesting species on the Rondo
Plateau, which lies just a hundred miles south of the type locality, is
principally of interest as furnishing further proof of the similarity of
the plateau forest fauna with that of the Uluguru Mountains, the only
other place from which methneri has been recorded.

Sie. Length 25-50 mm., average 37.9 mm.

Habitat. The entire series were collected by my boys and me from
beneath logs in the forest or along the forest-edge, the ground being
kept moist by frequent downpours. In some cases colonies of termites
were located under the same logs as those sheltering these black and
scarlet toads.

Probreviceps macrodactylus macrodactylus (Nieden)

Breviceps macrodactylus Nieden, 1926, Das Tierreich, 49, Anura, 2, p. 6:
Usambara Mountains, Tanganyika Territory.

4 (M. C. Z. 25501-4) Magrotto Mtn., T. T. 28. vi. 39.

Native name. Kikorowe (Kisambara, but applied also to Callulina).
Range. Magrotto is but twenty miles from the type locality.
Si::e. Length c? cT 40-40 mm., 9 58 mm.

Breviceps mossambicus Peters

Breviceps mossambicus Peters, 1854, Monatsb. Akad. Wiss. Berlin, p. 628:
Mozambique Island, and Sena, Mozambique.

9 (M. C. Z. 25505-9) Mikindani, T. T. 18. iv. 39.
7 (M. C. Z. 25510-4) Nchingidi, T. T. 11. v. 39.

loyeridge: African amphibians 435

Native names. Neither the Makonde nor Mawiha distinguish this
species from Hemisus m. mormoratum , which see.

Size. Length 20-47 mm., average 29 mm.

Enemies. On two occasions recovered from the stomachs of bird
snakes (Thelotornis k. capensis).

Habitat. The Mikindani series were obtained by turning over piles
of rubbish, often gathered by rainstorms, in the red roadside ditches.

Hoplophryne rogersi Barbour & Loveridge

HopJophryne rogersi Barbour & Loveridge, 1928, Mem. Mus. Comp. Zool.,
50, p. 258, pi. ii, fig. 5: Mount Bomoli, near Amani, Usambara Mountains,
Tanganyika Territory.

5 (M. C. Z. 25515-9) Magrotto Mtn., T. T. 3-6. vii. 39.

Range. Heretofore known only from the type locality which is about
twenty miles from Magrotto.

Color. In life. Pupil black, roundish.

Size. Length cfcf 25-27 mm., 9 9 22-25 mm.

Habitat. The first pair, the male of which had lost a foot, wTere found
beneath logs beside a sawpit in a section of forest where there were
neither wild bananas nor bamboos in which they could breed! Two
days later I searched through the bamboos growing beside the stream
which one crosses on entering Magrotto Plantation from Muheza, but
found only one pair and these beneath a rotted log lying among the
bamboos. The only wild bananas said to be surviving on the mountain
are those at Kitulwe which I visited on July 6, a search of fifty wild
bananas, however, resulted only in the capture of a single male!

PHRYNOMERIDAE

Phrynomerus bifasciatus (Smith)

Brachymerus bifasciatus A. Smith, 1849, 111. Zool. S. Africa, Rept., pi. lxiii:
"Country to the east and northeast of Cape Colony.-'

35 (M. C. Z. 25520-5) Lindi, T. T. 31. v. 39.

Size. Length 25-49 mm., average 31.8 mm.

Habitat. The majority of these half grown toads were taken beneath
the thatching of some collapsed huts on the edge of a swampy area in
process of dessicating.

436 bulletin: museum of comparative zoology

BIBLIOGRAPHY
Ahl, Ernst

1931c. "Anura III," in Das Tierreich, No. 55, pp. i-xvi + 1-462, figs.
1-320. (This actually has a month's priority over the article in
Mitt. Zool. Mus. Berlin, 17, pp. 1-132).

Barbour, Thomas & Loveridge, Arthur

1928c. "A Comparative Study of the Herpetological Fauna of the Uluguru
and Usambara Mountains, Tanganyika Territory, with Descrip-
tions of new Species. Mem. Mus. Comp. Zool., 60, pp. 87-265,
pis. i-iv.

BOULENGER, G. A.

1882b. "Catalogue of the Batrachia Salientia s. Ecaudata in the Collection
of the British Museum." Ed. 2, London, pp. i-xvi + 1-305, figs.,
pis. i-xxx.

Laurent, R.

1941b. "Contribution a l'Osteologie et a la Systematique des Ranides

africaines. IV-VTI." Revue Zool. Bot. Africaine, 34, pp. 192-230,

figs. 1-7, pis. vi-vii.
1941c. "Les Megalixalus (Batraciens) du Musee du Congo." Revue

Zool. Bot. Africaine, 35, pp. 119-132.

Loveridge, Arthur

1933h. "Reports on the Scientific Results of an Expedition to the South-
western Highlands of Tanganyika Territory. VII. Herpetology."
Bull. Mus. Comp. Zool., 74, pp. 197-416, pis. i-iii.

1936h. "African Reptiles and Amphibians in Field Museum of Natural
History." Field Mus. Nat. Hist. Zool. Series 22, pp. 1-111.

1936k. "Scientific Results of an Expedition to Rain Forest Regions in
Eastern Africa. VII. Amphibians." Bull. Mus. Comp. Zool., 79,
pp. 369-430, pis. i-iii.

1941h. "South African Frogs of the Genus Hyperolius in the Museum of
Comparative Zoology, Cambridge, Massachusetts." Ann. Trans-
vaal Mus., 20, pp. 283-291.

Proctor, J. B.

1920. "On a Collection of Tailless Batrachians from East Africa made
by Mr. A. Loveridge in the years 1914-1919." Proc. Zool. Soc.
London, pp. 411-420, figs. 1-4.

EXPLANATION OF PLATES

Lovebidge — African Arrphibians

PLATE 1

Map showing Principal Collecting Localities

1938

Landing at Mombasa (25.x), except for a stopover at Naivasha and Kinan-
gop (26-31.x), Loveridge proceeded by rail direct to Jinja (l-5.xi). Thence to
Mabira Forest (5-21 .xi), Budongo Forest (22.xi-7.xii), Kibale Forest (8-19.xii),
Bundibugyo near Bwamba Forest (19-26.xii), Bugoye, foot of Ruwenzori
Mountains (26-28.xii) and Mubuku Valley at 7000 ft. (29.xii-).

1939

On leaving Mubuku (-9.i) Loveridge descended down the valley to Mihunga,
circa 6000 ft. (9-21.1) then back to Bugoye (21-24.i), Nyakabande (25-30.i),
Mushongero (30.i-4.ii), returned to Nyakabande (4-8.ii), Kisenyi (8-13.ii),
Goma (13-14.ii), Mamvu on Idjwi Island (14-16.ii), Upper Mulinga River on
Idjwi (16.ii-6.iii), Uvira (7-8.iii), Ujiji (9-16.iii), Dar es Salaam (18-19.iii),
Mikindani (22-24.iii), Kitaya (24.iii-7.iv), Mikindani (7-24.iv), Mbanja
(25.iv-6.v), LakeRutamba (6-8.v), Nchingidi (9-21.v), Lindi (22.v-4.vi),Siga
Caves (7-17.vi), Amboni Estate (17-27.vi), Magrotto Mountain (27.vi-21.vii),
Tanga (21-23.vii), Likoni opposite Kilindini (24-26.vii).

BULL. MUS. COMP. ZOOL.

Loverioge. African Amphibians. Plate 1

MubuKu Valley §■, Jr '$

W^Kibalt r°

UGANDA,.-'

rarest" [&g

"5<^03* Mabira
Jo? %5b Forest?

Qum

o-TM-J-

Lake KivJpdSt
CostcrmansfM' '
Uvira.

BELGIAN

CONGO

TAN GANY-UA
TER RI TORY

-v ( 9

*

A

*

V

9~

$

<

Ju\

-~,

t h'

co

)5

UJ

; p

J o

q

\ M

o

.> o

*

)

<fr

/ %

Scale in Miles

0 100 200

U I I I I 1

PLATE 2

Loveridge — African Amphibians

PLATE 2
Age Dichromatism in Frogs of the Genus Leptopelis

Fig. 1. L. notatus christyi (Boulenger) young

One (M. C. Z. 25118) of several young frogs which I found on a forest trail —
apparently dislodged by heavy rain from the canopy — or seated on the leaves
of freshly-felled trees in the Kibale Forest.

Fig. 2. L. notatus christyi (Boulenger) adult

This topotypic 9 (M. C. Z. 25114), gravid on November 7, was found be-
neath loose bark on the trunk of a tree growing in the garden of the late Major
C. C. Christy's house at Mubango, in the vast Mabira Forest, Chagwe.

Fig. 3. L. johnstoni (Boulenger) adult

A 68 mm. gravid 9 (M. C. Z. 25125) was one of a series of these tree frogs
taken on Magrotto Mountain, July 7. The males are sometimes green, and
many half-grown green specimens were taken as they squatted on broad-leaved
plants in a forest clearing.

Fig. 4. L. johnstoni (Boulenger) young
I netted this froglet (M. C. Z. 25135), measuring 18 + 25 mm., in the
swamp immediately below my camp on Magrotto Mountain. En route to the
swamp I had brushed against a luxurious growth of ferns projecting from the
stem of an oil palm, and shortly afterwards discovered a 48 mm. johnstoni
squatting on my sleeve!

Fig. 5. L. concolor Ahl. adult
The frog (M. C. Z. 20563), one of a series taken at Malindi on the coast of
Kenya during the course of a former expedition, was selected because a de-
tailed description of its color in life was available. Malindi lies just south of
the type locality which I visited during the breeding season when the males,
their vocal sacs inflated to great white bubbles, were calling on every side.

Fig. 6. L. concolor Ahl. young

This frog (M. C. Z. 25124) was one of three recovered from the stomach of
a spotted wood snake in the vicinity of Siga Caves. Two of the frogs were too
macerated to be worth preserving, and the preservation of the figured specimen
leaves much to be desired. Its morphological characters appear to be those of
concolor, but its coloration raises doubts as to whether it may not be a young
johnstoni, for slightly smaller concolor, taken at Mikindani, are uniformly
green.

BULL. MUS. COMP. ZOOL.

Loveridge. African Amphibians. Plate 2

2

\

'%

X

PLATE 3

Loveridge — African Amphibians

PLATE 3

Sexual Dichromatism in Frogs of the Genus Hyperolius

Fig. 1. H. ahli Loveridge. cf Type
While the type was green, other males varied from a paler green through
straw to flesh color. Apparently these frogs remain under water during the day,
for only one was seen during half-an-hour spent wading among the lily pads.
Continuing over the same area immediately after sunset, however, sent more
than a dozen leaping away.

Fig. 2. H. ahli Loveridge. 9 Paratype

Though the spots of the figured female were pure white, those of others were
pale flesh color and, at Kitaya, were pale red.

Fig. 3. H. p. parkeri Loveridge. cT Paratype
The type series of this species came from near Bagamoyo where their ring-
ing calls, like a clear "pop-pop", came from a swamp bordering the Ngerengere
Road. The sedges, on which they were found squatting, were so sharp that it
was next to impossible for my bare-legged assistants to wade among them.

Fig. 4. H. p. parkeri Loveridge. 9 Type
The females deposited their white eggs, ranging from 69 to 110 in number,
on these sedges at a point just above the water level so that, provided the
rains continued as heavy as they had begun, the spawn would be submerged
within the course of a day or two.

Fig. 5. H. p. rovumae Loveridge. d" Paratype
The males of this slightly larger southern form differ also in bearing black
breeding spinosities on the belly as well as on the thighs and feet. In H. p.
parkeri, on the other hand, such spinosities are confined to the feet.

Fig. 6. H. p. rovumae Loveridge. 9 Type
At Kitaya these females were assembling to spawn in early April. Apart
from color differences, such females were readily recognizable from the ova
being clearly visible through the delicate and transparent skin of the abdomen

Fig. 7. H. milnei Loveridge. c? Paratype
Even by day these diminutive frogs may be found squatting on grass and
lily pads, at night their numbers are augmented by others coming out of the
water. Be it day or night they are constantly on the alert, leaping, diving, or
swimming away before one can come within netting distance of them.

Fig. 8. H. milnei Loveridge. 9 Type = H. pusillus (Cope)
Despite the type localities of milnei (Kenya Colony) and pusillus (Natal)
being nearly 2500 miles apart, I believe that they should be regarded as one
species. It is true that milnei is usually more spotted, but there is much
variability and if a northern form is to be recognized then microps would have
precedence over milnei.

BULL. MUS. COMP. ZOOL

Loveridge. African Amphibians. Plate 3

PLATE 4

Loveridge — African Amphibians

PLATE 4

f

African Frogs of arboreal, aquatic, and terrestrial habits.

Fig. 1. Leptopelis notatus chrislyi (Boulenger)

A green, gravid, 46 mm. female found squatting on a green sedge in Mihunga
Swamp, Ruwenzori Mountains. Its color, in marked contrast to that of a
brown, gravid, 48 mm. female found clambering over fallen sedges that had
been cut down earlier in the day.

Fig. 2. Kassina senegalensis (Dumeril & Bibron)

A gravid, 44 mm. female found hiding beneath a log on Idjwi Island, Lake
Kivu. Of the many examples of this species, which should be arboreal, that I
have encountered, not one was found off the ground. When the rains break
they emerge from their places of concealment and assemble at the pools, at
such a time their ringing calls, like the sound of bursting bubbles of water,
may be heard a mile away.

Fig. 3. Rana albolabris albolabris Hallowell

In contrast to the subject of Fig. 2, the white-lipped frog, though belonging
to a genus whose members are terrestrial, fossorial, or aquatic, has developed
slight digital expansions and taken to living in trees or, as on Idjwi Island, in
bananas.

Fig. 4. Rana occipitalis Gunther

In habits this five-inch bullfrog is probably the most aquatic African mem-
ber of the genus. Like R. edulis it is also cannabalistic for, in addition to a
large Dytiscid beetle, the stomach of one of these Ujiji specimens held the
bones of a small frog. Apparently, however, they do not indulge in scorpions,
centipedes, biting ants, and other poisonous creatures which make a meal in-
teresting for edulis.

BULL. MUS. COMP. ZOOL.

Loveridge. African Amphibians. Plate 4

>s.

"»

20G

71 270S

C 70 1 «-

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE
Vol. XCI, No. 6

THE AMERICAN CAECILIAXS

By Emmett Reid Dunn

CAMBRIDGE, MASS., U.S.A.

PRINTED FOR THE MUSEUM

December, 1942

PUBLICATIONS
OF THE

MUSEUM OF COMPARATIVE ZOOLOGY
AT HARVARD COLLEGE

The Bulletin and Memoirs are devoted to the publication of
investigations by the Staff of the Museum or of reports by spec-
ialists upon the Museum collections or explorations.

Of the Bulletin, Vols. I to XC, Vol. XCI, Nos. 1, 2, 3, 4 and
5, and Vol. XCII, No. 1 have appeared and of the Memoirs,
Vol. I to LVI.

These publications are issued in numbers at irregular intervals.
Each number of the Bulletin and of the Memoirs is sold separately.
A price list of the publications of the Museum will be sent upon
application to the Director of the Museum of Comparative
Zoology, Cambridge, Massachusetts.

After 1941 no more Memoirs are to be published.

Bulletin of the Museum of Comparative Zoology
AT HARVARD COLLEGE
Vol. XCI, No. 6

THE AMERICAN CAECILIAXS

By Emmett Reid Dunn

CAMBRIDGE, MASS., U.S.A.

PRINTED FOR THE MUSEUM

December, 1942

No. 6. — The American Caecilians1
By Emmett Reid Dunn*

An interest in American Caecilians, begun in 1920 when I found a
few specimens of Gymnopis in Costa Rica, was enhanced when I took
a specimen of a new species in wes'tern Panama in 1923. In 1928 I
attempted, rather unsuccessfully, to list the North American forms.
In Europe in 1929, as a holder of a John Simon Guggenheim Memorial
Fellowship, I took the opportunity to examine the American Cae-
cilians in the collection of the British Museum of Natural History
and in the principal museums of the continent. Since my return I
have examined practically all the material in the United States, in
Panama, and in Costa Rica, and have been sent extensive collections
by the Institute La Salle in Bogota, by the Museu Paulista in
Sao Paulo, and by the Museu Nacional in Rio de Janeiro.

Systematic treatment of American Caecilians since 1895 has been
based almost entirely upon the work of Boulenger and upon the col-
lections of the British Museum. This institution contained, in 1929,
103 American Caecilians (28 species, 6 genera, and the types of 15
described forms). ^Tiile it is the best single collection, it is far from
complete.

My present treatment is based on the examination of 850 American
Caecilians (44 species, 6 genera, the types of 39 described forms, and
the types of nine forms thought to be new). I have not been able to
examine the types of 14 described forms. I consider one of these valid
and can place it in its genus. I suspect that another may be valid
but as I cannot place it in any known genus it must remain incerta
sedis. I therefore recognize 6 genera, and 44 species, of which I have
seen specimens of all but one species.

genus

specimens

types seen

types not seen

species

Rhinatrema

19

3 + 2 new

1

6

Gymnopis

157

11

2

11

Siphonops

253

4

5

5

Caecilia

324

15+7 new

2

16

Chthonerpeton

39

2

1

3

Typhi onectes

58

4

2

3

850

39 + 9 new

13

44

The type of Siphonops syntremus Cope is another I have not been
able to examine, but I cannot place it in any known American genus
or species.

1 Contributions from the Biology Department, Haverford College, No. 8.

440 bulletin: museum of comparative zoology

Distribution

American Caecilians range from latitude 20 north (Vera Cruz and
Guerrero in Mexico) to latitude 35 south (Buenos Aires, Argentina)
on the Atlantic side, and to latitude 3 south (Guayaquil, Ecuador)
on the Pacific side. They range from sea level to 4500 feet (Cartago,
Costa Rica) and to 6200 feet (Milligalli, Ecuador).

They occur on the following islands : Saboga and San Miguel in the
Gulf of Panama; Gorgona off the Pacific coast of Colombia; Trinidad;
Victoria and Sao Sebastiao off the coast of Sao Paulo, Brazil.

Mexico to Costa Rica inclusive have only the genus Gymnopis.
Bolivia and Paraguay have only Siphonops. Argentina and Uruguay
have only Chthonerpeton. Panama has Gymnopis and Caecilia.
Colombia has 5 genera and 18 species; Ecuador, 4 genera and 11
species; Peru, 4 genera and S species; the Guianas, 5 genera and 8
species; Brazil, 4 genera and 13 species.

Gymnopis and Siphonops form a pair of allied genera, the former
northern, the latter southern. Typhlonectes and Chthonerpeton form
another such pair of genera, the former northern, the latter southern.

Rhinatrema and Caecilia occupy the center of the group range,
northwestern South America, which is also the area of greatest
abundance of genera and species.

There would seem to be a minor center of development in the
south (Chthonerpeton and Siphonops) and perhaps another in Cen-
tral America (Gymnopis).

It may be inferred from the distribution that Caecilians have in-
habited South America since pre Tertiary times, and that they have
only entered North America since the midTertiary. Only two genera
reach Panama, only one reaches Costa Rica, and the northern limit
is 15 degrees of latitude short of the southern limit.

Generic assignments and affinities

A primitive Caecilian should, theoretically, have the following
characteristics:

1. A definite tail.

2. Secondaries all complete and equal in number to the primaries.

3. Two complete rings of scales to each segment, one for the
primary and one for the secondary.

4. Inner mandibular tooth row well developed.

5. Teeth of any given row uniform in size.

DUNN: THE AMERICAN CAECILIANS 441

6. Tentacular aperture close to eye.

7. Eye well developed and in an open orbit.

8. Body approximately cylindrical, short and fairly stout, without
dorsal fin.

9. Anus not surrounded by a sucking disk.

10. Oviparous.

11. Aquatic larvae, gill slit open.

12. Jaw muscles not roofed by bony contact between parietal and
squamosal.

13. Skull with more rather than fewer separate bones.

The specimens here assigned to the genus Rhinatrema agree in all
respects with the above criteria. Specimens assigned to other genera
differ more or less, and are presumably less primitive.

Specimens assigned to the genus Gymnopis have no tail; the secon-
daries are less in number than the primaries and are not all complete;
scales are absent anteriorly; the inner mandibular tooth row is poorly
developed or absent; the tentacular aperture may be some distance
anterior to the eye; the eye is, in some species, invisible, and the orbit
is sometimes roofed over by bone; they are viviparous and have no
aquatic larval stage.

Specimens assigned to the genus Siphonops agree on the whole
with Gymnopis but lack secondaries and scales completely; the ani-
mals are oviparous; but there is not known to be an aquatic larval
stage. These are all the differences I can find between such species
as Gymnopis mexicanus and Siphonops annulatus. A common ances-
tor for these two genera may be inferred to have existed, with the
secondaries and scales of Gymnopis and the breeding habits of Sipho-
nops, and thus closer to Rhinatrema than either of the two.

The species assigned to Chthonerpeton have no tail; they lack
secondaries and scales entirely; the tentacular aperture is always
some distance anterior to the eye and may be just behind the nostril;
the anus is surrounded by a sucking disk; the animals are viviparous
and the embryos have a single pair of large allantoic gills; it may be
inferred that an aquatic larval stage is absent.

The species assigned to Typhlonectes agree on the whole with
Chthonerpeton, but the tentacular aperture is always just behind
the nostril; the body is flattened laterally, with a dorsal fin. These
are all the differences I can find between such species as Chthonerpeton
indist Inchon and Typhlonectes compressicauda. Chthonerpeton may be
inferred to be ancestral to Typhlonectes.

442 bulletin: museum of comparative zoology

These two genera agree with Rhinatrema in having a well developed
inner mandibular tooth row.

The species assigned to Caecilia have no tail; the secondaries are
reduced in number and sometimes entirely absent; scalation is reduced
or entirely absent; the inner mandibular tooth row is reduced or
absent; the tentacular aperture is remote from the eye, being under
the tip of the snout, below and somewhat posterior to the nostril;
the eye may be invisible and the orbit roofed by bone; the body may
be excessively attenuated; the animals may be inferred to be ovi-
parous, with an aquatic larval stage.

There are thus the following groups of genera in America: Rhina-
trema; Gymnopis and its ally Siphonops; Chthonerpeton and its
derivative Typhlonectes; Caecilia. Of these four groups, Rhina-
trema occupies an isolated and a primitive position. The other three
exhibit characters which preclude any linear arrangements of them.
It is not impossible that each has been derived independently from a
more primitive common ancestor. There is nothing known to prevent
this common ancestor from having the characters of Rhinatrema.

The species here assigned to Caecilia have been listed as three
genera; Amphiumophis, Herpele, and Caecilia. The unique type
specimen of Amphiumophis is a Caecilia tentaculata. The only differ-
entiating character given for the genus was the absence of the inner
mandibular tooth row, which is poorly developed in some Caecilia.
The roofed orbit and invisible eye of C. ochrocephala and C. polyzona
have caused their reference to Herpele, but the eye is frequently
invisible in other species of Caecilia, and ochrocephala and polyzona
are so similar to the other forms of Caecilia that I cannot but regard
them as congeneric.

The species here assigned to Gymnopis are usually listed as two
genera; Gymnopis and Dermophis. The only difference given is the
roofed orbit and invisible eye of Gymnopis. The variability and un-
certainty of this condition in Gymnopis multiplicata oaxacae and in
Gymnopis nicefori make a generic division of the species impractical.

I gather from the literature that four genera and sLx species occur
in Southeast Asia; one genus with six species in the Seychelles Islands;
six genera and 17 species in tropical Africa.

The degree of affinity between Rhinatrema and the genera Ich-
thyophis and Uraeotyphlus of southeastern Asia remains to be deter-
mined. Statements in literature would seem to indicate a fairly close
relationship.

Parker (1941, Ann. Mag. Nat. Hist. (11), 7 pp. 1-17), has shown

DUNN: THE AMERICAN CAECILIANS 443

that African and Seychelles Islands forms, formerly referred to Der-
mophis [ = Gymnopis of this paper] are not congeneric with American
species.

The African Herpele squahstoma, the type of Herpele, is not con-
generic with any American form, although two have been referred to
Herpele from time to time. The American forms in question are*
Caecilia. Whether or not the Indian "Herpele" fulleri is congeneric
with either remains to be determined.

As matters stand it is not safe to consider that any genus of Ameri-
can Caecilians has representatives in the Old World, or, indeed, that
any genus of Caecilians occurs in more than one of the four areas
(southeast Asia, Seychelles Islands, .African tropics, American tropics)
inhabited by these animals.

The eye

Normally and primitively the eye is in an open orbit and visible
through the skin. At the opposite extreme the orbit may be closed
over by bone, and the eye may be invisible. In some forms the orbit
may be open but the eye may be concealed by the thickness or the
opacity of the skin. It is also possible that the eye may remain visible
externally even after the orbit is roofed by bone. In many forms,
known only from a few rare or unique specimens which it is not
possible to dissect, the exact condition of the eye is not yet known.
It is therefore often impossible to say more than that the eye is or is
not visible externally.

It is so visible in all Rhinatrema, Chthonerpeton, and Typhlonectes,
and the orbit is not known to be roofed over in any of these.

In Siphonops the orbit is not known to be roofed over in any form,
but the eye is invisible externally in half the 5. insulanus seen. Of
21 >S. brasiliensis seen the eye is very indistinct in one and invisible in
four.

In Gymnopis the eye is invisible externally in all unicolor, oligozona
and multiplicata multiplicata seen. The orbit is known to be roofed
by bone in some specimens of unicolor and multiplicata multiplicata.
In nicefori the eye is invisible in 4 specimens out of 6. In one of these
four the orbit is not roofed by bone. In multiplicata proximo the eye
is visible externally in a single specimen (of 38 examined), and in this
one the orbit is open. The eye is visible in 13 out of 15 multiplicata
oaxacae, but the condition of the orbit is not known. In other forms
of Gymnopis the eye is always visible and the orbit is not known to
be roofed over.

444 bulletin: museum of comparative zoology

In Caecilia the eye is invisible externally in all known specimens of
ochrocephala, polyzona, and elongata. The orbit is known to be roofed
over in some ochrocephala. In the following species the eye is occa-
sionally invisible externally; gracilis, one of 31; dunni, one of 19;
thovipsoni, one of 9; tentaculata, three of 26; bassleri, three of twelve.
The orbit was open in the specimen of gracilis.

The eye is always visible in the other forms of Caecilia, and the
orbit is not known to be roofed over in any of them.

Cranial characters

I have examined specimens of Rhinatrema bi-color, Gymnopis mexi-
canus me.vicanus (2), Gymnopis unicolor, Siphonops annulatus, Siph-
onops brasilicnsis, Caecilia ochrocephala, Chthonerpeton indistinctum,
Typhlonectes comprcssicauda natans, Typhlonectes kaupii.

The cranial characters confirm the position of Rhinatrema as primi-
tive; the alliance between Gymnopis and Siphonops; the alliance
between Chthonerpeton and Typhlonectes. .

Rhinatrema bicolor has the premaxillae separate from the nasals.
In the other genera the premaxilla and nasal are fused. Rhinatrema
bicolor has a large flat bone posterior to the combined maxilla-palatine.
What is obviously the same bone (but smaller) can be found in
Gymnopis and in Siphonops. Xo such bone exists in Caecilia, Chthon-
erpeton, or Typhlonectes. This bone has the relationships of an ectop-
terygoid more than that of a pterygoid. In the literature it has gone by
both names. I think that some Caecilians have an ectopterygoid, thus
differing from all other living Amphibians, and that no Caecilians
have a pterygoid. There has been much confusion in literature, be-
cause a forward extension of the quadrate (coossified in cartilage)
has been called a "pterygoid bone" by many investigators.

In Rhinatrema, in Gymnopis, and in Siphonops the internal naris
is enclosed by the maxilla-palatine. In Caecilia, Chthonerpeton, and
Typhlonectes the internal naris is enclosed on the outer side by the
maxilla-palatine and on the inner by the prevomer.

The frontals are in contact in Rhinatrema, in Gymnopis, in Chthon-
erpeton, and in Typhlonectes. They are separated by the "ethmoid"
in Siphonops and in Caecilia. The former condition would seem primi-
tive.

In Rhinatrema, in Chthonerpeton, and in Typhlonectes there is a
wide gap between the squamosal and the parietal, and the temporal
muscles are not covered by bone. In Gymnopis, Siphonops, and

DUNN: THE AMERICAN* CAECILIANS 445

( aecilia squamosal and parietal are in contact, and the temporal
muscles are roofed by bone. The former condition would appear to be
primitive.

The three genera with a gap between squamosal and parietal have
markedly "kinetic" skulls, with considerable movement between the
'■maxillary segment" and the "occipital segment." They are "moni-
mostylic" as the quadrate is firmly attached to the squamosal. The
three genera without a gap between squamosal and parietal have
much less movement between the segments of the skull, and are less
"kinetic" but are just as much "monimostylic." The former condi-
tion would appear to be primitive.

On these characters, Rhinatrema is alone. Its skull characters, as
well as its other characters, seem to me to be primitive.

Gynmopis differs in skull characters from Siphonops only in having
the frontals in contact, in which trait as in its other characters it seems
to me to be more primitive.

( 'hthonerpeton and Typhlonectes agree in all significant cranial
characters.

Caecilia stands alone, and is the most specialized of the genera in
cranial characters.

The cranial characters of American Caecilians align them in rela-
tion to each other in the same way and the same order as do their other
characters.

While I am quite aware of previous remarks on the cranial char-
acters of American Caecilians, and aware that the above remarks dis-
agree with some of them, I offer no apologies. The statements given
above result from examination of all the American genera at the same
time, and consequent comparison of one with another. All the state-
ments are from my own observations and none are from any other
sources.

The tentacle

Statements in the literature give the impression that the tentacle of
American Caecilians is present in two quite different conditions: a
valvular or flap-shaped tentacle, in a horseshoe-shaped groove or
aperture, attached posteriorly to the skin of the head; a globular
tentacle in a circular aperture or groove. This is erroneous, as all
American Caecilians have a quite similar tentacle and aperture, all of
the first type. In American Caecilians the second type is an occasional
consequence of unusual retraction of the organ, and careful observa-
tion will disclose the posterior attachment. This occurs more often in

446 bulletin: museum of comparative zoology

specimens of Gymnopis. The two appearances may be present on
opposite sides of the same individual. The tentacular aperture is the
posterior end of the naso-lachrymal duct.

The anatomical base of the tentacle is, in all forms, the anterior
border of the eye socket, and this is also the place of origin of the
organ embryologieally. It may therefore be inferred that the original
position of the aperture was on the side of the head, just anterior to
the eye. This is the position in all Rhinatrema and in most forms of
Gymnopis and Siphonops. In the races of G. mexicanus, in G. albiceps
and in G. parviccps the aperture is further forward, but nearer the eye
than the nostril. In a single specimen of G. m. mexicanus (of 66 ex-
amined) the aperture is exactly equidistant between nostril and eye.
In 8 specimens of Siphonops annulatus (of 175 examined) the aperture
is further forward, in one nearer the nostril than the eye.

In Chthonerpeton the aperture is, in viviparum, slightly nearer the
eye than the nostril; in indistinctum it is slightly nearer the nostril
than the eye; in peter si and in all forms of Typhlonectes it is directly
behind the nostril.

In all forms of Caecilia the aperture is on the under side of the
snout, below and slightly posterior to the nostril.

The vent

The vent is an unmodified opening except in Chthonerpeton and in
Typhlonectes, where the area surrounding it becomes modified into a
sucking disk. Every stage in this transition may be seen in the three
species of Chthonerpeton. The disk is slightly developed in C. vivi-
parum, intermediate in C. peter si, and large in C. indistinctum and in
all Typhlonectes.

Sex

American Caecilians have no external signs by which they may be
sexed. Males have a median intromittent organ, which is occasionally
extruded, perhaps during the death throes. Pregnant females of vivi-
parous species are quite stout, and may have the hinder portion of the
body enlarged. It is usually necessary to dissect in order to determine
the sex. No variation in number of segments, of secondaries, or of
scale rings has so far been found correlated with sex.

DUNN: THE AMERICAN CAECILIANS 44^

Annular grooves

In all American Caecilians the muscle segmentation is marked ex-
ternally by grooves, the "primaries." These correspond in position to
the ends of ribs and therefore to vertebrae. A count of them gives the
number of vertebrae. They are precisely identical to the "costal
grooves" of salamanders. They may extend completely around the
body, but are frequently incomplete dorsally and, less often, ventrally.

In American Caecilians the number of these primary grooves ranges
from 76 (in Chthonerpeton indistinctum) to 285 (in Caecilia bassleri).
The range 76-166 covers all specimens of Gymnopis, Siphonops,
Chthonerpeton, and Typhlonectes. Rhinatrema has 108-198 prim-
aries, and Caecilia has 110-285.

Individual variation is, of course, greater in forms with a high count.
No age variation appears or is to be expected. No sexual variation
has been discovered.

In Rhinatrema, in Gymnopis, and in most Caecilia, some or all of
the segments are partly or completely divided by secondary grooves in
the middle of the segment. In Rhinatrema these are present and
complete in each segment, and it is impossible, without dissection, to
distinguish between primary and secondary grooves. In this genus the
number of vertebrae equals half the number of superficial rings. In
Gymnopis and in Caecilia the secondaries are absent from the more
anterior segments. In these two genera the secondary rings appear at
first anteriorly as two unconnected grooves, between the primaries,
and parallel to them, in the dorsolateral area. The first appearance is
often asymetrical. They increase in length in the more posterior
segments, the two join first dorsally, and then, towards the posterior
end, ventrally. At the hind end they are exactly like the primaries,
but as they rapidly become incomplete anteriorly on the under side it
is not hard to make a separate count of the two sets. It is extremely
important in these two genera to keep the primary and secondary
counts separate.

These secondary grooves are an outward and visible sign of the
presence of bony scales in the anterior half of the segment. The
secondary counts given in this paper are all taken by beginning with
the first incomplete (dor so-lateral) secondary groove to appear, and
counting all the segments posterior to it.

Secondary grooves are present in all species of Rhinatrema (equal
in number to the primaries and all complete) ; all species of Gymnopis
(from a minimum of 10 anterior segments without them in G. multi-

448 bulletin: museum of comparative zoology

plicata oaxacae to a maximum of 87 in G. nicefori; a maximum count
of 121 in G. multiplicata oaxacae, a minimum of 13 in G. paniceps;
anterior secondaries always incomplete, maximum complete 67 in
G. nicefori); most species of Caecilia (from a minimum of 55 anterior
segments without them in C. dunni to a maximum of 268 in C. bassleri;
a maximum count of 94 in C. armata; anterior secondaries always in-
complete, maximum complete 26 in C. dunni).

Secondary grooves are present or absent in two species of Caecilia
(C. guntheri, 8-0); C. pachynema, 11-0).

Secondary grooves are unknown in three species of Caecilia (C.
caribea, C. degencrata, C. elongata) in all species of Siphonops, of
Chthonerpeton, and of Typhlonectes.

The individual variation in number of secondaries, and in number
of complete secondaries, shows no correlation with age or sex.

Scalation

Bony cycloid scales are concealed beneath the skin anterior to both
primary and secondary grooves in all Rhinatrema, all Gymnopis, and
in most Caecilia. They are absent in all Siphonops, in all Chthoner-
peton, and in all Typhlonectes. They invariably accompany secon-
dary grooves. In Gymnopis and in Caecilia the first secondary con-
ceals a single scale. A complete secondary conceals a complete ring of
scales. Wherever secondaries are present there are scales present an-
terior to the primaries. They appear first in the dorsolateral area and
extend further dorsally and ventrally as one passes back along the
body. At the hind end each segment contains two complete rings of
bony scales. In Rhinatrema, every segment of the body contains two
complete rings of scales.

In some (but not in all) specimens of Caecilia without secondaries
scales may be found in connection with the hindmost primaries.
Ordinarily, lack of secondaries indicates lack of scales; presence of
secondaries always indicates presence of scales.

Xieden (1913, Gymnophiona, p. 2) says: "scales . . . are in most
genera restricted to the back (only Ichthyophis and Herpele have
scales on the belly also) and are besides arranged in many rows in the
hinder half only of each of the epidermal folds limited by two circular
grooves." As may be seen from the foregoing remarks, none of the
statements made by Xieden are correct. Scales are on the belly in
Rhinatrema, Gymnopis, and Caecilia ; there are never more than two
rows or rings to a segment ; they are usually in both halves of a segment.

DUNN: THE AMERICAN CAECILIANS 449

The statement about "many rows" is obviously reached by examina-
tion of microscopic sections, as the scales of any one ring overlap each
other considerably. There is no overlapping of the scales of one ring
by those of another. My statements concerning scalation are derived
from examining the scales in situ on the animals.

Dentition

American Caecilians bear teeth on the premaxillary and maxillary
bones as an outer, upper row; on the pre vomers and palatines as an
inner, upper row; on the dentaries as an outer, lower row. An inner,
lower row, sometimes present, has been considered splenial.

At one extreme of American variation the teeth are all similar, and
relatively numerous in all rows. It is legitimate to infer that this is
the primitive condition.

At the other extreme the teeth of the premaxilla-maxilla set and of
the dentary set are progressively enlarged anteriorly into big hooked
fangs, and are reduced in number. The inner mandibular row may be
entirely absent. This condition is probably secondary.

The species of Rhinatrema, Chthonerpeton, and Typhlonectes have
the presumably primitive condition, and no generic distinctions in
dentition have been observed.

In Gymnopis and in Siphonops the teeth on the lower jaw are
uniform but larger than those on the upper. The inner mandibular
row is reduced to one tooth on a side (in oligozona and in muttiplicata)
or is entirely absent.

In Caecilia the anterior teeth on the lower jaw are much enlarged
and sharply pointed; to a less degree this is true of the maxillary teeth.
The inner mandibular row may consist of as many as four teeth on a
side (five or six were reported for the types of polyzona); they may be
reduced to one on a side or may be entirely absent.

Accurate counts of the number of teeth in any given row are well
nigh impossible to make unless the specimen is stained and cleared, or
unless it is made into a skull. Either of these two operations enables
one to count the teeth and the sockets, and thus arrive at an accurate
statement of the total dentition. Such treatment is obviously im-
possible for most of the specimens. I am profoundly skeptical of
dental characters in these animals as a basis for specific discrimina-
tion, having found considerable variation in count between the two
sides of the same individual in skulls and in cleared specimens.

The presence of enlarged, sharply pointed, anterior teeth in all

450 BULLETIN1: MUSEUM OF COMPARATIVE ZOOLOGY

American Caecilians with the tentacular aperture under the nostril
(and only in these) tends to establish the genus Caecilia as here
treated.

The great reduction or absence of the inner mandibular row in the
species here considered Gymnopis and Siphonops (in connection with
other characters) confirms their alliance.

Chthonerpcton viviparum (with 3-4 teeth in the inner mandibular
row), and Siphonops brasiliensis (with none), are otherwise so similar
that they have been confused. Aside from this I know of no case
where it is necessary to examine dentition in order to arrive at specific
or generic identification, and it is not absolutely necessary even in
this case.

Dimensions

The smallest individual seen is a specimen of Gymnopis nicefori
100 mm. long. Perfectly formed young 76 mm. long have been taken
from the oviduct of a pregnant Gymnopis parviceps. The smallest
species are: Siphonops hardy i (nine specimens with a maximum length
of 178 mm.) and Gymnopis parviceps (a single pregnant female 180
mm. long).

Eleven species (5 Rhinatrema, 2 Siphonops, 4 Gymnopis) have
their maximum recorded lengths under 251 mm. The maximum
length recorded outside the genus Caecilia is 695 mm. Six species of
Caecilia exceed this length, and three (tentaculata 1075 mm., abitaguae
1200, thorn psoni 1375) exceed a meter. The maximum length attained
by Caecilians in the Old World is 500 mm.

A diameter of 30 mm. is attained by Gymnopis m. mexica?ia, by
Caecilia tentaculata, and by Typhlonectes compressicauda natans.

If Caecilians were represented in collections only by specimens
ideally collected and preserved, accurate measurements of length and
of diameter could be taken with little difficultv, and the ratio of
length to diameter would be very reliable. Actually, specimens have
to be measured in every conceivable state of preservation and dis-
tortion. There is wide discrepancy in the length measurement of a
number of specimens as taken by different observers, it is impossible
to avoid a possible error of as much as a millimeter in diameter mea-
surements, and 1/d ratios presented here are in no case carried into
decimals, and in most cases give a range of variation which exceeds
that of the animals in life.

Stout species are often slimmer when young and vice versa. Many
seem to retain the same proportions throughout life.

DUNN: THE AMERICAN" CAECILIAXS 451

A small Typhlottcctcs c. compressicaudiis has an 1/d ratio of 12, and
a pregnant female Gymnopis m. mexicanus one of 14. Xo Rhinatrema
seen has an 1/d ratio of over 30. Outside of Caeeilia the slimmest
specimen seen is a Gymnopis nicefori with a ratio of 67. Seven species
of Caeeilia may be more attenuate than this, and Caeeilia bassleri
may be 160 times as long as wide.

The most elongate forms have the most vertebrae, but otherwise
there is not too much correlation, and there is a wide range of verte-
bral count among equally stout forms.

The body is roughly cylindrical in most forms, and the diameter is
the same from neck to vent. This statement is not true for Typh-
lonectes, which is compressed laterally, and has the posterior part of
the body much deeper than the anterior. In this genus there is also a
dermal dorsal fin fold, restricted to the posterior third in eompressi-
eanda and extending nearly to the head in kaupii.

In Rhinatrema there is a tail. In other genera the body ends bluntly
just behind the vent.

Coloration

The majority of the forms have no definite markings, being dull
blackish above, somewhat lighter below. The head is usually some-
what lighter than the dorsal surface, and the anal region is usually
whitish.

The ventral surface is much lighter than the dorsal surface in some
Gymnopis, and spotted or mottled with white in some Caeeilia.

The primary grooves are white in two species of Siphonops (aiuiul-
ata# and paidensis), in marked contrast to the dark background of the
segmental folds which they delimit. The reverse of this is seen in some
Caeeilia (principally ochrocephala and polyzona). In these the grooves
are black and the folds are of a lighter color.

Yellow spots, one on each side on the segmental folds, are quite
usual in Caeeilia pachynema, and occur sporadically in a few other
species of Caeeilia.

Vivid yellow lateral stripes, one on each side, from jaw to vent, are
present in three species of Rhinatrema (bititatum, parkeri, and bicolor).

Habitat and habits

Something about the habitat may be inferred from the range given
for a species. I have included the few ecological notes under the
specific headings.

452 bulletin: museum of comparative zoology

The climatic and botanic areas inhabited are: tropical rain forest;
tropical deciduous forest; tropical savanna; temperate forest; temper-
ate savanna. In North America the only temperate areas inhabited
are montane cloud forest; it is probable that the animals occur in
savanna only in galeria forest along rivers.

Except for the aquatic, river-dwelling Typhlonectes it is probable
that all are terrestrial and burrowing. I have seen only three forms
(Gymnopis multiplicata proximo,, Gymnopis parviceps, and Caecilia
ochrocephala) alive in the field, and the literature is singularly un-
informative.

The animals are unquestionably carnivorous, but the precise ali-
ment is not known.

Notes in literature indicate that they are preyed on by snakes;
Ninia atrata, Pscudoboa clclia, Sordellina brandonjonesi, and several
species of Micrurus being mentioned.

Males have a median intromittent organ and fertilization is pre-
sumably internal.

Published observations would indicate that Rhinatrema and Sipho-
nops are oviparous, and that Gymnopis, Chthonerpeton, and Typh-
lonectes are viviparous. The behavior of Caecilia is not positively
known, but as no embryos have been found in females it is probably
oviparous.

External gills have only once been reported for larvae (Rhina-
trema. They have been reported for embryos in the eggs of Rhina-
trema, and of Siphonops, and for embryos in the oviducts of Chthon-
erpeton and Typhlonectes. They have been reported absent in em-
bryos in the oviducts of Gymnopis. In Rhinatrema these gills are in
three pairs, the two anterior fimbriated, but with rather few filaments.
The gills of Siphonops are similar but the posterior may be absent.

The gills of Chthonerpeton and of Typhlonectes are a single pair
of large, flat, leaf-like structures. It is probable that these are entirely
embryonic, and that the one case of birth with persistent gills was
premature.

Free living (? aquatic) larvae without gills, but with a single pair
of gill slits have been noted in Rhinatrema and in Caecilia. Well
formed embryos in the oviducts of Gymnopis do not have gill slits.
Normally born young of Typhlonectes lack gill slits, and gill slits
were not reported for embryos of Chthonerpeton or of Typhlonectes,
although external gills were present.

Gymnopis, Chthonerpeton, and Typhlonectes normally give birth
to small replicas of the adult. Rhinatrema has a larval stage, which

DUNN: THE AMERICAN CAECILIANS 453

is presumably aquatic, and which emerges from eggs laid by the
mother. Siphonops lays eggs, but whether there is a larval stage is
not known. Caecilia was reported long ago to have a larval stage. It
is not certain that this is correct. It is not certain, but it is probable,
that Caecilia lays eggs. It is peculiar that less is known of the breeding
habits of Caecilia than of the other five genera, since specimens of
Caecilia make up nearly 40% of those in collections.

Acknowledgments

I wish first of all to express my gratitude to the John Simon Guggen-
heim Memorial Foundation, as a Fellow of which I was enabled to
examine material in the Old World; and to the authorities of the British
Museum of Natural History, of the Museum of the Jardin des Plantes
in Paris ; of the Senckenberg Museum in Frankfort ; of the Museums
of Berlin, Dresden, Munich, and Vienna; and of the Jagellonian Uni-
versity in Krakau, for the privilege of examining their specimens.

I am indebted to the authorities of the following institutions in the'
United States for the same privilege : Museum of Comparative Zoology;
Smith College; American Museum of Natural History; Academy of
Natural Sciences in Philadelphia; Carnegie Museum; National Mu-
seum; Museum of Zoology of the University of Michigan; Field
Museum; Museum of Leland Stanford University; California Academy
of Science.

Dr. Harvey Bassler sent me his Peruvian collection of 31 specimens,
including 4 genera and 6 species, probably the best lot ever gathered
by one man. He has my especial thanks. Dr. E. H. Taylor sent me
two Mexican specimens from his private collection for identification.

I have seen a specimen from the Museo Nacional of Salvador.

To my friend Prof. Juvenal Valerio, Director of the Museo Nacional
of Costa Rica, I am indebted for the privilege of examining the speci-
mens in that collection, and those of the Seminario de San Jose and of
the Collegio de San Luis Gonzaga in Cartago. I am also indebted to
Dr. Picado and Prof. Manuel Valerio of San Jose, and to Prof. Torres
of Cartago for additional Costa Rican material.

Mr. Lindsay, of the Summit Gardens in the Canal Zone, let me ex-
amine a specimen he had, and the authorities of the Collegio La Salle,
in Panama City, showed me their specimens. Mr. James Zetek, of
Balboa, has passed over to me a number of Canal Zone specimens.

Hermano Niceforo Maria, of the Institute La Salle in Bogota,
Colombia, has forwarded me a number of Colombian specimens.

454 bulletin: museum of comparative zoology

My friend, Dr. Afranio Amaral, sent me the collection of the Museu
Paulista in Sao Paulo, Brazil.

I was sent for identification most of those in the Museu Nacional.

I wish to express my thanks to all these individuals and to the
authorities of all these institutions.

Mr. H. W. Parker, of the British Museum of Natural History,
Mr. Arthur Loveridge, of the Museum of Comparative Zoology, and
Mr. Charles Bogert, of the American Museum of Natural History,
have been extremely courteous and kind to me in the course of this
work.

Dr. Leonhard Stejneger, of the United States National Museum,
has, as usual, given me good advice on nomenclatorial problems.

No single museum has anything like a complete set of the forms of
American Caecilians (6 genera, 44 forms). The three best are: British
Museum (6 genera, 28 forms); American Museum (6 genera, 25 forms);
Museum of Comparative Zoology (5 genera, 22 forms). These three
collections together contain 38 forms, lacking only Rhinatrema bivi-
tatum, Rhinatrema colombianum, Siphonops insulanus, Caecilia sub-
nigricans, Caecilia abitaguae, and Caecilia armata.

Seven forms are not represented by specimens in any museum in
this country (Rhinatrema bivitatum, R. columbianum, R. parkeri,
Siphonops insulanus, Caecilia guntheri, C. armata, Chthonerpeton
petersi).

Identification and methods

Measurements are given in millimeters. In Rhinatrema body length
is from tip of snout to posterior end of anus; tail length is from latter
point to tip of tail. In other genera only total length is given.

Primaries are counted first, then secondaries, then number of
complete secondaries. It is sometimes easier, and just as accurate, to
count the primaries down to the first secondary, and get the number
of secondaries by subtracting this count from the total number, of
primaries. The foremost secondary is to be found in the dorso-lateral
area, sometimes isolated from its successors. The primary and secon-
dary counts must be kept separate.

The position of the tentacle should be noted.

Tentacle position, primary and secondary count, and 1/d ratio will
ordinarily serve to identify any American Caecilian. It may some-
times be necessary to examine the dentition. In Caecilia attention
should be paid to color and condition of eye and geographical prob-
ability.

DUNN: THE AMERICAN CAECILIANS 455

The inner mandibular teeth, if present, are barely anterior to the
edge of the tongue, the anterior teeth of the two rows close together.
They are usually rather concealed in the gums, and their tips are on
quite a different level from those in the outer mandibular row, so that
it is not hard to overlook them.

Most descriptions of Caecilians in literature are much too long,
repeating for the individual or for the species statements true of every
member of the genus. This serves no useful purpose and may be con-
fusing.

In cross-section the animals (except for the laterally compressed
Typhlonectes) are circular when alive. Various preservatives and
death throes may cause muscular contractions which materially alter
this. Most prominent of these is contraction of the obliquus externus
muscle, which makes a more or less marked dorso-lateral fold appear.
This has fortunately not caused any difficulty in America, but one Old
World form has been described as a new species on this basis, the fold
being imagined to be "glandular".

The synonymies give, I hope, most of the papers with information
on range, habits, relationships, and systematic treatment of each
species. No attempt whatever has been made to include the multifari-
ous references in anatomical papers to a few of the species, principally
to SipJionops annulatus. Much of this material may be found in the
bibliography given by the Sarasins (1890, Erg. nat. Forsch. Ceylon,
2, pts. 3-4) and in Werner's compilation (1931, in Kukenthal, Handb.
Zool. 6, 2, pp. 143-208, ff. 231-338, bibliography). The latter work is
extremely good and useful, but the systematic section extends only as
far as genera, and is taken from Nieden (1913, Die Gymnophiona,
section 37 of Das Tierreich), and Nieden's work, while the latest
treatment of the Caecilians of the world, extending down to species,
is compilation pure and simple.

Key to genera of American Caecilians

A. Secondaries as numerous as primaries; a tail (tentacle im-
mediately anterior to eye; teeth uniform, inner row of mandibu-
lar teeth well developed) Rhinatrema (p. 45G)

A A. Secondaries (if present) not as numerous as primaries; no tail.
B. Tentacle under nostril, under tip of snout (anterior maxil-
lary and outer mandibular teeth enlarged and pointed ; inner
mandibular row well developed to absent; secondaries from
half as numerous as primaries to entirely absent; scales
present or absent) Caecilia (p. 494)

456 bulletin: museum of comparative zoology

BB. Tentacle on side of head.

C. Secondaries present (teeth of a row uniform, maxillary teeth
smaller than outer mandibular; inner mandibular row one
tooth or absent; tentacle nearer eye than nostril)

Gymnopis (p. 461)
CC. No secondaries (no scales).
D. No dorsal fin.

E. Tentacle nearer eye than nostril; no inner mandibular
tooth row; primaries mostly complete; no anal disk

Siphonops (p. 479)
EE. Tentacle usually nearer nostril than eye; inner mandi-
bular tooth row well developed; primaries usually in-
complete dorsally ; an anal disk . . Chthonerpeton (p. 527)
DD. A dorsal fin (tentacle close to nostril; inner mandibular
tooth row well developed; primaries incomplete dorsally;
an anal disk) Typhlonectes (p. 532)

This key may not serve to separate all specimens of Siphonops from
Chthonerpeton (especially <S. brasiliensis and C. viviparum). It would
be advisable to consult the specific descriptions in the case of speci-
mens from Southeastern Brazil.

"Siphonops syntremus" of Cope, from Northern Central America,
was said to have: secondaries not so numerous as primaries; a tail;
tentacle on side of head just posterior to nostril; mandibular teeth
large and few. The unique type is lost. This combination of characters
is otherwise unknown.

Rhinatrema Dumeril and Bibron

1841. Rhinatrema Dumeril and Bibron, Erpet. Gen. 8, p. 288 (monotype
Caecilia bivitata Cuvier).

1883. Epicrionops Boulenger, Ann. Mag. Nat. Hist. (5), 11, p. 202 (mono-
type E. bicolor Boulenger).

Diagnosis. Caecilians with a distinct pointed and flattened tail;
primaries 108-198 on body; secondaries as numerous as primaries,
all complete; body and tail with two complete rings of bony scales
in each segment; l/d 20-30; tentacle in horseshoe-shaped groove,
immediately anterior to eye and very small; eyes visible; teeth of
any row uniform; inner mandibular tooth row well developed; length
145-370 mm.; six forms.

Range. Colombia, Ecuador and Peru. The Guianas. Sea level (?) to
3900 feet elevation.

DUNN: THE AMERICAN CAECILIANS 457

Key to forms of Rhinatrema

A. Tail very short, much less than 5 mm. long; (striped;

primaries 167-181; Guiana) bivitatum

AA. Tail 8-20 mm. long; tail segments 5-25.

B. Primaries on body 198; striped; Colombia parkeri

BB. Primaries on body 191 ; uniform; Guiana nigrum

BBB. Primaries on body 140-175; uniform; Peru. . .peruvianum
BBBB. Primaries on body 117-135; striped; Ecuador and Peru.

bicolor
BBBBB. Primaries on body 108; uniform; Colombia . . columbianum

Remarks. I keep all these forms in the original genus Rhinatrema
in spite of the fact that there is an obvious dichotomy between the
type of Rhinatrema with scarcely any tail, and the other forms (includ-
ing the type of Epicrionops) which have a well developed tail. Young
specimens are difficult to count with accuracy, and sometimes seem
to differ in color from adults (cf . uniform young from the type locality
of parkeri, and uniform young from an area inhabited only by the
striped bicolor). I have seen 19 specimens and know of three that I
have not seen. I have examined the types of all the species with the
exception of columbianum.

Rhinatrema bivitatum (Cuvier)

1829 Caccilia bivitata Cuvier, Regne Animal, (2), 2, p. 100; Guerin-Meneville

1829-1838, Iconogr. Regne Animal, 3, Rept., pi. 25, f. 2.
1831. Caecilia bivittata Gray, in Griffith's Cuvier's Animal Kingdom 9, app.,

p. 110.
1841. Rhinatrema bivittaturn Dumdril and Bibron, Erpet, Gen. 8, p. 288,

pi. 85, f. 4; Dumeril 1863, Mem. Soc. Cherbourg 9, pi. 1, f. 5, 12;

Vaillant 1895, CR. Acad. Sci. 120, p. 460; Boulenger 1895, Proc.

Zool. Soc. London, p. 407; Nieden 1913, Gymnophiona, p. 14.
1879 Ichthyophis glutinosus (part), Peters, Mon. Ak. Berlin, p. 928, 931, f. 2.

Type. Paris No. 8.
Type locality. Cayenne.
Range. French Guiana.

Diagnosis. A striped Rhinatrema, with tail at most 2 mm. long;
primaries 167-181; 1/d 24-30; 195-300 mm. long.
Description. Nothing can be added to the diagnosis.
Remarks. This species was originally of rather uncertain locality,

458 bulletin: museum of comparative zoology

and was confused with Ichthyophis glutinosus of southeastern Asia
in the literature. The type is fortunately preserved and shows that
this confusion was baseless. As the first known species of the most
primitive American genus, the confusion with the most primitive
Old World genus is not incomprehensible.
Specimens seen. 3 as follows:

length

prim.

length

of tail

diam.

1/d

Cayenne,

Paris 8 181

210

7

30

Guiana,

Paris 8a 167

195

2

8

24

Sao Paulo

Hamburg 5268 169

300

10

30

N.B. Quite likely Sao Paulo is an erroneous locality.

Rhixatrema parkeri spec, now

Type. BMNH 97-11-12, 23.

Type locality. Medellin, Colombia.

Range. Known only from type locality.

Diagnosis. A striped Rhinatrema with well developed tail (10 mm.);
body primaries 198; 1/d 26.5; 212 mm. long.

Description. The type has two light stripes; 198 primaries on body
and 12 on tail; total length 212 mm., tail 10 mm.; diameter 8 mm.

Remarks. The type has more body primaries than any other Rhina-
trema examined, the nearest approaches being made by the striped
bivitatum of Guiana (181) which has a very short tail; the uniform
nigrum of Guiana (191); the uniform peruvianum of Peru (175). The
other species known from Colombia, columbianum from the Province
of Cauca, is uniform and has only 108 body primaries.

A larval Rhinatrema from Medellin, AMXH 1380, may possibly
belong to this species. It is not in the best of condition and I cannot
count its annuli. It is uniform black; has one gill slit but no external
gills; length 162 mm., tail 10 mm.; diameter 8 mm.; 1/d 20.

I take great pleasure in naming this form after my friend H. W.
Parker, of the British Museum of Natural History, to whom I am
vastly indebted for help and advice, and to whom all herpetologists
are indebted for his papers on Caecilians.

Specimens seen. Two, the type and one larva.

Rhixatrema nigrum spec, now

Type. AMXH (specimen mislaid, and number in my notes, 34088,
either incorrect or duplicated).

DUNN: THE AMERICAN CAECILIANS

459

Type locality. Arundabara, British Guiana, elevation 2200 feet.

Range. Known only from type locality.

Diagnosis. Uniformly dark; tail well developed (11 mm.); 191 body
primaries; 1/d 23; length 211 mm.

Description. The type has, in addition to the diagnostic characters,
13 tail primaries; a diameter of 9 mm.

Remarks. This form differs from parkeri of Colombia, which has a
similar segment count (198) in color; it differs from the other Guiana
species, bimtatum, in color, in having a well developed tail, and in
having ten more body segments; it differs from peruvianum in having
sixteen more body segments.

Specimens seen. One, the type.

Rhinatrema peruvianum Boulenger

1902. Rhinatrema peruvianum Boulenger, Ann. Mag. Nat. Hist. (7), 10,
p. 153; Xieden 1913, Gymnophiona, p. 15; Noble 1927, Ann. New
York Acad. Sci., p. 59, f. 5.

Type. BMXH 1902-5-29, 207.

Type locality. Marcapata Valley, southeastern Peru.
Range. Southeastern Peru.

Diagnosis. A uniformly colored Rhinatrema; tail 17-20 mm. long;
primaries 140-175; 1/d 20-23; 2S0-370 mm. long.
Description. Uniform brown.
Specimens seen. 4, as follows:

Peru:

Marcapata Valley
BMXH 1902-5-29,"^ 207
"Juliaca," AMNH 1454
AMNH 1457

No data, Vienna

body

prim.

175

140

tail
prim.

14
15

152

body

length
280
320
65

370

tail
length
917

20
6.5

20

diam
12
16
3

17

1/d
23
20
19

embrvo

21

Remarks. Noble (1927) has pointed out that this species is "ovi-
parous, and has figured an encapsuled embryo, with external gills.

The locality Juliaca is probably erroneous. Dr. Harvey Bassler has
suggested that the specimens so labeled, sent by a member of the Inca
Mining Co., came from the vicinity of the mine at Santo Domingo,
north of the shipping station of Juliaca, and at about 3000 feet above
sea level.

460 bulletin: museum of comparative zoology

The embryo has branchial structures as described by Parker for
bicolor. Its tail has a dorsal and a ventral finfold.

Rhixatrema bicolor (Boulenger)

1883. Epicrionops bicolor Boulenger, Ann. Mag. Nat. Hist. (5), 11, p. 203.

1895. Rhinatrema bicolor Vaillant, C. R. Ac. Sci., 120, p. 461 ; Boulenger 1895,
Proc. Zool. Soc. London, p. 407, pi. 23, f. 2; Nieden 1913, Gymno-
phiona, p. 15; Parker 1934, Ann. Mag. Nat. Hist. (10), 14, p. 265.

Type. BMNH 78-1-25, 110.

Type locality. Intac, Ecuador [3900 feet elevation in western
Ecuador].

Range. Western Ecuador and the eastern part of Ecuador and of
Peru.

Diagnosis. A striped Rhinatrema; with tail 8-15 mm. long; prim-
aries 117-135; 1/d 16-27; 145-250 mm. long.

Description. The color of the La Merced specimens was a dark
purplish brown ; on each side a ventrolateral yellow band from jaw to
vent.

Specimens seen; 9, as follows:

body

tail

tail

Ecuador :

prim.

prim.

length

length

diam.

1/d

Intac: BMNH 78-1-25,

110

117

12

225

8

9

25

East Ecuador AMNH

46205

130

10

210

14

13

16

Peru :

La Merced, Chanchamayo Valley,

3000-3500'

AMNH 42858

135

25

241

15

9

27

" 42859

124

22

194

12

8

24

" 42860

124

19

250

14

10

25

" 42861

130

22

225

13

9

25

Chanchamayo or Perene

AMNH 17304

US

12

220

13

10

22

" 17305

119

11

230

13

11

21

" 17306

128

12

■ —

—

—

Remarks. Parker (1934) has recorded a larva, probably of this
species, from Zamora, Ecuador (3250 feet, east of the Andes). It had
about 135 primaries, uniform body color, length 145 mm., tail 9.

DUNN: THE AMERICAN CAECILIANS 461

"Three pairs of small external gills are persistent, the two anterior
pairs fimbriated, with five or six finger-like processes, and the last
reduced to a mere knob ; ventral to this last is a single oblique slit-like
gill-cleft equipped with a large valvular flap on each side and lying in a
circular depression."

It would seem that this species occurs on both sides of the Andes.
It is thus either able to cross the passes, some of which are fairly low
in Ecuador, or, and more probably, it antedates the present elevation
of these mountains.

Rhinatrema columbianum Rendahl and Vestergren

1938. Rhinatrema columbianum Rendahl and Vestergren, Arkiv f. zool.
31A, 3, p. 1, ff. 1-3.

Type. Stockholm 19, collected by Kjell von Sneidern.

Type locality. El Tambo, Pro v. Cauca, Colombia, about 1000 m.
elevation.

Range. Known only from type locality.

Diagnosis. A Rhinatrema without stripes; tail well developed
(8.7 mm. long); primaries 108; 1/d 20.; total length 161 mm.

Description. "227 . . . skinfolds, of which 11 are on the tail"; 108
primaries and 108 secondaries on body; 5-6 primaries on tail; "great-
est body diameter 20.1 times in total length; tail length 18.5 in total
length"; black, uniform; anal region whitish; total length 161 mm.
The tail length would appear to have been 8.7 mm.; the diameter to
have been 8 mm.

Remarks. This is the shortest bodied member of the genus. The
next shortest, bicolor of Ecuador and Peru, is striped and has 117-
135 body primaries. The other Colombian species, parkeri, is striped,
and is the longest bodied member of the genus, with 198 body primaries.

Gymnopis Peters

1874. Gymnopis Peters, Mon. Berlin Ak., p. 616 (monotype Gymnopis
muUiplicata Peters).

1879. Dermophis Peters, Mon. Berlin Ak., pp. 930, 937 (genus based on
Siphonops mexicanus Dumeril and Bibron and Dermophis brevirostris
Peters. Four species inquirenda were also included. Noble designated
mexicanus as type in 1924, in Bull, Amer. Mus. Nat. Hist. 49, 11,
p. 305).

1883. Cryptopsophis Boulenger, Ann. Mag. Nat. Hist. (5), 12, p. 166 (mono-
type Cryptopsophis multiplicatus Boulenger).

1924. Gymnophis Barbour, Proc. Biol. Soc. Washington 37, p. 125 (pro
Gymnopis Peters).

462 bulletin: mltseum of comparative zoology

Diagnosis. Caecilians without a tail; primaries 95-158; secondaries
13-121; scales always present; 10-87 primary folds without secondar-
ies; 1/d 14-67; tentacle in horseshoe-shaped groove on side of head
between eye and nostril and nearer the former; few or no teeth in
inner mandibular row; mandibular teeth larger than maxillary or
palatine; teeth of a row uniform; eye visible or invisible, in orbit or
under bone; length 100-510 mm.; eleven forms.

Range. Vera Cruz and Guerrero, Mexico, to western Panama. Ap-
parently absent from Yucatan Peninsula. Cauca and Magdalena
Valleys, Colombia. Ecuador. French Guiana. Sea level to 4500 feet.

Key to forms of Gymnopis

A. North American species.

B. Eye visible; tentacle slightly nearer eye than nostril; primaries
110 or less.
C. Secondaries 32-80.
D. 1/d 14-26.

E. Secondaries 51-80 mexicana

EE. Secondaries 41 clarkii

DD. 1/d 25-32 gracilior

CC. Secondaries 13 parviceps

BB. Eye invisible (if visible tentacle extremely close to eye); pri-
maries 112 or more.
C. Secondaries 84-121.
D. Eye usually visible; primaries 121-137, secondaries 101—

121 oaxacae

DD. Eye invisible.

E. Primaries 128-132 multiplicata

EE. Primaries 112-126 proxima

CC. Secondaries 62-74 oligozona

AA. South American species.

B. Eye visible; tentacle slightly nearer eye than nostril ; primaries

124-125 albiceps

BB. Eye invisible (if visible tentacle extremely close to eye).

C. Primaries 100-120 unicolor

CC. Primaries 133-15 nicefori

DUNN: THE AMERICAN CAECTLIANS

463

A tabular list of counts of Gymnopis

Primaries

Specimens

minus

seen

Form

Primaries Secondaries

Secondaries

1/d

1

parviceps

96

13

83

22

4

gracilior

95-102

32-78

22-68

25-32

66

mexicana

97-110

51-80

26-59

14-26

4

clarkii

101-107

41

60-66

16-19

15

unicolor

100-120

22-74

41-87

27^0

38

proxima

112-126

84-104

15-36

23-34

2

albiceps

124-125

45-55

70-89

35-^6

10

oaxacae

121-137

101-121

10-26

26-40

8

multiplicata

128-132

101-111

17-28

25-35

3

oligozona

128-135

62-74

57-68

44-64

6

nicefori

133-158

45-104

43-87

39-67

157

Remarks. I have examined 157 specimens of this genus, including
the types of eleven names. I have not examined the type of Cryptop-
sophis multiplicatus Boulenger 1883 in the British Museum (= Gym-
nokis multiplicata proxima) or the type of Gymnopis multiplicata
oaxacae Mertens 1930 (Senck. Mus. 22130).

The generic name Cryptopsophis appears in the synonymy of
Gymnopis because the type species and specimen is identical with an
earlier described American form. The specimen was erroneously sup-
posed to have come from the Seychelles Islands, and geography,
rather than anatomy, seems to have prompted its description.

The generic name Dermophis appears in the synonymy of Gym-
nopis because the type species of Dermophis (mexicanus) appears to
me to be congeneric with the type species of Gymnopis (multiplicata).
Dermophis was characterized by having the eye visible, in an open
orbit, whereas Gymnopis had the eye invisible and roofed over by
bone. This difference exists as far as the two type species are con-
cerned, but these two extremes are so bridged, in other forms, that
generic distinction is impractical.

In general forms with visible eyes have the tentacular aperture
slightly nearer the eye than the nostril. Forms with invisible eyes have
the tentacular aperture very close to the eye and further from the
nostril. Forms with visible eyes have, on the whole, no teeth in the

464 bulletin: museum of comparative zoology

inner mandibular row; fewer primaries and secondaries; greater differ-
ence between primary and secondary count (= less of the body with
bony scales). These criteria, which are not sufficiently clear cut to
serve for dichotomy of the genus, indicate a vague division into two
groups. One of these retains a well developed eye in an open orbit,
but tends to a loss of scalation, a loss of the inner mandibular tooth
row, and a more anterior position of the tentacular aperture. This
group, containing vicxicana and its races, parviceps and albiceps, is
primitive in the retention of the eye, but presumably secondary in the
other characters. The other group tends to reduction of the eye, but
to retention of the inner mandibular tooth row, of the secondaries
and scales, and of the posterior position of the tentacular aperture.
This group contains multiplicata and its races, oligozona, unicolor and
nicefori.

These two groups are more distinct in North America than they
are in South America.

Specimens of oaxacae show the most complete scalation; specimens
of parviceps show the most reduced scalation.

The "Dermophis crassus" of previous lists is, as appears from exami-
nation of the types, a straight synonym of Siphonops annulatus.

The difference in the present treatment of mexicana and of multi-
plicata and their races from that of Boulenger is a natural consequence
of the examination of 157 specimens and 11 types by me, as against
the examination of 13 specimens and 3 types by him.

Gtmnopis multiplicata multiplicata Peters

1874. Gymnopis multiplicata Peters, Mon Ak. Berlin, p. 616, pi. 1, f. 1; 1879
Mon. Ak. Berlin, p. 939, f. 7; Boulenger 1882, Cat. Batr. Grad. Brit.
Mus. (2), p. 100; Cope 1885, Proc. Amer. Phil. Soc, 22, p. 171;
Boulenger 1895, Proc. Zool. Soc. London, p. 410; Gunther 1902,
Biol. Cent. Amer., Batr., p. 308; Nieden 1913, Gymnophiona, p. 21,
f. 11; Dunn 1928 (in part), Proc. New England Zool. Club, 10, p. 75.

1877. Siphonops simus Cope, Proc. Amer. Phil. Soc. 17, p. 91; Brocchi 1883,
Miss. Sci. Mex., p'. 121.

1879. tDermophis simus Peters, Mon. Ak. Berlin, p. 938; Boulenger 1882,
Cat. Batr. Grad. Brit. Mus. (2), p. 99.

1885. Gymnopis sima Cope, Proc. Amer. Phil. Soc 22, p. 171; 1887, Bull.
U. S. Nat. Mus. 32, p. 9.

Type. Berlin No. 3705, collected by Warszewicz.

Type locality. Veragua.

DUNN: THE AMERICAN CAECILIAXS

465

Range. Pacific side, western Panama and Costa Rica ; Atlantic side,
Honduras. Sea level to 4500 feet.

Diagnosis. A Gymnopis with invisible eyes; primaries 128-132;
secondaries 101-111; difference 17-2S; 1/d 25-35; length 35S-510 mm.

Description. The few specimens seen afford no points other than
those given in the diagnosis. The color is black, lighter below. Peters
(1874) says there are 18 teeth on each side of the upper jaw.

Remarks. The species Gymnopis multiplicata may be divided into
three races; multiplicata from the Pacific side, proximo, from the
Atlantic side, and oaxacae from Mexico. The differences are not great:

prim.

sec.

diff.

proximo

112-126

84-104

15-36

no eyes

oaxacae

121-137

101-121

10-26

eyes

multiplicata

128-132

101-111

17-2S

no eyes

131 104/10 365 12 30

The criteria given above will serve to distinguish all proximo from the
other two races, and almost all oaxacae from almost all multiplicata.
The ranges are quite intelligible save for the single Honduras locality
for multiplicata, which would seem to indicate that proximo holds
territory between two areas of multiplicata. Specimens seen, eight, as
follows :

prim. sec. length diam. 1/d
Honduras :

Progreso Dist. MCZ 11048
Costa Rica:

Tilaran USNM 70656

San Mateo USNM 37761

Cartago Coll. St. Luis Gonzaga

Taboga MNCR

No locality MNCR

" USNM 29765

Panama :

Veragua Berlin 3705
Peters (1879) records it from Antioquia, Colombia, but it is very prob-
able that this record was based on Berlin Xo. 9524 from Caceres,
which is a specimen of G. nicefori.

129
129

101/16
111/17

358
380

12

12

30
32

128

103

370

11

—

128

111

34

132

110/8

390

11

35

type
sima

131

105/9

510

20

25

466 BULLETIN': MUSEUM OF COMPARATIVE ZOOLOGY

Gymnopis multiplicata proxima (Cope)

1875. Siphonops mexicanns Cope, Journ. Acad. Nat. Sci. Philadelphia (2),

8, p. 96.
1877. Siphonops proximus Cope, Proc. Aruer. Phil. Soc 17, p. 90; Brocchi

1883, Miss Sci. Mex., p. 121.
1879. Dermophis ? proximus Peters, Mon. Ak. Berlin, p. 938; Boulenger 1882,

Cat. Batr. Grad. Brit, Mus. (2), p. 99.
1883. Cryptopsophis multiplicata Boulenger, Ann. Mag. Xat. Hist. (5), 12,

p. 166 (Seychelles Is. in errore).
1885. Gymnopis proxima Cope, Proc. Amer. Phil. Soc. 22. p. 171; 1887,

Bull. U. S. Xat. Mus. 32, p. 9; Boulenger 1895, Proc. Zool. Soc.

London, p. 410; Gunther 1902, Biol. Cent. Amer., p. 308; Nieden

1913, Gymnophiona, p. 21; Noble 1918. Bull. Amer. Mus. Nat. Hist.

38, p. 346.
1928. Gymnopis multiplicata (part) Dunn, Proc. New England Zool. Club

10, p. 75 (breeding habits); Parker 1936, Trans. Linn. Soc. London,

19, 4, p. 455.

Type. USNM 29762-3, collected by Gabb.

Type locality. Eastern Costa Rica [= Limon].

Range. Nicaragua. Eastern Costa Rica, Pro v. Bocas del Toro,
Panama. Sea level to 4500 feet.

Diagnosis. A Gymnopis with eyes usually invisible; primaries 112—
126; secondaries 84-104; 15-36 primary folds without secondaries; 1/d
23-34; length 190-480 mm.

Description . Most specimens are distinctly lighter (even white)
on the belly. The primaries are somewdiat interrupted in the anterior
dorsal region. AXS 492S, without locality, has the eye visible and
not completely under bone. USNM 19614 has two inner mandibular
teeth ; the first scale appears on the side, under the primary, four seg-
ments anterior to the first secondary. The tentacle is much closer to
the position of the eye than to the nostril.

Remarks. Specimens of my own collecting, from Guapiles, Monte-
verde and Suretka, Costa Rica; Farm Six near Almirante, Panama;
were under logs in damp pastures. One of the last was a pregnant fe-
male with one perfectly formed young in the oviduct. The embryo
was 131 mm. long, 1/d 22., the mother was 375 mm. long, 1/d 25.

DUNN: THE AMERICAN CAECILIANS 46*3

Specimens seen, 38, as follows :

prim. sec. length diam. 1/d
Panama:

Bocas del Toro USNM 38754 124 101/16 225 7 32

Coco Plum Estate, near Bocas

MCZ 7990
Farm Six MCZ 9931
" " " 9932

Costa Rica:

Suretka MCZ 9934
Limon USNM 297(12
" 29763
Salvadora Farm USNM 84241
Monteverde MCZ 7987
" 7988
Reventazon USNM 38144

38145
38146
Guapiles MCZ 7989
Cariblanco BMXH 1907-10-9, 10
Peralta MNCR

( lartago Colleg. St. Luis Gonzaga
Parismina, M. Valerio coll.
5 km. Xortk of Cartago

MCZ 24526
No locality MCZ 24527
" " Seminario de San Jose
" MNCR

Nicaragua :

Rio San Juan (Colorado Jet.)

USNM 19612
Rio San Juan USNM 19613
" " " " 19614

San Juan del Xorte USNM 15630
" " " USNM 15643
Bluefields USNM 37351
Escondido R. (50 mi. above
Bluefields) USNM 20704

119

97/10

365

12

30

115

93/11

375

15

9 25

116

85/9

250

11

23

119

97/11

313

12

26

117

91/8

430

15

29

115

88/11

480

21

23

116

92

—

—

120

84/7

395

16

25

123

. 98/10

215

8

27

120

95/8

300

12

25

119

92/5

320

13

25

117

92/3

212

9

23

124

104/9

380

15

25

116

94/9

470

425

18
18

26

122

94/8

—

112

91/9

24

121

'.17 10

320

14

23

123

96/10

—

—

121

98

—

—

118

86

—

—

124

100/9

395

12

33

126

99/14

380

11

34

119

97/9

300

9

33

118

92/8

340

15

23

124

99/9

300

10

30

122

102/16

280

8

35

120

98/11

322

12

27

468 bulletin: museum of comparative zoology

prim.

sec. k

sngth

diam.

1/d

El Bluffs, Bluefields AMNH 8397

121

99/8

—

—

Eden Mine AMNH 8399

116

101/16

375

11

34

Hac. Valencia, San Miguelito,

Chontales Mts. AMNH 8396

122

103/11

470

19

26

Cape Gracias USNM 15311

122

102/16

280

8

35

San Ramon, 125 mi. up Rio

Wanks BMNH 1908-5-29,

122

121

98/11

190

6

31

Boquete I. AMNH 8398

122

95/10

335

13

26

No locality USNM 15199

124

101/4

383

12

32

ANS 4928

118

97/10

277

9

30

Parker (1936) gives 119 primaries; 97 secondaries, last 10 complete;
and 1/d 24 for the type of Cryptopsophis multiplicatus.

Gtmnopis multiplicata oaxacae Mertens

1930. Gymnopis multiplicata oaxacae Mertens, Abh. Ber. Mus. Magdeburg 6,
2, p. 153, f. 14.

Type. Senckenberg 22130, Dr. K. Lafrentz, Dec. 1927.

Type locality. Cafetal Concordia (900 m. alt., between Puerto Angel
and Salina Cruz), Oaxaca, Mexico.

Range. Guerrero, Oaxaca, and Chiapas, Mexico.

Diagnosis. Eyes usually visible; tentacular aperture very close to
eye; primaries 121-137; secondaries 101-121; difference 10-26; 1/d
26-40; length 153-430 mm.

Description. The eye is visible in nearly all specimens. I could not
make it out in the Mirador specimen, and Mertens failed to see it in
one of the type series of five, so that two out of 15 lack eyes. Only five
have the secondary count below 111, and three have the difference
over 16.

Remarks. Lafrentz (1928, Blatt. Aquar. Terr. 39, 6, p. 115) says
that the type series came from the "dungheap of the mule stable" and
that the native name is "metlapil." He gives a photograph.

USNM 115058 contained four well formed young 104 mm. long, and
4 mm. in diameter, 1/d 26. The eye was conspicuous in all.

This form with its visible eye, inner mandibular teeth, posterior ten-
tacle position, and nearly complete scalation, is presumably the most
primitive member of the genus.

Its relations are obviously with multiplicata, as all the characters

DUNN: THE AMERICAN CAECILIANS

469

overlap, although it has been possible to allocate all specimens without
recourse to locality.

Specimens seen, ten, as follows:

Mexico :
Guerrero :

Xaltianguis USNM 115057
El Limoncito, 15 km. N.
Acapulco, EHT 16869

Oaxaca :
Mirador AMNH 13448

Cafetal Concordia
Berlin 31696

" 31696

" 31696
No locality Vienna

Chiapas :

Tonala, E. H. Taylor
La Esperanza USNM 115058
30 km. N.E. Escuintla, 900 m.,
Mich 88205

Mertens' counts:

prim. sec. length diam. 1/d

133

121/11

153

4

38

127

106/8

275

7

39

128

114

373

11

no

34

eyes

127
125
130
137

101/7
111/8
117/13
121

275
350
430

8
10
15

34
35
29

121
121

106/16
103/11

331
283

9

7

37
40

121 106/11 335 12 28

124

111/10

295

13

26

type

125

114/9

390

15

26

126

112/10

430

15

29

125

111/12

400

14

28

121

111/9

295

11

27

no eyes

Gymnopis oligozona (Cope)

1877. Siphonops oligozonus Cope, Proc. Amer. Phil. Soc. 17, p. 91.

1879. Gymnopis oligozona Peters, Mon. Ak. Berlin, p. 939; Cope 1885, Proc.

Amer. Phil. Soc. 22, p. 171; Dunn 1928, Proc. New England Zool.

Club 10, p. 76.

Type. USNM 25187.

Type locality. Unknown.

Range. Known only from Guatemala.

Diagnosis. A Gymnopis without visible eyes; tentacular aperture

470 bulletin: museum of comparative zoology

remote from nostril; primaries 128-135; secondaries 62-74; 1/d 44-64;
length 255 mm. to 305 mm.

Description. Rather uniform dark, the primary grooves lighter
and the top of the head lighter. The type has 12 teeth on a side in the
upper jaw, ten on a side in the prevomero-palatine series, nine outer
mandibular and one inner mandibular. The tentacular aperture is
horseshoe-shaped, concave posteriorly, and quite far back as in multi-
pi i rata.

Specimens seen, three, as follows:

prim.

sec.

length

diam. 1/d

Guatemala:

Finca El Volean, Alta Vera

Paz U. Mich., Field No. 224

128

71/20

305

7 44

No locality. BMXH 87-4-12, 2

135

74/15

292

6.5 45

No data. USXM 25187 130 62/11 255 4 64

Remarks. The type is absolutely without any data at all. The Brit-
ish Museum specimen was received from the Basle Museum, and said
to have been collected in Guatemala by Bernouilli. A note in the
British Museum catalog says that Bernouilli visited Palenque, Flores,
and Lake Itza. The Michigan specimen, collected by L. C. Stuart,
gives at last a definite locality.

It is possible that the type of oligozona also served as the type of
Si phonops syntremus Cope. The geographic and anatomical relation-
ships of this form are clearly with G. multiplicata.

Gtmnopis xicefori Barbour
1924. Gymnophis nicefori Barbour, Proc. Biol. Soc. Washington 37, p. 125.

Type. MCZ 9609, collected by Hermano Xiceforo Maria, March
1924.

Type locality. Honda, Magdalena Valley, Colombia.

Range. Known from Honda, Girardot, and San Juan de Rio Seco
in the Magdalena Valley, and from Caceres in the Cauca Valley,
Colombia.

Diagnosis. A Gymnopis with eyes usually invisible, apparently not
under bone; tentacle very close to eye; primaries 133-158; secondaries
45-104; 43-88 primary folds without secondaries; 1/d 39-67; length
100-245 mm.

Description. The eye is visible in the specimens from Girardot and

DUNN: THE AMERICAN CAECILIANS 471

San Juan. The color is "dark slate color, head a little lighter" (Bar-
bour I.e.). The dentition of the type is given by Barbour as "maxillary
teeth many, apparently about thirty; mandibular probably about equal
in number, in two rows [I find no inner row in the type]." The Girardot
specimen has 7 premaxillary-maxillary teeth on a side; 12 palatine;
10 outer mandibular; 0 inner mandibular. The mandibular teeth are
twice the size of the maxillary or the palatine.

The Honda and Girardot specimens (4) have 150-158 primaries;
97-104 secondaries; 1/d 39-67. The Caceres specimen has only 138
primaries. The San Juan specimen is tiny, ill-preserved for counting,
but appears to have 133 primaries and 45 secondaries.

Remarks. Probably directly allied to unicolor of Guiana, but just as
similar to oligozona of Guatemala. Its relationships with albiceps are
obscure.

Specimens seen, six, as follows :

prim. sec. length diam.l/d

Colombia :

Honda MCZ 9609
" AMNH 23387

153
158

104/60
102/63

193
233

5
5

39

Type

47

" AMNH 23388

150

97

178

4

44

Girardot Inst. La Salle

152

100

200

3

67

San Juan de Rio Seco

MCZ 16089

133

45

100

2

50

Caceres Berlin 9524

138

95/67

245

6

41

Gymnopis unicolor (Dumeril)

1863. Rhinatrema unicolor Dumeril, Mem. Soc. Cherbourg, 9, p. 321.

1863. Rhinatrema concolor Dumeril, I.e., pi. 1, f. 6-7.

1879. Gymnopis unicolor Peters, Mon. Berlin Ak., p. 939; Boulenger 1882,

Cat. Batr. Grad. Brit. Mus. (2), p. 100; 1895, Proc. Zool. Soc.

London, p. 410; Nieden 1913, Gymnophiona, p. 21.

Type. Paris 6 (three specimens one of which is labeled "type").

Type locality. Cayenne.

Range. Known only from Guiana.

Diagnosis. A Gymnopis with invisible eyes; primaries 100-120;
secondaries 22-74; 41-87 primary folds without secondaries; 1/d 27-
40; length 108-235 mm.

Description. Uniform dark.

472

bulletin: museum of comparative zoology

Remarks. AMXH 1335, from "S. Amer.," has 22 fewer secondaries
than have nine Guiana specimens, lowers their secondary count from
44 to 22, and raises the number of primary folds without secondaries
from 64 to 87. Additional similar specimens, with locality, might be
the basis for a different form.

Specimens seen, 15, as follows:
Cavenne Paris 6

BMXH 84-12-8, 5
Berlin 9600

Guiana

Paris 6a

"

" 6b

(<

" 6c

u

" 6c

a

" 6c

a

" 6c

<<

" 6c

a

" 6d

Oke R., Cuvuni Trib., Brit. Guiana

Field 35117
"S. Amer." AMNH 1335

>rim.

sec.

length

diam.

1/d

108

67/45

195

6

32

113

63/45

205

6.5

31

type

110

55/44

185

5.5

34

115

68

230

6

38

109

58

235

7

33

160

5

32

114
120

187
200

6
6

33
33

56/22

118

74

195

5

39

113

59

200

6

33

154

5

30

109
108

4
3

27
36

100

44/13

162

4

40

109

22/5

189

5

34

Gymxopis albiceps (Boulenger)

1882. Dermophis albiceps Boulenger, Cat. Batr. Grad. Brit. Mus. (2), p. 98,
pi. 8, f. 1.

Type. BMXH 80-12-5, 147.

Type locality. Ecuador.

Range. Known only from Prov. Santiago Zamora in the Oriente.

Diagnosis. A Gymnopis with visible eyes; tentacle between eye
and nostril, slightly nearer eye; primaries 124-125; secondaries 45-55;
1/d 35-46; length 177-210 mm.

Description. "Blackish gray, the head white" (Boulenger, I.e.).

Remarks. This is the only South American Gymnopis with the ten-
tacle remote from the eye. It has more primaries than any other form
with a similar tentacle position. In counts of rings and in proportion

DUNN: THE AMERICAN CAECILIANS 473

it is close to and somewhat intermediate between the Guianan unicolor
and the Colombian nicefori, which have the eye usually invisible and
the tentacle very close to the eye.
Specimens seen, two, as follows:

prim. sec. length diam. 1/d
Ecuador:

No data, BMNH 80-12-5, 147 125 55 210 4.6 46

Prov. Santiago Zamora. Michigan
83051 124 45/15 177 5 35

Gymnopis mexicana mexicana (Dumeril and Bibron)

1841. Siphonops mexicanus Dumeril and Bibron, Erpet. Gen. 8, p. 284;

Cuvier 1849, Regne Animal (3), pi. 36, f. 1, 6; Dumeril 1863, Mem.

Soc. Cherbourg 9, p. 318, pi. 1, f. 10; Brocchi 1882, Miss. Sei. Mex.,

p. 120, pi. 21, f. 2.
1850. Siphonops mexicana Gray, Cat. Batr. Grad. Brit. Mus., p. 59.
1879. Dermophis mexicanus Peters, Mon. Ak. Berlin, p. 927, f. 6; Cope 1879,

Proc. Amer. Phil. Soc. 18, p. 265; Boulenger 1882, Cat. Batr. Grad.

Brit. Mus. (2), p. 98, pi. 8, f. 2; Cope 1885, Proe. Amer. Phil. Soc. 22,

p. 171; 1887, Bull. U. S. Nat. Mus. 32, p. 9; 1888, Journ. Morph. 2,

2, p. 300, pi. 22, f. 6 (otic region); 1889, Bull. U. S. Nat. Mus. 34, pi.

51, f. 21 (hyoid); Boulenger 1895, Proc. Zool. Soc. London, p. 404;

Giinther 1902, Biol. Centr. Amer., p. 305; Nieden 1913, Gymno-

phiona, p. 8; Ochoterena 1932, Ann. Inst. Biol. [Mexico] 3, 4, p. 363

(integument).
1928. Dermophis mexicanus mexicanus Dunn, Proc. New England Zool.

Club 10, p. 74, pi. 5 (breeding habits).

Type. Paris 5c

Type locality. Mexico.

Range. Oaxaca and Vera Cruz, Mexico, to western Nicaragua.

Diagnosis. A Gymnopis with visible eyes; tentacle between eye and
nostril, slightly nearer eye; primaries 97-110; secondaries 51-80; 1/d
14-26; length 152-485 mm.

Description. The belly is usually light in color. The scales appear
first in the posterior half of the segmental folds on the sides (after the
eleventh primary in U. Mich. 64354a from Guatemala). They are
present in both halves, dorsally, laterally and ventrally in the posterior
part of the body where the secondaries are present and complete.
Mich. 64354a has 7-8 premaxillary teeth ; 11-12 maxillary; 17-18 pala-
tine; 15 mandibular; no inner mandibular. USNM 51380 has the

474 bulletin: museum of comparative zoology

tentacle equidistant from eye and nostril; U. Mich. 64354a has the
tentacle 3 mm. from the eye and 4 mm. from the nostril, which is the
usual position.

Sixty-two specimens have been counted for primaries and second-
aries. The extremes are: five specimens with 97, 100, 110, 110, 110
primaries from Nicaragua, Central America, Nicaragua, Guatemala,
and Mexico respectively; 57 specimens fall into the narrow range
of 101-109 primaries. The extremes in secondary count are: 51, 51, 52,
78, 80, from "N. E. Mexico," Tabasco, Vera Cruz, Tehuantepec and
Chiapas respectively; 57 specimens have from 55-75 secondaries. No
sexual difference has been found in the counts of secondaries, pri-
maries, or complete secondaries.

The length-diameter ratio (always somewhat untrustworthy) has
been computed for 51 specimens. An adult pregnant female is, natur-
ally, the fattest, with the low ratio of 14. The slimmest are 23, 24, 25,
26 from Mexico, and one with 24 from Salvador. The four fattest are
two from Mexico and two from Guatemala. Aside from the pregnant
female, no sexual difference can be made out, and no changes in pro-
portions with age are apparent.

None of the figures on proportions or segment counts give any indi-
cation of a geographical trend.

Habits. The animal is viviparous. MCZ 12122 from Guatemala
was a pregnant female 430 mm. long (Dunn 1928, pi. 5). It had six
young in the left oviduct and four in the right. The young were
145 mm. long.

Remarks. This form with its three races resembles in tentacle posi-
tion parviccps from Panama and albiccps from Ecuador.

Specimens seen, 66, as follows :

prim. sec. length diam. 1/d

Mexico :

Vera Cruz:

Cuatotolapam Mich. 41571

107

72/12

327

19

17

41572

105

70/12

374

21

18

Vera Cruz ANS 4886

105

55/6

355

17

21

4887

105

55/6

228

11

21

4888

105

55/6

—

—

4889

106

59/6

340

19

18

4890

104

59/6

232

11

21

" AMNH6306

102

58/12

354

18

20

DUNN: THE AMERICAN CAECILIANS 475

prim. sec. length diam. l/d

Oaxaca :

Tehuantepec MCZ 1604a

108

69/15

281

17

16

1604b

109

73/11

425

17

25

Hamburg 310

103

61

—

—

Barrios DSNM 30535

106

78/11

378

24

16

30536

105

60/0

170

7

24

30537

—

Tabasco :

Tabasco USNM 25102

105

61/9

365

18

20

Teapa BMNH 1907-12-19, 135

104

51/7

—

—

Chiapas:

La Zacualpa AMNH 897

102

72/5

—

—

898

107

74/9

—

—

899

105

69/9

—

—

23^ km. W. Soconusco, 50 m.

Mich. 88203

102

73/8

280

15

19

6 mi. NE. Escuintla, 150

m. Mich. 88202

103

74/14

350

23

15

88204

106

80/10

310

17

18

? State ?

Finca Berlin 24051

—

460

30

15

St. Augustin Paris 5

104

69/10

475

27

17

a

101

62/6

387

19

20

a

106

73/10

388

17

23

"N.E. Mexico" ERD

110

51/5

395

15

26

"Mexico" Paris 5c

105

60/5

365

20

18

Berlin 9104

—

—

AMNH 13445

104

58/2

210

12

17

13446

105

61/2

283

13

22

13813

104

63/2

320

16

20

Guatemala :

No locality USNM 25641

105

74/10

440

23

19

Hamburg 1926

103

70/6

—

—

Senck, 2098b

103

73/9

—

—

MCZ 12121

108

62/9

420

23

19

12122

109

69/10

430

30

14

12123

10S

70/11

240

15

16

476 bulletin: museum of comparative zoology

Pacific side BMNH 64-1-26, 397
64-1-26, 152
Escuintla USNM 12691
Retalhuleu Senck, 2098a 107 72/16

Finca El Cipres, Volcan Suchil,

Prov. Suchetepequez Mich.

64354

prim.

sec.

length

diam.

1/d

106

68/10

463

24

19

110

62/7

390

18

22

MCZ 11222

11223

Salvador :

Volcan Isalco ANS 4925
No data Mus. Na. Salvador
BMNH 1906-11-8, 2

Honduras :

Amapala USNM 51380
No data Berlin 13207

Nicaragua :

Pol von MCZ 1491a
1491b
2165a
2165b
AMNH 1153
18667
No data USNM 16147
Mich. 65674

(?) Panama:

' No data BMNH 67-9-23, 3 104 63/10 430 21 20

Central America:
No data USNM 30008 100 64/9 345 18 19

It has also been recorded from Atitlan, Guatemala, by Brocchi

(1883).

105

65/9

350

—

—

102

65/14

485

29

17

105

65/10

210

13

16

104

66/11

222

14

15

106

73/9

160

10

16

106

73/8

172

10

17

101

64/8

280

15

19

108

67/8

460

22

21

105

65/5

395

21

18

110

75/8

165

7

24

105

73/15

152 '

7

22

107

64/7

395

20

20

107

68/7

350

18

19

103

61

105

60/11

107

57/7

350

16

22

104

54/10

185

9

20

106

70/15

395

20

20

97

55/7

261

16

16

105

59/12

305

16

19

110

62/13

356

19

19

106

66/5

336

19

17

DUNN: THE AMERICAN CAECILIANS 477

Gymnopis mexicana clarkii (Barbour)

1926. Gymnophis clarkii Barbour, Occ. Papers Boston Soc. Nat. Hist. 5,

p. 191.
1928. Dermophis mexicanus clarkii Dunn, Proc. New England Zool. Club

10, p. 73.

Type. MCZ No. 11047, collected by Dr. Herbert Clark, June, 1925.

Type locality. Tela, Honduras.

Range. Known only from Tela and San Pedro Sula, Honduras.

Diagnosis. A Gymnopis with visible eyes; primaries 101-107;
secondaries 41; 1/d 16-19; length 145-420.

Description. Only four specimens are known, so that little can be
added to the diagnosis save that the color is "as usual, plumbeous,"
and that the tentacle is well in advance of the eye.

Remarks. The type was 145 mm. long, and since this is the length
of an unborn embryo of mexicana mexicana, it must have been very
young. The primary count is that of mexicana, and only the low
secondary count (41 as against 51-80) distinguishes it.

Specimens seen, four, as follows:

Honduras :

Tela, MCZ 11047

" 11779
San Pedro Sula, AMNH 33386
No locality AMNH 49953

Gymnopis mexicana gracilior (Giinther)

1902. Dermophis gracilior Glinther, Biol. Centr. Amer., Amph., p. 306, pi.

76, f. B; Nieden 1913, Gymnophiona, p. 9.
1928. Dermophis mexicanus gracilior Dunn, Proc. New England Zool. Club,

10, p. 73.

Type. BMNH 1901-12-19-137.

Type locality. Chiriqui, Panama.

Range. Pacific slope of Costa Rica; Chiriqui, Panama. Sea level to
4000 feet.

Diagnosis. A Gymnopis with visible eyes; primaries 95-102;
secondaries 32-78; 1/d 25-32; length 192-345 mm.

prim.

sec.

lengtn

diam. i/a

107

41/0

145

9 16

TYPE

107

41/4

420

25 17

101

41/5

350

18 19

104

41

380

— —

478 bulletin: museum of comparative zoology

Description. The color and the tentacle position are as in G. m.
mexicana.

Remarks. Three out of four specimens of this species are slimmer
than any mexicana seen. The exception is a pregnant female which is
as stout as the slimmest mexicana seen. This individual contained six
well formed young which measured 100-106 mm., and about 6 in
diameter, yolk still being noticeable. The Panama specimens have the
usual secondary count of mexicana. but the single Costa Rican one
has a very low count.

Specimens seen. four, as follows:
Costa Rica: prim. sec. length diam. 1/d

Pozo Azul BMNH

1907-6-28, 27 100 32 192 6 32

Panama :

( hiriqui BMNH

1901-12-19, 137 95 73/10 343 11 31

Boquete Cal. Acad.+Sci.

79463 99 68/8 325 13 25
Boquete Cal. Acad.+Sci.

79464 102 78/9 345 11 31

Gtmnopis parviceps (Dunn)

1924. Siphonops parviceps Dunn, Occ. Papers Boston Soc. Nat. Hist. 5, p. 93;
1928, Proc. New England Zool. Club, 10, p. 74 (breeding habits).

Type. MCZ 9407, collected by E. R. Dunn and Chester Duryea,
Aug. 6, 1923.

Type locality. La Loma (or Buena vista, another name), at elevation
of 1200 feet (erroneously 2000 in original description), on the trail from
Chiriqui Lagoon to David, Atlantic slope in Province of Bocas del
Toro, Panama.

Range. Known only from type locality.

Diagnosis. 96 primaries; 13 secondaries; 1/d 22.

Description. Primary folds all complete, extending to anus; second-
ary folds 13, first three incomplete; scales present all over in region of
complete secondaries; maxillary teeth 13, palatine teeth 10, mandibu-
lar teeth 10; tentacle between eye and nostril, nearer to lip than to
either, slightly nearer to eye than to nostril; eye nearer to lip than to

DUNN: THE AMERICAN CAECILIANS 479

tentacle, nearer to lip than is the nostril; eyes farther apart than length
of snout. Black; head lighter, tinged with brown. Length 180 mm.;
diameter of head 5 mm., neck 5 mm., body 8 mm.; posterior angle of
mouth to tip of snout 6 mm. ; ratio of length to diameter 22.

Habits. We were eating breakfast in a palm thatch hut when one
of our men called attention to a "snake" which was coming out of the
ground under the raised platform on which we slept. The whole ter-
rain was steep slopes. The animal was impossible to extricate from its
burrow by pulling, and was dug out. The peculiar bottle-shape of the
beast (possibly because it was a pregnant female) was immediately
noticeable and was the cause of the difficulty of extraction. Later,
three perfectly formed young were found in the right oviduct. They
measure 76 mm., and the diameter is about 3.5 mm., 1/d 21.

Remarks. My lack of knowledge of South American forms and of
the correlations of scales and secondaries in Caecilians led me to place
this form originally in Siphonops. The eye and the tentacle are nearer
the lip than in Gymnopis mexicanus mexicanus, but the relative dis-
tance of tentacle, eye, and nostril is the same in both.

The low secondary count makes the species remarkably distinct.

Siphonops "Wagler
1828. Siphonops Wagler, Isis 21, p. 742 (monotype Caecilia annulata Mikan).

Diagnosis. Caecilians with no tail; no secondaries; no scales; ten-
tacle on side of head, between eye and nostril, nearer to eye and usually
very close to it; no inner mandibular tooth row; no dorsal fin; anal
region not a sucking disk; eye usually visible; primaries 81-133; 1/d
ratio 16-54; length 126-535 mm.; five forms.

Range. Colombia to southern Brazil and Paraguay. Argentina (?)

Key to species of Siphonops

A. Large species; primaries grooves white.

B. Primaries 81-100; black and white in preserved specimens.

annulatus
BB. Primaries 102-118; brown and white in preserved specimens.

paulensis
AA. Small species; unicolor.

B. Primaries 95-104 hardy i

BB. Primaries 108-112 insulanns

BBB. Primaries 120-133 brasiliensis

•480 bulletin: museum of comparative zoology

Remarks. S. -paulensis occurs within the range of annulatus, and
occupies dry regions back of the coast range.

S. insulanus, which is intermediate between hardyi and brasiliensis,
is an insular form, while the two more extreme mainland forms may
occur together.

I have examined 253 specimens of Siphonops, and these include the
types of annulatus, crassus, paulensis, and hardyi. I have not been
able to examine the types of interrupta, brasiliensis, insulanus, macula-
tus, or marmoratus.

Siphoxops annulatus (Mikan)

1820. Caecilia annulata Mikan, Delect. Flor. Faun. Bras., pi. 11; 1924 Spix,
Serp. Bras., p. 74, pi. 26, f. 1; 1829 Cuvier, Regne Anim. (2), 2,
p. 100; 1831 Gray, in Griffith's Cuvier's Anim. King. 9, App., 110.

1828. Siphonops annulatus Wagler, Isis 21, p. 742, pi. 10, f. 1, 2; 1830, Nat.

Syst. Amphib., p. 198; 1838 Tschudi, Mem. Soc. Sci. Nat. Neuchatel
2, p. 90; 1841 Dumeril and Bibron, Erp. Gen. 8. p. 282, pi. 85, f. 1;
1863 Dumeril, Mem. Soc. Sci. Nat. Cherbourg 9, p. 317, pi. 1, f. 2;
1868 Cope, Proc. Acad. Nat. Sci. Philadelphia, p. 118; 1879 Peters,
Mon. Ak. Berlin, p. 940, f. 10: 1879 Wiedersheim, Anat. Gym. pi. 1,
f. 1-13, pi. 2, f. 27, 32-34, pi. 3, f. 37-44, pi. 7, f. 82, pi. 9, f. 83;
1882 Boulenger, Cat. Bat. Grad. Brit. Mus. (2), p. 102, pi. 8, f. 4;
1889 Cope, Bull. U. S. Nat. Mus. 34, pi. 53, f. 1, pi. 56, f. 3; 1892
Boettger, Kat. Amph. Mus. Senckenberg. p. 62; 1895 Boulenger,
Proc. Zool. Soc. London, p. 412; 1899 Goeldi, Zool. Jarhb. Syst. 12,
p. 120, pi. 9, f. 1-4; 1911 Ihering, Rev. Mus. Paulista 8, p. 108, f. 1,
2, 3, 6, 7; 1912 Phisalix, Cong. Int. Zool 8 (Graz), pi. 4, f. 5 (integ.);
1913 Nieden, Gymno., p. 25, f. 1; 1915 Spengel, Blatt. Aq. Terr. 26,
p. 220; 1927 Muller, Abh. Senckenberg. Mus. 40, p. 260; 1936 Sawaya,
Rev. biol. hyg. (2) 7, p. 80, pi. 7; 1937 Bull. Univ. Sao Paulo 1,
Zool. 1, pi. 30, f. 1-2, pi. 32, f. 13-15.

1829. Caecilia interrupta Cuvier, Reg. Anim. (2), 2, p. 100.

1863. Siphonops indistinctus Dumeril, Mem. Soc. Sci. Nat. Cherbourg 9,

p. 318 (in part, the dried specimen).
1885. Dermophis crassus Cope, Proc. Amer. Phil. Soc. 22, p. 184 (upper Beni

R., Bolivia).
1937. Siphonops annulatus marmoratus Sawaya, Bull. Univ. Sao Paulo 1,

Zool. 1, p. 238, pi. 30, f. 4-5; pi. 31, f. 7 (Theresopolis, Rio de Janeiro,

Brazil).

Type. Paris 15.

Type locality. Sebastianopolis, Brazil.

Range. From "Argentina or Paraguay" and Rio Grande do Sul,

DUNN: THE AMERICAN CAECILIANS 481

Brazil, to Bolivia, Guiana, Venezuela, and the eastern part of Peru,
Ecuador, and Colombia. Upper Cauca River, Colombia.

Diagnosis. A Siphonops with white primary grooves; eye distinct;
primaries 81-100; 1/d 16-43; length 126-535 mm.

Variation. In Mus. Nac. Brazil 831, from Theresopolis, Rio de
Janeiro, the tentacle is slightly nearer the nostril than the eye, the
snout is unusually long, and the hind end of the body is acuminate.
In six from Serra de Mache, Rio de Janeiro, Mus. Paul. 940 A-E, and
in one from "Brazil," Paris 17, the tentacle is almost equidistant be-
tween eye and nostril. In proportions and primary count these speci-
mens do not differ from others which have the tentacle in the normal
position closer to the eye.

The primary count ranges from 85 to 95 in 171 specimens (160 seen
and 11 reported). Six specimens (four seen and two reported) have
81-84, and these are all from the southern part of the range. Seven
specimens (five seen and two reported) have 96-100, and five of these
are from the western and northern parts of the range.

The majority of the measured specimens (79) have the 1/d ratio
from 20-30. The 20 stouter specimens ((1/d 16-19) include the five
smallest and three of the twelve largest; and seven out of eleven
Colombian specimens. The seven slim specimens (33-43) are all from
the south, and the slimmest is dried somewhat.

Remarks. Sawaya's marmoratus is a color variety, not a geographical
race.

Goeldi (1899) says it lives by preference in dry localities. He speaks
of a female found rolled up under an old stump in a very dry place at
Colonia Alpina, near Theresopolis, in the Organ Mts., Rio de Janeiro.
In the middle of the coil was a clump of six eggs, in a continuous string,
from each end of which there was a free, thread-like, projection. The
eggs measured 10 by 8.5 mm. The contained embryos were 4 mm. in
diameter, with two external gills on the left side and three on the right.
The find was made in December.

Sawaya (1936) speaks of this species being eaten by the snake
Pseudoboa deli a.

The only close ally is paulensis. The two are known to occur to-
gether in two localities. S. annulatus has a vastly wider range, which
practically encloses that of paulensis.

Specimens seen, 175, as follows:

prim. length • diam. 1/d
Argentina or Paraguay:

Xo locality, Hamburg 1064 87 175 5 35

482

bulletin: museum of comparative zoology

Brazil :

prim. length diam. 1/d

banta (_ athanna, Joinv

nle, Viem

la 87

—

Joinville AMNH 23693

88

322

15

21

prim, length

diam.

1/d

Sao Paulo,

Taubate, Mus. Paul.

942

86

278

14

20

Franca "

953

95

400

24

17

Interior MCZ 10782

93

201

7

29

No locality

Hamburg 911-912

85(4)

(35 spec.)

86 (4)
87(5)

it

88 (7)
90(8)
91(5)

tt

92 (2)

AMNH 23470

86

23471

86

23472

86

335

23473

90

350

23474

87

23475

88

275

7

39

23476

87

23477

86

Berlin 5968
Munich 140/1912

Rio de Janeiro :
No locality
MCZ 290
" 328
Berlin 3704
Krakau 14671

BMNH 74-5-21-7
AMNH 23503
Serra de Macahe,
Mus. Paul. 940
940A
940B
940C

87

185

9

20

92

375

14

27

84

310

17

19

91

405

12

34

94

—

—

92

—

—

89

325

13

25

84

327

18

18

85

242

10

24

82

179

9

19

85

185

9

20

85

253

12

21

DUNN: THE AMERICAN CAECILIANS

483

Mus. Paul. 940D
940E
Petropolis
Vienna
MCZ 2481

Munich 140/1912

Theresopolis
Munich

Mus. Nac. Brazil 540
831
Neu Friburgo
Hamburg 1093
USNM
USNM
.USNM
USNM

Espiritu Santo:

Sta. Tereza, 700 m.
Mus. Nac. Brazil 842
Sta. Tereza, 760 m.,
Mus. Nac. Brazil 843
Pau Gigante, Mus. Nac.
Brazil 847

No locality

Hamburg 1353

a it

Frankfort 21026
Berlin 14043
Vienna

a

Mts. between Espiritu Santo
and Minas Garaes,
Hamburg 1354

prim.

length

diam.

1/d

86

225 ,

10

22

91

378

14

27

86

—

—

88

145

9

16

85

205

13

16

83

—

—

89

325

13

25

93

350

17.5

20

93

290

11

26

91

305

13

23

94

320

15

21

89

—

—

87

—

—

92

88

—

—

91

370

18

21

92

147

8

18

88

295

16

18

90

350

11

33

89

171

6

28

89

—

—

87

385

10

38

89

160

6

26

91

190

8

23

93

395

14

28

89

355

15

24

91

350

15

23

93

360

12

30

91

300

9

33

484

bulletin: museum of comparative zoology

prim.

length

diam.

1/d

Hamburg 1354

94

245

9

27

a n

90

210

8

28

It It

90

178

8

22

it a

93

165

5.5

30

Minas Geraes,

Mendez, on Rio Jequitinhonha,

Vienna

92

—

■ — -

Bahia.

No locality,

MCZ 1528

95

367

16

23

AMNH 18668

88

356

14

25

23502

93

283

16

18

Vienna

93

((

90

Berlin 9526

93

315

11

29

Hamburg 1355

92

382

13

29

Paris 15e

100

157

7

22

BMNH 61-3-23-20

92

62-1-30-62

91

171

7

24

69-2-22-6

95

225

9

25

1924-9-20-0

93

210

10

21

Amazonas, Tabatinga,

Vienna

97

—

■ —

Lagoa Japaranao, near Teffe

MCZ 1520

88

360

16

22

? State?

Tozuzu,

Berlin 7169

95

128

6

21

Brazil, no locality:

Paris 15

86

—

Type

" 15d

94

450

24

19

" 17

91

261

6

43

dry

AMNH 23501

84

284

13

22

Hamburg 1094

90

323

13

24

Frankfort 2102a

89

—

—

a a

88

364

17.5

20

USNM 58749

93

190

18.5

22

DUNN: THE AMERICAN CAECILIANS

485

Bolivia:

Near Riberalta AMNH 15000
Upper Beni R.
MCZ 6636

AXS 11344

Peru:

No locality

ANS 11346

Moyabamba

BMNH 74-8-4,5
74-8^4, 6
Iquitos, AMNH 42850
E. of Contamna, Peru
Brazil frontier
AMNH 42835
San Antonio, Rio Itaya

prim. length diam. l/d
93

98 300 12 25

cotvpe Dermophis crassus

94 348 13 27

eotype Dermophis crassus

92

420 17 25

eotype Dermophis crassus

94

425 18 24

94

385 16 24

94

420 21 20

94

280

11

25

AMNH 42842

91

337

13

26

42843

92

378

17

23

42844

92

310

13

24

42845

91

413

15

21

42846

93

380

15

25

42847

91

410

18

23

42848

90

380

16

24

42849

89

263

11

24

Pampa Hermosa, middle Ucayali,

mouth of Cushatabay

AMNH 42838

96

185

7

26

42839

93

300

14

21

Rio Cenipa, upper Marafion

AMXH 42836

89

295

13

23

42837

90

225

10

22

Mouth of Santiago, upper

Marafion

AMNH 42833

89

280

11

25

42834

93

425

17

25

Ecuador :

Sarayacu U. Michigan 89460

91

—

—

BMNH

92

148

9

16

486

bulletin: museum of comparative zoology

Pastaza R. (Canelos to
Marafion R.)
MCZ

prim. length diam. l/d

it

No locality

Berlin 9814
AMNH 17448
U. Michigan (5)

Colombia :

Villavicencio Inst. La Salle

(C (t 11

AMNH 23270
23171
Medina Mts., N. E. of Villa-
vicencio
AMNH 49955
49956
49957
49958
Guaicaramo

AMNH 23384
23385
23386

90

160

9

18

87

245

13

19

90

290

14

21

91

385

13

30

92
93
96

345

14

25

—

—

94

375

19

20

90

360

17

21

91

345

18

20

94

200

10

20

87

535

20

27

93

126

7

18

89

290

18

16

85

355

19

19

91

415

20

21

95

365

15

24

94

340

15

23

91

347

18

19

92

257

13

20

92

430

23

19

87

144

8

18

85

171

10

17

Cayenne :

No locality,
Paris 15c

86

Surinam:

No locality,
Paris 15b

90

DUNN: THE AMERICAN CAECILIANS

487

prim. length diam. 1/d
Venezuela :

Barinas, Zamora Pro v.,

Munich 94 320 12 27

South America :
No locality

AMNH 49975 91 336 15 22

Besides the localities listed above, Ihering (1911) has recorded
annulatus from the following places in Brazil: Rio Doce, Espiritu
Santo; Pelotas, Rio Grande do Sul; the State of Matto Grosso.
Spengel (1915) records it from Para.

Siphonops paulensis Boettger

1892. Siphonops paulensis boettger, Kat. Batr. Mus. Senckenbergianum.

p. 62; Boulenger 1896. Proc. Zool. Soc. London, p. 412; Ihering 1911.

Rev. Mus. Paulista 8, pp. 91, 92, 109; Nieden 1913. Gymnopiona,

p. 25; Serie 1918-19, Physis, 4, 17, p. 361; Sawaya 1937, Bull Univ.

Sao Paulo 1, Zool. 1, p. 238, pi. 31, f. 11.
1937. Siphonops paulensis maculatus Sawaya; Bull, Univ. Sao Paulo, 1,

Zool. 1, p. 240 (Theresopolis, Rio de Janeiro, Brazil).

Type. Mus. Senck. 2102, lb.

Type locality. Sao Paulo, Brazil.

Range. States of Rio Grande do Norte, Goyaz, Matto Grosso, Rio
de Janeiro, and Sao Paulo, Brazil; Villarica, Paraguay; Sta. Cruz, and
Buenavista, Bolivia; San Ignacio, Missiones, Argentina.

Diagnosis. A Siphonops with white primary grooves; primaries
102-118; 1/d 22-39; tentacle anterior to and a little below eye; eye
distinct; length 139 to 480 mm.

Description. Boettger described paulensis as slimmer than annu-
latus; with more primaries; smaller head; different tentacle position;
different color.

Eight specimens under 300 mm. in length have the 1/d ratio 22-28;
sixteen between 300-350 have it 23-38; eleven between 350-400 have it
25-38; nine over 400 have it 27-39. Forty-two annulatus from the
south have it 16-39, six above 30. This character is not diagnostic.
Boettger gives an 1/d of 32 for paulensis.

Boettger gives a range of primaries of 110-115. Ihering says 20 Sao
Paulo specimens had 114-116, except for one with 111. I count 106-117
on eighteen Sao Paulo specimens and 104-110 on seven from Goyaz.

488 bulletin: museum of comparative zoology

Nine from Paraguay have 104-116; seven from Bolivia have 102-113.
The maximum count in southern annulatus is 96, the minimum 81.

Sawaya (1937) has recorded a maximum of 118 for paulensis.

The primaries are interrupted dorsally in some of the Paraguay
series.

Preserved specimens are brown, while annulatus is black.

The size of the head I cannot see to be different from that of annula-
tus.

The position of the tentacle seems to me to be exactly that of
annulatus, with the exception that some annulatus have it quite far
from the eye.

The number of rings is the best differentiating character, but since
three annulatus from Bahia, Upper Beni R., and Tabatinga, have
100, 98, and 97 annuli respectively, this difference may be very slight.

Habits. Ihering (1911) says it is the commonest species in the
environs of Sao Paulo City, is found in "dry places such as the range
of Ypiranga," lives in ant hills but does not eat the insects, and has had
the egg capsule of a spider in its stomach.

Remarks. Its only ally is annulatus, to which it is remarkably
close. Both occur in the states of Sao Paulo and Matto Grosso, and
in Paraguay and Bolivia. The only locality from which I have seen
both is Taubate in Sao Paulo. The two specimens recorded from there
were approximately of the same proportions; annulatus 278/14= 19.8;
paulensis 286/13 = 22. The annulatus had 86 primaries ; the -paulensis
109. Both are reported by Sawaya from Theresopolis, Rio de Janeiro.
Paulensis seems to be more an inhabitant of the high interior, although
its range seems to be completely surrounded by the range of annulatus.
The range has a remarkable similarity to that of Cnemidophorus
ocellatus, and it is probable that it inhabits the savanna country which
stretches back of the coast from Rio Grande do Norte southwest to
Sao Paulo.

Sawaya's maculatus is not a race but a variety, as it is an occasional
occurrence in the midst of normal paulensis.

Specimens seen, 44, as follows:

Brazil: prim. length diam. 1/d

Sao Paulo:

Taubate, Mus. Paul. 1013

109

286

13

22

Ypiranga " " 939

112

398

16

25

" " 947

110

363

13

28

" " 947A

114

360

16

22

DUNN: THE AMERICAN CAECILIANS

489

prim.

length

diam.

1/d

Ypiranga Mus. Paul. 947B

114

380

11

34

" " 949

110

317

13

24

" " 949A

114

334

13

26

u a 951

106

326

14

23

" " 951A

108

341

9

38

" " 951B

110

330

13

25

" " 956

117

303

12

25

" " 956A

112

227

11

24

Sao Paulo AMNH 23624

111

373

12

31

Frankfort 2102, la

115

418

13

32

2102, lb

110

340

12

28

TYPE

2102, lc

115

380

10

38

BMNH 94-7-25, 9

111

420

12

35

Munich 283/1920

111

330

10

33

Matto Grosso:

Corumba BMXH 92-4-20, 23

109

190

7

27

Goyaz :

Annapolis, 1000 m.,

AMNH 43855

106

470

15

31

43856

105

440

15

29

43858

106

139

5

28

43859

104

455

17

27c/1

43860

110

480

13

37

43861

108

362

14

26

43862

106

153

6

25

Rio Grande do Norte:

Ceara Mirim, Cal. Acad. Sci. 49897 113

290

12

24

Paraguay:

Villarica AMNH 19920

116

395

16

25

19921

115

380

13

29

19922

116

358

13

27

19923

115

314

12

26

19924

113

398

13

31

19925

107

304

10

30

19926

116

305

10

30

19927

115

340

12

28

No locality AMNH 23433

104

461

16

39

490 bulletin: museum of comparative zoology

T, ,. . prim, length diam. l/d

.Bolivia : '

Sta.Cruz. 500 m., Carnegie 11598 108 217 9 24

" 11599 106 166 7 24

2643 102 302 9 33

Buenavista BMNH 1927-8-1, 135 105 332 10 33

136 113 435 15 29

137 109 328 9 36

138 110 317 9 35

No data.

Berlin 4 112 450 13 34

Ihering (1911) has recorded a specimen in the Museu Paulista from
Raiz de Serra, Sao Paulo. Sawaya (1937) has recorded paulensis from
Theresopolis, Rio de Janeiro, Brazil. Serie (1918-19) has recorded
paulensis from San Ignacio, Missiones, Argentina.

Siphoxops hardyi Boulenger

1888. Siphonops hardyi Boulenger, Ann. Mag. Nat, Hist. (6), 1, p. 189;
1891, Ann. Mag. Nat. Hist. (6), 8, p. 457; 1895, Proc. Zool. Soc.
London, p. 412, pi. 24, f. 3; Ihering 1911, Rev. Mus. Paulista, 8, p.
109; Nieden 1913, Gymnophiona, p. 26.

Type. BMNH No. 87-12-29-39, collected by M. F. Hardy de
Drenduf.

Type locality. Porto Real, Rio de Janeiro, Brazil.

Range. The states of Rio de Janeiro and Sao Paulo, Brazil, in moun-
tains of the coast ranges.

Diagnosis. A Siphonops of uniform color; primaries 95-104; l/d
27-45; eye distinct; tentacle a little anterior to and below the eye;
length 136-178 mm.

Description. "Teeth small, subequal"; "uniform blackish" (the
Ypiranga specimen is gray, lighter below); tentacle very near eye;
primaries complete; Ihering gives 100 primaries for a specimen from
Ypiranga, and 95 for one from Serra de Macahe. Boulenger (1895)
says "eye more or less distinct"; and "tentacle close to and very slightly
below eye."

Habits. Not known.

Remarks. This is the shortest species of the hardyi-insulanus-
brasiliensis group. The proportions are much the same in all three, the
tentacle position is not diagnostic, although Boulenger (1891) states

DUNN: THE AMERICAN CAECILIANS

491

that the tentacle is closer to the eye in hardyi than it is in brasiliensis.
The only real differences between the three are the number of pri-
maries. The maximum number in hardyi is 104, insulanus has 108-112,
and brasiliensis has 120-133. Both hardyi and insulanus are quite
small (max. length 178 mm., and 200 mm. respectively) while brasil-
iensis reaches a length of 312 mm.
Specimens seen, nine, as follows:

Rio de Janeiro :
Porto Real
BMXH no.

87-12-29-39

prim.

104

91-6-16-14 102

91-6-16-15 100

Mambucaba, Mus. Nac. Brazil 841 103

Organ Mts. BMXH

1902-11-25-11 97

Tijuca, Fed. Dist. M, C. Z. 24954

Serra de Macahe

Mus. Paul 962
962 A
Sao Paulo:
Ypiranga

Mus. Paul. 944 96

ngth

diam.

1/d

145

4

36

TYPE

150

4

37

145

4

36

136

3

45

178

o.o

32

95

152

4

38

99

160

6

27

97

174

6

29

170

34

Siphonops insulanus Ihering

1911. Siphonops insulanus Ihering, Rev. Mus. Paulista, 8, p. 109; Nieden
1913, Gymnophiona, p. 26.

Type. In Museu Paulista, not seen.

Type locality. I. Victoria and I. Sao Sebastiao, off coast of Sao
Paulo, Brazil.

Range. Known only from the type localities.

Diagnosis. A Siphonops of uniform color, primaries 108-112; 1/d
31-41; eye indistinct; tentacle a little anterior to and below the eye;
length 152-200 mm.

Description. Little can be added to the diagnosis. Ihering says all the
rings "are interrupted in the dorsal region and at times a little on the
ventral line." The color is uniform light gray. The tentacle is very

492 bulletin: museum of comparative zoology

close to the eye and a little below. Of the four specimens seen the eye
was invisible in two. The primaries were interrupted dorsally in the
middle of the body in one out of four. Ihering mentions a length of
200 mm., and a length-diameter ratio of 41.

Habits. Not known.

Remarks. Allied to hardyi and to brasiliensis, and apparently be-
tween the two.

prim. length diam. 1/d

10S

157

5 31

110

194

5 39

eye invisible

111

152

4 38

eye invisible

112

162

4 40

Specimens seen, four, as follows :

Isla Victoria:
Mus. Paul 946
946A

916B

Isla S. Sebastiao
Mus. Paul. 945

Siphonops brasiliensis Liitken

1851. Siphonops brasiliensis Liitken, Vid. Meddel., p. 52; Reinhardt and
Liitken 1861, Vid. Meddel. p. 202; Boulenger 1891, Ann. Mag. Nat.
Hist. (6), 8, p. 457; 1895, Proc. Zool. Soc. London, p. 412; Ihering
1911, Rev. Mus. Paulista, 8, p. 110; Nieden 1913, Gymnophiona,
p. 25; Parker and Wettstein 1929, Ann. Mag. Nat. Hist. (10), 4,
p. 594.

1879. Dermophis ? brasiliensis Peters, Mon. Ak. Berlin, p. 938.

Type. In Copenhagen Museum, collected by Langgaard. Not seen.

Type locality. Brazil.

Range. Known from the states of Santa Catharina, Sao Paulo,
Minas Geraes, and Rio de Janeiro, Brazil.

Diagnosis. A Siphonops with uniform color; primaries 115-133;
1/d 31-54; tentacle somewhat anterior to and below the eye; eye in-
distinct; primaries frequently interrupted; length 167-312 mm.

Description. The original description gives 133 primaries; the 20
first and the last 13 complete; "gray"; 1/d 46. The eye may be dis-
tinct, indistinct, or invisible. The primaries may be mostly interrupted
or all complete. There seems to be no change in proportions with age.
Parker and Wettstein (1929) state that the premaxillary-maxillary
teeth are 6-8 on a side; total 12 in the type.

Habits. Not known.

DUNN: THE AMERICAN CAECTLIANS

493

Remarks. The relationships of this set of species have been dealt
with under hardyi. The two mainland forms occur together at Ypi-
ranga, Sao Paulo, where hardyi has 96-100 primaries and brasiliensis
has 122.

Specimens seen, 21, as follows:

Santa Catharina:
Colonia Hansa
Hamburg 1807
Frankfort 2102, Id
Joinville

Vienna

it

Mus. Xac. Brazil 542

840

Sao Paulo:

Pernahyba
Vienna

Franca

Mus. Paul. 960

Ypiranga

Mus. Paul. 961

Rio de Janeiro :
Rio

Paris 15 m.
Petropolis

MCZ 24829
" 24826

South Brazil:
Hamburg 1927
Hamburg 1927

1927

Brazil :
Frankfort

prim. length diam. 1/d

124

213

5

43

122

235

6.5

36

124

131

115

215

5

43

no eve

125

235

6

39

no eye

126
127
126

130
122c?

120

217
205

190

6 36

prim. int.

6

36

prim. comp.

47

130
129

312
167

9 35
5 33

121
121
121

268
268
260

5 54
7 39
5 52

121

265

8 44

eye indist.

494 bulletin: museum of comparative zoology

prim. length diam. 1/d
No locality :

BMNH 98-6-27, 3

Vienna

123

245

8 31

122

247

6.5 38

eye invisible

123

230

5 46

eye invisible

Mus. Xae. Brazil 543

Ihering (1911) has recorded specimens in the Museu Paulista from
Rio Fieo, Sao Paulo, and says it occurs in the State of Minas Geraes.

Note for identification: This animal has been confused with Chthon-
erpeton viviparum (q. v.). The Siphonops has: no inner mandibular
teeth, the Chthonerpeton has 3-4; the tentacular aperture in the
Siphonops is much closer to eye than to nostril, in the Chthonerpeton
it is only slightly closer to eye than to nostril; the Siphonops has
primaries 115-133, the Chthonerpeton has primaries 133-166; the
Siphonops has a normal vent, the Chthonerpeton has a small sucking
disk around the vent; the Siphonops is presumably oviparous, the
Chthonerpeton is known to be viviparous.

Caecilia Linne

1758. Caecilia Linne, Syst. Nat. (10) 1, p. 229 (included species tentaculata
Linne and glutinosa Linne). Fitzinger (1843, Syst. Rept., p. 34)
designated C. lumbricoidea [lombricoidaea] Daudin 1803 (=C.
gracilis Shaw = C. tentaculata Linne in part) as type. Shaw in 1802
(Gen. Zool. 3, 595) restricted tentaculata when describing gracilis,
and I designate tentaculata Linne as restricted by him as type of
Caecilia. Daudin 's species was not in the content of the original
genus.

1802. Coecilia Latreille, in Sonnini and Latreille, Hist. Rept. 4, p. 237 (pro
Caecilia Linne).

1901. Amphiumophis Werner, Abh. Mus. Dresden 9, 2, p. 14 (monotype
Amphiumophis andicola Werner, 1. c).

*

Diagnosis. Caecilians without a tail; primaries 110-285; secondaries
0-94; scales usually present; 55-268 primary folds without secondar-
ies; 1/d 26-160; snout projecting; tentacle in horseshoe-shaped groove
on under surface of snout, below and slightly posterior to nostril ; eye
visible or invisible, in open orbit or roofed by bone; anterior teeth on
both jaws enlarged, especially on lower; inner mandibular tooth row
well developed to absent; length 126-1375 mm.; 16 forms.

Range. Code, Panama, to Guayaquil, Ecuador, and Carabaya
and Chanchomayo, Peru. The Guianas. Brazil. Sea level to 6200 feet.

DUNN: THE AMERICAN CAECILIAXS 495

Key to forms of Caecilia

A. Secondaries present.
B. Primaries 110-150.

( \ Secondaries 38-83 dunni

CC. Secondaries 12-37 tentaculata

CCC. Secondaries 8 or less.

D. Primaries 110-119 guntheri

DD. Primaries 139-150 abitaguae

BB. Primaries 154-285.

C. Primary count minus secondary count plus 1/d ratio less
than 282; primaries less than 239; 1/d ratio less than 94.
D. Xo color markings.
E. Secondaries 28-94.

F. Primaries 185; 91 without secondaries armata

FF. Primaries 155-190; 108-138 without secondaries. .

nigricans
FFF. Primaries 187-238; at least 152 without secondaries

thompsoni
EE. Secondaries 8-25.

F. Primaries 154-161 subnigricans

FF. Primaries 185-214 gracilis

DD. Color markings usually present.

E. Eyes visible; usually a pair of yellow spots on each

segment; primaries 154-199; secondaries 2-11

pachynema
EE. Eyes invisible; gray with black primary grooves;
secondaries 7-29.

F. Primaries 171-192 ochrocephala

FF. Primaries 204-209 polyzona

CC. Primary count minus secondary count plus 1/d ratio
more than 291; primaries over 205; 1/d ratio usually

over 100 bassleri

AA. Xo secondaries.

B. Primaries less than 200; eyes visible.
C. Primaries less than 150.

D. Primaries 110-119 guntheri

DD. Primaries 125-139 'degenerata

DDD. Primaries 145-146 caribea

CC. Primaries 154-199 pachynema

BB. Primaries 226-231 ; eyes invisible elongata

496

bulletin: museum of comparative zoology

Tabular list of counts of Caecilia

Specimens
seen

4
27
63
19

2

3

2

1

19

25

101

31

2

10

3

12

species primaries secondaries

guntheri

tentaculata

degenerata

dunni

caribea

abitaguae

subnigricans

armata

nigricans

pachynema

ochrocephala

gracilis

polyzona

thompsoni

elongata

bassleri

110-119

112-147

125-139

123-150

145-146

139-150

154-161

185

155-190

154-199

171-192

185-214

204-209

187-238

226-231

206-285

0-8
12-37

0
38-83

0

5-6
17-18
94
28-62

0-11

7-29

8-25
10-17
29-41

0
14-41

primaries

minus
secondaries

110-119

79-133

125-139

55-85

145-146

134-144

137-143

91

108-138
154-199
149-179
167-193
187-199
152-200
226-231
174-268

1/d

27-31

22-52

31-76

32-57

53-55

43-59

58-62

56

37-66

37-84

39-87

48-93

43-61

45-92

83-89

80-160

324

1

99

3

1
2
1
1

2
15

7

Caecilia by areas
Caecilia of Panama

prim.

131

171-192

226-231

sec.
12

9-29
0

tentaculata

ochrocephala

elongata

Caecilia of Atrato drainage, Colombia

guntheri 119 0

dunni 132-133 50-61

nigricans 190 52

ochrocephala 185 23

Caecilia of Colombian Choco

guntheri 110-115 0

dunni 128-150 50-83

nigricans 159-188 36-47

1/d
28
39-87

83-89

29

32-35
57
50

27

37-57

37-58

DUNN: THE AMERICAN CAECILIANS 497

Caecilia of Pacific slope of Ecuador

1
1

9
3

12

guntheri

dunni

nigricans

bassleri

pachynema

prim.

118

123

155-180

206-251

158-183

sec.

8
38

28-62
14-32

0-10

1/d
31
41

42-66

119-130

40-81

Caecilia of the Cauca Valley,

Colombia

1
2
2

1

caribea
pachynema
polyzona
thompsoni (?)

145

159-166
204-209
212

0

2-7
10-17
35

(Rio Coqueta,

55

52-78
43-61
84

Cauca Valley?)

Caecilia of

Magdalena \

'alley,

Colombia

2
6

subnigricans
thompsoni

154-161
187-238

17-18
29-39

58-62
45-92

Caecilia of Barranquilla and Santa Marta region, Colombia

2

1

tentaculata
caribea

116-147
146

14-21
0

31-38
63

Caecilia

of the Colombian Oriente

51
2

1

1
2

degenerata
tentaculata
thompsoni (?)

bassleri
pachynema

128-139
113-146
212

244
156-180

0

29-31
35

25
0

31-76
31

84

(Rio Caqueta?)

80
38-54

Caecilia of the Ecuadorian Oriente

1
3

3

2

1

dunni

tentaculata

abitaguae

bassleri

pachynema

123

115-122
139-150
254-271
174

67
29-33

5-6
28-41

0

35

30-35
43-59
124-160
73

Caecilia of the Peruvian Oriente

2

1
6
4

tentaculata
gracilis
bassleri
pachynema

120-129
188

230-285
165-199

28-31
21

17-30
0-11

36-39
56

80-124
59-70

49S

bulletin: museum of comparative zoology

prim.

sec.

13

tentaculata

112-146

13-37

25

gracilis

185-207
Caecilia of Brazil

9-23

1

tentaculata

130

15

1

armata

185

94

1

gracilis

214

25

Caecilia of the Guianas

1/d
27-52
48-93

46
56
92

This list by areas shows clearly that as many as five perfectly dis-
tinguishable forms may occur in a single geographical area. Colombia
has twelve forms, Ecuador seven, Peru four, Panama and Brazil
three, and the Guianas two. No specimens of Caecilia have been seen
or reported from Venezuela, but at least two (tentaculata and gracilis)
must occur there.

Remarks. A diagram of the forms, with dunni, nigricans, and
armata, thompsoni, and bassleri arranged in order of increasing num-
ber of primaries and increasing slimness, and with the other forms
appended as they seem to fit, is given here as a possible scheme of
relationships.

Two forms, C. dunni and C. armata, retain more of the scalation
than do the others. C. nigricans is a close third. As dunni has a com-
bination of few primaries and many secondaries it may be assumed
to be the most primitive existing form. Other forms show: an increase
in primaries; a decrease in secondaries; extreme attenuation; degenera-
tion of the eye; a combination of these characters, and may be as-
sumed to be more specialized.

dunni.

- tentaculata

armata
nigricans .

.subnigricans

guntheri

degenerata

caribea

abitaguae

thompsoni-

-polyzona .

.gracilis

.elongata

.ochrocephala
pachynema

bassleri-

DUNN: THE AMERICAN CAECILIAXS 499

The genus could, very plausibly, be regarded as monotypic with 16
races, so closely do allied forms resemble each other. But as many as
five different forms may occur together and remain distinct, and at
present it seems best to treat each recognizable form as a species.
The difficulty of treating them in any other way may be illustrated
by the fact that one can start with nigricans of the Choco, and, by a
series of easy transitions (via subnigricans of the Magdalena and
tentaculata of the Oriente) arrive at guntheri, also of the Choco. Also,
in this set of forms, dunni is about as good an intermediate between
nigricans and tentaculata and occurs with both, while nigricans itself
is intermediate between the two forms of the Magdalena Valley,
subnigricans and thompsoni. One could, perhaps consider ochroce-
phala and polyzona as races of a species, and the guntheri-degenerata-
abitaguac-caribea set as races of another species.

Linne (1758) used the spelling Caecilia. The first occurrence of
the emended spelling Coecilia that I have noted is Latreille (1802).
There have been so many writings of the generic name in type which
does not differentiate the diphthong "ae" from the diphthong "oe"
that I have given up any attempt to differentiate between them in
synonymies, and have used Caecilia throughout.

Cope, in 1885, put his Caecilia ochroccphala into the genus Herpele.
I see no reason why ochroccphala should be placed in a different genus
from tentaculata and gracilis. The West African squalostoma, the
type of Herpele, has the tentacular aperture more posterior than any
American species, and in it the anterior maxillary and dentary teeth
are not enlarged. I do not consider squalostoma as congeneric with
any American species.

The single specimen in Dresden upon which Werner, in 1901,
founded his genus and species Amphiumophis andicola, is, in my
opinion, conspecific with Caecilia tentaculata.

The following list contains described species which I think valid :

tentaculata Linne 1758.

gracilis Shaw 1802.

pachynema Gunther 1859.

ochroccphala Cope 1866.

guntheri Peters 1S79.

polyzona Fischer 1S79.

nigricans Boulenger 1902.

thompsoni Boulenger 1902.

dunni Hershkovitz 1938.
I have examined the types of all of these except tentaculata and gracilis.

500 bulletin: museum of comparative zoology

I have seen a specimen of gracilis which was so named by the describer
of the species.

The following list contains described species which I regard as
invalid. I have examined the types of all of these.

albivcntris Daudin 1803 = tentaculata.

lombricoidaea Daudin 1803 = gracilis.

isthmica Cope 1877 = tentaculata.

buckleyi Boulenger 1884 = pachynema.

andicola Werner 1900 = tentaculata.

sabogae Barbour 1906 = ochroccphala.

intermedia Boulenger 1913 = nigricans.

palmeri Boulenger 1913 = nigricans.

The following names are substitutes:

ibiara Daudin 1803 for tentaculata.

vermiformis Gray 1850 for gracilis.
I describe hereinafter seven forms in addition to the nine recognized
above as valid, making a total of 16 forms.

The most recent systematic treatment of these species is that of
Nieden in "Gymnophiona" (1913). This is based on previous work by
Boulenger. As my treatment differs very considerably, I should explain
why. In the first place Boulenger was able to examine very little ma-
terial aside from that in the British Museum. This contained, in 1929,
54 specimens of Caecilia, seven of them being the types of described
forms; I have been able to examine 324 specimens and the types of 15
described forms. In the second place, Boulenger lumped primaries and
secondaries together into one count, and thus a Caecilia with many
vertebrae and few scales would appear statistically similar to one with
few vertebrae and many scales. By keeping these two independent
variables separate, I arrive at results which are frequently different
from Boulenger's.

Caecilia dunni Hershkovitz

1913. Caecilia intermedia Boulenger (in part, numbers 5-6), Proc. Zool. Soc.

London, p. 1020.
1913. Caecilia nigricans Boulenger, 1. c, p. 1022 (not C. nigricans Boulenger

1902).
1938. Caecilia dunni Hershkovitz, Occ. Papers Mus. Zool. U. Michigan 370,

p. 2, f. 1.

Type. Mus. Zool. U. Michigan 82901, collected by Philip Hershko-
vitz, Dec. 1935.

DUNN: THE AMERICAN CAECILIANS 501

Type locality. Near Tena, Province of Napo-Pastaza (Oriente) Ecua-
dor, 1700 feet above sea level.

Range. Atrato Valley, Colombia; Colombian Choco; Cachabe,
northwest Ecuador; Tena, Ecuadorian Oriente. Sea level to 1700 feet.

Diagnosis. A Caecilia with 123-150 primaries; 38-83 secondaries;
55-S5 primary folds without secondaries; 1/d 32/57; eye visible in
most specimens; no markings; length 147-450 mm.

Description. The type has nine teeth in each of the long rows and
two on each side in the inner mandibular row. The eye is invisible in
the Cachabe specimen. There is some variation which may be geo-
graphic.

Prim.
1 E. Ecuador 123

1 N. W. Ecuador 123
15 Colombian Choco 128-150

2 Atrato Valley 132-133

With more material the form might be divided.

Remarks. This form is allied only to tentaculata, from which it differs
in having a higher secondary count. C. tentaculata is absent from the
Pacific slope and from the Atrato Valley, but apparently occurs with
dunni in the Oriente of Ecuador.

BMNH 1913-11-12, 134 from Pefia Lisa was "taken from the stom-
ach of a Streptophorus atratus swallowed by an Elaps corallinus"
(Boulenger 1913).

Sec.

1/d

67

35

38

41

50-83

37-57

50-61

32-35

Specimens seen 19, as follows:

Colombia:

Las Animas Cr., Quito R.,

Atrato system, AMNH 13678
Quibdo on Atrato, Inst. La Salle

Anda Goya, BMNH 1915-10-21, 73 133

prim. sec. length diam. 1/d

1916-4-25,
Pefia Lisa, Condoto, 300',

133

61/6

290

9

32

132

50

210

6

35

73 133

78/15

188

5

38

74 136

80/11

210

5.5

38

75 142

64

147

4

37

76 139

63/20

435

8

54

77 136

77/26

300

6

50

78 -

79 -

31 131

65/4

409

9

45

32 134

77/11

375

9

41

67/8

280

6

56

83/10

270

5

54

62/8

395

7

57

54/5

240

6

40

50/20

186

5

37

78/21

350

5

50

502 bulletin: museum of comparative zoology

prim. sec. length diam. 1/d
BMNH 1913-11-12, 134 146

135 150

136 147
1914-5-21, 93 129

Condoto, BMNH 1910-7-11, 73 128

74 146

Ecuador :
Cachabe, BMNH 98-3-1, 36 123 38/6 290 7 41

Tena, U. Michigan S2901 123 67/5 450 13 35

Caecilia tentaculata Linne

1758. Caecilia tentaculata Linne (except reference to pi. 5, f. 2, Mus. Adolph,
Frid.) Syst. Nat. (10), p. 229; Shaw 1802, Gen. Zool. 3, 595; Latreille
1802, in Sonnini and Latreille, Hist Rept. 4, p. 237, pi. 22, f. 2
Cuvier 1817, Regn. Anim. 2, p. 87; Goldfuss 1820, Handb. Zool. 2
p. 138; Merrein 1820, Vers. Syst. Amph., p. 168; Cuvier 1829, Regn
Anim. (2), 2, p. 100; Gray 1831, in Griffith's Cuvier's Anim. King
9, App., p. 110; Gray 1850, Cat. Batr. Grad. Brit, Mus., p. 58
Peters 1879, Mon. Berlin Ak., p. 934, f. 5; Boulenger 1882, Cat.
Batr. Grad. Brit. Mus. (2), p. 93; Boulenger 1895, Proc. Zool. Soc
London, p. 406; Phisalix 1912, Congr. Int. Zool. 8 (Graz 1910)
pi. 4, f. 3, 8, 11 (integ.); Nieden 1913, Gymnophiona, p. 12, f. 3, 4, 10

1838. Caecilia lenticulata Tschudi, Mem. Soc. Sci. Neufchatel 2, p. 90 (typ
error)

1803. Caecilia albiventris Daudin, Nat. Hist. Rept. 7, p. 423, pi. 92, f. 2
(Surinam, type Paris 9); Cuvier 1829, Regn. Anim. (2), 2, p. 100
Gray 1831, in Griffith's Cuvier's Anim. King. 9, App,, p. 119
Dumeril and Bibron, 1841, Erp. Gen. 8, p. 276, pi. 85, f. .3; Tschudi
1845, Faun. Peru. p. 80; Dumeril 1863, Mem. Soc. Sci. Cherbourg 9,
p. 313, pi. 1, f. 1, 9.

1820. Caecilia albiuentris Merrem, Syst. Amph., p. 169 (emendation).

1803. Caecilia ibiara Daudin, Nat. Hist. Rept. 7, p. 427 (substitute for
tentaculata Linne).

1877. Caecilia isthmica Cope, Proc. Amer. Phil. Soc. 17, p. 91 (Atlantic side
isthmus of Darien, type USNM 25188); Dunn 1928, Proc. New
England Zool. Club 10, p. 73.

1900. Amphiumophis andicola Werner, Abh. Mus. Dresden (2), p. 14 (Chan-
chamayo, Peru, type Dresden 1689).

Type. Not known to exist.

Type locality. "America" = Surinam (cf. Amoen. Acad. 1, p. 498,
pi. 17, f.,1, Linne, 1749).

DUNN: THE AMERICAN CAECILIANS 503

Range. Darien to Brazil and to eastern Peru. Sea level to 2800 feet.

Diagnosis. A Caeeilia with 112-147 primaries; 12-37 secondaries;
79-133 primary folds without secondaries; eye usually visible; 1/d
22-52; length 126-1075 mm.; belly usually with white blotches.

Description. Boulenger (1882) gives the dentition of the Shaw speci-
men as "teeth moderately large; on each side maxillaries 6 to

8, vomero-palatines 5, outer mandibulars 6 or 7, inner mandibulars
very small, few." The type of andicola has no inner mandibular teeth.
The three specimens from Demarara and Mazaruni River have no visi-
ble eyes. White blotches are present on the bellies of most Guiana
specimens irrespective of the primary count. The diagnosis above
could have been made out entirely from Guiana specimens save for one
more primary in a Colombian specimen, one less secondary in the
type of isthmica, and for a stouter Ecuadorian specimen.

Remarks. Linne first mentioned this form in 1749 (Amoen. Acad. 1,
p. 498, pi. 17, f. 1) as from Surinam and as having 135 rings, in 1754
(Mus. Adolph. Frid., p. 19) he abbreviates his 1749 description, and
adds a figure (pi. 5, f. 2) of a much slimmer animal (gracilis of this
paper). His 1758 description is very brief. In it he refers to his two
previous papers, citing the former as page 489, a mistake that has
been widely copied.

There is a rather wide range of variation, and possibly this species is
composite, especially as the Guiana specimens are easily separable into
two sets. I therefore list possible divisions.

1. low primary, low secondary, stout.

116 14 1/d 38 A single Colombian.

2. low primary, high secondary, stout.

112-129 24-37 22-39 This includes 19 specimens,

10 Guianan (type of albi-
ventris), one Colombian,
four Ecuadorian, two Pe-
ruvian (type of andicola),
two without data.

3. high primary, low secondary, stout.

131-147 12-21 28-31 First numbers in each case

are the Panamanian type of
isthmica; second numbers a
Colombian specimen.

504

bulletin: museum of comparative zoology

4. high primary, low secondary, slim.
130-146 13-15 44-52

5. high primary, high secondary, slim.
140 29 40

This includes three Guiana
specimens and the Brazilian
one. The type of tentacu-
lata probably belonged here.

A single Colombian speci-
men.

2 and 4 occur together in Guiana; 1 and 3 occur together in northern
Colombia; 2 and 5 occur together in the Colombian Oriente. Additional
material may in time afford some clarification of this puzzle, but I do
not wish to divide the 27 specimens into five species under the existing
conditions of knowledge.

The series as a whole has fewer secondaries than dunni, and more
secondaries than the guntheri-degcnerata-caribea-abitaguae series which,
with tentaculata, comprise those Caecilia with less than 151 primaries.
C. tentaculata occurs with degcnerata at Garagoa in the Colombian
Oriente, near dunni aod abitaguae in the Ecuadorian Oriente, and near
caribea in northern Colombia.

Specimens seen,

27, as follows:

Panama :

prim.

sec.

length

diam.

1/d

Atlantic side Darien

USNM 25188

131

12/0

570

20

28

Colombia :

Rio Frio

MCZ

17376

147

21/0

330

8

31

Sabana arga

AMNH

14032

116

14

430

15

38

Garagoa

MCZ

17384

146

29/0

260

6.5

40

Pto. Asis, R. Putomayo

Inst. La Salle

113

31/0

470

15

31

British Guiana:

Marudi Mts.

AMNH

49470

121

37

126

4

31

49471

121

26

145

5

29

49472

120

33

135

5

27

49473

120

34

140

5

28

49474

119

34

128

4

32

49475

115

29

442

13

34

do

do

49476

119

27

205

7

29

Demarara

BMNH

89-9-30, 16

136

13/3

313

6

52

DUNN: THE AMERICAN CAECILIANS

505

prim. sec. length diam. 1/d

Kamakusa

AMXH 49962

112

33

500

15

33

Mazaruni R.

AMNH 20079

146

15/5

307

44

20080

146

13/5

340

7

48

Dutch Guiana:

Surinam

Paris 9

120

29

600

18

33

do

BMNH

58-6-1, 36

120

24/4

350

10

35

Brazil:

No data

Hamburg 1717

130

15/8

502

11

46

Eastern Ecuador:

Tuvola, 2800'

AMNH 23421

115

29/0

275

9

30

Copatava R.

AMNH 49961

122

29/2

640

18

35

Rio Suno 300'

Mich.

121

32/5

155

7

22

No data

Mich. 89459

121

33/4

1075

30

36

Eastern Peru:

Chanchamayo

Dresden 1689

129

28/4

350

9

39

Monte Alegre, R.

Pachitea

750-1000'

AMNH 42855

120

31/2

365

10

36

South America:

No data

Berlin 3901

115

31/4

565

15

37

No data:

Shaw coll.

BMNH

1929-5-16, 1

114

28/4

510

19

27

Tschudi (1845) records tentaculata from Vitoc in middle Peru. He
says that the young have gill slits.

The measurement of length in the giant Ecuadorian specimen is
purely approximate. It could be stretched to 960 mm.; using a string
it measured 1190 mm.

Caecilia degenerata spec. nov.

Type. MCZ 17384.

Type locality. Garagoa, eastern Colombia.

Range. Eastern Colombia.

Diagnosis. A Caecilia without secondaries; primaries 125-139; 1/d
31-76; length 132-555 mm.

Description. The primaries have been counted in 48 specimens. Of
these 35 came from either Choachi or Tomaque, nearby localities in the

506 bulletin: museum of comparative zoology

Oriente of Colombia. The range of these 35 is 128-139, and only three
specimens are outside the range of 130-138. Specimens from elsewhere
have similar counts except the Rio de Pache specimen with 125 prima-
ries. The primaries are interrupted dorsally and ventrally (AMNH
22584-6, 22588-92).

The length-diameter ratio has been computed for 51 specimens. The
range is 31-76 in 40 Choachi and Tomaque specimens. The larger ani-
mals seem to be slightly slimmer, since for the same 40 specimens the
highest ratio below a length of 300 mm. is 45, and the lowest above a
length of 400 mm. is 37. For these two places the four ratios above 60
are for animals of 362 mm. long and over, and the ratios below 35 are
for animals of 300 mm. long and under. The Rio de Pache specimens is
in contrast to this with a ratio of 31 (as stout as any) and a length of
525 mm.

The Rio de Pache specimen has yellow spotting laterally; 5-6
maxillary-premaxillary teeth on a side ; 9 palatine teeth ; 8 left mandi-
bular teeth with the four first enlarged; 6 right mandibular teeth with
the two first enlarged; 3-4 inner mandibular teeth; the eye definitely
visible. A Choachi specimen has two inner mandibular teeth.

Remarks. The large series from Choachi and Tomaque, in the Ameri-
can Museum, gives the range of variation and the characters. The
specimens from "Colombia," "Bogota," and Garagoa are within this
range of variation.

The specimen from Rio de Pache is so close in primary count that it
is best placed here. I have not been able to place the locality. It was
from one of Eigenmann's collections, and no data save "Rio de Pache,
Porte" were with it. The University of Michigan staff, Dr. Barbour,
Dr. Chapman, and myself, have been unable to find the Rio de Pache,
Porte. Barbour suggested that it is Lima, near Peru, which would
complicate the situation considerably.

Scales are definitely not present in the Garagoa specimen, the Rio
de Pache specimen, AMNH 23355 from Colombia, the La Salle one
from Choachi, and AMNH 23270 from Choachi. They are present in
AMNH 23271 from Choachi.

Specimens seen, 63, as follows:

Colombia:

prim.

length diam.

1/d

Choachi

AMNH 23259

131

384 9

42

23260

132

326 9

36

23261

133

400 10

40

DUNN: THE AMERICAN CAECTLIAXS

507

prim.

Choachi

AMNH

23262

23263

136

23264

11

a

23265
23266

—

23267

131

11

ii

23268

23269

—

Choachi

AMNH

23270

— —

23271

—

23272

134

ii

ii

23273

138

23274

131

Choachi, Inst.

La Salle

135

Choachi and

Tomaque

AMNH

22560
22561
22562

138

it

<<

22563

139

22564

133

22565

tt

ii

22566

136

22567

138

22568

136

tt

it

22569

22570

130

22571

136

It

tt

22572

135

22573

134

22574

129

tt

a

22575

130

22576

137

22577

130

ii

a

22578

134

tt

tt

22579
22580

135

22581

131

22582

length

diam.

1/d

398

8

50

425

8

54

443

8

55

415

10

41

441

9

49

378

9

42

451

8

58

367

10

37

360

10

36

390

8

49

512

11

46

550

13

42

466

12

39

450

12

37

467

8

58

406

9

45

380

8

47

433

8

54

153

5

31

260

7

37

400

9

44

321

8

40

455

6

76

132

4

33

365

8

45

213

6

35

420

7

60

408

8

51

508 bulletin: museum of comparative zoology

length diam. 1/d

Choachi and

prim.

Tomaque

AMNH

22583

136

22584

134

22585

134

•

22586

135

22587

134

22588

128

22589

132

22590

138

22591

138

22592

130

Garagoa

MCZ

17383

134

? Bogota

AMNH

23421

132

no locality

AMNH

23348

132

23349

137

23350

136

23351

135

23352

134

23353

127

23354

133

23355

137

23356

138

no locality

AMNH

34254

133

??? Rio de Pache,

Porte

U. Mich.65263

125

182

4

45

380

10

38

300

9

33

485

8

60

154

4

38

445

9

49

362

6

60

555

10

55

555

10

55

350

10

35

470

10

47

450

10

45

380

7

54

470

10

47

352

9

40

298

7

42

335

9

37

400

8

40

159

5

32

373

9

52

520

17

31

Caecilia abitaguae spec. nov.

Type. Mus. Univ. Michigan 89930.

Type locality. Abitagua, Oriente, Ecuador, 1100 m. elevation.

Range. Known only from type locality.

Diagnosis. A Caecilia with 139-150 primaries; secondaries 5-6;
1/d 43-59; length 300-1200 mm.; eye visible; no markings.

Description. Nothing of importance can be added to the diagnosis
and the characters of the individual specimens.

Remarks. This form is close to degenerata of the Colombian Oriente,
but these Ecuadorian specimens have a higher primary count, and all
three have a few secondaries. It is also related to C. guntheri of western
Ecuador, but has a much higher primary count, and is somewhat

89929

150

6

1200

22

59

89930

145

6

780

18

55

5061

139

5

300

8

43

DUNN: THE AMERICAN CAECILIANS 509

slimmer. It is extremely similar to C. caribea of northern Colombia,
differing only in having secondaries.

It occurs with C. tentaculata in the Oriente of Ecuador, but tentacu-
lata has at least 17 fewer primaries and 23 more secondaries in the
region where the two are together.

Specimens seen, 3, as follows:

^ , ~. prim. sec. length diam. 1/d

Ecuador, Oriente:

Abitagua Mich.

a (i

Stanford

Caecilia caribea spec, now

Type. MCZ 24520.

Type locality. Pensilvania (Cauca valley south of Medellin), Co-
lombia.

Range. Known only from type locality and from Barranquilla,
Colombia.

Diagnosis. A Caecilia with 145-146 primaries; no secondaries; eye
visible; 1/d 53-55; no distinctive markings; length 390-585 mm.

Description. Nothing can be added to the diagnosis and the charac-
ters of the individual specimens.

Remarks. This form is similar to degenerate/, of the Colombian
Oriente, to abitaguae of the Ecuadorian Oriente, and to guntheri of the
Atrato valley and the Pacific coast. Strangely enough, it is most simi-
lar to abitaguae, differing only in lacking secondaries. It has a higher
primary count than degenerata and a much higher one than guntheri.

C. caribea occurs with C. tentaculata in northern Colombia, but
tentaculata there has 14-21 secondaries and a length-diameter ratio of
31-38.

Specimens seen, 2, as follows :

„ , , . prim, length diam. 1/d

Colombia:

Pensilvania MCZ 24520 145 390 7 55

Barranquilla Senckenberg 3095a 146 585 11 53

510 bulletin: museum of comparative zoology

Caecilia guntheri Peters

1859. Caecilia rostrata Gimther, Proc. Zool. Soc. London, p. 417 (not Caecilia
rostrata Cuvier = Hypogeophis rostratus).

1879. Caecilia guntheri Peters, Mon. Berlin Ak., p. 936 (substitute name).

1880. 1 Caecilia pachynema Boulenger, Bull. Soc. Zool. France 5, p. 48 (two

specimens in Brussels from "Andes of Ecuador," not C. pachynema
Giinther).
1882. Caecilia isthmica Boulenger (at least in part) Cat. Batr. Grad. Brit.
Mus. (2), p. 94, pi. 6, f. 1 (not Caecilia isthmica Cope); Boulenger
1895 Proc. Zool. Soc. London, p. 406; Boulenger 1913, Proc. Zool.
Soc. London, p. 1020.

Type. BMNH 60-6-16, 58.

Type locality. West Ecuador.

Range. Western Ecuador and western Colombia. The Atrato Val-
ley, Colombia.

Diagnosis. A Caecilia with primaries 110-119; secondaries 0-8;
1/d 27-31; eye visible; no markings; 260-630 mm.

Description. "Teeth moderately large, on each side maxillaries

11, vomero-palatines 5, outer mandibulars 8; inner mandibulars very
small, few" (Boulenger 1882). The Urrao specimen has 3 inner man-
dibular teeth. Mr. H. W. Parker kindly informs me that both the
Pefia Lisa specimens have scales; the Urrao specimen has none.

Remarks. The two Brussels specimens, first called pachynema and
then isthmica by Boulenger, have not been seen by me. They are
probably what is here called guntheri. They had 119 and 124 "circular
folds." The largest was 750 mm. long. They had 6 maxillary, 6-7
palatine, and 5-7 mandibular teeth.

This species is close to degenerata, having fewer primaries, and to
tentaculata, having fewer secondaries. It does not occur with either.

prim. sec. length diam. 1/d

Specimens seen, 4, as follows:

Colombia:

Pefia Lisa, Condoto

BMNH 1913-11-12, 131

BMNH 1913-11-12, 132
Urrao on Atrato Inst. La Salle

West Ecuador:

BMNH 60-6-16, 85 118 8 630 20 31

115

0

268

10

27

110

0

330

12

27

119

0

260

9

29

DUNN: THE AMERICAN CAECILIANS 511

Caecilia subnigricans spec, now

Type. ANSP4821.

Type locality. Magdalena River, Colombia.

Range. Known only from type locality.

Diagnosis. A Caecilia with 154-161 primaries; 17-18 secondaries;
137-143 primary folds without secondaries; eye visible; 1/d 58-62;
length 350-370 mm. ; no distinctive markings.

Description. Nothing can be added to the diagnosis.

Remarks. This form has fewer secondaries than nigricans and more
primaries than tentaculata. It is anatomically between these two forms,
neither of which occur in the Magdalena Valley, as nigricans is west of
this area and tentaculata is east of it. It probably occupies the lower
part of the valley, as the closely allied thompsoni, which has more
primaries and more secondaries, is the only form known from the upper
Magdalena.

Specimens seen, 2, as follows:

Colombia: Prim- sec" length diam. 1/d

Magdalena River ANSP 4921 161 18/8 370 6 62

4922 154 17/4 350 6 58

Caecilia armata spec. nov.

Type. Mus. Nac. Brazil 832.

Type locality. No data, probably Brazil.

Range. Unknown.

Diagnosis. A Caecilia with 185 primaries ; 94 secondaries ; 91 primary
folds without secondaries; eye visible; 1/d 56; length 390 mm. ; no color
markings.

Description. It may be added to the diagnosis that the diameter is
7 mm., and that the last 12 of the secondaries are complete.

Remarks. This remarkable form has the hind half of the body with
bony scales, and in that respect agrees with dunni. But the latter is a
much shorter (123-150 primaries) form, and is usually stouter. In
primary count and in proportions it falls close to nigricans, some speci-
mens of which have the hind third of the body scaled. But nigricans
has at most 62 secondaries and is a Pacific coast form, while armata
may be presumed to be Brazilian. I offer the suggestion that the primi-
tive scalation may have persisted at the eastern as well as the western
periphery of the range of the genus. In this case the alliance might be

512 bulletin: museum of comparative zoology

with another Brazilian Caecilia (gracilis), a species whose primary
counts and proportions are also like those olarmata. The only Brazilian
gracilis has the highest secondary count (25) for that species.

It is a great pity that the specimen has no data, but its characters
are such that it must be described as a new form.

Caecilia nigricans Boulenger

1902. Caecilia nigricans Boulenger, Ann. Mag. Nat. Hist. (7), 9, p. 51;

Nieden 1913, Gymnophiona, p. 13.
1913. Caecilia intermedia Boulenger (in part, numbers 1-4), Proc. Zool. Soc.

London, p. 1026, f. 174 (St. Javier, N. W. Ecuador, type BMNH

1907-3-29, 69); Parker 1926, Ann. Mag. Nat. Hist. (9), 17, p. 549.
1913. Caecilia 'palmeri Boulenger, 1. c, p. 1021, f. 175 (Novita, Rio San

Juan, Colombia, type BMNH 1910-7-11, 72).

Type. BMNH 1901-3-29, 88.

Type locality. Rio Lita, 3000 feet, N. W. Ecuador or S. W. Colom-
bia [= Ecuador].

Range. West coast of Colombia; Atrato valley, Colombia; Gorgona
I.; west coast of Ecuador.

Diagnosis. A Caecilia with primaries 155-190; secondaries 28-62;
108-138 primary folds without secondaries; eye visible; 1/d 37-66;
length 147-950 mm. ; no distinctive markings.

Description. The specimens are uniform blackish. Boulenger (1902)
says the type had 8 maxillary and 6 mandibular teeth. He says (1913)
of the type of palmeri "dentition as in C. pachynema," and of the type
of intermedia "outer mandibular teeth smaller than" pachy-
nema, but his figures show larger teeth in intermedia than in palmeri.
He also states that the snout of palmeri is like that of pachynema;
and that of intermedia is more strongly projecting. His figures show
palmeri with a more prominent snout than intermedia.

The intromittent organ of BMNH 1913-11-12, 133 is extruded and is
"10 mm. in length and terminates in a fourdobed 'glans' '' (Boulenger
1913).

Remarks. The male just mentioned was "swallowed by an Elaps
rosenbergii."

The types of Boulenger's three species, which I think synonymous,

have: .

prim. sec. 1/d

168

47

65

intermedia

174

43

58

palmeri

177

32

60

nigricans

DUXX: THE AMERICAN CAECILIANS

513

C. pahncri and C. intermedia were described in the same publication.
The only differences in the descriptions refer to minor discrepancies
in dentition and snout shape, and these are directly contradicted by
the figures. No comparison with Boulenger's earlier nigricans was
given.

This form of the Pacific coast differs from subnigricans of the Magda-
lena valley in the much higher secondary count. It differs from thomp-
soni of the upper Magdalena in lower primary and higher secondary
counts; and from the more eastern gracilis in lower primary and higher
secondary counts.

Specimens seen, 19, as follows:

Colombia:

Quesada River, Atrato valley

prim. sec. length diam. 1/d

AMNH

13679

190

52

850

15

57

Anda Goya

BMNH

1916-1-25, 30

175

37/6

845

17

50c?

Novita, R.

San Juan

BMNH

1910-7-11, 72

174

43/6

700

12

58

Pefia Lisa,

Condoto

BMNH

1913-11-12, 133

174

47

720

14

51c?

it a

BMNH

1914-5-21, 91

159

40

625

11

57

tt if

BMNH

1914-5-21, 92

171

Gorgona I.

BMNH

1926-1-20, 145

166

36/3

680

12.5

57

Choco, Inst. La Salle

188

37/10

147

4

37

No locality

BMNH

1923-7-11, 72

174

43/6

600

12.5

48

M

Hamburg 384

172

33/0

638

11

58

?uador:
Rio Lita

BMNH

1901-3-29, 88

177

32/7

600

10

60

Manabi

AMNH 3872

166

43/7

485

10

48

Salidero

Vienna

173

59/7

455

7

65

St. Javier

«(

180

62/3

800

12

66

«<

BMNH

168

47/8

950

17

65

Pambelar

<«

168

28/8

705

11

64

514 bulletin: museum of comparative zoology

prim. sec. length diam. 1/d

Paramba BMNH 166 47/8 640 14 46

155 47/7 820 19 42
Plaza d'Oro, Santiago

USNM 20590 162 53/5 930 20 46

Caecilia thompsoni Boulenger

1899. Caecilia gracilis Cope, Sci. Bull. Philadelphia Commer. Mus. 1, p. 8

(not Caecilia gracilis Shaw).
1902. Caecilia thompsoni Boulenger, Ann. Mag. Nat. Hist. (7), 10, p. 152;

Nieden 1913, Gymnophiona, p. 14.

Type. BMNH 1902-5-15, 26.

Type locality. Villeta [between Honda and Bogota], 3500', Colombia.

Range. Upper Magdalena valley and Rio Caqueta, Colombia.

Diagnosis. A Caecilia with primaries 188-238; secondaries 29-41;
152-200 primary folds without secondaries; 1/d 45-92; eye usually
visible; no distinctive markings; length 345-1375 mm.

Description. MCZ 9726 has 8 maxillary, 4 outer mandibular and 2
inner mandibular teeth. Boulenger (1902) says the type had "teeth
very large in front, 6 or 7 on a side in upper jaw, 15 or 16 in lower, 14
vomero-palatines on each side, 8 small inner mandibular teeth," and
"blackish speckled with yellow on the sides." The eye is invisible in
the specimen from Muzo.

Counts taken on specimens from definite Magdalena valley localities
are altered by others as follows : the La Esperanza specimen raises the
secondary count from 39 to 41, and lowers the difference between pri-
mary and secondary counts from 157 to 152; the Rio Caqueta speci-
men raises the 1/d ratio from 79 to 84, and the specimen reported as
gracilis (AMNH 49976) from "probably near Bogota" raises it to 92.

Remarks. In each individual respect my diagnosis of thompsoni, over-
laps my diagnosis of bassleri, but all the specimens can be allocated by
combining characters. It is distinguished from gracilis by higher sec-
ondary count, and from nigricans by higher primary count.

Boulenger measured the type as 1170 mm., diameter 13. I measure
it as 1000 mm., diameter 15. Cope measured AMNH 49976 as 1300
mm. I measure it as 1375 mm. This is the largest American Caecilian.

Specimens seen, 10, as follows:
Colombia: prim. sec. length diam. 1/d

Villeta BMNH

1902-5-15,26 192 29/0 1000 16 62

TYPE

DUNN: THE AMERICAN CAECILIANS 515

prim. sec. length diam. l/d

Honda or Bogota MCZ 9726 188 29/8 670 15 45

Ibaque MCZ 24522 207 39/7 370 6 62

Muzo MCZ 24521 238 38/8 790 10 79
La Mesa near Bogota,

Inst. La Salle 187 30/0 550 11 50

Bogota ? AMNH 49976 217 37 1375 15 92
Rio Caqueta BMNH

1902-5-29, 179 212 35 760 9 84

no data MCZ 24523 193 29/0 680 12 57

La Esperanza, Brussels 193 41/7 490 10 49

N. S. Amer. Hamburg 1936 197 39 345 o.o 63

The British Museum Rio Caqueta specimen seems to be this species
but has confusing locality data. Additional information gives "Cauca
Valley, S. E. Colombia, collected by Dr. M. D. Eder, purchased
through Rosenberg." The Rio Caqueta is in southeast Colombia, is a
tributary of the Amazon, is not an unlikely place for thompsoni as it
heads near the head of the Magdalena, but it is not in the Cauca
valley.

The Cauca valley has a Rio Coqueta, but this is in northeast Colom-
bia, and is a very unlikely place for thompsoni.

Caecilia gracilis Shaw

1758. Caecilia tentaculata Linne (in part, the reference to pi. 5, f. 2, Mus.
Adolph. Frid.) Syst. Nat. (10), p. 229.

1802. Caecilia gracilis Shaw, Gen. Zool. 3, 2, p. 597; GraA- 1850, Cat. Batr.

Grad. Brit. Mus., p. 57; Dumeril 1863, Mem. Soc. Sci. Nat. Cher-
bourg 9, p. 313; Peters 1879, Mon. Ak. Berlin, p. 935; Boulenger 1882,
Cat. Batr. Grad. Brit. Mus. (2), p. 75; Nieden 1913, Gyranophiona,
p. 13.

1803. Caecilia lombricoidaea Daudin, Hist. Nat. Rept. 7, p. 420, pi. 92, f. 2

(Surinam, types Paris 12); Dumeril and Bibron 1841, Erp. Gen. 8,

p. 275, pi. 85, f. 2.
1820. Caecilia lumbricoides Merrem, Vers. Syst. Amph., p. 168 (emendation);

Cuvier 1829, Regn. Anim. (2), p. 100; Gray 1831, in Griffith's

Cuvier's Anim. King. 9, App., p. 110; Wiedersheim 1879, Anat.

Gymn., pi. 2, f. 14, 19, 20, 22, pi. 6, f. 61, 65-7, pi. 7, f. 72-4, 76-9,

81, pi. 9, f. 89.
1820. Caecilia lumbricoidea Goldfuss, Handb. Zool. 2, p. 138 (emendation);

Wagler 1830, Nat. Syst. Amph., p. 198; Tschudi 1838, Mem. Soc.

Sci. Neufchatel 2, p. 90.
1850. Caecilia vermiformis Gray, Cat. Batr. Grad. Brit. Mus., p. 57 (MSS

name of Shaw, quoted in synonymy of C. gracilis).

516 bulletin: museum of comparative zoology

Type. Not known to exist. BMNH 1929-5-16, 2 is from the Shaw
collection and was named by Dr. Shaw (Boulenger 1882, p. 75, spec.
"g") but does not agree with Shaw's measurements, which were 13%"
long and \/%' in diameter. Since no type was named this may be a
co type.

Type locality. "America."

Range. The Guianas; Para, Brazil; Iquitos, Peru. Sea level to 500
feet.

Diagnosis. A Caecilia with primaries 185-214; secondaries 8-25;
1/d 48-93; eyes usually visible; no markings; length 165-680 mm.

Description. The measurements show that this form becomes slim-
mer with age. BMNH 66-8-14, 341, the largest specimen, has the eye
invisible but not covered by bone.

Data taken from Guianan specimens is altered by others as follows :

The primary count is raised from 207 to 214 by the Para specimen,
and the secondary count is raised by it from 23 to 25. The Vienna
specimen from "S. Amer." lowers the secondary count from 9 to 8.

Remarks. The Guiana population to which the name gracilis applies
is abundantly distinct from any other Guiana form, but is very confus-
ingly allied to some of the western forms of the genus.

C. thompsoni of the upper Magdalena is larger, and has more second-
aries, but the primary counts overlap those of gracilis in the range
188-214.

C. bassleri of Ecuador and Peru overlaps gracilis in all the numerical
counts (primaries from 206 to 214, secondaries from 14 to 25, 1/d from
85 to 93). On combining characters, all specimens can be placed in one
of the two forms. The two occur together at Iquitos, Peru.

C. pachynema of Ecuador and Peru overlaps gracilis in all the numeri-
cal counts (primaries 185-199, secondaries 8-11, 1/d 48-84). I have
seen seven specimens within this range of overlap in all counts. These
can only be allocated on the basis of color (when present in pachynema)
and by locality.

C. polyzona of the Cauca Valley, has all its numerical counts within
the range of those of gracilis. Specimens can be distinguished by color,
by visibility of the eye, and by locality.

C. ochrocephala of Panama and northwestern Colombia overlaps
gracilis in all the numerical counts (primaries from 185 to 192, second-
aries from 8 to 25, 1/d from 48 to 87). I have seen 26 specimens (8
gracilis and 18 ochrocephala) within this range of overlap in all counts.
These specimens have been allocated by color, by visibility of the eye,
and by locality.

DUNN: THE AMERICAN CAECILIANS

517

Specimens seen,

31, as follows:

prim.

sec.

length

diam.

1/d

British Guiana:

Dunoon

Michigan 47410

185

16

328

6

55

47411

187

10

385

5

77

47411

188

14

285

5

57

47411

192

13

200

4

50

47411

197

9

165

3

55

52507

187

14

321

6

53

Wismar

Michigar

l 76676

189

10/3

295

4.5

66

Maccasseema

BMXH

87-1-22,

30

199

14/3

330

5

66

Demarara

(t

—

—

No locality

USXM

58750

204

11/3

245

5'

49

Oronoque R. Field

35116

198

16/5

27s

3

93

Dutch Guiana:

Surinam

ANS

4923

207

23/0

440

5

88

4924

190

12/4

490

7

70

Surinam

BMXH

66-8-14,

341

195

22/9

680

11

62

70-3-10,

58

203

16/6

444

5

89

70-3-10,

60

202

17/5

472

6

79

it

Berlin

5826

199

15/2

367

4

92

Vienna

200

165

3

55

Munich

310

4

77

MCZ

6637

201

22/0

370

7.5

48

Paris

12

197

405

5

81

t(

Paris

12

204

•

540

7

77

French Guiana:

Cayenne

Paris

12b

—

—

—

a

tt

12c

—

—

—

"G

uiana

Paris

12d

Brazil:
Para

Vienna

214 25/6 370 4 92

Peru:

Iquitos

AMXH 42851 188 21/8 390 7 56

518 bulletin: museum of comparative zoology

South America:

prim.

sec.

lgth.

diam.

1/d

No locality

AMNH 23658
BMNH

190

14/5

398

6

66

1929-5-16, 2

184

14/7

500

6

83

Berlin 3700

197-

■ 9/2

500

6

83

a

Vienna

19S

8/4

420

5

84

Caecilia bassleri spec. nov.

Type. MCZ 19401.

Type locality. Pastaza R., Ecuador (Canelos to Maranon).

Range. Eastern Colombia; eastern and western Ecuador, eastern
Peru. Sea level to 500 feet.

Diagnosis. A Caecilia with primaries 206-285; secondaries 14-41;
1/d 80-160; 495-865 mm.; eyes visible or invisible.

Description. Uniform dark, head a little lighter. The eyes are invisi-
ble in AMNH 3874, in the Colombian specimen, and in MCZ 19401.
Three have the 1/d below 105. A single specimen has the primary
count below 227. The other eight have the 1/d over 104 (no other Cae-
cilia has the 1/d over 93) and primaries over 226 (no other Caecilia has
a primary count above 217, except the Panamanian elongata which has
no secondaries and the Colombian thompsoni, which is stouter).

Remarks. Probably allied to thompsoni and to gracilis. Both gracilis
and bassleri occur at Iquitos, Peru.

It is a pleasure to name this form, extreme alike in slimness and in
number of vertebrae, for my friend Dr. Harvey Bassler, whose collec-
tion of Peruvian Caecilians included five of this species.

Specimens seen, 12, as follows:
Western Ecuador: Prim' sec" length diam. 1/d

RioCayapas AMNH 3874 206 32/13 725 6 121
St. Javier BMNH

1901-3-29,66 251 14/3 832 7 119
BMNH
"60-6-16,86 227 14/4 650 5 130
No locality

Eastern Ecuador:
Rio Pastaza MCZ 19401 271 41/8 800 5 160
Canelos BMNH

80-12-8,141 254 28/0 495 4 124

DUNN: THE AMERICAN CAECILIANS 519

Eastern Peru: prim' SeC" length diam- 1/d

Iquitos AMNH 42852 285 17/0 865 7 124

Monte Carmelo, nr. Requena

lower Ucayali AMNH 45327 257 25 840 7 120

Pampa Hermosa, mouth of
Cushabatay, Mid. Ucayali

AMNH 42840 230 28/9 630 6 105
42841 232 30/6 655 6 109
Chaquimayo, Carabaya
BMNH

1908-3-11, 1 231 21/4 770 9 85
Mouth Rio Santiago

AMNH 42832 234 17 640 8 80

Eastern Colombia:

Rio Putumayo, Punto Asis

Inst. La Salle 244 25/6 800 10 80

Caecilia ochrocephala Cope

1866. Caecilia ochrocephala Cope, Proc. Acad. Nat. Sei. Philadelphia 18,
p. 132; Peters 1879, Mon. Ak. Berlin, p. 935; Boulenger 1882, Cat.
Batr. Grad. Brit. Mus. (2), p. 94; Brocchi 1883, Miss. Sci. Mex.,
Batr., p. 119, pi. 21, f. 1; Dunn 1931; Occ. Papers Boston Soc. Nat.
Hist. 5, p. 408.

1885. Herpele ochrocephala Cope, Proc. Amer. Phil. Soc. 22, p. 171; 1885,
Proc. Amer. Phil. Soc. 23, p. 279; 1887, Bull. U. S. Nat. Mus. 32, p.
9; Boulenger 1895, Proc. Zool. Soc. London, p. 409; Gunther 1902,
Biol. Centr. Amer., Rept., p. 307; Nieden 1913 (in part) Gymnophi-
ona, p. 20; Dunn 1928, Proc. New England Zool. Club 10, p. 73.

1876. Caecilia gracilis Garman, Proc. Boston. Soc. Nat. Hist. 18, 412.

1906. Caecilia sabogae Barbour, Bull. Mus. Comp. Zool. 46, p. 228 (Saboga
Island, Panama. Types MCZ 2425).

Type. USNM 29764, collected by Gallaer and LeConte.

Type locality. Atlantic side Isthmus of Darien.

Range. Province of Code, Panama to Turbo, Colombia. Sea level
to 2000 feet.

Diagnosis. A Caecilia with 171-192 primaries; 7-29 secondaries;
149-179 primary folds without secondaries; eyes invisible; pale gray,
with black primary grooves; 1/d 39-87; length 151-610 mm.

520 bulletin: museum of comparative zoology

Description. The eyes are invisible in all specimens seen. Nearly all
specimens are colored as Cope described the type, "yellowish plumbe-
ous. The plicae dark; head and throat ochre yellow." A specimen from
Panama Sabanas in the MCZ is pale and uniform; USNM 52486 from
"Panama" has paired light dorsolateral spots on each segment, thus
resembling C. pachynema in color.

Outer mandibular teeth 8-9 on a side, inner 3-4, maxillary teeth
7-9 on a side; palatine teeth 10-13 on a side.

The range of variation in primary and secondary count found in 71
Canal Zone specimens is slightly exceeded by one from Cana which
raises the primary count from 189 to 190; by one from San Miguel
Island which raises the primary count to 192; by the type and the Cana
specimen which raise the secondary count from 28 to 29.

Of 100 primary counts, 89 are from 174-188, six are below this range
and five above. Of 98 secondary counts, 83 are from 10-25, seven are
below this and eight above. Of 94 1/d ratios, 84 are from 41-65, four
are 39-40, six (four in one lot and poorly preserved) are 66-87.

Remarks. C. ochrocephala has been taken in excavations on Barro
Colorado Island at a depth of some ten feet below the surface.

In 1928 I reported finding eggs 3 mm. in diameter in a female, four
on the right side and five on the left.

C. ochrocephala is similar to C. polyzona of the Cauca valley in color
and in condition of the eye, but has a lower primary count.

C. ochrocephala is similar to C. gracilis of Guiana, Brazil and eastern
Peru in numerical counts (8 gracilis and 18 ochrocephala fit in the region
of overlap of the three counts), but differs in color and in condition of
eye.

C. ochrocephala is also similar to C. pachynema of Colombia, Ecuador
and Peru in numerical counts (6 pachynema and 19 ochrocephala fit in
the region of overlapping of all three counts; primaries 171-192, sec-
ondaries 7-11, 1/d 40-84). It differs from pachynema in color (usually)
and in condition of eye.

Specimens seen, 101, as follows:

Panama Canal Zone:

prim.

sec. 1

ength

cliam.

V<

Fort Sherman MCZ

9610

183

10/2

382

9

42

10665

176

13/0

233

6

39

10671

173

22/1

610

12

51

Cristobol Iowa State

174

15/0

250

5

50

DUNN: THE AMERICAN CAECILIANS

021

prim,

sec.

length

diam. l/d

Gatun

AMNH

6644

178

21/4

414

7

59

MCZ

9589

176

23/5

490

—

—

9590

183

16/0

410

7

58cf

9591

183

12/0

402

8

50

9592

173

18/0

430

10

43

9593

178

23/3

355

9

39

9594

179

15/3

382

8

48

«

tt

9595

181

29/3

:;25

7

46

San Pablo

MCZ

1306

1S3

24/4

242

5

48

tt

AMNH

18670

184

10/0

310

6

51

Gorgona

MCZ

1493

175

23/3

375

7

53

Monte Lirio

MCZ

14816

171

20/2

515

10

51

Chagres River

MCZ

16289

182

26/4

285

6

64

Indio on Chagres

USNM

102850

184

16/4

211

4

53

Barro Colorado I.

MCZ

11855

173

18/2

375

i

53

11856

172

23/2

310

t

44

Majagual

tt

10672

174

20/2

382

7

54

10673

175

25/5

244

6

41

10674

179

25/5

328

7

47

Summit Lindsay coll.

—

—

— -

Albrook Field E. R. Dunn coll.

185

17/0

261

6

43

186

8/8

302

6

50

«

a

182

13/0

294

6

49

Corozal

Carnegie

8698

184

13/3

335

6

56

tt

(<

8699

179

10/0

340

<

49

a

MCZ

16290

185

12/3

370

7

53

a

i *

16291

179

12/3

365

i

52

16292

177

12/0

330

6

55

16293

182

16/0

328

6

55

16294

175

17/0

248

5

50

16295

186

12/2

355

6

59

16296

185

10/0

350

7

50

16297

185

13/4

320

6

53

16298

185

9/0

208

4

52

tt

it

17888

184

9/0

390

8

49

Fort Clayton

USNM

65845

186

10/1

400

7

57

tt

a

65846

181

28/1

390

7

56

522

bulletin: museum of comparative zoology

prim.

sec.

length

diam

. 1/d

Fort Clayton MCZ

14821

177

19/3

291

6

48

14822

179

11/3

354

6

59

14823

183

17/5

395

7

56

14824

179

8/0

360

7

51

14825

175

7/1

344

8

43

14826

179

12/0

265

5

53

14827

183

21/3

436

7.5

58

14828

176

17/0

308

7

44

14833

185

17/0

262

5

52

14834

185

25/0

190

4

47

14835

188

9/1

260

4

65

14836

183

17/0

210

4

52

14837

183

13/3

151

3

50

14838

184

19/0

290

o

58

15721

180

22/3

295

6

49

a <<

15722

184

24/5

258

5

52

Balboa Carnegie

8490

184

24/4

457

8

57 cT

a (I

8491

181

20/4

356

6

59

Ancon MCZ

8600

173

14/0

545

12

45

10675

178

19/2

382

7

54

14817

189

27/5

525

8

66

14818

180

17/4

435

5

87

14819

178

20/3

472

7

67

Aneon MCZ

14820

183

23/2

522

t

*7C

id

BMNH

1926-1-1

>0, 72

177

16/3

350

7

50

Ancon or Balboa

MCZ

14829

183

25/4

420

10

42

Madden Dam AXS

21825

185

16/3

325

5.5

59

21826

183

26/4

335

5

67

"Canal Zone" USNM

37857

189

17/2

428

7

61

37858

182

11/3

407

8

51

Panama

Rio Grande to Rio Code,

Prov. Code MCZ

181

13/2

321

8

40

Xombre de Dios

MCZ

14832

182

10/0

290

7

41

Panama City

4268

181

15/0

331

9

46

15719

183

10/0

410

—

—

15720

180

21/3

450

11

41

DUNN: THE AMERICAN CAECILIANS

523

Panama-Sabanas MCZ

prim. sec. length diam. 1/d

177 10/0 370 9 44
179 15 0 352 8 53

178 8/0 420 8 40

Saboga I.

«

2425

181

24/4

375

6

50

<<

<<

2425

181

11/3

—

—

San Miguel I.

n

2503

192

21/4

470

11

43

Cafia, 2000'

USNM

50249

190

29/2

490

10

49

Atlantic side Darien

USNM

28185

180

10/2

420

6

70

a

USNM

29704

185

29/2

330

6.5

50

Darien Isthniu

s Vienna

175

—

—

—

a

a

189

—

—

— ■

"Panama"

Paris 7 i
BMXH

ilpha

184

21

—

—

87-12-1!

2,1

182

IS 2

531

10

53

BMXH

94-5-9, 6

185

21 4

450

9

50

USNM

14116

181

20/5

428

8

53

(<

52495

187

27/3

550

10

55

<<

52496

185

12/4

505

11

46c?

a

52497

180

18/5

480

9

53

tt

181

10/3

490

12

41

MCZ

1521

187

23/0

535

10

53

tt

tt

178

19/4

355

—

—

AXS

4919

185

12/2

425

7

60

MCZ

2502

185

20 7

432

7

61

Colombia :

Turbo

MCZ

1492

185

23/3

355

7

50

"Brazil"

MCZ

188

10/0

495

9

55

( Iaecilia polyzoxa Fischer

1879. Caecilia polyzona Fischer, in Peters, Mon. Berlin Ak., p. 936; Fischer
1880, Arch. Xaturg. 46, 1, p. 215, pi. 8, f. 1-4; Boulenger 1882, Cat.
Batr. Grad. Brit, Mus. (2), p. 94; 1895, Proc. Zool. Soc. London,
p. 407.

1913. Herpele ochrocephala Xieden (in part), Gymnophinna, p. 20.

524 bulletin: museum of comparative zoology

Type. Originally two specimens in the Berlin Museum, collected by
Grosskopf. One is now AMXH 23449, the whereabouts of the other is
not known.

Type locality. Caceres, Prov. Antioquia, Colombia.

Range. Cauca Valley, Colombia.

Diagnosis. A Caecilia with 204-209 primaries; 10-17 secondaries;
1/d 43-67; eyes invisible; "brownish gray, head little lighter, grooves
black, light gray below"; length 560-670 mm.

Description. Fischer (1880) gives a count of the dentition which ap-
parently includes the teeth of both sides; 22-25 maxillary; 20-22
vomerine; 20 outer mandibular; 10-12 inner mandibular. The missing
type had 209 primaries, 10 secondaries, 1/d 59, and was 650 mm. long.

Remarks. C. polyzona is very close to ochrocephala, the only differ-
ence being the higher primary count.

Specimens seen, 2, as follows:

prim. sec. length diam. 1/d

Colombia :

Cauca Valley Vienna 204 17/0 560 13 43

Caceres AMXH 23449 207 12 670 10 67

Caecilia pachynema Gunther

1859. Caecilia pachynema Gunther, Proc. Zool. Soc. London, p. 417; Cope
1868, Proc. Acad. Nat, Sci. Philadelphia, p. 118; Peters 1879, Mem.
* Ak. Berlin, p. 935; Boulenger 1882, Cat. Batr. Grad. Brit. Mus. (2),
p. 95, pi. 6, f. 2; 1895, Proc. Zool. Soc. London, p. 407; Nieden 1913,
Gymnophiona, p. 13; Parker 1934, Ann. Mag. Nat. Hist. (10), 14,
p. 265.

1884. Caecilia buckleyi Boulenger, Ann. Mag. Nat. Hist. (5), 13, p. 398
(Intac, Ecuador, type BMNH 78-1-25, 47); 1895, Proc. Zool. Soc.
London, p. 407, pi. 23, f. 1; Nieden 1913, Gymnophiona, p. 13.

Type. BMXH 66-6-16, 87, collected by Fraser.

Type locality. Western Ecuador.

Range. Colombia; Ecuador; Peru. Sea level to 6200 feet.

Diagnosis. A Caecilia with primaries 154-199; secondaries 0-11;
1/d 38-84; eyes visible; usually with a lateral yellow spot on each seg-
ment; length 145-900 mm.

Description. MCZ 16288 has 6 maxillary teeth, 4 palatine teeth, 4
large mandibular teeth, and 1 inner mandibular tooth. Gunther (1859)

DUNN: THE AMERICAN CAECILIANS 525

says that the type has "5 hook-like teeth on each side, the anterior
larger, three palatal teeth." Boulenger (1882) says of the same speci-
men "teeth large, few, widely separated; inner mandibulars very small
few;, outer mandibulars very large, anterior largest, 5; maxillaries 6;
vomeropalatines 9." Cope (1868) says that two specimens from Guay-
aquil had 8 maxillary teeth, 6 mandibulars, and 5 palatine. The type
of buckleyi had (Boulenger 1884) "maxillary teeth large, 10 on each
side. Vomeropalatines 8 on each side. Inner mandibulars small, few.
Outer large — 9 on each side."

All specimens with secondaries have scales. Those without second-
aries have scales or lack them in about equal numbers (cf. Parker 1934).

Xo scales, 6: Zamora; type pachynema; type buckleyi; Normandia,
Villa vicencio; Medina Mts.

Scales, 4: Milligalli, Intac (2), Pallatanga.

Colombian specimens lower the Ecuadorian primary range from 160
to 159 and 154, and lower the 1/d ratio from 40 to 38; Peruvian speci-
mens raise the Ecuadorian primary range from 192 to 194 and 199,
and raise the Ecuadorian secondary count from 10 to 11.

The primaries are usually interrupted dorsally and ventrally. Most
specimens have a large yellow spot on each side of each segment.

Remarks. Two specimens in the Berlin Museum (3716 and 3722)
from Guayaquil are labeled as types of C. guntheri Peters. This is in-
correct as the type of guntheri Peters (a substitute name for C. rostrata
Gunther, not C. rostrata Cuvier which is now called Hypogeophis ros-
tratus) is BMNH 60-6-16, 85, the specimen erroneously called C.
rostrata by Gunther.

The type of C. buckleyi Boulenger seems to me a very young speci-
men of pachynema, which is stouter than larger individuals. A still
smaller individual from Colombia is even stouter.

Specimens with primaries 185-199, secondaries 8-11, and 1/d 49-
84 fit into the diagnostic counts of both pachynema and gracilis, and
have been allocated by color and by locality.

Specimens with secondaries 7-11 agree in almost all counts with
ochroccphala. These have been allocated on color, condition of eye, and
locality.

This species occurs on both sides of the Andes, but is known to reach
an altitude of 6200 feet at Milligalli (record altitude for an American
Caecilian) and, if the Quito locality is correct, which I doubt, it reaches
9274 feet.

526

bulletin: museum of comparative zoology

Specimens seen, 25, as follows:

Colombia:

prim.

sec.

length

diam

. 1/d

Medellin

ANS 12980

159

2

519

10

52

AMNH 49973

166

7

467

6

78

Villavicencio

Inst. La Salle

180

0

270

5

54

Medina Mts.

N. E. Villavicencio

AMNH 49959

156

0

145

4

38

Xo data

M( Z 16288

154

0

900

15

60

Ecuador :

Guayaquil

Berlin 3716

168

0

468

7

66

Berlin 3722

172

10/7

490

6

81

USNM 12353

183

6/2

380

4.5

84

BMNH

85-2-23, 15

172

8/4

380

6

63

Quito

Paris 11

181

5

780

11

71

Intac, 3000'

BMNH

78-1-25, 46

162

0

476

10

47

47

181

0

160

4

40

Michigan

158

8/0

750

12

63

a

172

0

745

12

62

Milligalli, 6200

' BMNH

85-2-23, 14

164

0

820

12

68

Pallatanga, 5000'

AMNH 16986

160

0

630

12

52

Normandia, Zunia, Upana R.,

1400-1800 m

.AMNH 23434

174

0

726

10

73

"Western Ec."

BMNH

66-6-16, 87

168

0

468

7

66

No data

Munich 148/1912

169

0

550

12

46

t(

Vienna

192

10

375

6

62

Peru:
No data

Vienna

169

11/0

415

*7

i

59

It

«<

194

0

410

6

68

a

it

199

0

385

6

64

u

ANS 16129

165

11/0

211

3

70

No data:

Smith College

190

2

485

8

60

Parker (1934) has recorded pachynema from Zamora, Ecuador, 3250
feet, a locality which is like Normandia, on the east face of the Andes.

DUNN: THE AMERICAN CAECILIANS 527

Caectlia eloxgata spec. nov.

Type. Munich 1327/0.

Type locality. Panama.

Range. Known only from Yavisa, Darien, Panama.

Diagnosis. A Caecilia without secondaries; primaries 226-231; 1/d
83-89; no scales; eye invisible; no markings; 500-620 mm.

Remarks. The high primary count and the complete absence of
secondaries render this form quite distinct. The head-and-neck from
Yavisa looks quite different from ochrocephala.

Specimens

seen

, 3 ,as follows:

Panama :

prim.

length diam. 1/

Yavisa

MCZ (head and neck
only)

No locality

Munich 1327/0

226

(320 7 89

a

1324/0

231

500 6 83

Chthonerpeton Peters

1879. Chthonerpeton Peters, Mon. Berlin Ak., p. 930, 940 (monotype Sipho-
nops indistinctus Reinhardt and Ltitken).

Diagnosis. Caecilians without secondaries or scales; no tail; tentacle
in a horseshoe-shaped groove, on side of head between eye and nostril ;
eyes visible; two rows of teeth in lower jaw; anus usually a well devel-
oped sucking disk; primaries 76-166; 1/d 23-57; length 170-620 mm.;
three species.

Range. Argentina, Uruguay and Brazil.

Key to species of Chthonerpeton

A. Primaries 76-87; (anal disk large; tentacle slightly nearer nostril
than eye) ; Argentina, Uruguay, and southern Brazil . indistincfum
AA. Primaries 133-166.

B. Tentacle very close to nostril; anal disk medium; 145 prim-
aries ; upper Amazon petersi

BB. Tentacle nearer eye than nostril; anal disk small; 133-166
primaries; southern Brazil viviparum .

Remarks. In number of primaries and size of anal disk indistinctum
is at one extreme of the genus and resembles the closely allied forms of

528 bulletin: museum of comparative zoology

the more northern genus Typhlonectes. In position of the tentacle,
however, petersi is extreme and in this character it is the closest to
Typhlonectes.

I have examined 39 specimens, including the types of viviparum
and petersii. I have not seen the type of indistinctum.

Chthonerpeton viviparum Parker and Wettstein

1907. Siphonops braziliensis (non Lutken) V&vra, Vesmir, Prag, 36, 1, p. 11,
f. 10 (not seen); Nieden 1913 (in part) Gymnophiona, p. 25.

1929. Chthonerpeton viviparum Parker and Wettstein, Ann. Mag. Nat. Hist.
(10) 4, p. 594.

Type. BMNH 1907-8-28, 1.

Type locality. State of Santa Catharina, Brazil.

Range. States of Santa Catharina and Sao Paulo, Brazil.

Diagnosis. A Chthonerpeton with 133-166 primaries; tentacle nearer
eye than nostril; anal disk small; 1/d 36-57; 170-510 mm.

Description. (Original from type, eight from Joinville in Vienna
Museum and one without data in Vienna. I have seen only the type)
"primaries "143-144 and 147-149," "possibly sexual fe-
males lower" [of my four additional one has 133 and one 166,

so that the range is 133-166]; "largest 404, 339, and 333, mm., with
diameters of 8, 8, and 7.5 mm. respectively; a half-grown specimen
170 mm diameter 4, 5 mm." These figures indi-
cate 1/d ratios of 51, 42, 44, and 36 respectively; the range of specimens
I have seen being 37-57. Obviously the older are slimmer. "Tentacle

nearer the eye than the nostril, and a little below the straight

line connecting these two"; "premaxilla and maxilla 12-13 a side;
palatopterygoid 9-10 a side ; mandible, outer row 11-12, inner row 3-4" ;
"greyish-olive with a purplish tinge"; embryo with a "single pair of
plate-like gills which are closely apposed to the inner walls of the ovi-
ducts, no trace of an egg-capsule being found." "Each gill is, in reality,
an oval plate lying parallel to the sagittal plane of the embryo, and con-
nected to its neck by an exceedingly short peduncle which arises from
the gill at a point rather dorsal and anterior to its geometric centre."
Length of embryo 62 mm., diameter 2.5 mm., length of gill-plate 14
mm.

Remarks. Siphonops brasiliensis occurs with this form in the state of
Santa Catharina, Brazil, and the two animals have been confused by

DUNN: THE AMERICAN CAECILIANS 529

Vavra, by Nieden and by me. The Hamburg specimen was actually
in my notes as the Siphonops. Except for one specimen of each form
(with 133 primaries) the Siphonops has fewer primaries and the Chtho-
nerpeton has more. The skull and the dentition afford perfectly diag-
nostic characters, and the anal disk of the Chthonerpeton is usually
distinct enough.

Specimens seen, 5, as follows:

prim, length diam. 1/d

166

510

9

57

144

215

5

43

148

355

8

44

144

335

9

37

Brazil:

Sao Paulo:

Franca Mus. Paul. 950 133 3S0 8 47

Santa Catharina:
Joinville MCZ 24593

No locality Mus. Nac. Braz. 829

Hamburg 1937
BMNH 1907-8-21, 1

Chthonerpeton petersii Boulenger

1882. Chthonerpeton petersii Boulenger, Cat. Batr. Brit. Mus. (2), p. 104,
pi. 9, f. 2; 1895, Proc. Zool. Soc. London, p. 411; Nieden 1913,
Gymnophiona, p. 24.

Type. BMNH 51-9-2-6.

Type locality. Upper Amazon.

Range. Known only from type locality.

Diagnosis. A Chthonerpeton with 145 primaries; 1/d 39; tentacle
much closer to nostril than eye.

Description. I see no reason to alter Boulenger's original description,
which follows. "Teeth small, numerous, subequal. Snout rounded,
moderately prominent; eyes not distinct through the skin; tentacle
close to and behind the nostril. Body elongate; 145 circular folds, com-
plete except the anterior 28, which are interrupted on the dorsal and
ventral line. Tail indistinct, rounded. Uniform dark olive-grey, the
eyes indicated by a whitish spot. Total length 620 millim.; greatest
diameter of body 16 millim." "Anus largish, but smaller than in
indistinctum." (Parker, in litt.)

Specimens seen, 1, the type.

530 bulletin: museum of comparative zoology

Chthonerpeton indistinctum (Reinhardt and Liitken)

1861. Siphonops indistinctus Reinhardt and Liitken, Vid. Meddel. Kjoben-
havn, p. 203; Dumeril 1863, Mem. Soc. Sci.Nat. Cherbourg, p. 318,
p. 1, f. 3; Wiedersheim 1879, Anat. Gymnophiona, pi. 2, f. 13, 15,
16, pi. 6, f. 68, pi. 9, f. 84-7.

1879. Chthonerpeton indistinctum Peters, Mon. Ak. Berlin, p. 929, 940, f. 9!
Boulenger 1882, Cat. Batr. Brit. Mus. (2), p. 104; Cope 1889, Bull-
USNM 34, pi. 11, f. 1-6; Boulenger 1895, Proc. Zool. Soc. London,
p. 411; Ihering 1911, Rev. Mus. Paulista, p. 107; Nieden 1913,
Gymnophiona, p. 24, f. 8; Procter, 1923. Ann. Mag. Nat. Hist. (9),
11, p. 230; Gliesch 1929, Blatt. Aqua. Terr. 40, 13, p. 229, pi. 31;
Gaggero 1934, Prelim. Ann. Mus. La*Plata 3, 1, p. 173.

Type. In Copenhagen Museum. Not seen. Collected by Prof.
Kroyer of the frigate Bellona.

Type locality. Buenos Aires, Argentina.

Range. Argentina (Buenos Aires), Uruguay, Southern Brazil as
far as Parana.

Diagnosis. A Chthonerpeton with 7G-S7 primaries; 1/d 20-45; anus
much enlarged; tentacle slightly nearer nostril than eye; length 119—
593 mm.

Description. Primaries 76-87; only three out of 27 specimens over 83;
primaries indistinct and incomplete, distinct only on belly; tentacle a
flap in a horseshoe-shaped groove, posterior to nostril, nearer to it
than to eye (not so close to nostril as in petersii)', uniform dark in color;
anus in a large sucking disk; length from 119-593 mm.; no apparent
change in length-diameter ratio with increase in size. A single -105 mm.
specimen has 1/d 45; 20 others have 23-36. Uniform dark in color.
According to Peters (1879) and Wiedersheim (1879) the teeth are as
follows: vomerine 5-3; palatine 8-5; premaxillary 5-6; maxillary 9-8;
mandibular, 13 outer and 4 inner. Argentine specimens (10) have 76-
81 primaries (the tvpe had 78); specimens from Brazil and Uruguay
(12) have 76-87.

Habits. Peters (1879) speaks of its being taken from "deep in the
earth"; while Gleisch (1929) tells of a 405 mm. specimen in Porto
Alegre, during a rain, being on the surface, apparently in a gutter, and
noticed the enlarged anal disk functioning as a holdfast or sucker.

Remarks. This, the most common and best known of the species, is
in some wTays the most extreme. Reinhardt and Liitken, in the original
description, mention a specimen in the Paris Museum, from Brazil,
which they considered this species. They also had one from Buenos

DUXX: THE AMERICAN" CAECTLIAXS

531

Aires with 78 primaries. The Paris specimen was said to have 100
primaries, and to this day indistinctum is always said to have 78-100
primaries. I found two specimens in Paris labeled Cthonerpeton indis-
tinctum; Paris 17, Brazil, primaries 91 or 92, length 261 mm., diameter
6 mm., 1/d 43; and Paris 17a, Buenos Aires, primaries 78. Paris 17 is,
I imagine, the specimen referred to by Reinhardt and Liitken. It is a
young Siphonops annulatus, very dry, and with the tentacle much
further from the eye than usual in that species, and I imagine that it
was this feature which misled Reinhardt and Liitken. Dumeril (1863)
counted 98 primaries and mentions the white grooves and the tentacle
position.

Specimens seen, 33, as follows :

Argentina

j

^llUl.

LV, Alt, til

/

Southern Argent.

Munich

81

593

20

30

Buenos Aires

Paris 17a

78

405

9

45

(c (t

AMNH 11949

79

251

9

28

tt a

Frankfort 2104a

81

430

15

29

Xo locality

Mus. Paul. 959

78

317

10

32

959a

77

247

9

27

Berlin 26340

78

—

—

Berlin 26340

76

— ■

—

it n

a tt

—

119

4

30

AMNH 23508

80

188

5.5

34

Isla Ella, R. Parana

BMXH 1926-5-29-17

80

— ■

—

Uruguay:

Durazno

USNM 65538

78

240

9

27

Brazil :

Rio Grande do Sul

Porto Alegre

Berlin 9559

80

355

13

27

it tt

" 6803

80

160

6

27

tt a

a tt

—

160

5

32

tt tt

a it

86

210

7

30

tt tt

a a

80

330

11

30

tt a

BMXH 83-1-19-2

83

189

6

31

it it

AMXH 23507

76

163

4.5

36

Camaquam R.

BMXH 89-8-24-4

87

365

14

26

532 bulletin: museum of comparative zoology

prim, length diam. 1/d

No locality AMNH 23506

Berlin 10458

MCZ 1501
Santa Catharina Hamburg 5

Joinville Mus. Nac. 844
Parana " "846

Castro Tibeira, Rio Parana

BMNH 1922-11-23-10
No state or locality Berlin 16445

77

165

6

27

79

165

7

24

80

270

9

30

76

260

11

24

82

390

13

30

82

300

13

23

83.

300

13

23

1

^head only)

11

Central America" Hamburg 1715 76 278 11 25

No data ANS 13948 76 205 9 23

Gaggero (1934) has recorded it from the mouth of the Rio Santiago,
Argentina.

Procter (1923) states that the Castro Tibeira specimen was taken
from the belly of a Sorellina brandon-jonesii.

Typhlonectes Peters

1879. Typhlonectes Peters, Mon. Berlin Ak., p. 930, 941 (type compressicauda) .
1912. Thyphlonectes Peracca, Mem. Soc. Sci. Neufchatel, 5, p. 111.

Xote. I hereby designate Caecilia compressicauda Dumeril and Bib-
ron as the type of Typhlonectes Peters. The genus as originally de-
scribed contained compressicauda, dorsalis, natans, and, with a query,
sipitremus.

Diagnosis. Caecilians without scales or secondaries; primaries
poorly developed; eyes visible; two rows of teeth in lower jaw; no tail;
anus in a well developed sucking disk; tentacle in a horseshoe-shaped
groove, on side of head very close to and posterior to nostril ; laterally
flattened with a dorsal keel and fin in the posterior part of body;
aquatic; primaries 77-105; 1/d 12-41; length 140-695 mm.; two spe-
cies; three forms.

Range. Colombia, Venezuela, Guiana, Brazil; Atrato, Magdalena,
Orinoco and Amazon svstems.

DUNN: THE AMERICAN CAECILIANS 533

Key to forms of Typhlonectes

A. Somewhat compressed; dorsal fin restricted to posterior; head
large.

B. Primaries 77-87; Guiana and Brazil

compression uda compressica uda

BB. Primaries 86-105; Venezuela and Colombia

compressicauda natans
AA. Extremely compressed; dorsal fin nearly to head; head small;
primaries 88-104; Venezuela, Guiana, and Brazil kaupii

Remarks. The genus is very close to Chthonerpeton. Among the
forms, kaupii is decidedly the most specialized.

I have examined 5S specimens, including types of compressicauda,
natans, venezuelense, and dorsalis. I have not seen the type of kaupii,
or of microcephala.

Typhlonectes compressicauda compressicauda (Dumeril and

Bibron)

1841. Caecilia compressicauda Dumeril and Bibron, Erp. Gen. 8, p. 278;

Gray 1850, Cat. Batr. Grad. Brit. Mus. p. 57; Dumeril 1863, Mem.

Soc. Sci. Nat. Cherbourg 9, p. 316; Peters 1874, Mon. Berlin Ak.,

p. 45 (habits); 1875 idem, p. 683, f. 1-4 (habits).
1879. Typhlonectes compressicauda Peters, idem, p. 941, f. 11 ; Boulenger 1882,

Cat. Batr. Grad. Brit. Mus. (2), p. 102; Sarasin and Sarasin 1887,

Ergeb. Forsch. Ceylon, 2, p. 26, fig.; Boulenger 1895, Proc. Zool.

Soc. London, p. 4111; Nieden 1913, Gymnophiona, p. 22, f. 16.
1912. Thyphlonectes compressicauda Fuhrmann, Mem. Soc. Sci. Neufchatel,

5, p. 119, f. 4.

Type. Paris 18.

Type locality. Cayenne.

Range. British Guiana to Para, Rio Purus, and Rio Solimoes.

Diagnosis. A Typhlonectes with fin in posterior third of body;
primaries 77-87; 1/d 12-2-1; length 14S-515 mm.

Description. In general the larger specimens seem to be slimmer, but
in these more or less flattened forms the length-diameter ratio is very
uncertain. Only two specimens have the primaries below 83. Second-
aries have been recorded by Fuhrmann (1912) on material which I have
not seen. In this species, and still more in natans, the primaries are

534 bulletin: museum of comparative zoology

very hard to count, and indistinct. At the same time adventitious
folds appear and are liable to be taken for primaries or secondaries or
both. The primaries are all incomplete on the back, and no true second-
aries are present. The color is uniform blackish.

Habits. Peters (1874) states from a communication of Jelski that an
adult female was taken in a fishing net in the Kaw river, eastern
Guiana. Between the river and the house she gave birth to a single
young. She was immediately killed, and five more young were found in
the oviducts. The mother measures 500 mm., the young one was 157
mm. long, and an embryo measured 136 mm. No gill slits were ob-
served, but there were two large, flattened, allantoic gills which meas-
ured 55 mm. Peters (1875) figures one of these, and so do the Sarasins
(1887).

A specimen in the American Museum, from Manaos, was found "in
a dead log come up out of the water."

Remarks. This beast is closely related to ■natans. The ranges are
adjacent, and the relationship so close that I regard the two as races.
Fuhrmann (1912) came to the conclusion that Guiana compressicauda
and Venezuelan animals were racially different, since according to him
a specimen from Guiana had 84 primaries, and two Venezuelan speci-
mens had 94 and 95 primaries. He therefore called the Venezuelan
form Typhlonectes compressicauda venezuelense. His primary count for
this form does not differentiate it from natans, for which he himself
gives 90-95. He regarded the difference between compressicauda and
nata?is to be the higher head and close approximation of nostril and
tentacle in compressicauda, as against the flatter head and less approxi-
mated tentacle and nostril of natans. These differences are scarcely
appreciable, although Fuhrmann figures them. His figure of venezuel-
ense, furthermore, shows a quite intermediate condition in these two
respects between his figures of compressicauda and of natans. I regard
the primary count as of more importance, and judged by that criterion,
venezuelense is indistinguishable from natans.

Specimens seen, 17, as follows:

prim, length diam. 1/d

British Guiana:

Harauruni Cr., Demarara R.

U.Mich. 82854 ? 79 — — —

French Guiana:

Kaw River Berlin 8170 157 10 16

larva

DUNN: THE AMERICAN CAECILIANS

535

No data

No locality

Paris 18

18

18

18a
Vienna

prim. lgth. diam. 1/d
83 — —

TYPE
TYPE

9 15

TYPE

20 23

83
85

Brazil :
Para MCZ 289

Monte Alegre, Grande Para

84 150

455
148

85 495 25 20

BMNH 1926-10-28-7

84

375

23

16

11

n

BMNH 1926-10-28-7

— ■

175

15

12

Manaos

AMNH 12979

87

19

—

«

BMNH 93-4-24-2

84

515

24

21

«<

1913-3-11-1

84

215

10

21

H

1916-4-12-1

77?

—

—

a

" 1916-4-12-2

85

391

27

14

Ayapua,

Rio Purus

Berlin 31991a

86

345

20

17

Solimoes

, Amazonas

Berlin 30991b

—

435

18

24

Typhlonectes compressicauda natans (Fischer)

1879. Caecilia natans Fischer, in Peters, Mon. Berlin Ak., p. 941; 1880 Arch.

Naturg. 46, 1, p. 217, pi. 8, f. 5-7.
1879. Typhlonectes natans Peters, loc. cit., p. 941; Boulenger 1882, Cat. Batr.

Grad. Brit. Mus. (2), p. 103, pi. 9, f. 3; 1895, Proc. Zool. Soc. London

p. 411; Nieden 1913, Gymnophiona, p. 23, f. 17; Heimroth 1915,

Blatt. Aqua. Terr. 26, p. 34 (habits).
1912. Thyphlonectes natans Peracca, Mem. Soc. Sci. Neufchatel, 5, p. Ill;

Fuhrmann 1912, t. c, p. 113, f. 1-3.
1888. Typhlonectes compressicaudus Cope, Journ. Morph. 2, 2, 1889, pi. 22,

f. 5 (otic region); 1889, Bull. U. S. Nat. Mus. 34, pi. 51, f. 20 (hyoid).
1912. Thyphlonectes compressicauda venezuelense Fuhrmann, loc. cit., p. 124.

f. 5-6 (Maracaibo, Venezuela, COTYPE Hamburg 823).

Types. Berlin 9522-3, 3772; AMNH 23486; BMNH 81-4-9, 5.
Type locality. Cauca R., Colombia.
Range. Colombia, Venezuela, Trinidad.

Diagnosis. A Typhlonectes with dorsal fin in posterior part of body;
primaries 86-105, incomplete; 1/d 15-41; length 140-615.

536

bulletin: museum OF COMPARATIVE ZOOLOGY

Description. Only a single specimen out of 18 has the primaries above
the range of 86-97. Peters and Fuhrmann have both counted second-
aries in this form. I have been unable to make any out. There is no
clear indication of any change of proportions with age. I find four
inner mandibular teeth in U. Mich. 60881. Fischer (18S0) gives for
this row 14, and for the outer mandibular 38, for the maxillary 40-42,
and for the vomerine 34-36. These are total counts, and the Michigan
specimen then has 8 inner mandibular teeth as against 14 in the type.

The color is rather uniformly dark. Fischer says a little lighter
below. The anal disk is white. Its diameter in the two types was 7 and
6 mm. respectively.

Habits. The types were taken by fishing. A specimen from Quesada
R. was "floundering in test pit." Peracca (1914) says "caught on a
line at Puerto Berrio in the Magdalena." Heimroth (1915) says he re-
ceived a 4S0 mm. female July 14, 1914. On Jan. 16, 1915, four young
were born to her. They measured 190-200 mm. and had no sign of
gills. MCZ 24525 is 140 mm. long. It has no sign of gills or gill slits.

Remarks. The relationship of natans to comprcssicauda, and the
status of venezuehnse as a synonym of natans have been dealt with
under comprcssicauda. The Bogota specimens must be mislabelled.

Specimens seen, 22, as follows :

Colombia:

prim.

Igth.

diam. 1/d

Caceres

BMNH 81

369

11 33

TYPE

(t

Berlin 95

22

93

462

15 30

TYPE

(<

" 37

72

90

475

25 19

TYPE

a

AMXH 5

!3496

94

511

18 28

TYPE

Cauca

Berlin 95

23

103

525

15 35

TYPE

Cauca R., W.

of Medellin

Inst. La Salle

88

410

10 41

Honda

AMNH

22592

89

485

18 27

a

MCZ

9316

92

325

12 27

Medellin

AMNH

12978

96

540

25 22

Ciicuta

MCZ

24524

91

580

30 19

tt

MCZ

24525

92

140

8 17

Bogota (?)

AMNH

23418

86

440

14 31

a

<(

23419

—

470

15 31

a

tt

23420

445

18 25

DTJNN: THE AMERICAN' CAECILIANS 537

Quesada R., Atrato R.

prim.

lgth.

diam.

1/d

AMNH

49978

90

440

18

24

Sopleviento

U. Mich.

60881

87

310

20

15

it

tt

it

88

250

14

18

a

(i

a

86

285

14

20

Barranquilla

Hamburg

335

87

285

13

22

Venezuela :

Maracaibo

Hamburg

823

97

345

15

23

COTYPE renezuelense

Trinidad :

Dresden

639

South America:

\ lPTinn

105

615

23

26

"Belize"

USMN

30534

92

325

12

27

It has been reported from Puerto Berrio, on the Magdalena, by
Peracca (1914).

Typhlonectes kaupii (Berthold)

1859. Caecilia Kaupii Berthold, Nachr. Ges. Gottingen, p. 181.

1867. Siphonops Kaupii Keferstein, idem, p. 361.

1891. Typhlonectes kaupii Boulenger, Ann. Mag. Nat. Hist. (6), 8, p. 457;

1895, Proc. Zool. Soc. London, p. 411; Nieden 1913, Gymnophiona,

p. 23.
1877. Caecilia dorsalis Peters, Mon. Ak. Berlin, p. 459, f. 1-3 (Angostura,

Ciudad Bolivar, Venezuela).
1879. Typhlonectes dorsalis Peters, Mon. Ak. Berlin, p. 941; Boulenger, Cat.

Batr. Grad. Brit. Mus. (2), p. 103.
1912. Thyphlonectes dorsalis Fuhrmann, Mem. Soc. Sci. Neufchatel 5,

p. 124, f. 7.
1937. Chthonerpeton microcephalum Miranda Ribeiro, O Campo, May, p. 66.

Type. Not seen; in Gottingen Museum.

Type locality. Angostura, Venezuela [now Ciudad Bolivar].

Range. Venezuela to Brazil and Peru; specifically, from Ciudad
Bolivar to Para, Matto Grosso, Iquitos, and middle Ucayali.

Diagnosis. A Typhlonectes with dorsal fin almost to head; head very
small; tentacle close behind nostril; anal disk very large; primaries
88-104. •

Description . No trace of secondaries ; primaries 88-104, all complete ;
eye visible; dorsal fin or keel beginning on the neck; body very com-

538

bulletin: museum of comparative zoology

pressed posteriorly; anus in the hinder part of a large sucking disk
(much larger than head in ANS 4926; 5^ mm. in Berlin 10104); 1/d
difficult to measure, approximately 20-36; smallest seen 167 mm. long;
largest 695 mm. long; light brown, primaries blackish.

Habits. The specimen from Manaos was taken "under stones in 4
inches of water near rock ledge of river." That from Belem was from
"docks in river."

Remarks. This is the most specialized of the group in compression
of body, extent of fin, and size of anal disk. The complete folds seem
more primitive than those of the others, but in all other ways it is
much the most specialized. The type of kaupii had 104 primaries, two
more than any I have seen.

Specimens seen, 19, as follows:

Venezuela :

pi i in.

lengui

uiaiii.

vu

Ciudad Bolivar

Berlin

10104

96

600

25

24

TYPE dorsalis?

it a

Hamburg-

336 (head only)

Angostura

Berlin

9092

102

270

8

35

TYPE dorsalis"!

Orinoco R.

Hamburg

489

101

420

?

?

Guiana:

ANS

4927

98

300

15

20

Brazil :

^\J

Para

Hamburg

1928

98

335

13

25

Para, Belem

Carnegie

2908

92

410

7-14

28

R. Negro, Manaos

it

2906

88

172

6

28

"Brazil"

Vienna

?

480

?

?

11

ANS

4926

92

405

14

29

No locality

BMNH 98-10-17-7

93

167

6

28

U ti

98-10-17-8

96

260

11

24

Peru:

R. Ampiyacu

, near

Pebas

USNM

101105

95

425

16

27

Iquitos

AMNH

42853

99

545

15

36

n

AMNH

42854

98

545

19

29

San Antonio,

Rio Itaya

AMNH

42857

96

695

20

35

Rio Pisqui (mid. Ucayali)

AMNH

42856

98

505

14

36

DUNTN: THE AMERICAN CAECILIANS 539

No locality:

prim.

lgth. diam.

1/d

MNH 98-10-17, 7

93

167 6

28

98-10-17, 8

96

260 11

24

Xote. Peters (1877) gives 99 primaries for the type of dorsalis, and
says it was 265 mm. long, diameter 7 mm. This fits pretty well to my
count and measurements for Berlin 9092 and would seem to make that
specimen the type were it not that Berlin 10104 is labeled "type."
Probably it is best to regard them as cotypes.

Dr. Joseph Bailey very kindly furnished me with a copy of the de-
scription of Chthonerpeton microcephalum Miranda Ribeiro, and having
examined the type, wrote me that it was a Typhlonectes. The descrip-
tion fully confirms this and indicates 92 primaries, a length of 560 mm.,
a diameter of 23 mm., and a length/diameter ratio of 24. These counts
fall within the known variation of kaupii. The description mentions
the small head, the complete dorsal fin fold, and the very large anal
disk.

The type came from Matto Grosso collected by Rondon. Bailey
writes me that "the Matto Grosso material all came from the northern
and western sections of the state, and the snake material has a large
number of Amazonian elements in it. I think most of it came from
the Serra de Parecis or along what is now Rio Roosevelt."

INCERTAE SEDIS

"SlPHONOPS SYNTREMUS" Cope

1866. Siphonops syntremus Cope, Proc. Acad. Nat. Sci. Philadelphia, p. 129.
1879.? Typhlonectes syntremus Peters, Mon. Ak. Berlin, p. 942.
1885.? Dermophis syntremus Cope, Proc. Amer. Phil. Soc. 22, p. 171.

As Cope is the only herpetologist known to have examined the
unique type of his Siphonops syntremus, I quote his remarks:

"A collection from Belize from Dr. Parsons." "The same

correspondent sends from the neighboring region of Honduras

Ninia collaris and Rhegnops visonimis." "Siphonops syntremus

sp. nov. This species differs from the four hitherto known [Siphonops
annulatus, Siphonops brasiliensis, Chthonerpeton indistinctum, Syn-
noj.is mexicanus, all considered as Siphonops in 1866] in the close ap-
proximation of the narial and tentacular openings ; the latter lie a little
behind the former, and are slightly larger. The minute eyes are just
visible; the internal nares are some distance behind the palatine arch.

540 bulletin: museum of comparative zoology

Muzzle projecting, obtuse in profile; from above narrowed, rounded.
Teeth large, five on each ramus mandibuli. A gular, and strong post-
gular fold; 130 annular plicae, which are complete, except some slight
ventral interruption anteriorly; the posterior third of the length with
intermediate annuli, which are at first lateral only, then complete
above, entirely complete on the terminal inch; the whole number will
then be about 170 annuli. Form of body rather slender; tail depressed
at end, short, acuminate. Color dark plumbeous, annuli yellow lined;
head yellowish brown. This species resembles the Caedlia ochrocephala
[described in the same paper], but is primarily distinguished by the
position of the foramen, and of the inner nares, also by the color and
character of the annuli."

Under Dermophus syntremis, in 1885, Cope says "I refer this species
here provisionally only, as I have not been able to find the type speci-
men. Belize."

If it were not for this second statement the arrangement of the
species in the original paper would tend to give the impression that
the provenance of syntremusv;as"the neighboring region of Honduras."

If the description was accurate syntremns belongs to none of the
species (or, indeed, none of the genera) listed in this paper. No other
American form has the combination of the tentacle position of Typh-
lonectes or Chthonerpeton, with the primaries (130) and secondaries
(40) of a Gymnopis or a Caecilia, the teeth of a Caecilia, and the "short,
acuminate" tail of a Rhinatrema. As described, the species demands a
new genus for its reception.

We have no right to assume that the description was inaccurate.
Caecilia ochroccphala was described in the same paper, the type is ex-
tant, and the description is very accurate.

If the description was inaccurate, it is possible that it dealt with a
specimen of Gymnopis oligozona (primaries 130-135, secondaries 62-
74) from the same general area. G. oligozona was described from a
specimen without data, by Cope, in 1877, and it is barely possible
that the type of oligozona was previously the type of syntremus.

The National Museum (which contains the Parsons collection) has
a specimen of Typhlonectes eompressieauda nutans labelled "Belize."
This might, as a remote possibility, have been the type of syntremus.

No measurements were given for syntremus.

I cannot place this species in any genus known to me. I do not wish
to name a new genus on the basis of a single unexamined specimen. I
therefore merely set down the pertinent facts and refrain from any
action.

?'

9-

6

3)

ACME
Bo .. INC.

AUC 2 3 1984

no IOG£

Harvard MCZ Library

III 1 1 II III II I ill

3 2044 066 303 645

Date Due