STUDIES ON THE FAUNA OF CURAÇAO AND OTHER
CARIBBEAN
ISLANDS:
No. 92.
Distribution
and
geographic variation
of
the Ameivas
of Puerto
Rico and the Virgin
Islands
by
Harold
Heatwole
and
Frank Torres
(Museum of Biology, University of Puerto Rico, Rio Piedras)
INTRODUCTION
64
MATERIALS AND METHODS
66
Effect of size and sex on characters used
68
Effect of time on scale characters
71
Ameiva exsul Cope
1862
72
Habits
72
Distribution on the Puerto Rican mainland
74
Distribution on other islands
77
Variation on the Puerto Rico mainland
79
Variation on the smaller islands
83
Coloration
on the Puerto
Rican
mainland
86
Coloration
on small islands
87
Relationship of eastern populations and status of Ameiva birdorum
....
89
Ameiva alboguttataBoulenger 1896
92
Habits
92
Distribution
92
Variation
and Coloration
93
Ameiva desechensis sp. nov
95
Description
95
Variation
96
Relationships of Ameivas of Mona, Desecheo
and Puerto Rico
97
Ameiva wetmorei Stejneger
1913
97
Distribution
97
Habits
97
Variation
98
1)
HERPETOGEOGRAPHY
OF
PUERTO
RICO.
III.
This
paper
is part of
a symposium
on Biogeography, held at the University of
Puerto Rico, Rfo Piedras, P.R., December 8,
1963.
64
Ameiva polops Cope
1862
102
Distribution
102
Coloration
103
Habits
104
Variation
104
LITERATURE
109
DISCUSSION
105
Key to the
Ameivas of Puerto Rico and the Virgin Islands
105
INTRODUCTION
COCHRAN (1941)
in a review of the Hispaniolan
forms modified
this
scheme
by
(1)
considering
A.
wetmorei closely
related
to
*) Since
this
paper
was
finished,
BASKIN
&'
WILLIAMS (1966)
have
discussed
the
movement of South American ameivas
up the Lesser Antillean chain. They suggest three
separate
invasions resulting in three species
groups.
Our
usage
of
the term
"Ameiva
ameiva group" includes
these subgroups.
group
-
see below).
A. undulata
are
considered allied to the
A.
polops
and
Ameiva
wetmorei
Accordingly
to
BARBOUR
&
NOBLE’S view,
then,
the
Antilles
comprise
two
main
groups
which
have
invaded
the
area
from
opposite
directions
and
which
overlap
on
Hispaniola
(or
from
Hispaniola
to
St.
Croix
if
group or from still a fourth stock.
Ameiva ameiva
and its subspecies.
They postulated
that it either
arose
from the
Ameiva bifrontata
group.
A
final,
somewhat
problematic group consists of the South American
A.
ameiva
than to
the
A.
undulata
but which they allied more closely to the
A. polops),
and
(A. maynardi,
A. wetmorei
group which originated
in and
dispersed
from
Central America,
moving into
northwestern South America
and into
the
Greater
Antilles
as far eastward
as Hispaniola.
In addition to these two
main stocks, they postulated still a third origin for a small group of
species in the Bahamas - Puerto Rico
- St. Croix area
Ameiva
undulata
groupwhose center of origin and dispersal
was northeastern South America,
and
which extended westward
into Central America and also up the Lesser Antilles to Puerto Rico
and Hispaniola.¹
(2)
The other
gave rise
to the
Ameiva ameiva
The broad
outlines of
the
taxonomy
and
distribution
of
the
Antillean
ameivas
have
been sketched
by
BARBOUR
&
NOBLE
(1915). Twoprincipal ancestral stocks
were recognized:
(1) One gave
rise to the
65
A. lineolata of Hispaniola,
thus allying the former with one of the
species BARBOUR & NOBLE had definitely placed in the A. undulata
group, (2) pointing
out that A. polops
is related to A. taeniura of
Hispaniola
rather than
to
A.
wetmorei of
Puerto
Rico,
and (3)
synonomizing A. vittipunctata (considered by BARBOUR & NOBLE as
belonging
to the
South American stock)
with A.
chrysolaema
(of
Central American stock). Clearly,
a re-study
of the relationships
of
the Antillean ameivas is in order. This paper deals with those forms
existing
from Mona Island to Anegada inclusive (Figs.
17 and 24).
In
the
Puerto Rican
area,
three
described
forms
belong
to
BARBOUR & NOBLE'S South American stock. They
are Ameiva exsul
Cope from Puerto Rico and the Virgin Islands, Ameiva alboguttata
Boulenger
from
Mona
Island and Ameiva
birdorum Grant
from
Diablo Key.
In this paper
we place the latter in the synonomy of
Ameiva exsul and also describe
a new species, Ameiva
desechensis,
from Desecheo Island. We shall subsequently refer to these forms as
the Ameiva exsul subgroup.
The Ameiva undulata group is represented in the
area by
three
named forms, A. wetmorei from Puerto Rico, A. eleanorae from Caja
de Muertos Island (the latter is placed in the synonomy of the former
Fig.
17. Map of
the
PUERTO
RICO-VIRGIN
ISLAND
area showing
distribution
of
Ameiva
wetmorei
(dots)
and
Ameiva
polops
(squares).
Solid
symbols
represent
localities from which specimens have
been examined; open symbol represents
a
literature record (COPE 1862). Line surrounding most of the A.
wetmorei
localities
is the 30 inch isohyet. A. indicates the
area shown in Fig. 25
and B. that shown
in Fig. 26.
66
in this paper) and A. polops from St. Croix. The last is more closely
related to
a Hispaniolan
form than to any
on
the Puerto Rican
island shelf (COCHRAN
1941) and probably represents an independent
invasion of the A. undulata group eastward.
We
are grateful to
the following persons
and institutions
for
the privilege
of
examining specimens: ERNEST E. WILLIAMS of the Museum of Comparative Zoology
at Harvard,
DORIS
COCHRAN
of the
U.
S.
National Museum, the
late
NORMAN
HARTWEG of the University of Michigan Museum of Zoology, CHARLES M. BOGERT
of the American
Museum of NaturalHistory, WILLIAM DUELLMAN of the Museum of
Natural
History
of the
University of
Kansas
and
R.
F.
INGER
of the Chicago
Natural History Museum.
CHARLES
WALKER facilitated
our work
in
many
ways
while we examined specimens at the University of Michigan.
The following people aided
our expeditions to St. Croix
and surrounding keys:
BILL MILLER, GEORGE SEAMAN,
JOHN WOODSON
and
DONNA CRANK of Christian-
sted, St. Croix, and HERMINIO LUGO LUGO of the University of Puerto Rico.
Tixo
WIRSHING of the
Don-Q
Rum distillery and
Prof. Luis
ESCAB£
of the Catholic
University of Puerto
Rico provided transportation and lodging on expeditions to
Caja de Muertos
and
Platillo
Islands. EUNICE BOARDMAN of Beef
Island provided
various kinds
of aid
on our expedition to
the British Virgin
Islands.
The police
department of Cabo Rojo, P. R., the Institute of Marine Biology of the University
of Puerto Rico at Mayagiiez, and the U. S. Coast Guard provided transportation and
lodging during various expeditions to Mona
and Monito
Islands and
to southwest
Puerto Rico. We
are gratefulto ALAN BATHAM, REGINALD PENN
and the
BIGELOW
family of the British Virgin Islands for permission to collect on their properties.
FAUSTINO MCKENZIE,
RICHARD
LEVINS,
MANUEL
VELEZ and
FRANCIS
ROLLE
assisted in the collection
of specimens. SHEILA BLASINI,
ISABEL COLORADO,
SARA
ARMSTRONG,
and
ANA
VASQUEZ performed
some
of the statistical
calculations;
AUDRY
HEATWOLE
and
OLGA TORRES
helped prepare
the manuscript. THOMAS
JACKSON did
the photography.
This project was supported by grant no. GB-2906 of the National
Science Foun-
dation of the United States of America.
MATERIALS
AND
METHODS
Various characters
used in separating
species
of
Ameiva
have
been the number of longitudinal
rows of ventral scales, whether or
not the caudal scales
are oblique,
the placement
of the nostril,
scutellation pattern
of
the
head and legs,
etc.
In
the Antillean
forms these
seem to
be relatively
constant within
a
species
and
hence
not
particularly
valuable for
an
analysis
of intraspecific
variation. Ameiva exsul,
Ameiva wetmorei, and Ameiva polops
can
be easily distinguished
from each other
on the basis of the
above
kinds of characteristics and no attempt is made here to repeat
the
67
previous
statements of
various authors who have outlined
these
differences.
Rather
attention is
concentrated
on
the
variation
among populations
within a particular
species,
and
on comparing
very
closely
related populations,
not
separable
by
the
above
categories of characters,
but which have been accorded specific
or
subspecific
rank
on
the basis
of number of femoral
pores and/or
color pattern.
These closely
related, previously
named forms
are:
(1) Ameiva wetmorei-Ameiva eleanorae and (2) Ameiva alboguttata
-
Ameiva exsul- Ameiva birdorum.
The following CHARACTERS were used:
1. Number
of femoral
pores:
"Number
of
femoral
pores" throughout
this paper
indicates
the sum of the number
of pores
on the right and left
sides.
Usually the counts of the two sides did not differ by more than
1
or 2. The femoral
pores
(more
appropriately called
glands)
do
not regenerate. Hence
data
were
discarded for animals in which some of the pores had been
obliterated by injuries.
2.
Number of scales
in
15th tail ring.
The caudal scales are arranged in
rings. Beginning just
behind
the cloaca,
the scale-rings
were counted posteriorly
until the
15th
one
was encountered; the number
of scales
in
this ring
was then
recorded.
These
are subsequently referred to as "caudals."
Data
were discarded for
all
animals which
had
lost their
tails
anterior
to this region
even if subsequent
Fig.
18.
Scutellation
of ventral
area anterior to vent in Ameiva exsul showing the
preanal plates surrounded by preanal scales. The latter are stippled and outlined in
heavy lines.
68
regeneration had
occurred;
scales
of regenerated tails
are smaller and; greater in
number.
3.
Preanal scales:
Antillean
ameivas have a series of 3-5 flat enlarged scales
just
anterior
to the cloaca known
as preanal plates. Encircling these
is a row of
smaller scales which we have chosen to call the preanal scales (Fig. 18). The number
of these served
as one of the
characters analyzed.
4. Transverse
rows
of ventrals:
The
number
of longitudinal rows
of
ventral scales is rather constant within
a species of Ameiva and
has been used
as a
character for the separation of species. However, the number of transverse
rows of
scales
varies and it was thought that it might be
a useful
character in analyzing
infra-specific
variation. The number of rows were counted beginningwith
the first
one behind
the prepectoral fold
and proceeding posteriorly
to the abrupt point
where the
scales are no longer arranged in rows.
5.
Number
of brachials:
The
number of brachial
scales on the right and
left forelimbs were counted and the sum of berth
sides used in analyzing patterns of
variation. Difficulty was sometimes experienced in decidingwhich scale represented
the first brachial.
6.
Ratio
of snout length to body length: Body length was considered
the distance from the tip of the snout to the anterior edge of the vent. Snout length
was taken from the posterior edge of the nostril to the anterior
border of the eye.
All measurements were made with
Vernier Calipers.
7. Color
pattern:
All
species
and
subspecies either
now
or previously
re-
cognized for the area under consideration, have been collected
alive by the authors.
Color pattern as well
as scutellation
has
been used
in defining species althoughno
form has been recognized solely on the basis of color pattern. Use of color pattern as
a taxonomic character in Ameiva is complicated by
the great ontogenetic changes
that take place and
also by the wide variability even within the same locality, in
some species. For example, BEEBE (1919) collected "perfectly typical specimens" of
4 named subspecies of Ameiva ameiva within an area of about 500 yards around
a
research
station
in
British
Guiana.
Clearly,
care
should
be
exercised
in
basing
diagnoses on coloration.
All animals except small
ones classed
as juveniles, were sexed by examination
of
the gonads unless the hemipenes were protruding.
The names of Museums will
be
abbreviated
in the following manner throughout
the remainder
of the
paper:
AMNH
-
American
Museum
of Natural
History;
CNHM
- Chicago Natural History Museum;
KU
-
Museum
of Natural History,
University of Kansas; MCZ
- Museum of Comparative Zoology; UMMZ
- University
of Michigan
Museum
of Zoology; UPRRP
-
University of Puerto
Rico
at
Rio
Piedras; USNM
- United
States National Museum.
The
1,043 specimens which were examined
are listed in the chapters concerned.
EFFECT
OF SIZE AND SEX ON
CHARACTERS USED
Before comparison between populations
can be
made it is
necessary
to
know
whether
there is
an ontogenetic change or
a
difference
between
the
sexes in the
characters
used. Figure
19 shows
a plot of the
snout length to body length ratio
plotted against size for the sample of Ameiva exsul from the San Juan area. Figure 20
shows the same type of analysis for A. wetmorei from La Parguera, P. R. No onto-
69
genetic change in
the
ratio
is
evident
in
either
species.
Inasmuch
as the
scale
characters used
are fixed duringearly development (pre-hatching) such an analysis
was unnecessary
for them.
The same group of lizards
was also used for comparingmales and
females. Using
the graphic test of significance of DICE & LERAAS (1936), it was found that there was
no significant difference between the sexes in any of the characters used for Ameiva
exsul (Fig.
21).
Hence the
data for
both
sexes
were grouped for
the
analysis of
geographic variation.
The
same was true for A.
wetmorei except for femoral pores
which were significantly lower
in females (Fig. 22). In subsequent analyses
of this
character in A. wetmorei the sexes are treated separately.
Fig.
19. Relationship of ratio
of snout length and body length to
size in Ameiva
exsul from San Juan, P.R.
Fig. 20.
Relationship of ratio of snout length and body length to size
in
Ameiva
wetmorei from La Parguera, P.R.
70
Fig. 21.
Comparison of males
and
females of Ameiva
exsul from
San Juan, P.R.,
with respect to various characters. Central vertical bar represents the mean, terminal
vertical bars represent end points of the range
and rectangles represent 2 standard
errors
on each side of the mean. Failure
of these rectangles to overlap indicates
significant differences (DICE & LERAAS 1936). N
= 28 (￿), 48 (￿).
Fig. 22. Comparison of males
and
females
of Ameiva
wetmorei from La Parguera,
P.R., with respect to various characters. N
= 37 (￿),
26
(￿).
Symbols
the
same
as
in Fig.
21.
71
The similarity in relation of snout length to body length for the two sexes was not
expected.
Mr.
ANGEL MALDONADO,
who
had
been
doing experimental work
on
Ameiva exsul can sex them quite accurately (verified by examination
of the gonads)
by looking at the side of the head. He felt that the males had longer snouts than
females, although his subjective determinations apparently are based on some more
subtle cue.
Ontogenetic changes in color pattern are discussed later.
EFFECT
OF
TIME ON
SCALE CHARACTERS
Inasmuch as specimens collected
over a period of more than half a century were
lumped together and
used
for studying geographic variation, it was necessary to
ascertain whether
measurable
changes had
occurred
within
a locality during that
time. To do this, data from specimens collected in the San Juan area prior to
1900
were compared to those from lizards collected in the same region between
1960 and
1966. The results (Fig. 23) indicate
no significant differences in most
of the charac-
ters
used, i.e.,
no
observable evolution
had
occurred.
However,
there
was
one
exception (caudal scales in
15th row), indicatingthat at least for some characters,
60 years
is sufficient time for
measurable changes to
occur. Consequently, some of
the inter-locality differences in caudal
scale
number
subsequently described,
may
reflect temporal as well
as spatial
differences. In all
cases where
conclusions
are
Fig. 23.
Comparison of scale characters of a sample of Ameiva exsul collected prior
to
1900 (solid bar, N
=
13) with
a sample from the same locality collected
between
1960 and
1966 (open bar, N = 10). Symbols as in Fig. 21.
72
drawn from comparison of this character among localities,
the specimens from each
place involved
were collected at about the
same time.
Unfortunately, comparable data
are not available
for very
small
islands where
more rapid changes might be expected. However, in view of the above results, it is
reasonable to
assume that for most
characters patterns of variation reflect present
day inter-island
differences, rather
than discrepancies in
date of collection
of the
material.
Ameiva
ameiva
group
Ameiva exsul subgroup
HABITS
Ameiva exsul Cope 1862
STAHL (1883) gives the habitat of Ameiva exsul (under the name of A. plei)
as
"arenales
y pedregales"; STEJNEGER
(1904) lists it
as sandy
or gravelly soil,
and
SCHMIDT (1928) says “Ameiva exsul distinctlyprefers a sandy soil, and is everywhere
more abundant
on sand." It is true that this species is primarily
a beach form and
may have been restricted to this habitat in Pre-Colombian
times. Now, however, it
can also
be found
in a variety of open
areas with
low vegetation, such
as lawns,
vacant lots, roadsides, city parks and plazas, and around human dwellings. It is well
adapted to the conditions
created by
man and can frequently be seen on the
side-
walks near vacant lots in the San Juan metropolitan area. On a number of occasions
we have seenit wanderingabout on the floor of an open restaurant
on the University
of Puerto Rico
campus.
We have
never observed it in heavy forests although it is
encountered along roadsides
or on lawns which have scattered
trees.
When running rapidly, Ameiva
exsul frequently utilizes
a bipedal type of loco-
motion.
These lizards spend their inactive periods in burrows
which they dig. Usually
the burrow is beneath
a stone, piece of wood
or other loose object although some-
times the entrance is beside
a plant or less frequently in the open. SCHMIDT
(1 928)
mentions that they made burrows under tramway ties, sidewalks, and under stones
or other loose objects. Their main habitat requirements seem to be (a) an open area,
(b) soil suitable for digging, and (c) objects lying on the surface of the ground.
Ameiva exsul is strictly
diurnal
and
is usually active in direct sun. According
to WOLCOTT (1923) it feeds
at mid-day. However, GRANT
(1931a) indicates
that
there
are two activity periods,
one in late morning and another in early afternoon.
He
describes it in the following terms : "They never venture out in the cool of the
morning. If, when hunting, they suddenly
come to
a strip of sunlight, they will
flatten out, hind legs stretched out, and bask for a few minutes. During the extreme
heat of the day they are quiescent, seeking the shady places. As the afternoon cools,
they are againactive, but retire well before sunset into holes which they themselves
excavate
under
stones or boards,
or seek
a natural crevice
where available."
This
difference of opinion is probably explained by the observation of DANFORTH (1925)
that during the winter A.
exsul came out only in the middle
of the day,
but
after
73
April they could be found any time the sun was shining. A bimodal activity pattern
is also found in several other teids
(CARPENTER 1961, HIRTH
1963).
As is true of at least
some other
members of the genus (BEEBE 1945) this species
feeds by
foraging rather than sitting and waiting for prey
to approach within
reach. The lizards
walk along examining small objects
and poking into clumps of
grass, under leaves, etc., frequently extending their tongue in snake-like fashion
as
they do so. Theyare generalizedomnivores. WOLCOTT (1923) examined the stomachs
of 15 specimens and found that their food was 78% insects, with smaller proportions
of other arthropods, earthworms, and snails.
Of especial interest
was the fact that
non-movingfood
was also taken; 2 different individuals
had each eaten lizard eggs;
one had
eaten only mushrooms.
SCHMIDT (1920,
1928) examined
20
stomachs
and
found
that
11
contained
vegetable material, mostly red-coated
seeds.
Additional
items were insects, crabs, lizards
eggs, and the tail of an Anolis cristatellus. Thus they
may be
an important predator
on other
saurians. They also frequently eat table
scraps
put
out for
dogs
or cats,
and
GRANT
(1931a.)
has reported them feeding
primarily on "refuse and vegetable matter" including onion trimmings and potato
peelings.
We
have
observed
them
on
numerous occasions
eating cactus fruits,
bananas, apples,
and crawling into
recently discarded spaghetti, ravioli, and meat
cans and eating the remaining bits
of food. We suspect that
one of the
reasons
Palominitos
Island
can maintain its surprisingly large population of Ameiva
exsul
is
because
of the
energy entering the system
in
the
form of garbage scattered
unceremoniously on the
beach by
weekly infestations
of picnickers.
HEATWOLE
et al. (1963) reports that Ameiva exsul eats monkey-chow put out for monkeys on
Cayo Santiago, P. R., and suggests that this has resulted in anincrease in population
size onpart of the island.
Little
is
known
of
the
reproduction
in
this
species although NOBLE
&
BRADLEY (1933)
have published
a few
observations
of its courtship
and mating.
The tail
is apparently necessary
for
effective mating.
Field
observation indicates
that males which have recently lost heir tails (not yet regenerated) are unsuccessful
in attempts to copulate (MACKENZIE, personal communication). The pink
eggs are
Fig. 24.
Map of VIEQUES and PUERTO RICO and the latter’s northern, southern and
western outlying keys, showing distribution of the Ameiva exsul subgroup. Triangles
represent
Ameiva
alboguttata,
squares
represent
Ameiva
desechensis
and
dots
represent Ameiva exsul. Open circles
indicate sight records of Ameiva
exsul by
the
authors except for the one on Vieques which is a literature record (GRANT 1932e).
74
laid in loose soil several inches below the surface ofthe ground; they are 20-22 mm x
13-15.5
mm
and
are found
in groups of 4-7 (Wolcott
1923). A hatchling collected
on 5 February
1965 and still possessing its yolk-sac
scar,
measured 36 mm
s-v.
Growth
seems to occur long after achievement
of sexual maturity. Most adults
encountered
are not
100-120
snout to
vent. However, giant individuals
are
oc-
casionally encountered which are as large as 200 mm s-v. They are presumably very
old
individuals
which
have
continued
to
grow throughout their
life.
We
have
examined
several
such
specimens,
the largest of which
was
a
male
from Playa
Humacao (MCZ 58775; s-v 201 mm) and have
seen others about the same size in the
field.
SCHMIDT
(1920,
1928) lists
the
maximum
size
as "much
exceeding half
a
meter" in total length.
DISTRIBUTION
OF Ameiva exsul
ON THE
PUERTO
RICAN
MAINLAND
As
is
immediately
evident
from Figure
24,
Ameiva
exsul is
restricted to the lower elevations, not having been collected above
an altitude of 500 ft, although it has been observed in open areas and
along roadsides somewhat higher. WOLCOTT (1923)
gives its distri-
bution as "near the coast or in river valleys, but sometimes occurs
in hills of
no great elevation." It intuitively
appears unlikely that
altitude per
se should be
a factor limiting
the distribution of this
species
and
some
factor
correlated
with
altitude
appears
more
plausible.
Temperature, rainfall, vegetation, and soil vary through-
out
the
island
and
all
show
at least
a rough
correlation
with
altitude.
As
indicated in
the section
on
habits,
Ameiva exsul
seems
to
prefer
a sandy
soil and is most abundant in sandy habitats. Most
of the lowland soils
tend to be sands
or sandy
loam whereas the
upland
ones
are more often clay.
However,
this is not the factor
controlling
the lizards' distribution, for when the collection sites of
A. exsul were superimposed on a soil survey map, they fell on such
diverse soils
as sands, silty loams and clays (ROBERTS
1942). The
possibility
of
this lack
of
correspondence
to
soil type
being
an
artifact due to local pockets of sandy soil has been ruled out by our
direct observation of this species inhabiting Caguas clay. Geological
formations are similarly unimportant
as A. exsul is found in areas
characterized by limestones and other sedimentary rocks,
as well
as in those with substrate of volcanic origin (MITCHELL
1954).
Nature of vegetation
is probably
a factor in the distribution of
this species
inasmuch
as it is not found in heavy
forest.
This is
75
probably because of the reduced light intensity and/or temperature
below the level required for their activity period. They
are found in
the more open scrubby arborescent vegetation
in the hotter areas,
e.g., on Caja de Muertos island.
Inasmuch as cloudy
or rainy, cool weather inhibits the activity of
this species,
one might suspect that in the cooler areas of the island
with heavy rainfall, the activity period may be cut
so short that the
species
cannot
find enough food to maintain itself and reproduce.
The distributionof Ameiva exsul (shown in Figure 24) is delimited
on the inland side by the 76° isotherm for mean annual temperature
(Pico
1954, p.
116) almost as well as by the 500 ft countour line. In
fact the 76° F isotherm appears as a "smoothed out" diagram of the
500 ft contour line. The only A. exsul locality to fall in a cooler area
is Utuado which occurs between the 74° and 76° F isotherms; many
coastal localities
were
in
areas
with
mean
annual
temperature
greater than 76° F.
The
pattern
of
rainfall
is complex
in
Puerto Rico
with
the
greatest
amount occurring
at
the higher
altitudes
and
with
(at
comparable altitudes and exposures)
the eastern end of the island
receiving
more rain than the western
one. Distribution of Ameiva
exsul although
generally
in the
drier areas, shows
no
precise
cor-
relation with any isohyet (Pico
1954, p.
135).
In summary it
can be
stated that the distribution of Ameiva
exsul on Puerto Rico does not correlate with soil type
or with pat-
tern of rainfall, but does correlate with altitude and temperature.
The former is believed to be an indirect relation in that temperature
and
altitude
are
themselves
correlated.
The
hypothesis
is
put
forward that A. exsul is restricted to the lowlands in Puerto Rico
because the lower thermal level at higher altitudes shortens their
activity period to the extent that they cannot obtainsufficient food
to
subsist
over
a long-term
period.
The
final resolution
of the
problem awaits, of course, a detailed study of habitat selection and
activity pattern, tolerance limits, and eccritic limits of this species
to various environmental factors. Here we
are simply attempting
to
describe the geographic
distribution and its
correlation with
environmental patterns
in the hope it will suggest
experimental
approaches aimed at the solution of this problem.
Fig.
25.
Map
of
the
northeastern
keys
of
PUERTO
RICO
showing
distribution
of
Ameiva
exsul
(See
Section
A
of
Fig.
17).
Dots
represent
localities
from
which
specimens
were
examined,
circles
represent
sight
records
by
the
authors.
77
DISTRIBUTION
OF Ameiva exsul ON OTHER ISLANDS
In the Puerto Rico
- Virgin Island area there are several instances
of
more than
one island having
the
same
name, e.g.,
there is
a
Buck
Island
near
St.
Croix,
another
near
St.
Thomas and still
another
near
Tortola.
Where
such
possible
sources
of
confusion
exist, the name of the island is followed by that of its nearest large
island in parenthesis.
As seen in Figures 24, 25 and 26, Ameiva exsul
Fig. 26.
Map of the
VIRGIN ISLANDS showing
distribution
of Ameiva
exsul
(See
Section B of Fig.
17). Dots represent localities from which specimenswere examined,
circles represent literature records
(GRANT 1937), except for Marina Key which is a
sight
record by
the
authors.
1.
Dutchman’s
Cap;
2.
Salt Cay;
3. West Cay;
4.
Savanna Island; 5. Little Saba Island; 6. Outer Brass Island; 7. Inner Brass Island;
8. Water Island; 9. Hassel Island;
10. Hans Lollik;
11. Little Tobago; 12. Tobago;
13. Thatch Key;
14. Mingo Key;
15. Greater
St. James;
16. Jost Van Dyke;
17.
Great Thatch Island;
18. Little Thatch Island;
19. Frenchman’s Key;
20. Leduck
Island; 21. FlanniganIsland; 22. Norman Island; 23. Peter Island; 24. Dead Man’s
Chest; 25. Salt Island; 26. Cooper Island; 27. Ginger Island; 28. Buck Island; 29.
Beef Island; 30. Guana Island; 31. Great Camanoe; 32. Scrub Island, 33. Mosquito
Island; 34. Little Camanoe;
35. Little
Jost Van Dyke;
36. Marina Key;
37. Sail
Rock; 38. Buck Island; 39. Capella Island; 40. Cockroach Island; 41. Cricket Rock;
42. The Dog Islands (George Dog, West Dog,
and Great Dog); 43.
Seal Dogs;
44.
Prickly Pear Island; 45. Necker Island; 46. Fallen
Jerusalem; 47. Round Rock; 48.
Frenchman Cap;
49. Congo Key; 50. Little Hans Lollik.
78
occurs
on
all
of the larger islands from Puerto Rico
to Anegada
except
St. Croix.
It
was recorded from Desecheo (SCHMIDT
1928)
but the Desecheo form is not A. exsul and it is described in this paper
as
a new species.
Ameiva exsul is lacking
on
some
of the smaller
islands
within its
general
range.
In
some
cases
this
is probably
because
the island has not been herpetologically
explored
rather
than
an
actual absence of the species.
However,
on
a number of
these islands absence
of Ameiva is probably real.
GRANT
(1931b,
1932a, b,
1937) collected
on Cockroach Island (St. Thomas), Buck
Island (St. Thomas), Dog Island, Congo Key and Luis
Pena Key
and found no ameivas on any of them.
We
have examined
a number of
the
small keys
northeast of
Puerto
Rico
and
found
them to
lack
ameivas.
These
are
the
Cucarachas,
the
Farallones,
and
various
of
the
small
islands
around Diablo Key
and
Blanquilla
Key,
and between Culebrita
Island and Cayo Norte (Fig. 25). A few of
these lacked vegetation,
although
most had a herbaceous plant
cover and
some of
them,
arborescent vegetation. However,
with a few exceptions,
none had
beaches and the shorelines of most consisted of vertical cliffs, under
cut several feet at the base by
wave action. Off the southwestern
coast of Puerto Rico,
all keys
except
one (Spiney Butte)
lacked
A.
exsul.
Some of
these islands
were
surrounded by
mangroves,
others had sandy
beaches.
All
of
the small islands containing
Ameiva
exsul
have charac-
teristics similar to those of the coastal areas of Puerto Rico, i.e., low
altitude,
mean annual temperature above 76° F and open, i.e., not
heavily
forested,
areas. Rainfall is scant on most and the substrate
is
usually
either sand
or limestone. The larger islands
are
either
similar in their entirety
or have coastal
areas with these charac-
teristics.
This is also true of St. Croix and its outlying keys and yet Ameiva
exsul
is lacking there. Earlier reports of this species from St.
Croix
(BARBOUR 1914, BARBOUR & NOBLE
1915, SCHMIDT 1928) have been
questioned by
a number of authors (BARBOUR
1930,
1937) and were
apparently
based
on
confusion with A. polops
which superficially
resembles
the juveniles
of A.
exsul
(GRANT
1937).
Inasmuch
as
appropriate habitat for A. exsul occurs
on St. Croix, the lack of this
79
species
there is
probably
due to historical
considerations and
to
present
barriers
to
transport
between
St.
Croix
and
the
other
islands
(see
discussion).
Failure
to
compete
successfully
with
A. polops is a further possibility.
However, A. polops has now been
exterminated by
the mongoose on the mainland of St.
Croix, yet
A.
exsul has not become established there. A. exsul itself does not
seem to be greatly influenced by the mongoose as evidenced by its
abundance on mongoose-infested
islands.
BARBOUR (1930,
1937)
reports
Ameiva
exsul
as
occurring
on
Anguilla. However, UNDERWOOD (1962) lists the Anguilla species
as
Ameiva pleeii.
We have
not
examined any
specimens
from that
island.
Ameiva exsul is occasionally found on islands too small to support
permanent
populations.
For
example
on
Levin's
Rock
(Fig.
25)
which has an area of only 65 m2
, the total Ameiva
exsul population
consisted of one male, which had obviously
reached the island from
PuertoRico (a distance of only about 5 m at low tide). Similarly, one
individual
reached Cayo
Ahogado,
a
small
(3,000
m 2) sparsely
vegetated sand bar 0.8 km east of Puerto Rico (Fig. 25). This island
had been visited frequently
over
a period
of
18 months and
no
Ameiva found there until 23 June
1965,
when 2 individuals were
present.
Just previous to that date, the Fajardo River opposite the
key,
had flooded, casting adrift much debris. The lizards
had dis-
appeared again by 30 July 1965. After similar conditions of flooding
on
the mainland,
one individual was
found on Cayo Ahogado
on
17 October
1965.
These observations suggest that this species
can be dispersed
at
least
short
distances
via
sweepstakes
routes,
and
that it may
temporarily inhabit small islands not able to permanently support
a population.
VARIATION IN Ameiva exsul
ON THE
PUERTO RICAN MAINLAND
The Ameiva exsul on the Puerto Rican mainland do not constitute
a single homogeneous population but rather certain characters show
geographic
patterns
of
variation.
Disregarding
for
the
moment
those populations partially isolated in the inland river valleys
and
focusing
on
the
coastal localities
(represented
by heavy
lines in
80
Figs. 27 and 28), it is seen that the dominant pattern of variation is
an
east-west
cline along
the northern
coast.
This appears in
the
number of femoral pores with the northwestern ameivas having
a
significantly
lower number than those from the San Juan
area or
the
northeastern tip
of
Puerto Rico
(Fig.
28).
The
connecting
population in north central Puerto Rico is intermediate in number
of femoral pores and does not differ significantly
from those either
to the west or to the east. Thus the change in number of pores is
gradual and the extreme populations
at the two ends of the island
are
connected by
intermediate
ones.
The
eastern and southern
coastal areas are not significantly
different from each other or from
those of the eastern end of the cline. The number of caudals in the
15th row of scales around the tail shows the same geographic trend,
i.e.,
a
low number
of scales in northwestern and north
central
Puerto Rico
becoming significantly
higher
in
the San
Juan
area
and in northeastern Puerto Rico.
With respect
to this
character,
the
eastern and
southern populations
are identical to the north-
eastern
populations
and
all
are
significantly
higher
than
the
Fig. 27.
Variation
in
numbers of brachials, preanal scales,
and transverse
rows
of
ventrals of Ameiva
exsul from Puerto Rico, Cayo Santiago and Caja
de Muertos,
Ameiva alboguttata from Mona, and Ameiva
desechensis from
Desecheo.
Symbols
same as in Fig. 21. N varies between
10 and
115 specimens per locality.
81
northwestern ones. The cline seems to be steepest between the north
central populations
(Cibuco
and Dorado Beach)
and San
Juan
as
the
northwestern
and
north
central
populations
do
not
differ
significantly
from
each
other
although
both
are
significantly
different from San Juan and northeastern Puerto Rico. The latter
two areas are not significantly
different (Fig. 28).
With respect to preanal scales,
the northern populations
are not
significantly different, i.e., there is no northern coastal cline as in
the previous two characters. The southern and eastern populations
are not
significantly
different from
each other.
Except
for
the
southeastern specimens,
they have
a significantly
greater number
of preanal
scales than the San Juan population.
The northeastern
population
is
intermediate between
those
of
the
northern
and
eastern coasts and the character
can therefore be considered clinal
(Fig. 27).
Variability
in the number of longitudinal
ventral rows of scales
was
not
great and
most
of
the inter-population
differences not
Fig. 28.
Variation
in
numbers of femoral
pores, caudals
in
15th row,
and ratio of
snout to body length in Ameiva exsul from
Puerto Rico, Cayo Santiago and Caja de
Muertos,
Ameiva alboguttata from Mona,
and
Ameiva
desechensis
from
Desecheo.
Symbols same as in Fig. 21. N
=
10 to
115.
82
significant.
Those of the northern coast are the same except for the
San Juan area population which has a significantly larger number of
rows than adjacent populations
on the northern coast. This may be a
reflection of
gene-flow
from the
eastern
and
southeastern popu-
lations through the Caguas Valley, all of which tend to have
more
transverse rows than the northeastern and north central population
(however,
differences
are significant
only
for
the Caguas
Valley).
The two remaining
characters
which
were used
were relatively
uniform
throughout
the
mainland.
There
were
no
significant
differences among any of the populations
with respect
to numbers
of brachial scales, although the north central ameivas tendedto have
a greater number. The north central population
had significantly
shorter
snouts
than all
of
the other
populations
except
for
the
northwestern ones. None of the other populations
were significantly
different from each other with respect
to this character (Fig. 28).
Populations
were studied from two partly
isolated valleys.
One
of these,
the Caguas
Valley,
connects
the San
Juan
area
to the
eastern coast. From Figure 24 it appears that there is no connection
to the east at
an altitude below 500 ft. However, this is an artifact
due to lack
of detail on
the contour map. The road between the
Caguas Valley
and Naguabo
(approximately
upper dotted line of
Fig.
24)
follows
a
narrow
valley.
Altimeter readings
reached
a
maximum of 250 ft along this route. There is a similar connection
containing
the road
from
the
Caguas
Valley
to
Humacao
(ap-
proximately lower dotted line in Fig. 24) with a maximum elevation
just at 500 ft.
Because of partial isolation it was thought that the ameivas from
this
valley
might
differ
somewhat from
those of
coastal
areas.
Differences,
however,
were
slight.
In comparison
to
San
Juan
specimens,
those from the Caguas Valley
had a significantly lower
number of femoral pores and caudal scales,
a significantly
shorter
snout in relation to body length,
and greater numbers of preanal
scales.
The same relation
was shown with respect
to the eastern
populations
except that the differences were not
so great (although
statistically significant)
(Figs. 27-28).
As previously
mentioned,
the number of transverse ventral scale
83
rows
in the San
Juan
area
seems to be
influenced by
gene flow
through the Caguas Valley from the eastern populations.
The other area containing
a partly isolatedpopulation is the valley
of the Rio Grande de Arecibo and Rio Caonillas. It is a dead end,
opening only onto the northern coast between the northwestern and
north central collection areas (Fig. 24).
The specimens
studied all
came
from the
town
of
Utuado. This population
did
not
differ
significantly
from
those
of
the
northwestern and
north
central
areas
in any
character although
differences in number of preanal
scales
and ratio
of
snout
length
to body
length
were
near
the
borderline of significance.
With respect
to the cline in number of
femoral pores, the Utuado population
was intermediate between the
northwestern and north central population which is what would be
expected
for
a non-isolated population
in that geographic
region.
Hence, it seems that gene flow into the valley is great enough that
little divergence has taken place.
VARIATION IN Ameiva exsul
ON THE SMALLER ISLANDS
When
the
characteristics
of
the
ameivas
from
the
principal
Virgin
Islands
are compared
with their nearest neighbors
on the
Puerto Rican mainland (heavy lines
on Figs. 29-30) it is
seen that
the inter-island differences are not great. None of the larger islands
are
significantly
different
from
any
other with
respect
to
the
number of preanal scales of their resident ameivas. The same is true
of
number of
transverse
rows
of
ventral
scales
except
that the
St.
John population
has
a significantly
lower number than all
of
the
other populations
except
that
on
Tortola.
The
number
of
brachials is the same from Puerto Rico to, and including,
St. John;
then there is a cline in this character with Virgin Gorda having
a
significantly greater number than St. John. Tortola is intermediate
both
geographically
and
with
respect
to number
of
brachials.
Number of femoral pores was
constant in all
of the larger islands
except
that on St.
John the number tended to
be lower (signifi-
cantly
so only with respect to Tortola).
There was
a marked difference in number of caudals between the
western most of the larger Virgin Islands as compared to St. Thomas
and Puerto Rico.
The populations
on St.
John, Tortola and Virgin
84
Gorda all
had significantly
lower
numbers of caudal
scales
than
those
on
St.
Thomas
and in
eastern
Puerto
Rico which
were
practically
identical
with
each
other. In
this
case,
this may
re-
present temporal rather than spatial change.
The populations
on small keys
(thin
lines in Figs.
27-30)
were
usually similar to those on adjacent
larger
islands although
some
showed significant
variation in one or more characters. For example,
there was
a significantly
higher
number of caudals in the Cayo
Santiago population
than in those of either eastern or southeastern
Puerto Rico (Fig. 28). The Caja de Muertos population
differs from
the southwestern Puerto Rican
one (but not from the southeastern
one) in having significantly fewer preanal
scales (Fig. 27), Palominos
specimens
had relatively
shorter
snouts than the
nearest Puerto
Rico mainland population
(Fig.
28). Ameivas
on Little St. James
had significantly
fewer caudals than those on either St. Thomas or
St.
John between which Little
St.
James lies.
Those from Little
St.
James had fewer transverse
rows of
scales than those
on
St.
Fig. 29.
Variation
in numbers
of brachials, preanal scales
and transverse
rows
of
ventrals
in
Ameiva
exsul
from the
Virgin Islands,
various
of the
northeastern
Puerto
Rican keys
and 2 localities
in Puerto
Rico.
Symbols same
as in Fig.
21.
N
= 10—56 specimens per locality.
85
Thomas. The population
on Lovango key was intermediate between
that of St. Thomas and that of St. John and significantly
different
from both in number of transverse
rows of ventrals. It had fewer
caudals than the one on St. Thomas. Specimens on Culebrita have a
significantly
greater number of brachials
than those from either
St.
Thomas or
Puerto Rico.
We lacked
sufficient
material from
Culebra to make comparisons
with
Puerto Rico.
GRANT
(1931b)
however,
indicates
he found no
differences between Culebra and
Puerto Rican specimens.
In
some instances populations
on
the small keys
differ signifi-
cantly from those of any nearby large island, e.g., Sandy Key and
Culebrita in number of brachials and Little St. James in number of
caudals.
However,
in
all
the
other
instances
where
there
were
significant
differences, they occurred between a large island and its
eastern small neighbor. Small islands lying west of larger
ones tend
to have the
same characteristics
as the larger island even though
inter-island distances may be rather great (e.g., Palominos does not
differ from St. Thomas in number of femoral pores but does differ
Fig. 30.
Variation in numbers
of femoral pores and
caudals in
15th row in Ameiva
exsul from the Virgin Islands, various of the northeastern Puerto Rican
keys
and
2 localities in
Puerto Rico.
Symbols same as in Fig. 21.
N
=
10—56 per locality.
86
from northeast Puerto Rico). This is most clearly
seen with respect
to Little St.
James
and Lovango
Key,
both
of which
lie
about
halfway
between St.
Thomas and St.
John.
In
those characters
which
differ between St.
Thomas and St. John,
the
smaller two
islands resemble the latter more than the former (Figs. 29-30).
Of special
interest are
the characteristics
of the population
on
Cayo Diablo which GRANT (1932a)
described
as a separate species,
Ameiva birdorum,
on the basis of color pattern
(discussed
below)
and low femoral pore count. It is obvious from Figures
29 and 30
that the Diablo specimens
are not significantly
differentfrom those
of adjacent islands in number of femoral pores. In those characters
(preanals, brachials)
in which they
do differ from specimens
from
nearby Lobos and Hicacos, it is the latter 2 islands rather than Cayo
Diablo which depart
from the more general condition.
COLORATION OF Ameiva exsul ON THE PUERTO RICAN MAINLAND
A generalization
of the color pattern
of A. exsul
on the Puerto
Rican mainland is as follows. Head and back olive brown to olive,
becoming
olive gray to gray
on the tail, scattered black spots
on
dorsum, particularly posterior part. Sides with a darker ground color
containing irregular black markings
(sometimes
as vertical stripes)
interspersed with white dots. The same pattern continues onto the
legs and tail. On the side of the tail spots of blue become intermixed
with the black and white markings. The white dots on the sides of
the body rarely
coalesce to
form vertical stripes.
Sometimes the
lower ones fuse to form a white lateral stripe running from axilla to
groin. The tip of the snout and lower jaw is pink.
Venter whitish to
pearl gray with a bluish tinge
which becomes much lighter toward
the sides and on the underside of the tail and rear legs. The pupil is
biloculate,
surrounded by
an orange-brown iris.
Ontogenetic
variation in this pattern occurs. Young animals have
a dorsal series of white dots which usually begins
at the shoulder.
The size and prominence of these spots vary greatly and they tend
to be lost with age beginning anteriorally and proceding posteriorly.
Even quite large specimens still have them on the base of the tail.
The young also have prominent
dorsolateral stripes
on the body
beginning
at
the supraciliaries.
Usually
these stripes
are bright
87
yellow although
cream, buff, or yellowish-green
ones
are sometimes
encountered. Frequently they are bordered on each side by
a narrow
black line. They begin disappearing early in life but the process is
so gradual that it is difficult to indicate a size
at which they
are
completely
gone. Furthermore, individuals of the same size may be
at quite different stages of loss of stripes
(Fig. 31).
Occasionally,
quite young individuals are found with dim markings. The brightly
colored stripes first fade to a pale brown and gradually
assume the
background color of the dorsum,
as does the dark line medial to it.
In half grown specimens both the dorsolateralstripes and the dorsal
dots are frequently represented by brown areas, slightly paler than
the background color. In larger specimens the former location of the
dorsolateral stripes is marked by the transition from the background
color of the dorsum to the darker coloration of the sides.
Background
color is
also brighter
in the juveniles,
sometimes
approaching
a rusty
brown
in
very
small individuals.
Later
it
changes to olive and sometimes gray. The brown remains longest
on
the head and neck.
The bluish ventral coloration becomes
more
intense and
more
widely distributed with age. Very small individuals sometimes have
a coppery venter. The throat in larger specimens takes on
a pinkish
to rose color which increases in intensity with size.
These changes
in
color
are
general
trends
seen
on
the
Puerto
Rican mainland and
considerable variation among individuals
of
the
same
size is
encountered,
even
within
a particular
locality
(Fig. 31).
COLORATION
OF Ameiva exsul
ON SMALL ISLANDS
On the keys
and islands
east of Puerto Rico, there
is even greatervariation than
on the mainland.There is a tendency for the dorsolateral stripes to persist longer on
some of the
islands, particularly from Blanquilla through Culebra.
In specimens
from
the latter,
the lateral stripe resulting from
a coalescence
of lateral spots
(mentioned earlier)
is frequently prominent. There
is
also
a trend
toward larger
spots in the populationswith persistent dorsolateral stripes. Culebra specimens have
the largest dorsal spots of any we have examined. GRANT (1932C) who collected in
the Virgin Islands mentions that Guana Island, Peter Island and Anegada"produced
large, highly colored
specimens, with prominent dorsolateral stripes reminiscent
of
similar
ones on Culebra" although he reports those from Tortola and
Buck Island
(Tortola) as generally dark, losing their stripes early. Our own field work in the area
Fig.
31.
Ameiva
exsul
from
the
San
Juan,
Puerto
Rico,
area.
Note
extreme
variation.
Some
juveniles
have
stripes,
others
do
not;
stripes
persist
in
some
adults,
in
others
not;
amount
of
spoting
is
highly
variable
in
all
sizes.
89
has
confirmed
his
observations
and
added
some
details.
Guana
Island,
Little
Camanoe
and Greater Camanoe
are all characterized by ameivas with rusty red
on
the sides of the head
and with bright yellow dorsolateral
stripes which persist
even
in large adults. By contrast Scrub Island and Marina Key, slightly east of the above
mentioned islands,
have
ameivas with
duller
hues
on the head and in which the
dorsolateral
stripes disappear with age.
GRANT
(1932a)
indicates
that the back-
ground color of individuals
from Cayo Diablo is extremely dark.
GARMAN (1887)
mentioned (under the name of A. riisii) that St. Thomas specimens were darker,
more olive, and with
more conspicuous markings than those from Puerto Rico.
As mentioned previously many
of the larger individuals
from Puerto
Rico
and
elsewhere have
a pinkish to rose colored throat in life (fades in preservative). This is
not true of those from Tortola
(GRANT
1932c), although we
obtained
a specimen
from nearby Scrub Island
with
a pink throat.
GRANT (1932c)
also records
an in-
dividual from Peter
Island
with turquoise blue
on the
sides of the
neck and
head
(usually this
color is restricted to the belly, sides and underside
of the tail). Most
individuals
we
observed
on
Peter
Island
were
much
as
described
for
Greater
Camanoe.
However, the larger males had blue
heads; the same was true
onScrub
Island.
GRANT
(1932a)
lists
the
chin
color of
specimens from
Cayo
Diablo
as
"purple". However,
we have collected live specimens from there and found
their
chins to be only a slightly deeper shade of rose than those from the Puerto Rican
mainland.
The
difference
was not of the magnitude GRANT'S
usage
of the
terms
"red"
(P.R.)
and "purple" (Diablo) would suggest.
The
islands
south
of
Puerto
Rico
also
have
populations slightly
different
in
coloration from those on the main island. GRANT &
ROOSEVELT (1932) indicate that
the dorsolateral
stripe persists
longer on Caja
de Muertos
than
on Puerto Rico,
whereas
on Cardona Key,
there was a wide variety of color patterns ranging from
nearly unmarked
animals to those with
extra, white
lines
below
the
dorsolateral
stripes;
none had blue
or salmon
on the underside.
RELATIONSHIPS
OF
EASTERN
POPULATIONS
OF
Ameiva
exsul
AND
STATUS
OF Ameiva birdorum
As indicated above,
a number of authors have pointed out slight
color differences in populations of Ameiva exsul among the various
islands under consideration, and GRANT (1931a) stated, in reference
to Ameiva exsul, that "the color and markings vary considerably in
individuals but I have been unable to find any correlation between
this and their distribution". STEJNEGER
(1904) points out that there
is extreme variation in coloration on the mainland of Puerto Rico
even within one locality.
Our observations agree with this and we
feel that
assigning
these
various insular populations
to separate
species or subspecies
would not be realistic.
The diagnosis which GRANT (1932a) presents for Ameiva birdorum
is:
"Distinguished
from other
Ameivas
[sic]
of
the
Puerto Rico
region by its
black back,
blue undersides and low femoral pore
90
count, averaging
13.8" (on
one side).
BARBOUR
(1937)
also
con-
sidered
this
population
a
"good,
distinct
form".
However,
as
mentioned earlier it shares its melanistic tendencies with specimens
from
other islands,
e.g.
Tortola
and
Buck
Island (Tortola)
as
indicated by GRANT himself (1932c). The blue underside is distinc-
tive only in that the Cayo
Diablo animals have
a slightly
more
intense blue than other specimens
examined alive except for those
from Cayo
Blanquilla
which
are equally bright.
The
low femoral
pore count is not a valid distinguishing
feature,
as discussed in the
previous
section
(see
Fig.
30).
Other
characteristics
which
were
used
to
distinguish
A.
birdorum
from
A.
exsul include (1)
per-
sistence, and large size of dorsal dots which, however,
is also true of
Culebra specimens,
(2)
absence of
a median light band (which
is
just
another way of saying
the background
color
on
the back
is
black),
(3) minor details of the black bands,
(4) absence of the white
line from axilla to groin
(also
true of most specimens
from
most
localities)
and (5)
the purple chin (previously
discussed and shown
not
to be distinctive).
It is also mentioned in his description
that
the dorsolateral stripes
disappear early,
a character also found on
Tortola and highly
variable
on
the
Puerto Rican mainland and
elsewhere, but which is not shared by the Culebra population.
Inasmuch
as
there
is
condiderable variation
in
color pattern
within
an island as well as from island to island, and most of the
characteristics
of the Diablo Key population
are not unique
to it,
or to the combined populations
of Diablo and Blanquilla keys (e.g.
large, persistent spots also occur on Culebra specimens and melanism
with early loss of stripes
on those from Tortola and nearby Buck), it
seems unwarranted to maintain specific status for Ameiva birdorum
and hence we relegate
this name to the synonymy of Ameiva exsul.
In summarizing the variation in color pattern of ameivas east of
the Puerto Rican mainland, it can be stated that prominent
dorso-
lateral stripes
persist
longer
in
individuals
from
Guana Island,
Peter Island, Anegada and Culebra Island and its surrounding keys,
than those from other places. Melanism occurs on Tortola and Buck
(Tortola),
reaching
its
most
extreme condition
on
Cayo
Diablo.
Spots
are
larger
than usual
in
specimens
from
Culebra,
Cayo
Diablo and Blanquilla.
91
It is perhaps
true,
as BARBOUR
(1914) points out
in
respect to
these
insular
populations of Ameiva exsul,
that they "represent groups of individuals
strongly
tendingtoward the formation of distinct species by isolation". However, because of
the great intra-island
as well
as inter-island variation, observed differences in color
or scutellation are true only in a statistical
sense for a given character. Furthermore
it is
evident that
each
character
has its
own geographic distribution,relatively
independent of the others, and that they occur in various combinations with each
other. Despite the insular
nature of the range
of the species, geographic variation
appears
"clinal"
rather
than
a series
of discretely
different
populations; sharp
gradientsin one character do not correspond geographically to those in another
one.
It is difficult, therefore, to draw sharp boundaries delimiting subspecific categories.
Consequently, we consider subspecific recognition for these populations as obscuring
rather
than clarifying the relationships involved
and prefer simply
to
describe
various character gradients and interpret them
in terms of relative degree and time
of isolation
of different populations, and the influence
of
gene flow among them.
A detailed
analysis
of characteristics
of populations from
more of the
smaller
islands
would
undoubtedly further complicate the pattern of variation
just
des-
cribed. Certain of the small keys
differ significantly from
any of the other nearby
populations in
one or
more scale
characters
and by the
criteria used
by
various
authors could be considered
distinct subspecies, e.g., the population on Sandy Key
differs from those of nearby islands in number
of brachials.
As
a result of genetic
drift and/or an expression of the founder principle (MAYR 1963) many of the other
small islands
might be expected to
show similar phenomena. However, lack
of
a
sufficient
number of specimens from
some islands
and
fear
of unduly decimating
populationsthroughintensive collecting on some of the smallest ones, has prevented
as complete an analysis
as might otherwise have been desirable.
PUERTO
RICO: AMNH 5786, 5794-97, 12901, 12929-30,
12936-65, 13246,
13265,
13279-81,
13797-802, 13817;
CNHM
12450,
12467-68, 38550-51,
130258-76; MCZ 6082-83,
12110, 36468-72, 36478, 36488-505, 36508-19,
36539-43, 57853-55, 58775, 61499; UMMZ
53254, 55825,
73864
(9
indivi-
duals), 73878-81, 73887-91
(12), 73894, 78477 (3), 106173-74; USNM 25448-
50,
25521,
25581-85,
25608,
25635-38, 25640-41,
25654-59,
25660-70,
25720,
27046,
27146,
27221-22,
58795-96,
86547,
89232-33,
90349-50,
98968;
UPRRP
1,
24,
323-332,
1305-10,
1321-24,
1338, 1401-03,
1946,
2458,
2460,
2530-31,
2599, 3448-72,
3606, 3743,
3748,
7 uncataloged.
CAJA DE MUERTOS ISLAND:
MCZ 36473-77, 66424-28; UMMZ
73872
(6),
Field
Series
GH 2540-41; UPRRP 2569.
CARDONA KEY: MCZ 36479-83;
UMMZ
73867
(4).
CAYO SANTIAGO:
MCZ 36484-87; UMMZ
73865
(4);
USNM 58794; UPRRP 1815-27, 2542-57.
CAYO BATATA: UPRRP 2095.
HICACOS: MCZ 36522-25; UMMZ 73870
(3);
UPRRP 2700-01, 2723-24,
3953.
RATONES ISLAND (near Hicacos): MCZ 36532-33; UMMZ 73862 (2).
• KONYOKI: UPRRP 3981.
LOBOS
ISLAND: MCZ 36526; UMMZ
73863
(2); UPRRP 3543-64.
CAYO BLANQUILLA: MCZ 36534; UPRRP 3518-19.
CAYO DIABLO: UMMZ
73854 (type of A. birdorum), 73855 (8), 73895 (12),
73896-98
(16)
(last 3 entries paratypes of A. birdorum)); UPRRP 3506-14.—
LEPER ISLAND (Isla Cabras
near San
Juan): MCZ
36520-21;
UMMZ
73866
(2).
— SPINY BUTTE: UPRRP
1 uncataloged.
—• PALOMINOS : MCZ 36527-31;
UMMZ
73873
(5).
—
PALOMINITOS:
UPRRP
3603.
—
RAMOS
ISLAND:
92
UPRRP 3489.
PINEROS
ISLAND: UMMZ
73882 (2); UPRRP 3784, 3788-
89.
CABEZA
DE PERRO: MCZ 36467; UMMZ 73883
(2); UPRRP
4020.
ALGODONES
KEY:
UPRRP
3926.
LEVIN'S
ROCK:
UPRRP
3939.
CULEBRA
ISLAND:
USNM 49585; UPRRP
2374-75.
CULEBRITA ISLAND:
UMMZ
73892
(3),
80599
(5),
80788
(4).
CAYO NORTE: UPRRP
5067.
ST. THOMAS ISLAND: UMMZ 73877 (13), 80597-98.
ST. JOHN: KU 45465-
91, 45529-56; UMMZ 80596.
TORTOLA:
UMMZ 73885 (17),
80587 (3).
BEEF ISLAND:
UPRRP
4213-16.
VIRGIN
GORDA:
KU
45446; UMMZ
80595, 80600 (10).
ANEGADA: UMMZ 73874, 80592 (5),
WATER ISLAND :
USNM
30696
(type
of A. exsul).
SAVANNA
ISLAND:
UMMZ
80601.
OUTER
BRASS
ISLAND:
UMMZ
80602.
INNER
BRASS
ISLAND:
UMMZ
80588.
— • HASSEL
ISLAND: UMMZ
80591.
THATCH KEY: KU
45574.
MINGO KEY: KU
46687.
LOVANGO KEY:
KU 45557-73, 45575-88; MCZ
36506-07; UMMZ
73876.
GREATER ST.
JAMES: KU 45442-45. • LITTLE
ST.
JAMES: KU 45447-53, 45492-511.
LEDUCK ISLAND: KU 45454-58.
FLANNIGAN
ISLAND: KU
45459-64.
SANDY KEY: KU 45512-22, 45524-
28; UPRRP
1 uncataloged. —• BUCK
ISLAND (Tortola): UMMZ
73868.
GUANA ISLAND : UMMZ 73875.
SCRUB ISLAND : UPRRP 4302.
GREATER
CAMANOE
ISLAND:
UPRRP
4299.
LITTLE
CAMANOE
ISLAND:
UPRRP
4301. —PETER ISLAND: UMMZ 73886 (2), 80594; UPRRP 4241-42, 4262-63.
-
DEAD
MAN'S
CHEST:
UMMZ
80593.
SALT
ISLAND (near Tortola):
UMMZ 80590 (2).
COOPER ISLAND: UMMZ 80589.
Ameiva alboguttata Boulenger
1896
HABITS
Little is known
of the habits
of Ameiva alboguttata. It is
diurnal
and forages in
much
the same
way
as described
for A. exsul. At night we
have
found
inactive
individuals
under
limestone
boulders.
SCHMIDT
(1920,
1928)
indicates
that
20
stomachs yielded mostly vegetable matter (chiefly red-coated
seeds)
and
insects,
although land
snails
and
juvenile Anolis
were
also present. Alsophis
feeds
on
A. alboguttata.
DISTRIBUTION
This species is' endemic to Mona Island (Fig. 24) where it is found
on the beaches as well as on the limestone cap. The island is hot, dry,
and covered with a xeric scrub vegetation which grows up through
cracks in the limestone. There is an abundance of limestone boulders
as well as crevices in larger rock
masses which serve
as
cover.
The nearby
island of
Monito which is
also
a
limestone
mass
covered with xeric scrub has been examined and no Ameiva found
(ROLLE
et
al.
1964).
It is
perhaps
significant
that this
island is
completely surrounded by cliffs which
are undercut at the base by
93
wave action, thereby making repopulation following local extinction
difficult, particularly
for
a non-climbing form like Ameiva.
The relatively
great
distance of Mona from Puerto Rico,
the
absence of a Pleistocene land connection, and the fact that much of
the perimeter is composed of cliffs, probably
contributed to genetic
isolation of Ameiva alboguttata, resulting in divergence from A. exsul.
VARIATION IN SCALE
CHARACTERS AND COLORATION
Ameiva alboguttata
does not
differ significantly
from A. exsul in
numbers of brachials, preanal
scales,
transverse
rows
of ventrals
(Fig. 27)
or relative snout length (Fig. 28). However, the numberof
femoral pores and
number of
caudals in
the
15th
row
are
both
significantly lower than for any of the Puerto Rican localities. The
coloration of
A.
alboguttata
is
much
less
variable from
that of
A. exsul,
and differs from the latter in having
a gray (rather than
olive
or brown) background
color which
does not differ greatly in
intensity between the back and sides. The small white dorsal spots
begin
on the neck rather than the shoulder and persist throughout
life. The chin and iris color is the same as in A. exsul.
The coloration is uniform enough and sufficiently
distinct from
that
of Ameiva
exsul that the
two
species
can be
distinguished
immediately
from
either live
individuals
or preserved
specimens
(Fig.
32),
except
in the
case of rare
A.
exsul which
have
an
ex-
ceptionally
gray
background
color
and
are
unusually
heavily
spotted.
In such instances separation
can be made on the basis of
where the spotting
begins
anteriorly
(neck
in A.
alboguttata
and
shoulder in A. exsul).
There is
some ontogenetic
variation in Ameiva alboguttata.
The
dorsolateral lines are absent
even in
some young. Where they are
present they are indistinct and disappear early in age. A lateral line
running
from
axilla
to groin
sometimes
occurs
in young ones,
although it also
seems to disappear with age. The largest specimen
we have examined was
126.9 mm s-v; the smallest 34.6 mm s-v.
Ameiva alboguttata
and A. exsul
are clearly
very closely related.
It
is
perhaps
arbitrary
whether the
former be
considered
a full
species
or
a subspecies of the latter. We choose to follow previous
authors in maintaining it as a distinct species.
94
Ameiva desechensis.
Note absence of dorsolateral
stripes and presence of alternatinglight and
dark vertical bars
on side.
Fig. 33.
Side view of holotype of
A. alboguttata.
right,
A. desechensis;
; center,
A. exsul ;
Ameiva exsul subgroup: left,
Fig. 32.
Species of the
95
MONA
ISLAND:
AMNH
6447-49,
13704-06,
13738-40,
13742-44,
13746-54,
13756-65,
13767-70,
31890,
31894-95,
31899;
MCZ
7898;
36351-84;
UMMZ
73848-53
(35
inviduals);
USNM
29368-72,
29374;
UPRRP 3017,
3056,
3069.
Ameiva
desechensis sp.
nov.
Holotype: UPRRP 235 (subsequently donated to MCZ). Adult female collected
by H.
Heatwole, DESECHEO Island, P.R.,
28 March
1961.
Paratypes:
UPRRP
230-234, 236-238
(Fig. 32-33).
Diagnosis:
A species
of Ameiva with 10 longitudinal
rows
of
ventral plates,
32
or
more
femoral pores (sum
of those
on
both
sides),
41
or more caudals and 25 or more preanal
scales in 95% of
the individuals;
no
dorsolateral stripes
at
any
stage
of the life
history;
dorsal spots inconspicuous and mostly
on the tail, seldom
extending further anterior than the sacral region;
spots and black
markings
on sides
coalesced into a series of alternating
black and
bluish-white vertical bars.
Range:
Known only from Desecheo Island, Puerto Rico.
Description
of
Holotype:
Dorsum
nearly
uniformly
gray,
slightly lighter
on head and dorsal surface of tail; snout pink; iris
orange-brown;
sides with a pattern of alternating black and bluish-
white vertical stripes which fade out in dorsolateralregion; scattered
small white spots
on dorsal part of tail; becoming larger and more
pronounced
on sides of tail; venter bluish gray; front and rear legs
dark gray dorsally, white with a gray wash beneath, the latter with
small bluish white spots above;
10 longitudinal
and 35 transverse
rows
of ventrals;
3 preanal plates;
26 preanal
scales;
32 femoral
pores;
15 brachials; 41 caudals in the 15th row; upperlabials 6 + 6;
lower
labials
7 + 7;
supraoculars
4 + 5;
supraciliaries
6 + 7;
nostrils in suture between nasals;
dorsal scales small and granular
except
on tail where they
are platelike and keeled;
ventral plates
rectangular
and smooth except
on posterior
2/3 of tail where they
have weak keels;
scales of throat small and granular except for an
area
of
about
10
scales in
the center of the
throat and 4
rows
96
between the gular
and prepectoral
skin
folds;
these latter
scales
larger and almost plate-like;
snout vent length 90 mm.
VARIATION
AMONG
PARATYPES
Other specimens
collected at the same time and place as the type
were similar in most aspects
of the color pattern.
In the smaller
specimens, the black and white vertical markings tended to extend
across
the
dorsum
and
connect with
their counterparts
of
the
opposite side to form
a transversely striped
pattern. These dorsal
markings
were less distinct than
on
the sides, however,
and
were
lacking entirely in larger specimens.
No individuals showed a trace
of dorsolateral stripes,
even a very young specimen only
43.1 mm
snout
to
vent.
Coloration
seems
to
be
much
more
uniform in
Ameiva
desechensis
than in
A.
exsul,
its
nearest
living
relative.
Both A.
desechensis and A. alboguttata
differ in the same direction
from A. exsul in their reduction of the dorsolateral line and in the
grayness of background
color.
The latter is perhaps attributable to the fact that both A. alboguttata
and A.
desechensis live
on
islands consisting
principally
of
gray
limestone, whereas many of the Puerto Rican populations
occur in
areas of brown sand. The "grayness" of the ameivas has parallels in
other groups found on Puerto Rico and the two western islands, e.g.,
Anolis,
and hence may be convergent
responses to selective
pres-
sures common to both Mona and Desecheo (as well as to some of the
eastern keys). It is possible that degree of dorsolateral striping may
have quite different fitness values
on different islands and that the
extreme inter-island
variability
in this
character
as
well
as
in
grayness
can
be
related
to inter-island
differences
in
substrate
color.
The variability of scale characteristics and body proportions
can
be
seen from Figures
27 and 28. Ameiva desechensis differs signifi-
cantly
from all
mainland populations
of
A.
exsul in number of
preanal scales. Number of femoral pores is also significantly
higher
than in any Puerto Rican locality except northeastern Puerto Rico.
Numbers of transverse rows and brachials and ratio of snout length
to body length is the same as in A. exsul and A. alboguttata.
97
RELATIONSHIPS OF THE AMEIVAS OF MONA, DESECHEO
AND PUERTO
RICO
Examination of Figures
27
and 28
reveals that where Ameiva
alboguttata
and A.
desechensis differ from A. exsul, they diverge in
opposite
directions.
Thus A.
alboguttata
has
a lower
number
of
femoral pores than A.
exsul,
whereas A.
desechensis has
a higher
number. In addition, A. desechensis has a higher number of preanal
scales than A.
exsul;
the latter and A. alboguttata
do not differ in
this character.
The
same trend applies
to certain aspects
of
colo-
ration ; A. alboguttata
shows greater dorsal spotting than A. exsul,
whereas dorsal spotting is much reduced in A. desechensis (Fig. 32).
However, it seems that A. exsul is close to the ancestral type from
which A. alboguttata and A. desechensis separately diverged,
or was
perhaps
itself the ancestor
of
these
two species.
At
any rate A.
desechensis cannot be considered an intermediate between A. exsul
and A.
alboguttata
-
but
rather
has diverged
from
A.
exsul
in
different ways than has A. alboguttata. Both seem to be more closely
related to A. exsul than they are to each other.
DESECHEO
ISLAND:
USNM 59760-61; UPRRP
235 (type,
donated to MCZ),
230-234, 236-38 (last two entries paratypes).
Ameiva undulata group
Ameiva wetmorei Stejneger
1913
HABITS
Ameiva
wetmorei inhabits
open,
xeric scrub.
It is
diurnal
and
forages on the
groundin the same way as do other ameivas. It spends its nocturnal inactive period
in burrows under rocks.
The bright metallic blue tail is frequently lost by autotomy,
after which it writhes violently on the ground thereby distracting attention
from its
owner.
DISTRIBUTION
This species
has
a very restricted distribution, being found only
in southwestern Puerto Rico and
on Caja
de Muertos island and
Magueyes Island (Fig.
17).
The geological
formations,
soil
and vegetation
types
found in
these
areas
have
a wider
occurrence
in Puerto
Rico
than
does
98
Ameiva
wetmorei,
and
they
thus
can
probably
be
excluded
as
possible
factors restricting
the lizards' geographic
range. Tempe-
ratures
are
similar
all
along
the
southern
and
extreme
eastern
coasts with the annual mean temperature exceeding
78° F.
(Pico
1954). A.
wetmorei could probably not inhabit the cooler areas at
high
elevations
as
GRANT
(1931a)
mentions that individuals
are
almost torpid at
75° F.
However,
the
zone of high temperatures
extends further east along the southern coast than does the range of
A.
wetmorei. Hence,
temperature
does not
seem
to be the factor
confining it within the small area it occupies.
The most striking
feature of
southwestern Puerto
Rico
is
its
aridity
and it is the only
part
of the island with
a mean annual
rainfall of less than 30 inches. In fact the 30 inch isohyet
encloses
all of the recorded mainlandlocalities of this species except the one
on the
peninsula
of
Cabo Rojo
(Fig.
17).
Rainfall progressively
increases from the coast toward the interior mountains and also from
west
to
east
until
mean
annual
values
of
over
200
inches
are
recorded from some of the eastern peaks (Pico
1954). Although rain
data
are lacking
for Caja
de Muertos, it is
a relatively
hot, dry
island, similar in appearance to southwestern Puerto Rico.
Other
members of the species group also inhabit dry areas. For example
Ameiva maynardi occurs on Great Inagua and A. lineolata,
a species
which COCHRAN (1941)
indicates
as a close relative of A. wetmorei,
is restricted to Beata Island and the arid Cul-de Sac of Hispaniola.
VARIATION
The
range
of
Ameiva
wetmorei
was
earlier
considered
to
be
southwest Puerto Rico and Caja de Muertos Island (SCHMIDT 1928,
and Fig.
17).
However, GRANT & ROOSEVELT
(1932)
set
apart the
population
on Caja de Muertos as a separate subspecies, A. wetmorei
eleanorae,
on the basis of several minor color pattern
differences.
Later,
GRANT (1932d)
raised the Caja
de Muertos form to
a full
species, A. eleanorae, when he discovered that it had fewer femoral
pores
than
mainland populations.
BARBOUR
(1937)
considered
A. eleanorae as a "rather ill-definedform". We compared specimens
from both
areas and
have re-evaluated the status of the Caja
de
Muertos population.
99
Three
localities
were represented
by
sufficient
specimens
for
comparison
to be
made. These were
La Parguera
and Ensenada,
P. R., and Caja de Muertos.
There was
a west-east difference in a number of characters. The
snout
was relatively
longer in
the
eastern specimens
than those
from the west (Fig.
34). Although the Caja
de Muertos specimens
were significantly
different
from
those
from
La Parguera,
the
Ensenada
population
was
intermediate
and
not
significantly
different from either of the other two. The same was true for number
of brachials,
with the population
on Caja
de Muertos having the
highest number, that from La Parguera the lowest,
and that from
Ensenada being
intermediate. The two end localities
were signifi-
cantly different. Grant's observation that the Caja de Muertos form
had
a lower number of femoral pores than the mainland form was
verified by
the present study (Fig. 34).
However, there were also
differences between mainland populations
in
this
character,
the
number gradually
decreasing
from
west
to
east
with
Caja
de
Muertos simply being the extreme eastern population.
There
were
no
differences
among populations
of
the
various
localities in numberof caudals in the 15th row. Numberof transverse
Fig. 34.
Variation in
scale counts
and in ratio
of snout length to body length in
Ameiva wetmorei. Symbols as in Fig. 21. N
= 35—69 per locality.
100
rows varied widely between localities, all of which were significantly
different from each other, although there was no orderly geographic
sequence.
Both forms are very similar in coloration. The background
color
is black with a series of longitudinal
lines (Fig. 35) which are white,
cream or pinkish.
One is located middorsally
and extends from the
tip of
the snout onto the tail where it widens and becomes blue,
eventually being lost in the general blue color of the tail. At the base
of the tail the stripe has a central black line which is somewhat more
conspicuous
in the Caja
de Muertos specimens.
In these the white
line tends to be wider although there is a great amount of overlap
among populations. Lateralto the middorsal stripe is a pair of white
Left, specimen
from southwest
Puerto
Rico showing
typical persistence
of supraocular
stripe
on mid-body;
center,
specimen
from
southwestern
Puerto
Rico
which has coloration similar
to that typically found
on
Caja de Muertos, i.e., supraocular stripe disappears on body; right, specimen from
Caja
de Muertos.
Fig. 35.
Ameiva
wetmorei.
101
stripes
(supraocular
lines) beginning
between the nostril and
eye,
passing
across
the supraciliaries
and extending onto the tail. In the
Caja de Muertos population,
this stripe fades out and is lost on the
neck, not reappearing again until the base of the tail. In most of the
southwestern Puerto Rican individuals,
this line is not completely
lost but only fades out to
a rather inconspicuous
gray or brown.
This difference
was
considered by
GRANT & ROOSEVELT
(1932)
as
diagnostic
of the two
forms.
However,
some
of the mainland in-
dividuals also lose the supraocular stripe in the midbody region and
they
cannot be distinguished
from Caja
de Muertos ones (Fig. 35)
on the basis of this or other color characters (e.g., UPRRP 2116 and
2289 from
Tamarindo,
P.
R.).
Also
FOWLER
(1918)
presents
a
painting of an Ameiva wetmorei from Ensenada,
P. R., which lacks
the supraocular stripe at midbody. According to GRANT & ROOSEVELT
(1932)
some Caja
de Muertos individuals have a faint remnant of
this stripe at midbody,
thus resembling
the Puerto Rican form.
Lateral to the supraocular stripes
is a pair of stripes (subocular
lines) beginning just anterior to the eye and passing beneath it, over
the dorsal border of the tympanum and extending to about mid-tail.
Finally,
there is a line extending from the supralabials,
below the
tympanum, and along the sides of the groin. As GRANT & ROOSEVELT
(1932) point
out this tends to be
more conspicuous
in individuals
from Caja de Muertos.
In both
forms
the
dorsal surfaces of
the legs
are
black
with
conspicuous
white
spots,
sometimes coalescing
into short
bands.
The underside of
the
venter is salmon
to
coppery
(blue in
pre-
servative).
In life, the posterior part of the tail is metallic green or
metallic blue with broken,
circular black
bands surrounding
the
tail on the posterior part of each scale ring. In preservative
the tail
is always blue.
In view of the fact that the Caja de Muertos population simply
reflects
the extreme
eastern
end
of
gradual
west-east
character
gradients,
and is not sharply
set
off
from mainland
ones
on
the
basis of either scutellation or coloration,
we feel it does not warrant
specific
status. We also feel that the relationships
are most clearly
expressed by a description of the geographic trends in the characters
involved,
and
that
even
subspecific
status
overemphasizes
the
102
distinctiveness of the Caja
de Muertos population.
Consequently,
we place
Ameiva
eleanorae in the
synonomy of
Ameiva
wetmorei.
PUERTO
RICO: AMNH
13803-09, 13818, 13820-22, 13824-27,
13829-36,
13859; CNHM
12387-90, 12447-49; MCZ
36431-66
(69 specimens);
UMMZ
73856
(21), 73859-60 (45); UPRRP 1293, 1867-69, 2116, 2289.
MAGUEYES ISLAND : USNM 86548.
CAJA
DE
MUERTOS ISLAND: AMNH
32946-48; MCZ
36419-30 (paratypes
of
A.
eleanorae),
66429; UMMZ
73857-58
(17 individuals, paratypes
of
A. eleanorae) ;;
USNM
120785-86 (paratypes of A. eleanorae);
UPRRP
321,
2589.
Ameiva polops Cope
1862
DISTRIBUTION
This species
was
originally
found on St.
Croix
(COPE
1862). It
was not subsequently
found on
that island for
a number of
years
and was believed to be extinct
(BARBOUR
1937). However, it was
rediscovered
on
two
small keys,
Green Key
and Protestant Key
(Fig.
17) north of St. Croix by GRANT (1937) where it is still holding
out (SEAMAN
1961 and personal observation).
It
is
not
found
on
Buck
Island near
St. Croix,
and
the
two
previously mentioned keys
seem to constitute the entire range of
the species.
Although the mongoose has probably
served as a scape-goat and
received
an
undue share of blame for
extinction of West
Indian
reptiles,
its reputation
is probably
deserved in the
case of Ameiva
polops. The mongoose was introduced into St. Croix from Jamaica
in
1884 (SEAMAN
1961)
and
local inhabitants say it is
on
Buck
(St.
Croix).
It
is
not
on
the
two keys
where A.
polops
occurs.
Protestant Key has
a hotel which does a lively business and most
of the island's vegetation
has been modified.The fact that A. polops
has been able to maintainitself under these conditions would seem
to indicate that man has probably not been directly responsible
for
extinction of the St. Croix populations. l
1
Since this paper
was written, BASKIN & WILLIAMS (1966) have
discussed
in detail
the effect of the
mongoose
on A. polops. Their
conclusions
do not differ greatly from
those presented above. They quote Mr.
G. A.
SEAMAN
as indicating no Ameiva polops
on St. Croix.
However, recently
Mr.
SEAMAN (personal communication)
tells
he
has
discovered
a small population there. It is
not known whether this is
a relict population
which had been previously overlooked or whether it represents
a re-invasion.
103
COLORATION
The following description of coloration
was made from a live animal immediately
after capturing it
on Green Key, St. Croix, Virgin Islands,
on 17 April
1963
(see
also Fig. 36 and a figure presented by GRANT
1937): the chin, throat, chest, side of
snout and underside of arms are deep pinkish-red;
there is
a slight tinge of pink
beneath
the tail and on the underside
of the legs. The venter is
a light pearl gray.
The pupil is shaped like
a figure-eight, surrounded
by a light gray-brown iris.
The
legs are gray-brown with cream colored spots; the rear one has
a narrow white line
bordered by black
on the outer edge of the femoral region. The top of the
head is
uniform brown; the dorsal pattern consists of a series of longitudinal stripes. There
is a middorsal
one of light brown followed laterally on each side by a series of other
bands
in
the following order:
a wide
dark
brown
or almost black one,
a narrow
brown
one, another black one, a white one, and finally bordering the ventral scales
on each side, a dark brown one mottled with white. The narrow, dorsolateral brown
line extends forward to the posterior corner of the eye, passing over the top of the
tympanum where it becomes confluent with
a circle of the same color surrounding
the tympanum. The tail contains alternate rings of blue and
black. The blue rings
are about
2 scale
rows wide,
the black
ones are narrow and result
from the
dark
edges of two adjacent scale rings.
Variation
of color
among living individuals
consists
of differences
in
the
pro-
minence
and
color of these stripes. For example, the narrow stripe surrounding the
tympanum is sometimes nearly white and sometimes
fades out
near mid-body; at
other times
the stripe next to it laterally is light and the two
appear almost
as
a
single, wide, brown stripe.
One
adult male had
a rather bright blue color on the lateral part of the venter,
covering about
one third of the total area of the venter
on each side. Traces of this
same pattern could
also be detected in preserved specimens.
Fig. 36.
Ameiva polops.
104
HABITS
Like
most ameivas
this species is
diurnal.
On both
Green and Protestant Keys it
could
be
seen foraging actively
among
dead
leaves
on the
ground, frequently
extending its tongue, as though supplementing vision by
smell in detecting food.
This
is
further suggested by
the fact that various
individuals
came to
a piece of
waxed
paper
which had
contained
a ham sandwich
and repeatedly attempted to
eat it even though there were no insects
or other visible food
on it. They were also
observed
eating
ants
from
the ground.
GRANT
(1937) reports that
stomachs
of
Ameiva polops contained amphipods of the type found in beach debris.
VARIATION
Comparison
of the Green and Protestant populations
of Ameiva
polops reveals no significant
differences between them in any of the
characters studied, with the exception of number of preanal scales.
Individuals from Green Key
had a significantly
higher number of
preanal scales than did those from Protestant Key. The ranges did
not even overlap
(Fig. 37). Thus it appears that
some divergence
between these populations
has
taken place.
It
is probably
a
re-
flection of the smallness of both populations
and recent lack of gene
flow from the mainland of St.
Croix. The smallness of these keys
reduces
the
probability
of gene
flow between them via flotsam
transport of individuals.
Fig. 37.
Variation
in
scale counts in Ameiva polops. Symbols as in Fig.
21. N
= 6
(Green Key). N
= 7 (Protestant Key).
105
GREEN KEY: UMMZ 88236 (3 specimens); UPRRP 3677-79.
PROTESTANT KEY: UMMZ 80603-04
(6);
UPRRP 3675.
KEY
TO
THE AMEIVAS OF PUERTO RICO AND
THE VIRGIN
ISLANDS
la.
Eight longitudinal
rows of ventral scales
.
.
Ameiva
wetmorei
lb.
Ten (occasionally twelve) longitudinal rows of ventral scales
.
2
2a.
Dorsal scales of tail arranged in oblique rows
.
Ameiva polops
2b.
Dorsal scales of tail arranged in straight
rows
3
3a.
Dorsal spots prominent
and beginning
on neck
Ameiva alboguttata
3b.
Dorsal spots not prominent
or if so, beginning
on shoulder
.
.
4
4a.
Dorsolateral stripes absent; sides with alternating vertical light
and dark bands
Ameiva desechensis
4b.
Dorsolateral
stripes
present,
or
if
absent,
then
sides
with
irregular marks and spots, not arranged
as alternating bands
Ameiva exsul
DISCUSSION
The
historical
events
leading
to
the
observed
distribution of
organisms
among islands
of
the
Puerto
Rican
shelf
may
have
been (1)
fragmentation of
a
once larger land-mass into a series of
islands,
each containing
a population which then differentiated,
or
(2)
dispersal
by
improbable
means,
such
as
transportation
by
flotsam,
into
an
already
dissected archipelago.
In
either
case,
divergence of insular populations
would be expected,
the degree of
which would depend
on length of time of isolation and/or frequency
with
which interisland transport
of animals occurred.
These two
categories of events are not mutually exclusive and the distribution
and
speciation
of
Ameiva
on
the
Puerto
Rican
shelf
probably
reflects
a combination of the two.
According to HEATWOLE & MACKENZIE (in prep.), Mona, Monito,
Desecheo,
and
St.
Croix have been separated
from
Puerto Rico
longer than any of the other islands
on
the shelf, i.e., they
were
separated
even at maximum sea-level lowering of the Pleistocene.
MITCHELL (1954)
placed separation
sometime near the end of the
Pliocene.
By
contrast,
Culebra,
St.
Thomas,
St.
John,
Tortola,
106
Virgin Gorda and Anegada were all connected with each other and
with Puerto Rico
as late as about 9,000 years ago and Vieques and
Caja de Muertos lost their connections with Puerto Rico sometime
in
the
last
8,000
years
(HEATWOLE
&
MACKENZIE
MS).
Thus,
divergence of A. alboguttata and A. desechensis from the A. exsul-like
stock may reflect the long period during which Mona and Desecheo
have been isolated from Puerto Rico,
and/or
the relatively
great
inter-island distances involved (hence less opportunity for successful
rafting of individuals between islands).
The lack of speciation in the Ameiva exsul group on Puerto Rico
and the Virgin
Islands is probably
a reflection of the short time
islands have been isolated. However,
we believe that gene flow via
sweepstakes
dispersal
of
individuals
is
an
important
factor
in-
fluencing
distribution,
degree
of divergence
and pattern
of intra-
specific
variation. The lines of evidence are the following:
(1)
Waif
dispersal
does
occur
at
least
over
short
distances
as
evidenced
by
the
sudden
appearance
of
individuals
on
Cayo
Ahogado and the presence of
one male on Levin's Rock.
(2)
The form of small islands influences whether or not ameivas
are present.
Many
small keys
with gradually sloping
edges
have
ameivas,
whereas similar sized ones at comparable
distances from
large islands,
but which have vertical sides, undercut at the base
(e.g.
Monito and
various of the northeastern keys)
lack ameivas
even though what appears to be favorable habitat occurs
on top. It
is presumed that inasmuch as random extinction is more likely
on
small islands (small populations)
and the subsequent
re-colonization
difficult
due to its geomorphology
and the poor climbing ability of
ameivas, such islands remain devoidof these lizards most of the time.
(3)
Patterns
of geographic
variation in
some
characters
show
gradual changes
in almost clinal fashion
over
a series
of islands,
suggesting
influence of the
various populations
upon each other.
More importantly,
when relatively
sharp
character gradients
do
occur between populations
on different islands, these breaks occur at
different
places
rather
than coinciding
geographically,
i.e.,
the
various
characters
show
independent
patterns
of
variation,
a
condition suggesting
considerable gene flow among islands rather
than discrete gene pools.
107
(4)
Characteristics
of populations
on
various islands conform to
the pattern expected on the basis of wind and water currents. In the
Puerto Rico-Virgin
Island
area,
winds
are
from the east
a great
proportion
of the
time and the
ocean
current
runs
from
east
to
west (KAYE
1959). Thus, most flotsam transport would be east to
west.
Small islands
contain few individuals and
would therefore
contribute few waifs to be dispersed
to a larger western neighbor,
where the effect of such animals would, in any event, be minimal
because of the large resident population. Conversely,
a large island,
with a higher number of individuals available for transport
and a
greater perimeter
from
which
they
would
be
set
adrift,
might
contribute a significant
number of individuals to
a small island to
its west.
Such animals, inasmuch
as they
would contribute to
a
small gene pool
would have
a relatively
greater effect than in the
reverse
situation.
This effect
is observed
in
that populations
of
Ameiva exsul on small islands nearby, and to the east of, larger ones
tended to be divergent
from those of the larger
one (e.g. caudal
scales
on Cayo
Santiago).
Small islands between two
larger
ones
that differ share the characteristics
of the large island to the east
rather than the one to the west (e.g. Little St. James and Lovango
Key between St. Thomas and St. John). Finally,
where there was
a
group of small islands only, each
one tended to have slightly
dif-
ferent populations,
the direction of divergence seeming not to bear
any
correlation
to
characteristics
of
neighboring
islands
(e.g.
Hicacos, Lobos, Diablo).
It should be pointed out that a special
consideration applies
to
flotsam transport
from Puerto Rico
west to Mona and Desecheo.
The ocean current after passing the southwestern tip of Puerto Rico
swings
north through
the
Mona channel (KAYE
1959). Thus,
the
opportunity
for
flotsam
to
reach
the
western
islands
is
much
reduced and any which would arrive
there would most likely have
its
origin
in
southwest
Puero
Rico
rather
than
other
(closer)
localities. In the past
the configuration
of southwest Puerto Rico
was
such
that
it
would
have
deflected
the
current
toward
the
western islands (HEATWOLE & MACKENZIE). Hence, if the original
colonizers reached
Mona
and
Desecheo
via
flotsam
after
these
islands
were
isolated,
they
probably
diverged
after
sea
levels
108
rose
and
the pattern
of currents in
the Mona
channel changed.
(5)
A supplementary consideration, though not direct evidence, is
that the ecology
and habits of Ameiva exsul
are
such as to make
their being cast adrift highly probable. LEVINS & HEATWOLE (1963)
have pointed out that the probability of a species reaching
an island
by
the sweepstakes
method depends
on its population
density in
already
occupied
areas
from which it might
be
transported,
its
habits (likelihood of being foundin debris which could be washed out
to sea),
the
time required
for the trip
(in
turn depending
upon
strength
and direction of wind and water currents),
and tolerances
to environmental factors encountered in transit (salt water or spray,
high temperature,
desiccation, etc.). Not all of these factors
can at
present be evaluated for Ameiva exsul. However, it should be noted
that their habitat includes beach
areas where they
occur in dense
populations.
Thus, in close proximity
to the
sea,
there
are large
numbers available for transport. Also, any individuals successfully
reaching a new island will immediately encounter a favorablehabitat
andwill not need to journeyinland before finding suitable conditions.
Ameiva exsul spends its inactive periods in burrows in the soil,
beneath stones and in or under logs or other debris,
a habit which
would
increase
their
likelihood
of
being
cast
adrift.
Thus,
the
population
densities and habits of Ameiva exsul
are such
as would
favor waif
dispersal and subsequent
establishment
upon reaching
an island. These
same factors also favor gene flow between islands
and re-colonization of an island following local extinction.
Tolerance limits are not known. However, the fact that the usual
habitat is relatively hot and dry and in many instances subjected to
salt
spray,
would suggest
that tolerances to temperature,
desic-
cation and salinity might be high.
The history of the genus Ameiva in the Puerto Rican area can be
summarized in the following way:
(1)
The first group in the
area was probably
the ancestral stock
of A.
polops.
Present day distribution of members of the
group
probably represent relicts of a formerly more widespread
group.
(2) Following,
or
perhaps
contributing
to,
the
decline
of
this
group, two additional groups of the genus reached the Puerto Rican
109
area. One of these consisted of the eastward penetration of another
member of the A. undulata group, Ameiva wetmorei. The A. ameiva
group reached as far as Santo Domingo by "island-hopping"
up the
Lesser Antillean chain from South America. 1 On what was probably
at that time a continuous land mass containing Puerto Rico and all
of the Virgin Islands except St. Croix, this stock was represented by
an exsul-like form which then diverged into A. exsul on Puerto Rico,
A. alboguttata
on Mona and A.
desechensis
on Desecheo. Whether
this
ancestral
form reached Mona and
Desecheo before
or
after
these
islands
became
separated
is
problematical.
Geographic
patterns of variation within A. exsul are now probably influenced by
gene flow via flotsam transport of individuals from island to island.
LITERATURE
BARBOUR,
T.,
1914. A contribution
to the zoogeography of the West Indies, with
especial reference
to amphibians and reptiles. Mem. Mus.
Comp. Zool.
Harvard
44 :
209-360.
BARBOUR,
T.,
1930.
Some
faunistic
changes in the
Lesser
Antilles.
Proc.
New
England Zool.
Club n:
73-83.
BARBOUR,
T.,
1937.
Third
list
of Antillean reptiles
and amphibians. Bull.
Mus.
Comp. Zool. 82: 77-166.
BARBOUR, T. & NOBLE, G. K.,
1915. A revision
of the lizards
of the
genus Ameiva.
Bull.
Mus. Comp. Zool. 59:
417-479.
BASKIN,
J.
N.
& WILLIAMS, E. E.,
1966.
The
Lesser
Antillean
Ameiva
(Sauria,
Teiidae). Re-evaluation,
zoogeography and
the effects
of predation.
Studies
Fauna Curasao and other Carib. Isls. 23:
144-176.
BEEBE, W.,
1919. Lizards of the genus Ameiva in Bartica District. Notes
on their
color and pattern variation. Zoologica 2: 235-238.
BEEBE,
W.,
1945.
Field
notes
on the
lizards
of Kartabo, British
Guiana,
and
Caripito, Venezuela. Part 3. Teiidae, Amphisbaenidae and Scincidae. Zoologica30:
7-32.
CARPENTER,
C. C.,
1961.
Temperature relationships of two Oklahoma lizards. Proc.
Okla. Acad. Sci. 41: 72-77.
COCHRAN, D. M.,
1941. The herpetology of Hispaniola. Bull. U. S. Nat. Mus. iyy.
1-397.
1
See BASKIN &
WILLIAMS (1966) for
a
more detailed discussion
of this
dispersal.
110
COPE, E. D.,
1862. Synopsis
of the species of Holcosus
and Ameiva, with diagnoses
of new West Indian
and South American Colubridae.
Proc. Acad. Nat. Sci. Phila.
1862 : 60-82.
DANFORTH,
S. T.,
1925. Porto Rican herpetological notes. Copeia 147 : 76-79.
DICE, L. R. & LERAAS, H. J.,
1936. A graphic method for comparing several
sets of
measurements. Contrib. Lab.
Vert.
Genetics Univ.
Mich. 3:
1-3.
FOWLER,
H.
W.,
1918.
Some amphibians and reptiles from Porto
Rico
and the
Virgin Islands. Publ. Carnegie Inst.
Wash.
252:
1-15.
GARMAN, S.,
1887. On West Indian Teiidae in the Museum of Comparative Zoology.
Bull. Essex Inst, ig :
1-24.
GRANT, C.,
1931a. Notes
on the ameivas of Puerto
Rico. Copeia 1931 :
62.
GRANT, C.,
1931b. A revised list of the herpetological fauna
of Culebra
Island. J.
Dept. Agric. P. R.
15: 215.
GRANT, C.,
1932a. The large ameivas of the Puerto Rico region and one new species.
J. Dept. Agric. P. R.
16 :
157-160.
GRANT,
C.,
1932b.
The herpetology of St.
John
and
adjacent keys,
U.S. Virgin
Islands. J. Dept. Agric. P. R. i6 \ 331-334.
GRANT, C.,
1932c. Herpetology
of Tortola; notes
on Anegada and Virgin Gorda,
British Virgin Islands. J. Dept. Agric. P. R. 16: 339-346.
GRANT, C.,
1932d. Herpetological notes from the Puerto Rico
area. J. Dept. Agric.
P. R. 16 :
161-165.
GRANT, C.,
1932e. The herpetology of Vieques Island. /. Dept. Agric.P. R. 16 : 37-39.
GRANT,
C.,
1937. Herpetological notes with
new
species from the
American
and
British Virgin Islands,
1936. J. Agric. Univ.
P. R. 21: 503-522.
GRANT, C. & ROOSEVELT, C.,
1932. The herpetology of Caja de Muertos Island
and
Cardona Key, Porto Rico. J. Dept. Agric. P. R.
16 : 47-49.
HEATWOLE, H. & SADE, D.
S. & HILDRETH, R.,
1963. Herpetogeography of Puerto
Rico. I. Herpetofauna of Cayo Santiago and Cayo Batata.
Carib. ]. Sci. 3 :
1-5.
HEATWOLE,
H.
&
MACKENZIE,
F.,
Herpetogeography of
Puerto
Rico.
IV.
(in
preparation).
HIRTH, H. F.,
1963. The ecology of two lizards on a tropical beach. Ecol. Monographs
33:
83-112.
KAYE, C.,
1959. Shoreline
features and Quaternary shoreline
changes Puerto Rico.
Geol. Surv. Professional Pap. 317-B: 49-140. U.
S. Govt. Print. Off., Wash., D. C.
LEVINS,
R.
& HEATWOLE, H.,
1963. On
the distribution of organisms on islands.
Carib. J. Sci. 3 :
173-177.
MAYR, E.,
1963. Animal
Species and Evolution. Harvard
Univ. Press, Cambridge,
Mass. 797 pp.
MITCHELL, R. C.,
1954. A
survey of the geology of Porto Rico. Agric. Exp. Sta. Univ.
P. R. Techn. Pap.
13,
167 pp.
NOBLE, G. K. & BRADLEY, H. T.,
1933. The matingbehavior
of lizards; its bearing
on the theory of sexual selection.
Ann. N. Y. Acad. Sci. 35:
25-100.
111
PIC6, R.,
1954. Geografia de Puerto Rico. Editorial Universitaria, Rio Piedras, P. R.,
243 pp.
ROBERTS,
R.
C.,
1942.
Soil
Survey of Puerto
Rico.
U.
S. Dep.
Agr.
Bur.
Plant
Industry, Wash., 503 pp., maps.
ROLLE, F. J. & HEATWOLE, H. & LEVINS, R. & TORRES, F.,
1964. Faunal
notes on
Monito Island, Puerto Rico.
Carib.
J. Sci.
4 \ 321-322.
SCHMIDT,
K. P.,
1920. Contribution to the herpetology of Porto Rico. Ann. N.
Y.
Acad.
Sci.
28:
167—200.
SCHMIDT, K. P.,
1928. Amphibians and land reptiles of Porto Rico, with
a list of those
reported from the Virgin Islands.
Scient. Surv.
P. R. & Virg. Isl. 10,
160 pp.
SEAMAN, G. A.,
1961. Mammals, reptiles
&■amphibians of the Virgin Islands. Brod-
hurst's Printery,
Christiansted, St. Croix,
10 pp.
STAHL, A.,
1883 .Fauna
de Puerto-Rico. Clasificacion
sistematica
de los animates que
corresponden a
esta fauna y catdlogo
del gabinete zooldgico. Imprenta del
Boletin
Mercantil, San Juan, 249 pp.
STEJNEGER, L.,
1904. The herpetology of Porto Rico. Report U.
S. Nat. Mus. igo2\
553-724.
UNDERWOOD,
G.,
1962. Reptiles
of
the
eastern
Caribbean.
Carib.
Affairs
(n.s.)
1,
192 pp.
WOLCOTT, G. N.,
1923. The food of the Porto Rican lizards. J. Dept. Agric. P.R.J'.
5-37.