ECOLOGICAL SEPARATION OF ANOLIS LIZARDS
IN A COSTA RICAN RAIN FOREST
MICHAEL JON CORN
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
ECOLOGICAL SEPARATION OF ANOLIS LIZARDS
IN A COSTA RICAN RAIN FOREST
MICHAEL JON CORN
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
I would like to thank the people of Rio Frio, Costa
Rica, for their friendship and hospitality, especially the
McGinnis family and their employees at the Hotel Cabana.
I also appreciate the cooperation of Mr. Ed Pattimore of the
Standard Fruit Company and the hospitality of his family.
Mr. Jack D. DeMent, vice president of Castle & Cooke Foods,
provided weather data for their Costa Rican sites.
Support for the field research was provided in part by
a National Science Foundation Traineeship, by collection
funds from the Florida State Museum, and by the Organization
for Tropical Studies (O.T.S.). I am grateful to these organi
zations, but particularly to the Costa Rican office of O.T.S.
Sr. Jorge Campabadal and his staff were particularly helpful.
My introduction to tropical reptiles and their fasci
nating ecological relationships was in an O.T.S. ecology
course in the summer of 1967. Drs. Tom Emmel and Roy McDiar-
mid provided an especially sound beginning to my studies.
In addition I also benefitted from many discussion with
two other O.T.S. faculty members, Dr. Jay Savage and Norm
I would like to thank my many friends and colleagues
at the College of Lake County: the Sabbatical Leave
Committee for allowing me to return to Gainesville to finish
the prey analysis and begin writing; my friends in the Se
curity Office for providing me a quiet, secluded work area
and for some "friendly nagging"; the A-V department, espec
ially Mr. Bill Kniest, who photographed the figures; Mr.
Dan Ziembo, for the excellent background drawing in figures
3-1 and 5-26; Mrs. Lynn Floor, for typing the final draft;
and my fellow biology faculty members, for their continued
support and encouragement.
I am greatly appreciative of Drs. Archie Carr, Thomas
C. Emmel, Brian McNab and Hugh Popenoe, my graduate commit
tee. Through the many years since I started this project,
they have been supportive and cooperative beyond any reason
able expectation. Drs. Carr and Emmel have been especially
helpful in straightening out, or sometimes bending,
Finally my most important acknowledgment is to my
wife, Mary, and my children, Michael Jr., and Melanie.
Mary has helped in every stage of this project from the very
beginning to the final typing. They have all put up with
a seemingly endless, time-consuming project called "Daddy's
Dissertation" instead of doing many other activities.
I am most grateful for their love and support.
TABLE OF CONTENTS
I GENERAL INTRODUCTION 1
II DESCRIPTION OF THE STUDY AREA 3
III DESCRIPTION OF THE ANOLES 15
IVBODY SIZE 2 3
1 Introduction 23
2 Materials and Methods 25
3 Sexual Dimorphism 27
4 Interspecies Comparisons 30
5 Body Weight to Length Relationships .... 36
6 Seasonal Variation in Size 38
V FOOD 96
1 Introduction 96
2 Materials and Methods 97
3 Indices of Resource Use 99
4 Intraspecies Comparisons by Size 104
5 Intraspecies Comparisons by Sex 115
6 Intraspecies Comparisons by Season 118
7 Interspecies Comparisons 122
8 PreyUtilization by All Anoles 129
VI REPRODUCTION 245
1 Introduction 245
2 Materials and Methods 247
3 Results for Anolis humilis 250
4 Results for Anolis 1 iwi f rons 254
5 Results for Other Anolis 257
6 Female Reproductive Cycles 263
7 Interspecies Differences in Egg
Production Rate 268
8 Male Reproductive Cycles 269
9 Fat Cycles and Reproduction 275
VII POPULATION SIZE AND STRUCTURE 332
1 Introduction 332
2 Materials and Methods 333
3 Density and Biomass of Anoles 335
4 Population Structure of Anoles 341
5 Density, Biomass and Population Structure
of Other Lizards 346
6 Density of Frogs 349
VIII SUMMARY AND CONCLUSIONS 376
1 Definitions of Morphological Counts and
2 Specimens Used 385
LITERATURE CITED 397
BIOGRAPHICAL SKETCH 4 03
Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
ECOLOGICAL SEPARATION OF ANOLIS LIZARDS
IN A COSTA RICAN RAIN FOREST
Michael Jon Corn
Chairman: Archie Carr
Major Department: Zoology
Ecological separation in eight sympatric congeneric
lizard species in the genus Anolis was studied at Rio Frio,
a site in a lowland tropical rain forest on Costa Rica's
Caribbean slope. Four aspects of the ecology of the Rio
Fro anoles are considered: (1) body size, (2) food, (3)
reproduction, and (4) population size and structure.
Size and prey patterns have been thoroughly described
for island anoles by other authors, but the Rio Frio anole
community contains more species in a single habitat than do
the island communities. In many respects, these lizards
show size-prey patterns similar to island anoles, e.g.
ecological separation is achieved primarily through spatial
differences, both vertical and horizontal, and size differ
ence. But in contrast, differences in mainland anoles are
not as clear-cut, large male size is found in only two
species, body size differences are reduced, and prey size
differences are generally less than would be found within
or between island species.
Anoles that live without congeners on islands tend
toward certain size patterns. At Rio Frio, it was found
that anoles which are restricted to peripheral habitats,
e.g. streamside or exposed canopy, are little influenced by
congeners, and show the same size, sexual dimorphism and
prey size patterns as do solitary island anoles.
Anoles collected in May, just after the dry weason,
weigh less at any length than those collected at other
times. That this is a result of dry season food stress,
due to reduced populations of small arthropods, is further
corroborated by the finding that lizards collected in May
frequently had not eaten.
Reproduction in Ro Fro Anolis is continuous throughout
the year. There is some variation in some species, apparently
in response to environmental change, but in none is there a
complete cessation of reproduction. Young are present in
Density of low bush and ground amphibians and reptiles
at Ro Fro was estimated from fenced quadrats. For all
species collected, density is 3752-5378 individuals/ha; for
Anolis alone, the estimate is 688-1252 anoles/ha. Anoles
had a combined estimated biomass of 422-1026 g/ha.
I. GENERAL INTRODUCTION
This study concerns the ecology of the lizards of the
genus Anolis at a site in the lowland rain forest on Costa
Rica's Caribbean slope. Anoles are generally small, arboreal,
insectivorous lizards that make up a large, important part of
the fauna of the American tropics. In some areas, usually near
the geographic or climatic limits of the genus Anolis, only a
single species may be found, but in the rain forests many
species co-exist. Ro Fro, the site of this study, was
chosen because at least eight species of anoles live in close
contact, apparently dividing their resourcesspace, food,
time, etc.in a way that minimizes, without altogether
avoiding, the evolutionary friction of competition.
Rio Frio is the Standard Fruit Company's name for a very
large banana plantation. When I first visited there, most of
the area was rain forest, slightly disturbed and with some
second growth, but basically mature forest. The banana
company was in the process of cutting the forest and replacing
it with bananas. Most of the lizards collected for this
study were taken as the trees were felled; a few were collected
in still undisturbed areas. When I last visited Rio Frio in
May 1970, about two-thirds of the area had been cleared of
forest, and banana production was in full operation. At the
time of this writing, early 1979, the forest and most of its
animal inhabitants are gone.
So this study is about the ecological separation of the
anoles in a rain forest that used to be. Its conclusions
do apply, I believe, to the increasingly diminishing patches
of Costa Rican rain forest that remain.
Four aspects of the ecology of the Rio Frio anoles are
considered here: (1) body size, (2) food, (3) reproduction
and (4) population size and structure. Because of the dis
tinctness of the sections, each will have its own, more
II. DESCRIPTION OF THE STUDY AREA
Rio Frio is located at 10 20' N, 83 53' W in Heredia
Province, Costa Rica (Fig. 2-1). It is approximately 18 km
SE of Puerto Viejo and 16 km NW of Guapiles, and has an
elevation of approximately 100 M. All collections were made
on land belonging to the Standard Fruit Company, now a large
banana plantation. Most of my specimens were collected in the
area between the Rio Sucio and the Rio Chirripo. Population
quadrat collections were made on the NW side of the Rio Sucio.
Rio Frio is on a relatively flat alluvial plain below the
Cordillera Central. The area is shown as Premontane Wet
Forest, basal belt transition, on the Holdridge Life Zone map
of Tosi (1969), but rainfall data for 1961-65 (Table 2-1,
Figure 2-2) show more than 4000 mm of precipitation. This
would indicate a Holdridge Life Zone classification of Tropical
Wet Forest. The vegetation appeared to fit the description
(Holdridge et al., 1971) of Tropical Wet Forest more closely
than Premontane Wet Forest (Table 2-2). For a more detailed
description of a very similar forest, see the description of
Finca La Selva (Holdridge et al., 1971) which is located
about 10 km from Rio Frio.
While collections were being made at Rio Frio, the forest
was being replaced by bananas in the following sequence:
1. Selected trees were harvested for lumber. This
involved relatively few trees and caused no
appreciable difference in the structure of the
2. Some time after this lumbering (as long as many
months in some areas), bananas were planted in the
3. One to a few days after the bananas were planted,
all trees were felled and left to rot.
4. Bananas grew up among the rotting trees, and second
growth was cut back.
Most of my specimens were collected as the trees were
felled in step 3 above. Population quadrat collections were
made in an area that had been lumbered (step 1 above) many
months previously, but had not been otherwise disturbed. No
attempt was made to census or collect anoles in areas after
the forest was replaced by bananas.
It is generally accepted that wet tropical forests are
wet the year around, at least from a temperate point of view.
La Selva/Los Diamantes receives as much rain during the three
driest months, January through March, as Illinois does in an
entire year. Nevertheless, Rio Frio does have seasonality of
rainfall (Fig. 2-2). The wet season extends from May through
December, and the dry season from January through April. The
wet season has a characteristic rainfall depression, called
locally the veranillo, in late August and early September.
Even during this period, the forest usually receives regular
rainfall, and the forest floor is continuously damp. During
the dry season, however, there may be several days or even a
week or more with very little or no rain. Occasionally, the
dry season is interrupted by a period of rain, such as occurred
during February 1970. At Rio Frio, 495 mm of rain fell during
the 13 days from 5 February through 17 February, with only one
of these days receiving no rain. In contrast, the other 15
days of February received only 24 mm of rain. Because rainfall
data for Rio Frio were not kept prior to December 1969, the
average data from Puerto Viejo (18 km NW) and Los Diamantes
(15 km SE) are used.
Day length at the latitude of Rio Frio (Fig. 2-2) ranges
from 12 hours, 42 minutes at the end of June to 11 hours, 32
minutes in mid-December, a total variation of only 1 hour,
10 minutes (List. 1966).
Seasonal temperature variation is shown in Figure 2-3
and Table 2-3. Again, Rio Frio data are available only after
December, 1969. For a typical annual pattern, the Puerto
Viejo/Los Diamantes average is shown.
Anoles were collected during four periods (Fig. 2-2):
(1) September and early October, a dry period (veranillo)
during the wet season; (2) November and early December, the
very wet, last part of the wet season; (3) February and early
March, the driest part of the dry season; and (4) early May,
the last of the dry season or beginning of the wet season.
These four periods will be designated hereinafter simply as
September, November, February and May respectively. A few
anoles of some of the less abundant species were collected
during the preceding August and are used in some sections.
The fauna of lowland tropical forests is distinguished
for its; diversity, and the forest at Rio Frio was no exception.
Many other animals interact with the anoles as predators,
prey, or competitors. Predators include snakes, birds, small
mammals and occasionally larger lizards. Prey items include
most of the huge number of insect species, spiders, snails,
smaller lizards or frogs, and almost any other moving inverte
brate of appropriate size. Chief competitors, for food at
least, are probably birds and frogs, and possibly spiders;
and not other lizards. Most other lizards at Rio Frio are
either nocturnal or herbivorous, or are found primarily in
second growth (Table 2-4). Thus, the most prominent lizard
competitors of anoles are other anoles.
AND LA SELVA
NOTE: Average monthly rainfall (mm) for the years 1961-65 at
Puerto Viejo^and Los Diamantes, the two weather stations
closest to Rio Frio.
Comparison of Tropical Premontane Wet Forest
and Tropical Wet Forest Vegetation
Tropical Wet Forest
General. Tall, multistratal evergreen forest. A few canopy
species are briefly deciduous, usually when flowering, but this
does not affect the evergreen aspect of the forest as a whole.
Number of tree species very large.
Upper canopy. Trees 45-55 m tall, with occasional larger
emergents; crowns round to umbrella-shaped, usually not in
lateral contact with each other. Clear holes up to 30 m long
and 125-200 cm in diameter. Smooth, thin, light-colored bark
and high buttresses very common, but trees lacking buttresses
or with dark, rough, flaking or lissured bark also occur.
Leaves elliptical, usually lacking lobes or teeth, often glossy,
mostly less than 10 cm long.
Lower canopy. Trees 30-40 m tall, filling spaces between upper
canopy trees. Crowns round, trunks mostly slender, lacking
large buttresses. Bark dark or light, mostly smooth.
Understory. Trees 10-25 m tall, dense, commonly with narrow
conical crowns and slender stems, often leaning, twisted or
crooked, usually with smooth dark bark; flowers and fruits
sometimes produced directly from the trunk and lower branches
(cauliflory). Leaves or leaflets relatively large (up to 20
cm long), elliptical, long, pointed tips ("drip tips"). Stilt-
rooted palms often abundant.
Shrub layer. Dwarf palms, often with undivided leaves, usually
abundant, mostly 1.5-2.5 m tall. Miniature trees with unbranched
main stems present but usually less common. Giant herbs with
banana-like leaves often prevalent, especially in clearings
and disturbed places.
Ground layer. Often bare, or with a few ferns, Selaginellas,
or tree seedlings.
Epiphytes. Including orchids, bromeliads, and large-leafed
herbaceous climbers (aroids, Cyclanthaccae) common but often
not conspicuous; moss layer on tree trunks very thin or lacking.
Large bush ropes uncommon; epiphytic shrubs and strangling
trees mostly rare.
Tropical Wet Forest is distinguished from Premontane Wet Forest
by (1) larger buttresses of canopy trees; (2) larger trunk
volumes, especially in canopy trees; (3) bark generally smoother;
(4) palms much more common in understory and shrub strata;
(5) tree ferns less common; (6) more tree species per 0.1 ha;
(7) physiognomic differences between Tropical Wet and Premontane
Wet communities are subtle, but there are consistent differences
in floristic composition useful to the specialist; (8) more
luxuriant appearance than forests at higher elevations, which
feel cooler, darker, and more gloomy. (These subjective im
pressions, of course, may not be shared by some observers.)
TABLE 2-2 extended
Tropical Premontane Wet Forest
General. Tall to intermediate semi-evergreen forest with two
or three tree strata; emergent, canopy, and subcanopy not al
ways readily distinguished. In exceptionally dry years or on
edaphically dry sites, most canopy species may drop their leaves
for a few weeks. A few species are always dry-season decidu
ous. The subcanopy, small trees, and shrub layers are evergreen.
Canopy. Trees mostly 30-40 m tall (with occasional emergents
up to 50 or 55 m), mostly with round to spreading crowns and
slender to stout trunks with clear lengths of 25 m or less.
Buttresses are common but much smaller than in Tropical Moist
and Wet forests, except in warm transitional areas. Bark mostly
brown or gray, moderately thick, flaking or lissured, but smooth
and light-colored in a few species. Leaves are mostly simple,
elliptic, glossy, 5-10 cm long, with entire or minutely toothed
margins, often somewhat crowded at the tips of the branches.
Number of species large.
Small tree stratum. Composed of a dense layer 10-20 m tall of
sapling canopy trees mixed with small, slender-trunked trees
having relatively deep crowns and smooth, often dark bark. Stilt
roots and long, strapshaped leaves are common. Tree ferns are
occasional; stilt-rooted palms occur in warm transitional areas.
Shrub layer. A dense stratum of single-stemmed "miniature"
trees and young canopy trees 2-3 m tall. Small palms are
generally rare or lacking, but are abundant in warm transition
Ground layer. Generally bare except for ferns, which may be
abundant (especially along trails), and tree seedlings.
Epiphytes. Orchids and bromeliads mostly not conspicuous;
herbaceous vines are common, mostly climbing the larger trees.
A thick layer of moss covers the tree trunks.
Premontane Wet Forest is di-stinguished from Tropical Wet Forest
by (1) much lower frequency of palms in small tree and shrub
strata; (2) higher frequency of tree ferns; (3) smaller but
tresses in canopy trees, high buttresses rare; (4) bark gener
ally somewhat rougher; (5) trunk and crown volume of canopy
trees somewhat smaller; (6) fewer tree species per 0.1 ha.
SOURCE: Holdridge et al., 1971.
TABLE 2-3. Average Daily Temperature
AND LOS DIAMANTES MEAN
NOTE: Average daily maximum and minimum temperatures (C) for 1966 at Puerto Viej
and'Los Diamantes, the two weather stations closest to Rio Frio.
TABLE 2-4. Other Lizards at Rio Frio
nocturnal (?), arboreal
nocturnal (?), terrestrial
(Anolis 8 species)
diurnal, arboreal, riparian
diurnal, low arboreal
diurnal, arboreal, riparian
diurnal, high arboreal
FIGURE 2-1. Map of^Costa Rica. Shown are the location of
Rio Frio (circle), a Standard Fruit Company
banana plantation, and the two nearest towns
that are on most maps: Puerto Viejo (diamond)
and Guapiles (triangle).
MONTHLY RAINFALL (mm)
FIGURE 2-2. Rainfall and Day Length for Rio Frio. Rainfall (solid line) given as means
of the total monthly rainfall for the years 1961-1965. Values are the
averages of those of Puerto Viejo and Los Diamantes. Monthly total rainfall
at Rio Frio, from December 1969 to June 1970, is indicated by triangles.
Day length (dotted line) values are from the tables for 10 20' N (Rio Frio)
of List (1966). Shaded areas indicate collecting periods.
HOURS OF DAYLIGHT
AVERAGE DAILY TEMPERATURE
FIGURE 2-3. Average Temperatures for Rio Frio. Average daily maximum (solid line) and
minimum (dashed lined) temperature ( C). Values are the averages of those
of Puerto Viejo and Los Diamantes. Monthly averages of Rio Frio (December
1969 to June 1970) are indicated by dots (maxima) and circles (minima).
III. DESCRIPTION OF THE ANOLES
Part of the rationale for studying ecological inter
actions among sympatric congeners is that, being very close
in an evolutionary sense, they should show intense and, there
fore, important competition. In this study, ecological
relationships of eight species of the large iguanid genus
Anolis are discussed. These eight species, although diverse
in their ecological niches, are very close phyletically.
The anoles are a group of over 200 iguanid species, the
distribution of which centers in the American tropics and
extends into the subtropics. All eight species of anoles
present at Rio Frio belong to what Etheridge (1960) and others
have termed the Beta Section, as species derived from Central
American ancestry are designated (in contrast to the Alpha or
South American derived species). Three species groups, desig
nated as series by Etheridge (1960), are represented: petersi
series (A. biporcatus, A. capito and A. pentaprion), fuscoauratus
series (A. carpenteri, A. limifrons and A. lionotus) and
chrysolepis series (A. humilis and A lemurinus). Etheridge
(1960) suggested that the petersi series of large lizards is
a relatively primitive group, and that the fuscoauratus series
and chrysolepis series are its more advanced descendants.
The size, shape, color and gross habitat distribution are
ecologically important attributes of anoles. Some of these
will be discussed more thoroughly in later sections, but in
order to help the reader visualize these characteristics in
the eight species discussed here, a brief summary is given
Anolis humilis is the smallest and most numerous anole
at Rio Frio. Maximum snout-vent length (SVL) is only 40 mm
in my samples, although Fitch (1975) found them up to 43 mm,
and maximum weight is 1.7 g. A. humilis is a stout-bodied
anole of more or less uniform brown dorsal color; females
occasionally have a light stripe or a rhomboid pattern on the
back. Males have a relatively large dewlap, red bordered
with yellow, while females lack a dewlap, but may have a
small patch of red on the throat. A. humilis is usually found
close to the ground; Fitch (1975) found 92% on the ground or
below 20 cm. They rarely go above 50 cm on a tree trunk, even
Anolis lemurinus, the second member of the chry solepis
series, is considerably larger than A. humilis. My largest
specimen has a 60 mm SVL, and the heaviest is 5.06 g. Taylor
(1956) describes a specimen with a 63.5 mm SVL. Also a stout
bodied anole, A. lemurinus is dark gray in color usually with
a blotched dorsal pattern; females may have a light band or a
series of diamonds on the dorsal midline. Males have a rela
tively small dewlap, deep red in color, lighter at the edge.
The female dewlap is very small, but colored like that of the
male. A. lemurinus is found on the trunks of large trees,
usually on the lower part of the trunk, but when disturbed
will flee quite high.
Anolis limifrons is a slender, gray anole that is second
only to a. humilis in abundance at Rio Frio. Slightly longer
than A. humilis, A. limifrons has a maximum SVL of 42 mm in
the Rio Frio collection; Fitch (1975) lists 45 mm as the
maximum. Maximum weight in the Rio Frio sample was 1.5 g;
Fitch (1975) gives 1.7 g. Males are more or less uniform light
gray, lighter on the venter having a small white dewlap with
an orange spot in its center. Females are colored like the
males, or with a light dorsal band or diamond pattern. Females
lack a dewlap. a. limifrons is a shrub-level anole usually
found on vertical trunks of shrubs, saplings or larger trees,
almost always between 0.5 and 2 m above the ground. Fitch (1975)
found 87.3% of this species perching above ground level, with
a mean perch height of 0.78 m. More than any other Rio Frio
anole, A. limifrons is found commonly in second growth and
disturbed areas, but is still more abundant in less-disturbed
Anolis carpenteri is closely related to A. limifrons, and
is virtually the same length. The longest Rio Frio specimen
is 43.5 mm SVL; Fitch (1975) records it up to 45 mm SVL. It
is slightly more slender than A. limifrons, with the heaviest
A. carpenteri weighing only 1.2 g. A. carpenteri, which was
not described until 1971 (Echelle, Echelle and Fitch), is a
very cryptic, lichen-green anole with a light half-ring below
the eye and a white venter. Males have a large bright orange
dewlap; females lack a dewlap but may have a trace of orange
pigment on the throat. The habitat of A. carpenteri is
problematic. Fitch and his co-workers (Echelle, Echelle and
Fitch, 1971; Fitch, 1975) found their first ten specimens on
lichen-covered rocks at ground level in the Tropical Moist
Forest, premontane transition Life Zone. They then obtained
14 more lizards in 10 months at Finca La Selva in the Premon
tane Wet Forest, basal belt transition Life Zone. Of these
14, only 5 were from undisturbed rain forest; the rest were
from cacao or palm groves. Five of these lizards were found
at ground level, three on tree bases and two in leaf litter.
Fitch (1975) states that all were "on or near lichen-mantled
tree trunk or log which provided concealing background," and
that three of fourteen recorded "dropped to the ground almost
immediately after being flushed." When I first visited Rio
Frio in August 1969, I found A. carpenteri to be very common
among the just-felled trees. In fact, A. carpenteri was so
common that I collected several, decided it was a common
species that I didn't recognize, and released them all. Only
after a later conversation with N. J. Scott, did I realize
that the species was probably then undescribed. A. carpenteri
was also quite common among the cut trees in September and
October, but less common in November, December, February and
March. None were seen in May. Only one specimen, a 26 mm SVL
of was collected outside the area of newly felled trees. This
individual was collected from a small sapling after it had
been flushed out from a liana about 2.5 m above the ground.
I would suggest that the apparent rareness of A. carpenteri
reflects its preference for lichen-covered perches, probably
at moderate heights. The few specimens found on the ground
(with the exception of the type series) are there because of
a "drop-and-freeze" escape behavior. While not as common as
A. limifrons or A. humilis, it is like several other rain
forest reptiles (for example, see Corn, 1974), only encountered
when disturbed from its high arboreal habitat.
A. lionotus is the largest anole of the Rio Frio
fuscoauratus series, and the most restricted in habitat. It
is never found more than a few meters from the edge of a river
or stream. These lizards perch on the bare bank, rocks,
debris and occasionally on vegetation. When disturbed, they
hide under debris, or readily enter the water, where they
swim and hide under submerged material. The largest Rio Frio
A. lionotus is 71 mm SVL, weighing 6.6 g. Fitch (1975) found
much larger specimens, up to 85 mm SVL and 13.7 g. The color
of both sexes is chocolate-brown with cream-colored lateral
stripes and a lighter venter. Males have a large, pale orange
dewlap, which females lack.
Anolis capito is a rare, large anole. The largest Rio
Frio specimen was 87 mm SVL, weighing 13.0 g. Both Fitch
(1975) and Andrews (1971a) found them as long as 95 mm SVL.
The color of A. capito is a mottled olive-green to brown.
Females occasionally have a distinct brown mid-dorsal stripe.
The male dewlap is very small and greenish-white; an even
smaller, similarly colored dewlap occurs on females.
A. cap to is found on the ground or on tree trunk bases or
buttresses, usually lower than 1 m. Escape behavior involves
freezing, or, occasionally, running across the ground or onto
a tree trunk, never higher than two meters or so. a. cap to
was probably much more common at Rio Frio than my collecting
suggests, but its escape behavior made it very difficult to
find in the newly felled trees.
Anolis biporcatus is the largest Rio Frio anole, reaching
93.5 mm SVL, and weighing up to 20.0 g. Taylor (1956) states
that the maximum SVL is 102 mm. Dorsal color can be changed
from bright green through medium brown in both sexes. The
male dewlap is large, with a bright blue center and an orange-
red outer portion. The dewlap of the female is smaller.
A. biporcatus is a tree anole, found from lower sections of
trunks to high in the crown.
Anolis pentaprion is a moderately large, short-legged
anole. My longest specimen is 66.5 mm SVL, and the heaviest
4.3 g; Taylor (1956) and Fitch (1975) examined a specimen of
75 mm SVL. Dorsal coloration is light gray sometimes mottled
with dark patches. The dewlap is large and deep purplish-red
in color. The dewlap of the female is only slightly smaller
than that of the male. The tail is relatively short and
somewhat prehensile. A. pentaprion is the only truly open-
habitat heliotherm among the eight Rio Frio anoles. Most
specimens were found on exposed tree trunks in open country
or high in the crowns of forest trees. All of mine were taken
from just-felled trees. In forests A. pentaprion probably
spends most of its time high in the leaves and twigs, where
it can bask, or at least take advantage of the higher tempera
tures of the exposed tree crown. Campbell (1971) obtained a
mean of 33 C for 12 records of A. pentaprion in a thermal
gradient. The short legs and prehensile tail are undoubtedly
adaptations for movement on small twigs and branches.
From the preceding descriptions, it can be seen that the
eight anole species at Rio Frio are well distributed throughout
the habitat, with some important separation, both vertically
and horizontally (Fig. 3-1). Anolis capito, A. humilis and
A. lionotus form a group of highly terrestrial species, with
A. humilis and A. capito restricted to forest floor and tree
bases, and A. lionotus almost completely separated from all
other anoles in its riparian habitat. a. biporcatus and
A. lemurinus are usually found on tree trunks of moderate to
large size, and only rarely elsewhere. A. limifrons occurs
on a rather broad spectrum of perches, from ground to moderate
ly high trunks, but is primarily found in the shrub layer.
A. pentaprion and A. carpenteri occupy the highest habitats,
with A pentaprion preferring the sunny crown and A. carpenteri
restricted to dark, lichen-covered trunks, branches and lianas.
Perch Sites of Rio Frio Anolis
IV. BODY SIZE
The size of animals has always been of theoretical
interest to zoologists, and Hutchinson's (1959) classic
observations on size ratios of closely related species
pointed the way to an interesting and fruitful area of
investigation for ecologists. How do the sizes of sympatric,
closely related species increase or decrease their ability
to co-exist? From Hutchinson's (1959) own ratio suggestion,
several authors have proposed various models, and many others
have compared sizes within a variety of taxa (See Schoener,
1974, for an extensive review.). The sizes of anoles in
simple and more complex West Indian faunas have been the
subject of numerous studies (Schoener 1967, 1968, 1969a,
1969b, 1969c, 1970; Schoener and Gorman, 1978; Williams,
1972; and others). Only Fitch (1976) and Duellman (1978)
have discussed size in sympatric mainland species.
Two particularly interesting generalizations have come
from these studies. From the work of Schoener, have come
what Williams (1972)called Schoener Rules:
An anole occupying an island without congeneric
competition tends to a range of sizes with
maximal head lengths of between 20 and 25 mm
and maximal snout-vent lengths of between 65
and 96 mm. (p. 52)
If two anole species occur on an island,
one will be smaller and the other larger,
the ratio of the two sizes ranging from
1.5 to 2. (pp. 57-58)
The greater the diversity of island anole
faunas, the greater the disparity between
the largest and smallest species. (p. 48)
These are rules meant to apply only to anoles in the West
Indian islands, but they also appear to have some appli
cation to the more complex faunas of the mainland.
A second tentative generalization comes from Williams'
(1972) fascinating inferential model of the historical co
adaptation of the complex anole fauna of Puerto Rico. Build
ing upon a foundation of comparative osteology, karyotypes,
electrophoretic patterns and ecological information, Williams
has attempted to reconstruct the probable course of evolution
of the Puerto Rican anoles. From this and his wide knowledge
of other anole faunas, Williams (1972) concluded that Puerto
Rican anoles (and supposedly other anoles in complex island
are first able to utilize size differences as
a major means of syntopic co-existence. Beyond
the stage of the third species, however, size
ceases to have the same importance, and spatial
shift . and climatic shift . become
essential elements in the adaptations that
permit the addition of species to the fauna.
In light of these ideas, it would seem reasonable
that any discussion of the ecology of anoles should first
focus on size distribution among the species under consider
ation. That is the purpose of this section.
2. Materials and Methods
Times and places in which collections were made are
described above (section II). All anoles collected at Rio
Frio are considered in this section. Anoles were taken during
the morning and early afternoon, usually 9:00 A.M. to 2-3:00
P.M., and held in plastic bags until processing. As soon as
possible after collection, usually within six hours, each
lizard was weighed (alive) to the nearest 10 mg on an Ohaus
balance. Each was then killed by an injection of sodium nem
butal, and the snout-vent length and tail length were measured
to the nearest 0.5 mm with a plastic ruler. Head and hind leg
measurements (with a vernier caliper) and lamellae counts
(See Appendix 1) were made after the lizard had been preserved
in 10% formalin. Sex was determined by dissection (See
An outstanding problem of any morphological comparison
of different taxa is to determine which specimens should be
compared. I considered four possibilities:
1. All specimens of each species. The weakness of this
is that collecting bias might increase or decrease
the mean size data.
Sexually mature adults. Assessment of egg production
makes this method plausible for females, although
seasonal cessation of reproduction is a problem in
some species. For males, however, seasonal variation
in testis size and sperm production in relatively
small "subadult" size males (Fitch, 1956) makes
designation of a minimum adult male size extremely
3. The ten (or other arbitrary number) largest of each
species. This would be useful if maximum size were
desired, but sample size affects this choice too
greatly for its use here. For example, the ten
largest individuals of A. capita comprise 43% of my
collection of that species, while the ten largest
A. limifrons make up only 2.8% of my collection of
4. The largest third (or some other fraction) of each
species. This essentially eliminates the effect of
sample size, yet, if the fraction is appropriately
small, a reasonable sample of "largest adults" can
be compared. For a more thorough discussion of this
method, see Schoener (1969c).
Because an unbiased approximation of adult size is desirable
for size comparisons, I have chosen the fourth alternative,
carried out in the following manner. All specimens of each
category considered (species, sex or collecting period) were
ordered from longest to shortest. The length of the individual
that was one third of the number of specimens from the longest
was used as the lower limit for that sample. All specimens
of that length or greater are used in size comparisons for
that category, regardless of whether it is length or weight
being compared. When used hereinafter, the term "longest
third of" will refer to such a sample of the indicated
3. Sexual Dimorphism
Inasmuch as sexual dimorphism in body size is often strong
in anoles (Rand, 1964, 1967a; Schoener, 1967, 1968 1969c;
Schoener and Gorman, 1968; Fitch, 1976), it was deemed appro
priate to determine the pattern(s) and degree of intraspecific
sexual dimorphism before making size comparisons among the
species. Table 4-1 lists means and standard errors for snout-
vent length (SVL), head length (HL) and body weight (BW) for
all specimens of all eight species of Rio Frio anoles. Table
4-2 indicates the ratios and significance of differences
between those means. All eight species show a significant
(P<.05) degree of sexual dimorphism in all three parameters
except in the case of A. biporcatus for HL.
Two patterns of dimorphism are seen, one in which the
females are larger than the males (6 species), and the second
in which the males are larger than the females (2 species).
The patterns and relative positions (Fig. 4-1) are fairly
consistent for SVL, HL and BW. Using measurements from both
living and preserved specimens, Fitch (1976) examined sexual
dimorphism in 54 groups of anoles from Mexico to South America,
but primarily middle America. He found SVL ratios as low as
0.80 (Anolis vittigerus of Panama) and as high as 1.36
(a. cuprinus of xeric Mexico), but only in 34 of 54 was the
dimorphism significant. In his samples of A. biporcatus,
A. carpenteri and A. lemurinus, males and females were not
significantly different in SVL. Differences in our findings
may be due in part to inadequate sampling of large males
(lemurinus, lionotus, and pentaprion) by me and/or Fitch's
combining of samples from various localities, seasons or
years. SVL ratios in the six species studied by Duellman
(1978) at Santa Cecilia, Ecuador, ranged from 0.89 (a. trachy-
derma) to 1.10 (A. punctatus).
The pattern in which the males are larger than the females
is the rule in West Indian anoles, especially in solitary
species (Schoener, 1969a). This is apparently true also for
mainland Alpha anoles. Both species which have larger males
than females reported by Duellman (1978) are in the Alpha
section of Anolis. All seven of the Alpha anoles measured
by Fitch (1976) have larger males (significant in 6 of 7).
In Beta anoles, both patterns are common, but seem to
sort out on a climatic (faunal size?) basis. Seventeen popu
lations studied by Fitch had significantly larger males. He
considered 12 of the 17 (including A. lionotus and a. pentaprion)
to be primarily inhabitants of "severely seasonal" climates.
These are areas which also have reduced anole faunas, compared
to lowland rain forests. Only 3 of 21 Beta groups attributed
to "lowland rain forests" by Fitch have significantly larger
males. Of the 11 populations having significantly larger
females, eight are "lowland rain forest" forms, and the other
three inhabit "montane or cloud forest" areas, both of which
are "relatively aseasonal." No species which have the females
larger than the males inhabits severely seasonal areas. The
Rio Frio anoles fit this pattern precisely.
The development of sexual differences in solitary anoles,
in which males are always larger than females, is seen to be
a result of lack of congeneric competition (Schoener, 1967).
The two Rio Frio species with larger males are the two with
the least congeneric pressure. A. lionotus is the only anole
which normally occurs in the riparian microhabitat, i.e. it is
essentially a solitary anole, and it very closely fits the
solitary anole size pattern. Another point, which holds for
A. lionotus (but not for A. pentaprion) is Schoener's (1969b)
observation that relative head size is less in solitary anoles.
Table 4-3 indicates that A. lionotus, the "most nearly solitary"
anole at Rio Frio, has the shortest head relative to SVL.
A. pentaprion inhabits the outer branches and twigs of the
forest crowns. No other anole, with the possible exception
of the much smaller a. carpenteri, overlaps with it in this
exposed, climatically severe microhabitat.
The pattern in which the females are larger than the
male is seen by Fitch (1976) to be an asset to species in
seasonal habitats, allowing them to maximize egg/young pro-
duction. While the six Rio Frio species with this pattern
are all primarily inhabitants of aseasonal microhabitats, these
are also species which generally occur with the greater numbers
of congeneric competitors and can, therefore, least support
the food niche expansion that sexual dimorphism requires. Any
increase in the number of anole species in a fauna will result
in a general decrease in the availability of prey. Most par
ticularly this will reduce the density of the least common
large items, which are the prey of the large males in dimorphic
species (Schoener, 1967, 1968; Schoener and Gorman, 1968).
Schoener's (1969a, 1969b) models and examples (1969b, 1970)
suggest that such an increase in food competition might result
in a reduction of body size of all species, as smaller lizards
are better able to exploit the more numerous small prey items.
It would seem a reasonable explanation then, that the general
anole pattern in which males are larger than females is
abandoned in rain forest Betas because of the greater compe
tition in multi-anole fauna. Male size is more reduced than
is female size giving the size advantage to the reproductively
more important, egg producing females. The following section
(V) will further explore sexual differences in food.
4. Interspecies Comparisons
When an anole species exists without congeneric competitors,
Schoener (1967, 1968, 1969c) suggests that an optimum size
exists. When more than one species is present, sizes must
diverge from the solitary anole size to alleviate pressure on
particular prey sizes. Table 4-1 indicates the wide range of
sizes of Rio Frio Anolis (Fig. 4-2, 4-3). Note the range,
from a minimum mean SVL of 33 mm for a. humilis males to a
maximum of 91 mm for a. biporcatus females, an almost three
fold difference. As would be expected, the difference in
weight is even greater, from 0.65 g (a. carpenteri males) to
17.1 g (a. biporcatus females), a difference of over 26 times.
The expectation for a complex assemblage of species is
that each species will differ from the next by some constant
factor. Hutchinson (1959) found a range of 1.1 to 1.4 with an
average difference of 1.3 times, in a variety of taxa. Table
4-4 shows that this is not the case for Rio Frio anoles. Indeed,
if eight species could exist with the smallest 33 mm in SVL
and the length of each larger species increased by 1.3 times,
the largest in the sequence would be just over 207 mm SVL,
much larger than any mainland anole. To complicate this
possibility further, Schoener (1970) has shown that the
separation ratio should increase between larger species, so
the largest in an eight species sequence would be much larger
than any known anole. Since there is apparently not a constant
or increasing ratio between all eight species at Rio Frio,
is it possible to subdivide the assemblage into some logical
grouping that will show the expected pattern?
One possible sorting of species might be along phyletic
lines. The species groups (or series) of Etheridge (1960)
apparently represent well established evolutionary lines.
Table 4-5 shows that, within each of the three series represented
at Rio Frio, the separation ratios are not exactly in keeping
with Hutchinson's (1959) suggestion. The two members of the
chrysolepis series are too widely separated. In the fusco-
auratus series two species are too close, and the third is
too large. Finally, in the peter si anoles, the females of the
two smaller species are separated by too large a ratio, though
the males fall in the low end of Hutchinson's bracket, and
the larger two are just barely different enough to pass the
lower limit. In fact, only 1 out of 10 SVL comparisons and
5 of 10 for HL fall within the 1.1 to 1.4 range, even though
the average separation ratios are near 1.3 (1.32, SVL; 1.30 HL).
Though a slight expansion of the ratio range would encompass
most, if not all the ratios in Table 4-5, perhaps something
other than a phyletic grouping might yield a more logical,
consistent set of size relationships.
In an evolutionary sense, there seems to be no reason why
congeners should differ in size simply because of phyletic
relationship. Indeed, Hutchinson (1959) proposed that the
divergence of syntopic species pairs occurs because of their
ecological similarity. Therefore, ecological resemblances
(microhabitat, perch site or structural niche) provide a
better basis for grouping the species. Most of the anoles
which I collected or observed at Rio Frio were in the area of
newly felled trees. With this type of data, it is not possible
to establish numerical boundaries on the structural habitat
of each species as is frequently done (Rand, 1964; Schoener,
1967, 1968; Schoener and Gorman, 1968; Schoener and Schoener,
1971a, 1971b; Andrews, 1971a, 1971b). However, for the
discussion here, general descriptions of the habitats
(such as those given in Section III) will be quite use
ful. As Williams (1972) explains with regard to habitat
similarity in two Puerto Rican anoles:
They both may be crudely described as crown
animals of the shaded forest. It is
possible to quibble a bit about this but
the same modal situation describes both.
A giant anole such as cuvieri, though it
is, in fact, seen some of the time at every
level (personal observation), is most
often seen high in the crown. A dwarf
anole such as occultus is seen on branches
and twigs of small diameter . and
therefore not infrequently on bushes and
vines, but certainly its modal structural
niche includes the crown. (p. 74)
The eight Rio Frio species are best sorted into four
1. Ground dwellers. This set includes A. humilis
and A. capito, both of which inhabit tree
bases, logs and the intervening ground
2. Riparian species. Only A. lionotus fits here.
It is thus horizontally removed from the
other seven species.
3. Trunk dwellers. Included here are A.
limifrons, primarily an inhabitant of
lower trunks, smaller trees and low shrubs;
A. lemurinus, found on larger, higher trunks;
and A. biporcatus on large trunks up to, and
possibly into, the base of the crown.
4. Crown inhabitants. Although A. biporcatus might be
included here, I am using this to mean only the
small branches and twigs, the habitat of A. pentaprion
and probably A. carpenteri.
Table 4-6 gives the length ratio of the eight species grouped
by perch site.
The extension of the Schoener Rules (Williams, 1972)
and other findings of Schoener (1970) to Rio Frio anoles leads
to the expectation that: (1) when an anole is without congeners
in it's particular habitat, it will tend toward a particular,
sexually dimorphic size pattern (55-75 mm SVL for males,
40-60 mm SVL for females); (2) when two anoles share a micro
habitat, they will differ by a factor of 1.5 to 1, with one
larger and one smaller than the solitary anole size; and
(3) when three anoles occupy the same structural level, the
size ratio will be greater between larger species than between
smaller. Schoener (1970) found an average ratio of 1.5 between
smallest and 2.2 between largest species in 10 syntopic trios.
The Rio Frio anoles seem to fit these expectations.
The only anole that rarely shares it's habitat with other
anole species is A. lionotus, and it clearly fits the expected
pattern, except that there is not quite as much sexual di
morphism as expected. One factor that might push the female
size upward, i.e. limit the dimorphic spread, is the presence
of the smaller, more numerous A. humilis on the adjacent, non
riparian forest floor.
Two anole habitats, the ground and the crown, are also
relatively close to expectation. The two anoles with the
lowest perch sites, A. humilis and A. capito, are on either
end of the solitary anole size range, but they are separated
by more than the 1.5 to 2.0 suggested by rule two. The ex
ceptionally small size of A. humilis is perhaps explained by
its partial microhabitat overlap with A. limifrons. This
small, abundant species might compete in such a way as to
push the size of A. humilis down. In fact, both A. humilis
and A. limifrons may be reduced in size by any overlap (Schoener,
1969a, 1969b). The crown anoles, A. carpenteri and A. pen-
taprion, are separated by the required size range in males, but
not in females this being due to their having opposite
patterns of sexual dimorphism. Though a. carpenteri is below
the bottom end of the solitary anole size range, A. pentaprion
is not larger than the solitary size which rule two specifies
it should be. This may be due to the small sample size.
However, in equally small samples, Fitch (1976) found a mean
of 74.2 mm SVL for male and 60.0 mm for female A. pentaprion.
These measurements would make A. pentaprion males 2.03 times
and females 1.45 times the length of A. carpenteri and just
longer than the size range of solitary anoles. It may be that
the size of A. pentaprion is held down by the influence of the
larger a. biporcatus in the crown-- or even by the presence
of Polychrus gutturosus, a larger, although primarily vege-
tarian iguanid which was found in crowns at Rio Frio.
The only place where three species of anoles regularly
perch is the area broadly termed "trunk." Here the pattern
of increasing separation in larger species (Schoener, 1970) is
seen, although the magnitude is somewhat less than Schoener's
average for 10 islands. In both sexes, the difference in
length between A. biporcatus and A. lemurinus is greater than
the difference between A. lemurinus and A. limifrons.
In general then, the sizes of Rio Frio anoles are well
explained by Schoener's island patterns if habitat types are
considered separately. That small anole faunas on West Indian
islands are isolated by sea water, while the Rio Frio anoles
are segregated only by the much more nebulous boundaries of
perch site preference, should explain the slight variation.
5. Body Weight to Length Relationships
Schoener (1969a) predicted that weight could be related
to length by a power function (such as BW=a(SVL)^} and that
the power, b, ought to be close to 3, in groups such as Anolis
lizards. He was able to confirm this only with data for the
Puerto Rican A. gundlachi from Turner et al. (1965). He also
expected that b might prove adaptive where there is an advan
tage in increased length without increased weight, as in a
twig inhabiting anole. Data from Rio Frio Anolis fit these
expectations quite well (Table 4-7, Figs. 4-4 to 4-11). The
exponent b ranges from 2.46 to 3.27, with a mean of 2.91.
Because of the weight increase associated with the repro
ductive activity of larger females, the regression line for
females bends more than the curve for males. For females b
averaged 2.98, while the mean for males was only 2.83. In all
species except A. lemurinus males have a lower b than females,
reflecting the relative slimness of large males. This dif
ference was significant only for A. limifrons and A. biporcatus
(Table 4-8). Male A. limifrons often perch higher than females
(Fitch, 1975), so the "slimness" of males may be an additional
aid in moving along small stems and vines, or over leaves.
No information is available on sexual differences in perch
heights of a. biporcatus.
Though comparable reports are not available for other
Anolis, at least two other tropical iguanids seem to show the
same sexual difference. In two Costa Rican populations of
Basiliscus basiliscus, Van Devender (1978) found males to have
a significantly lower b than females. Recalculation of
weight-length regression equations for ctenosaura similis
given in a different form by Fitch and Henderson (1977) also
show a higher b for females than males.
Comparisons of b between pairs of species show few
differences (Table 4-8). The exceptionally slim A. carpenteri
have a lower b than males of all other species. This difference
is significant for comparisons with all species except A. hu-
milis and A. pentaprion. This may reflect a real adaptation
to the twig-vine habitat, or it may be an artifact of the
rather small, restricted sample of A. carpenteri since 14 of
21 weighed were within 5 mm in SVL and 12 were within 2 mm
(Fig. 4-6). Female A. biporcatus have a higher b than all
other females, but this was significant only for comparisons
with A. carpenteri and A. lemurinus. The large b may reflect
the fact that A. biporcatus perches only on the broad trunks
and large limbs of trees; but more likely it simply reflects
the fact that a. biporcatus females are by far the largest
anoles at Rio Frio, and are almost always gravid.
6. Seasonal Variation in Size
There are two ways in which seasonality may cause differ
ences in the sizes of lizards. A decrease in mean size may be
caused by changes in growth rate or by different reproductive
rates, which will alter the fraction of small lizards in the
population. A change in the weight-length regression power,
attributable to a size dependent change in food consumption,
could be the cause of a seasonal difference. Even though
little seasonality is apparent at Rio Frio (see Section II),
anoles may respond to minor variations. In this section
that possibility will be considered.
Table 4-9 shows SVL, HL and BW of four species of Rio
Frio Anolis for each of the four collection periods. Samples
of the other species were too small for adequate seasonal
comparison. Table 4-10 shows significance of seasonal dif
ferences in SVL and BW (see also Figs. 4-12, 13, 14, 15).
A. humilis shows very little seasonal change. Females
are largest (both SVL and BW) in the September sample, the
climax of the rainy season, but this is significant only when
compared to February, the driest season. There are no sig
nificant changes in the males of A. humilis.
Size of A. limifrons is affected by a decrease in the
number of large females and an increase in the number of small
males in the November sample, and by a decrease in the number
of juveniles and small of both sexes in the September and
May samples. The size then has an apparent "dip" at the end
of the rainy season. A similar pattern was seen by Fitch (1973a)
in A. limifrons from Turrialba, Costa Rica, which has a rain-
fall pattern similar to that of Rio Frio. In both sexes, May
specimens were the longest, but were not as heavy as the
shorter September lizards. This would seem an indication that
they are not feeding as well at the end of the dry season as
they did during the wet months. Even though the sample sizes
are very small, a. lemurinus and A. lionotus appear to be
similar to a. limifrons in seasonal size pattern. Eoth have
a late rainy season decrease in mean size, with an increase in
length through May.
There is one immediately obvious difference in the BW-SVL
regression curves (Figs. 4-16, 17, 18, 19)that for May is
well below all others. This is further evidence that, at any
SVL, A. humilis or A. limifrons will weigh less in May than
in any other period. The November curve is the highest in
three of the four indicating that anoles are heaviest at the
end of the rainy season.
The same pattern is approximately true for A. lemurinus
and a. lionotus. Even though sample sizes are small for some
periods and no A. lionotus were weighed in September, the
regression curves of the rainy season samples are highest, and
the May curves are lowest (Figs. 4-20, 21), as is true for
A. limifrons and A. humilis. The May curves are altered some
what by the absence of juvenile specimens of either a. lemurinus
or A. lionotus.
Differences in b, the exponent of weight-length relation
ships are not so easily explained (Tables 4-11, 12). First,
for A. humilis and A. limifrons in all periods, b is larger
for females than for males as found in the preceding section
for all species. This is significant only for the two rainy
season samples of a. limifrons. But there is some variation
within the sexes that is more puzzling. In three of the four
species, the May samples have the highest b, although in some
comparisons this difference is not significant. The regression
exponent is low in most February samples and low in September
samples of A. humilis.
The regression exponent, b, is essentially the factor deter
mining that' the big get bigger,, i.e.-, the.upward curvature of the
regression line. As such, it appears that even though most
groups are lightest in May, the longer individuals are com
paratively heavier than the shorter ones, an indication that
a late dry season food decline has a more pronounced effect on
smaller lizards. This may be caused by a variety of factors,
probably including actual physical struggle for food items,
size proportional territory (and, therefore, more potential
prey for larger individuals), and higher perches and a larger
scanning radius for larger anoles. The only exception is
A. 1 emurinus tve sample of May females has a heavier BW as
well as a longer SVL. However this is from a sample of two,
so the differences are not significant.
The only exception to the May peak of b is the case of
A. lionotus. In A. lionotus, the complete absence of small
individuals causes the regression line to rotate clockwise.
If the curve were drawn below the SVL of the smallest May
A. lionotus collected (56.5 mm), it would predict improbably
heavy juveniles (dotted line on Fig. 4-21). If normal-sized
juveniles had been collected, they might well have bent the
regression curve to the "normal" pattern.
TABLE 4-1. Sizes of Longest Third of Rio Frio Anolis
SVL : (mm)
33.22 ( 0.14)
36.50 ( 0.41)
4.00 ( 0.14)
4.95 ( 0.26)
75.80 ( 1.10)
18.52 ( 0.10)
9.16 ( 0.77)
23.4 5 ( 0.11)
all those in
- includes the
longest third plus any others of the same SVL as shortest specimen
in longest third. N (BW) may be less than NL (SVL and HL) if not
all of the longest third were weighed.
'Given as: mean(standard error)
Significance of Sexual Size Differences in Rio Frio nolis
larger than males:
it it it
it it it
Series: c, chrysolepis; f, fuscoauratus; p, petersi.
'^Ratios are mean of longest third of maletmean of longest third of female.
Significant (t test) in difference in male-female means: n.s.,
*, P<.05; **, P<.01; ***, P<.001.
TABLE 4-3. Relative Head Length
EMean HL mean SVL.
1Longer SVL shorter SVL.
test, comparing each with next longer; significance
as in Table 4-2.
TABLE 4-6. Length Ratios by Perch Site
TABLE 4-7. Weight to Length Relationships
+ j uv.
NOTE: Regressions are
the form of
'For A. humi1is
were used with
and A. limifrons small
F Test for Difference
is o vs.
right column is o vs. o.
NOTE: Degrees of freedom are 1, N +N -<
Sizes of Anoles by Collection Period
humi 1 is (
33.09 ( 0.30)
9.12 ( 0.09)
0.96 ( 0.04)
1.29 ( 0.05)
1.15 ( 0.06)
37.00 ( 0.30)
0.89 ( 0.04)
9.55 ( 0.08)
37.21 ( 0.13)
39.26 ( 0.30)
N=2; head length not measured on all a. 1ionotus.
Significance of Seasonal Differences Size
in Rio Frio Anolis
NOTE 1: t-test;
NOTE 2: Female comparisons are on the upper right, males on
the lower left
Seasonal Weight to Length Relationships
NOTE: Regressions as in Table 4-7
F Test for Seasonal Difference in b
MAY 6.11* 2.18
NOTE: Significance and arrangement as in Table 4-8.
4-1. Sexual Dimorphism in Rio Frio Anolis. Male:female ratios, for jDOdy weight
(BW), head length (HL) and snout-vent length (SVL) for Rio Frio anoles
and SVL for specimens measured by Fitch (1976). Bar below symbol indicate
male-female difference is not significant. Symbols represent:
^. biporcatus, solid triangle;
A. capito, hollow triangle;
A. carpenteri, x;
A. humilis, hollow square;
A. lemurinus, solid square;
A. limifrons, hollow circle;
A. lionotus, solid circle; and
A. pentaprion, hollow diamond.
I I I
1.1 1.2 1.3
FIGURE 4-2. Snout-Vent Lengths of Longest Third of Rio
Fro Anolis. Mean SVL indicated by a hori
zontal line, range by a vertical line, 95%
confidence interval by stippled bar (males,
on right) or cross-hatched bar (females, on
left), and sample size by number above each
UJ W ifci 4^ C/1 <_n
O U1 O U1 O U1
i > I I I I
A. J emu rinu s
O'! C7> I 00 00 VD O
o i_n o Ln o <_n o t_n
1*11 1 I 1 I
FIGURE 4-3. Body Weight of Longest Third of Rio Frio Anolis
Mean BW indicated by a horizontal line, range
by a vertical line, 95% confidence interval by
stippled bar (males, on right) or cross-hatched
bar (females, on left), and sample size by num
ber above each figure.
1 emurinu s
A. humi1 is
FIGURE 4-4. Weight Versus Length for nolis humilis. Males are indicated by solid
circles; females by hollow circles; unsexed juveniles by crosses. Curves
(solid line, males and juveniles; dashed line, females and juveniles)
Coefficients are given in table 4-7.
~i I r
30 35 40
FIGURE 4-5. Weight Versus Length for Anolis limifrons. Symbols as in figure 4-4.
FIGURE 4-6. Weight Versus Length for Anolis carpenter i. Symbols as in figure 4-4.
FIGURE 4-7. Weight Versus Length for Anolis lemurinus. Symbols as in figure 4-4.
FIGURE 4-8. Weight Versus Length for Anolis pentaprion. Symbols as in figure 4-4.
FIGURE 4-9. Weight Versus Length for Anolis 1ionotus. Symbols as in figure 4-4.
O -i i t r-
25 30 35 40 45
i l I
55 60 65 70 75
-10. Weight Versus Length for Anolis capito. Symbols as in figure 4-4.
I I I I ' 1 I 1
40 50 60 70 80 90
FIGURE 4-11. Weight Versus Length for nolis biporcatus. Symbols as in figure 4-4.
FIGURE 4-12. Seasonal Variation in Size of Anolis humilis.
For the longest third of each sample, mean
body weight (BW) and mean snout-vent length
(SVL) are given as horizontal lines.
Stippled bars indicate 95% confidence inter
vals; vertical lines show the range; and
sample sizes are given by numerals above the
SEP NOV FEB MAY SEP NOV FEB MAY
SEP NOV FEB MAY
SEP NOV FEB MAY
-13. Seasonal Variation in Size of Anolis limifrons.
Symbols are as in figure 4-12.
SEP NOV FEB MAY
SEP NOV FEB MAY
-14. Seasonal Variation in Size of Anolis lemurinus.
Symbols are as in Figure 4-12.
SEP NOV FEB MAY SEP NOV FEB MAY
SEP NOV FEB MAY
SEP NOV FEB MAY
-15. Seasonal Variation in Size of Anolis lionotus.
Symbols are as in figure 4-12.
NOV FEB MAY
NOV FEB MAY
NOV FEB MAY
NOV FEB MAY
FIGURE 4-16. Seasonal Variation in Weight-Length Curves for nolis humilis Males and
Juveniles. Curves were fit for each period by least squares to the
Coefficients are given in the table 4-11. September curve is a solid
line; November a dashed line; February a dotted line; and May a dot and
-17. Seasonal Variation in Weight-Length Curves for Anolis humilis Females
and Juveniles. Curves are as in figure 4-16.
-18. Seasonal Variation in Weight-Length Curves for Anolis 1imifrpns Males
and Juveniles. Curves are as in figure 4-16.
15 20 25
FIGURE 4-19, Seasonal Variation in Weight-Length Curves for Anolis limifrons Females
and Juveniles, Curves are as in figure 4-16,
-20. Seasonal Variation in Weight-Length Curves for Anolis lemurinus, Both
Sexes Combined. Surves are as in figure 4-16.