IJIII! liiiJtHtlHi
•'liHiifi
1
^^^
1
OCCASIONAL PAPERS * ®7'
of the ^^'VERsrm
MUSEUM OF NATURAL HISTORY
The University of Kansas
Lawrence, Kansas
NUMBER 3, PAGES 1-62 MAY 26, 1971
QUANTITATIVE ANALYSIS OF THE
ECOLOGICAL DISTRIBUTION OF A TROPICAL
HERPETOFAUNA
By
Martha L. Crump'
INTRODUCTION
Possibly the rainforest environment is not so constant, equitable,
and predictable as ecologists have assumed. Lloyd, Inger, and
King (1968) suggested this possibility as a result of studies on
amphibian and reptile diversity in tropical rainforests of Borneo.
The ways in which species utilize environmental resources have
long been of interest in ecology; recently some effort has been made
to analyze the inherent properties of the rainforests as they relate
to amphibians, and reptiles. Schoener (1970) studied nonsynchro-
nous spatial overlap of lizards, genus Anolis, in patchy habitats in
the West Indies. Schoener and Gorman (1968) studied niche dif-
ferences of three species of Anolis from the southern Lesser Antilles;
Schoener (1968) also studied resource partitioning among anoles
on South Rimini Island. Rand (1964) examined the ecological
distribution of anoles in Puerto Rico. Rand and Humphrey (1968)
studied ecological distribution and interspecific competition among
lizards in the rainforest at Belem, Rrasil. Duellman (1967) studied
isolating mechanisms and resource partitioning in tree frogs in
Costa Rica. Inger and Greenberg (1966) studied the relation be-
tween niche overlap and interspecific competition for three species
of frogs, genus Rana, in Sarawak. As indicated, the majority of
studies have been carried out on specific genera; no extensive, quan-
' Graduate Student, Museum of Natural History, University of Kansas.
2 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
titative ecological studies have been carried out in the New World
tropics on an entire reptilian or amphibian community.
This is a report on the ecological distribution of amphibians and
reptiles undertaken at the Guama Ecological Research Area near
Belem, Brasil. Field work was carried out from mid-January
through July 1969, in April 1970, and in June and July 1970. Part
of the resultant collection was given to the Museu Goeldi in Belem,
and part is catalogued in the Museum of Natural History at the
University of Kansas.
The objectives of the present study are threefold: 1) to deter-
mine the ecological distribution of 62 species of frogs, salamanders,
and lizards within the rainforest environment of Belem; 2) to ana-
lyze the environmental parameters affecting the distribution of
species; and 3) to compare and contrast the major areas with re-
gard to species composition. The ecological distribution of the
herpetofauna presented here is based on data obtained in one small
area in part of the year. A similar study carried out from August
through January or in a different area probably would yield some-
what different results.
Description of the Area
Belem is located about one degree south of the equator, in the
lower Amazon Basin, Estado do Para, Brasil; the elevation at the
highest point is 12 m above sea level. The mean annual temperature
is 26°C, and the average monthly temperature varies less than 2°C
throughout the year. Seasonality is reflected through the temporal
distribution of rainfall, yielding wet and dry seasons. The average
annual rainfall (44 years) for the wet season, January through
June, at Belem is 2028 mm, whereas that for the dry season, July
through December, is 830 mm (Belem Virus Laboratory, 1967 An-
nual Report ) .
Belem is the headquarters for the Instituto de Pesquisas e Ex-
perimenta^ao Agronomicas do Norte (IPEAN). An area of about
310 hectares of IPEAN property has been designated as the Guama
Ecological Research Area (APEG). Most of my study was carried
out in two of the APEG reserves. The Aura Reserve is part capoeira,
part terra firme (Fig. 1), and part varzea forest (Fig. 2); some
areas are transitional between terra firme and varzea forests. The
Catii Reserve is a transect of igapo forest (Fig. 3) 1000x200 m.
See figure 4 for spatial relationship of the reserves. The forest
types are defined below.
The reserves are divided into a network of 10x10 m quadrats,
each marked with a numbered stake. For each observation or in-
ANALYSIS OF ECOLOGICAL DISTRIBUTION
Fig. 1. Terra firme' forest (Aura reserve). Well-drained forest on relatively
high ground. Photo by Roger Arle.
OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
Fig. 2. Varzea forest (Aura reserve). Flooded daily by the back-up from
the Rio Guama; predominance of Acai palm trees (Euterpe oleracea). Photo
by Roger Arle.
ANALYSIS OF ECOLOGICAL DISTRIBUTION
:f9s'
^^nr
4 ** ^^'^^l
Fig. 3. Igapo forest (Catu reserve). Fermanently flooded forest. The
boardwalk provides easy access.
OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
Fig. 4. Map of Mocambo and Aura Reserves in relation to Belem and the
Rio Guama. The Mocamlio Reserxe consists of terra firme forest, surrounded
by the Catu resene of igapo forest (not indicated on map); the Aura Reserve
consists of terra firme, capoeira, \ar7ea, and transition forest areas. Some
studies were carried out in the vicinity of the IPEAN headquarters and at the
Agua Preta Reservoir ( Utinga Reserve ) .
dividual collected, the hectare and quadrat numbers were recorded,
thereby assuring that all data were collected in the same spatial
frame of reference. The distribution of water was determined and
mapped for the capoeira, terra firme, and varzea study areas ( Figs.
5-9); species dishibutions were superimposed on these maps to
determine the associations of species with standing water. For the
various quantitative analyses, 44 sampling plots, each 20x30 m,
from four of the major forest areas were studied. The location of
the 4 capoeira-terra firme transition, 19 terra firme-varzea transition,
and 9 varzea plots relative to each other and to the distribution of
water is shown in figures 10 and 11. The other 12 plots were in the
igapo forest. For the purpose of an analysis of ecological distribu-
tion, the rainforest at Belem was divided into seven major areas:
Terra firme forest. — Well-drained forest on relatively high
ground that is never subject to flooding is called terra firme forest.
It is a well-structured, complex, tropical rainforest. One 5.5 hectare
area of terra firme forest (Mocambo Reserve) has been studied
extensively by botanists. Cain et al (1956) found the area to be
extremely complex, both in vegetation species richness and in vegeta-
tion density; they estimated the density of trees exceeding 10 cm
ANALYSIS OF ECOLOGICAL DISTRIBUTION
Fig. 5. DistrilMition of Bolitoghssa altamazonica in relation to distribution
of water in terra firme, capoeira, and varzea transition forest. Each small
square represents a quadrat, 10 x 10 m. Cross-hatched quadrats are those
areas in which at least one frog, salamander, or lizard was observed by the
author. Stippled areas represent terra firme-varzea transition depressions filled
with standing water; non-stippled areas are better drained and usually are
located on higher groimd. Each dot represents the obser\ation of at least
one individual of Bolitoglossa altamazonica witliin the particular quadrat.
in diameter to be 594 trees per hectare. Dr. Murca Fires, a botanist
associated with IPEAN, identified 215 species of trees in this area.
Hatheway (1967) estimated the canopy to be 80 percent closed,
with an average canopy height of about 35 m. He distinguished
three strata of vegetation. Beneath the nearly closed canopy is a
OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
Fig. 6. Distribution of Boliioglossa altamazonica in relation to distriliution
of water in \'arzea forest. See figure 5 for explanation; in this figure stippled
areas represent varzea depressions filled with standing water.
deep layer of trees up to 20 m in height; the bottom, dense, scrubby
layer extends to a height of about 1.5 m from the ground.
Varzea forest. — Swamp forest bordering the rivers is known
locally as varzea. This forest is flooded daily by the back-up of the
Rio Guama, due to tidal effect. The degree of flooding \'aries
throughout the year and is correlated with rainfall. All aquatic
environments in the immediate vicinity of Belem seem to be fresh-
water (Humphrey, pers. com.). The "white water," so called be-
cause of the presence of sand, silt, and clay particles, yields a con-
ANALYSIS OF ECOLOGICAL DISTRIBUTION
^
Fig. 7. Distribution of Leptodacttjhis mannoraiiis in relation to distribu-
tion of water in terra firme, capoeira, and \arzea transition forest. See figure
5 for explanation of symbols.
tinual deposition of alluvium. The resultant alluvial varzea soil is
rich, but has a low permeability. During the rainy season, parts of
the varzea are flooded to a depth of 1 m or more. Depressions are
present, resulting in differential drainage. Tall woody plants,
palms, and giant aquatic herbs exist nearly side by side as a conse-
quence of drainage patterns (Hatheway, 1967). There is a pre-
dominance of palms in the varzea forest; the acai palm (Euterpe
oJeracea) is the most common tree. Lianas and epiphytes are com-
mon, and moss as thick as 1 cm covers the trunks of trees up to 2 m
10
OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
r;
i
'//
»
/J
«
#
2
r
^/<
L J
/,
14^
>
'/
#
'^
•"/-/
^/
/
>3
#
/K
'/
#
«]
zw:
B
»^-
/,
^, /
#
l«
/
/.
V
V
'/
'/
■)
4
/.
V
/^
»
1 —
J
V
/.
4
»
/'2
1
%
'A
; j
/
•
//(i»»»»
4
: j
■/y
/
1
/
/,
^/
!
^
/
/
t= —
»
4«
/
--/
to
'/
/
A
/
1^
4
r^'
4
K
X
/
1
*
E^
Y^
*L.
r — '
L^
2
#
/,
BC
*
4
V
.rr
V.
^ ! ^(
^M:-^
■
///
/"
h^
#
iTI-^
1
'/
i
I 1 j 1
/
#
_
■■
Sm
SZ9
tew
•^■f-
^
ily > yUi!y >
^71
Its
' i I
1 — ' — 1 — — r-
'v-
■^
^^
/
—
—
if
■
W!"
L^A Jl-Jl^
k^^y«g^«lg^
:^/ ''/ c
^
7
2
/
i
i
1
[
#
r
r/
■ ! j 1 '
, 4
r ^
/
^
i
P
4
1^x^/1
" ! ' ! '^
tj:' ':-|:':
7
«
/.
Ki
4
^
^1^
^(^
H^^
/a
^'^V ^ <2^r:::|:: t:::|
z.
'■ 1
/
• ^
^dx
>:j:_>I) ''/
%:\ >^-
W^Y ---
4
#
t-r^-
0' V ' -
i ^/*
/■I
^>:^:.^^v
■;:■:-[■:-:[;: f
i
«
.;
1
1 ^i-
-i':-:i t
:r
Six!:/^
Ty^^^^\^
V
o
•
i^:^
•i-O^
>:■: H. b
' /
mM\
7
#
%
/^ »
^^
^i^t^^.i^f^^
k^^i^TJww
^ /
ik
—
^
^
J
r"
-<
.-K
••i^
'4
^f
fe^g
^^m
— '
/
/
?i
—
'
7
b^
V-
Uj
—
r
pi
U
V
toff^
p^^f^yw\
A
/
\--
-r^
%>-
■J ;i , (Ty^iy i''^?/!^' '^
1 ■/
^.
/
. j«
:\
^2
K^ i .
V-t^-
Y
rx
" " rx
n' M tv! r/^''°
^■■■~
1
■t-.^r
/
41
J
__
\ r'-
ii
,
h^^
efc
S
'T — "7 vStt/- :>;i
4^
#-
,
_^_
^wm
^"7
^-
—
—
—
—
—
'^k
b^
^r
i^^
ll
i
fer^^
4^^
m^:^
^r
^.
^,
\'
'S
^ ■
n -'1
i^^
''■i n^^pf^^
:j
m/
r^
^
;-; ■:■
^■
:4'-v
'r^rl
--^ ■fejS*^ ^
4^-'^J?
;:
■--■?'
h "^
X '
Rr^ iTi 1^ J '-'K
J^^^iM*-
' 1^
/'■■
-
n
f"'-
■ -?
4
f ' ■■'
'-> V
pto^fnF f : ■ f»' ■ ■ >:^k
^ ■ -^ J
SV"
) ' (
■'■<'■
s ■'■■/ ^
<:'
\-
^7_
^
L, ■ "A
pU| R p- :-^N
.-r-"
i^
b^^
^
-fcf^
is-
i
i
'Li t
'^'f-
i
-A
?^-/7
f
ff
^•ifc
-i
'<-t
f
^
K
"%
H
k^
1 —
-
1
i
■■.itrj.;..
^r
4^
; r
^
l^t-
7
yi^---^
\ '=^>ifrj:.v
■^j_
7'
M.
V\
^Mo
^
f) i
.:^
r"
/: :■!
V
>:
■-■■^
''
'
j
~S'i^
/Jc--^
. r
■-VN
*-N
-'^^:
H
^^i
^ '
i-^
• ■
->
hs' ■
l^i
r?
r4o
^
1 :
^
-
Sh^
:';■' '
-t--
■:;£?
rfez
■^
1^
^
■^ "■
: ■<-
^
ES:
i:
^
^-
:^
^i':'
•:A
Lyl
r
■rr--
'*Ty*^
^'
'^
[^
^^^^
%.'^(
t-
-"
=1^
4^T^
'
;jr
^-^^
--
—
—
f-
t-
6«'^
fi
:::::^
/■:.
■^
i;
f.-
:;::/■:■
J 2
P*--
: \_^
t.<
rv
, IV
■':)
\
v^::
^/T
k
y,::i^:-\:.
A
'^i
J ^
/' ■■
^
q "~
"■:•
r
'^
'^\
%:
||-
c^<
%
^-
-
-
-
-q
^.^
-)^
-J
'■'1
V-
vi4
"
--
_
—
._
—
~\
xE
1
~"f^
!:-x
;^
-
-
—
~
-- -^
/■ ■-■-
;■;• -A
a
G:--
^':i
---
-
"
~
—
h-
-\:
m
Fig. 8. Distribution of Gonatodes hmneralis in relation to distribution of
water in terra firme, capoeira, and varzea transition forest. See figiue 5 for
e.xplanation of symbols.
from the ground. Hatheway ( 1967 ) proposed that epiphytic mosses
hkely indicate high humidity in tropical forests. In the varzea, this
high humidity probably results from constant evaporation from the
moist ground and water. Hatheway estimated that the total density
of trees over 10 cm in diameter is probably greater than 600 trees
per hectare. The canopy is about 50 percent open, and the canopy
trees are 30-35 m in height.
Igapo forest. — This is forest that is permanently flooded with
"black water," so called because of organic residues. Hatheway
ANALYSIS OF ECOLOGICAL DISTRIBUTION
11
Fig. 9. Distiilnition of Gonatodcs liumcniUs in relation to distribution of
water in varzea forest. See figure 6 for explanation of syml)ols.
(1967) estimated the average depth of water to be 25 cm; beneath
the water is another 25 cm of organic, water-logged muck, under-
neath which is white clay. The area consists of many stagnant,
foul-smelling, interconnected pools. Small islands of root masses
project from the pools; much of the vegetation in the swamp forest
is supported on these islands, although a few trees are rooted in
the muck. There is no well-formed canopy, and other distinct vege-
tational layers are difficult to distinguish. Most trees are small-
crowned, slender dicots, rising above the thick mesh of tangled
roots elevated to 3 m above the deep mud of the swamp.
12 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
Fig. 10. Distriliution of sampling plots in relation to distribution of water
in terra firnie, capoeira, and varzea transition forest. See figure 5 for explana-
tion of symbols. The numliered plots, each 20 x 30 m, are indicated by
heavy, straight lines; data from these plots were used in the contingency
table analysis.
Capoeira forest. — This is second growth forest on well-drained
ground. The capoeira areas studied had relatively open canopies
and fairly dense ground cover. Much of the area is composed of
tall grasses and ferns; the forest floor in some sections is covered
with brush and fallen logs.
Capoeira-terra firme transition. — Four plots were studied which
are intermediate between capoeira and terra firme forest with re-
ANALYSIS OF ECOLOGICAL DISTRIBUTION
13
Fig. 11. Distril)ution of sampling plots in relation to distribution of water
in varzea forest. See figines 6 and 10 for explanation.
gard to characteristics of canopy and ground cover. The plots ex-
hibit a greater vegetation density than typical capoeira, but less
than typical terra firme forest.
Terra firme-varzea transition. — Plots in one area exhibit some
characteristics of both terra firme and varzea forests, but differ
noticeably in other ways. For instance, on well-drained ground
there is a lower vegetation density than in typical terra firme forest.
The flooded portion lacks the predominance of palms, characteristic
of typical varzea forest. Corresponding to the varzea and terra
firme forests respectively, some of the soil in the transition area is
14 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
a poorly-drained, silty-clay alluvium, whereas other soil is a well-
drained, heavy, yellow laterite.
Open and edge. — All non-forest study sites are grouped in this
category. Observations and collections were made in swampy areas
in open fields, ponds along forest edges, and in second growth vege-
tation bordering the IPEAN reserves. During the rainy season, the
swamps and ponds contained water to a depth of about 1 m but
were usually less than half as full during June and July. Dirt roads
and roadside ditches on IPEAN property and sewage swamps within
the city were examined weekly.
Composition of the Herpetofauna
Three orders of amphibians: Gymnophiona ( caecilians ) , Cau-
data (salamanders), and Anura (frogs) and four orders of reptiles:
Amphisbaenia (amphisbaenids), Crocodilia ( crocodilians ) , Squa-
mata ( lizards and snakes ) , and Testudines ( turtles ) are represented
in the herpetofauna (116 species) of the Belem area; amphibians
represent 35.4 percent of the herpetofauna, and reptiles 64.6 percent.
The breakdown of species is as follows: caecilians — 3, salamanders
— 1, frogs — 37, amphisbaenids — 3, crocodilians — 1, lizards — 24,
snakes — 44, and turtles — 3. Further field work probably will reveal
several additional species of snakes, caecilians, and turtles, as well
as species of other groups.
Methods
Most observations and collections in the forests were made along
paths and boardwalks constructed several years previously. There-
fore, the data are biased to whatever extent the different species
are influenced by the narro\\', open areas maintained by continuous
human activity. Species distributions necessarily reflect my sam-
pling activity (Figs. 5-9).
An approximately equal amount of field work was done by day
and by night. Every frog, salamander, and lizard observed was
recorded by species, date, and locality (including hectare and
quadrat numbers from the labeled study areas). The distribution
of each of the 62 species was plotted on quadrat maps. Although
snakes, turtles, and caecilians were collected, the few numbers of
specimens of these groups precluded their inclusion in the analyses.
Environmental gradients aft'ecting the distribution of species within
four major forest areas was inferred by use of a contingency table
analysis. Resource partitioning was studied by means of field ob-
servations and analyzed by niche breadth and niche overlap anal-
ANALYSIS OF ECOLOGICAL DISTRIBUTION 15
yses. Following the analysis of species distribntions, the species
compositions of the major areas were compared and contrasted by
means of the Shannon species diversity formula, an equitability in-
dex, and coefficients of communities ( see appropriate sections ) .
Definitions of terms, as I am using them, and a brief discussion
of techniques of analysis are given below. The analytical techniques
are treated in detail in appropriate subsequent sections of this paper.
Major areas. — The region studied can be divided into several
geographical sections referred to as major areas. I have delimited
the artificial boundaries in such a way that each area possesses a
certain subjecti\'e uniformity with regard to physical environmental
parameters, such as xegetational physiognomy, light intensity, water,
and soil type. The quantitative analyses were carried out on data
obtained from four major forest areas: 1) capoeira-terra firme
transition; 2) terra firme-varzea transition; 3) varzea; and 4) igapo.
Resource partitioning observations were carried out in the following
major areas: 1) open and edge; 2) capoeira; 3) terra firme; 4)
varzea; and 5) igapo.
Habitat. — This term refers to the structural aspect of a niche;
it is that portion of the physical environment in which an organism
carries out its life processes. The physical environment supports
species in three major ways: 1) vertical zonatipn; 2) horizontal
distribution; and 3) temporal spacing.
Community. — A community consists of interacting populations
of animals. Each of the major areas included in this analysis has a
herpetofaunal community difi^erent from every other area. The
interaction and organization of each community is expressed in
terms of resource partitioning with regard to differential utilization
of the environment in space and time, species diversity including
both species richness and equitability components, and species com-
position and relative abundance.
Resource partitioning.— This term refers to the differential
utilization of the physical environment in space and time by dif-
ferent species. The result of resource partitioning is highly efficient
utiHzation of environmental resources.
Niche.— This is an abstract concept referring to the habitat and
biotic relationships of an animal. A niche can be thought of as a
hypervolume, consisting of numerous dimensions (Hutchinson,
1957 ) ; the dimensions are physical factors and biotic relationships
required by a species for survival. The physical factors of the en-
vironment making up the structural component (habitat) of the
niche exist independent of the species, but the entire niche, inclusive
of the position (biotic relationship) of the animal, does not exist
16 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
independent of the species. Therefore, the niche is a function of
the species. Formation of a particular niche is ultimately dependent
on the structural adaptations, physiological requirements and
capabilities, and correlated behavioral patterns of the species. No
two species have identical physical and biotic requirements, coupled
with identical structural, physiological, and behavioral attributes,
and therefore no two species have the same niche.
Niche breadfli. — This term is used to describe the spectrum of
any given dimension of the niche hypervolume. For instance, one
can speak of the food preference niche breadth of species A and B.
If species A eats 10 different kinds of insects and species B eats only
2 kinds of insects, species A is said to have a broad food preference
niche breadth and species B a narrow food preference niche breadth
relative to each other. Niche breadth as used in this paper refers
to the habitat niche breadth.
Niche overlap. — This term refers to the situation in which two
or more species have similar requirements with respect to some
dimension of the niche hypervolume. Niche overlap is a measure
of the association of two or more species. The measurements in this
study were obtained indirectly by the degree of coexistence of the
species in the various plots sampled.
Hahitat-generaUsts, intermediates, and specialists. — Habitat-gen-
eralists are species that utilize a broad spectrum of the environment,
as indicated by the contingency table indices; they are found in all
four major forest areas and have high habitat niche breadth scores
(16.0-32.0), as calculated from Levins' index. Habitat-specialists
are species apparently restricted in their distributions to one or two
of the major forest areas; they seem to \ive in a narrow range of the
environmental spectrum and have low niche breadth scores ( 1.0-
4.0). All other species are referred to as habitat-intermediates. In
most instances, the habitat-specialists are the least common species,
whereas the habitat-generalists are the most abundant.
Species diversity. — The concept of species diversity consists of
two components, species richness and equitability. The former is
the number of species, and the latter is the evenness with which the
individuals are distributed among the species. A community having
a large number of species in which the abundance decreases grad-
ually from the most to the least abundant species is considered to
have a high species diversity. According to Whittaker ( 1970), niche
differentiation results in greater species richness through time,
whereas a narrowing of habitat distributions tends to increase spe-
cies equitability. Some investigators propose that species richness
depends primarily on the structural diversity of the habitat, whereas
ANALYSIS OF ECOLOGICAL DISTRIBUTION 17
equitability is more dependent on the stability of physical condi-
tions. Apparently the more complex the \egetation is vertically, the
greater is bird species diversity (MacArthur and MacArthur, 1961;
MacArthur, MacArthur, and Freer, 1962; MacArthur, 1964, 1965; and
MacArthur, Recher, and Cody, 1966). Pianka (1967) proposed that
spatial heterogeneity is the most important single factor determin-
ing the number of species of lizards in any given area. One of the
most commonly accepted formulas to measure species diversity is
the Shannon function (Shannon, 1948). Pielou (1966) discussed
its use and disuse. The formula is used to describe an infinitely
large population and results in the average diversity per species.
Coefficient of community. — The coefficient of community (CC)
is a mathematical measure of relative similarity of samples from
two communities (Whittaker, 1970).
ANALYSIS OF ECOLOGICAL DISTRIBUTION
As discussed in the preceding section, each major area is a
complex of intrinsic physical environmental parameters, different
from those in other areas. Each species is adapted to a particular
range of each environmental gradient; the totality of environmental
gradients forms the structural niche, or habitat, of the species. One
must assume that habitat adaptation is based on the genetic make-up
of the indi\'iduals of the species in terms of morphology, physiology,
behavior, and life cycle. Based on the preceding assumptions, the
following hypothesis can be stated: The 62 species of frogs, sala-
manders, and lizards in the Belem area are distributed in such a
manner that environmental resources are partitioned; the conse-
quence of habitat differentiation is highly efficient utilization of the
environment.
Several techniques were used to study the ecological distribution
and to test the hypothesis; others were used to compare and con-
trast the species composition within each major area. To avoid con-
fusion, each analysis is presented separately. Included in each sec-
tion is an explanation of purpose and a presentation and discussion
of results; where applicable, advantages and limitations of the anal-
yses are indicated.
The distribution of frogs, salamanders, and lizards as taxonomic
groups within five of the major habitats is presented in table 1. The
varzea has the highest species richness, with 38 (61.3%) of the 62
species occurring there. Next in terms of species richness is terra
firme forest, with 36 species (58.1%). The area with the lowest
value is capoeira, with only 20 species ( 32.2% ) . The mature forest
18 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
Table 1. — Distribution of Amphibians and Lizards by Major Taxonomic
Groups in Five of the Major Areas. The top numliers are the number of spe-
cies of a taxonomic group in a given area; percentages are the proportion of
the taxonomic group in the area. Numbers in parentheses indicate the sum of
the coded relative abundance indices (Table 2) for the taxonomic group
in the area.
Terra Open &
Group Firme Varzea Igapo Capoeira Edge
Frogs and toads 20 22 13 8 24
37 species 54.0% 59.4% 35.1% 21.6% 64.9%
(36) (56) (23) (12) (80)
Salamanders 11110
1 species 100% 100% 100% 100%
(4) (4) (2) (3)
Lizards 15 15 8 11 10
24 species 62.5% 62.5% 33.3% 45.8% 41.7%
(27) (29) (16) (19) (23)
Total No. Species 36 38 22 20 34
% Total Species (62) .... 58.1% 61.3% 35.5% 32.2% 54.8%
Sum Abundance Indices 67 89 41 34 103
Average Species
Abundance Index .-.__ 1.86 2.34 1.86 1.70 3.03
areas likely are highest in species richness due to the greater vegeta-
tional diversity, yielding environmental heterogeneity, as contrasted
to second growth areas (capoeira) having less structural complexity.
Open and edge areas are relatively rich with 34 species (54.8%);
24 species of frogs (62.3% of the total anuran fauna) breed in the
numerous ponds in these areas. Abundance indices for each species
in each area were coded as follows: 0=apparently absent (none
observed ) ; l=not commonly seen ( 1-4 observations ) ; 2=moder-
ately common (5-15); 3=common (16-25); and 4=abundant (26 or
more observations). The average abundance index (obtained by
dividing the sum of the abundance indices for all the species in a
given area by the total number of species in that area) is much
higher in open and edge areas (3.03) than the next highest which
is the varzea forest ( 2.34 ) ; this is partially due to the large congrega-
tions of breeding frogs in open and edge areas. In addition, popula-
tion densities of lizards are higher in open areas than in the forest,
although this may be due to censusing methods; lizards are more
easily seen in open and edge areas than in the dense forest. The
ecological distribution and relative abundance of each species of
frog, salamander, and lizard are shown in table 2. It is evident that
certain species have a much broader range of ecological distribution
than do others. Figures 5-9 indicate the distribution of three species
relative to the distribution of water. The salamander, Bolitoglossa
ANALYSIS OF ECOLOGICAL DISTRIBUTION
19
Table 2. — Ecological Distribution of Frogs, Salamanders, and Lizards. Num-
bers indicate relative abundance of a species within an area, coded 0-4 as
follows: 0=Apparently absent (none observed), l=Not commonly seen (1-4
obsei-vations ) , 2=Moderately common (5-15), 3=Common (16-25), and
4=Abundant ( 26 or more observations ) .
Terra Open &
Species Firme Varzea Igapo Capoeira Edge
Pipa pipa 0 10 0 0
Eleiitherodactylus lacrhnosus .1 0 1 0 0
Leptodactijhts inarmoratus 4 4 0 3 0
Lcptodactt/lus uiystaceiis 10 0 0 0
Leptodactijlus oceUatiis 0 0 0 0 4
Leptodactyhi.s pentadactylus I 0 0 0 0
Leptodactijlus rliodomystax ___. 12 0 0 0
Leptodactyhis wagneri 14 4 0 2
Physalaemus ephippifer 4 4 114
Physalaemus pctersi 3 4 0 0 2
Biifo luariims 0 0 0 0 4
Biifo ty))honius 4 4 0 2 0
Dcndwhates trivittatus 2 0 0 10
Dendrohates vcntiimaculatiis ..0 2 3 0 0
Ilyla haumgardneri 0 0 10 4
Hyla boesemani 0 0 0 0 4
Hyh calcarata 13 0 0 0
Hyla egleri 2 3 2 14
Hyla geographica 0 4 0 0 2
Hyla goughi 12 10 4
Hyla granosa 1 2 3*0.0
Hyla Icticophijllata 0 2 10 4
Hyla mclanaigyrea 10 0 0 4
Hyla miinita 0 0 0 0 4
Hyla multifasciata 0 2 10 4
Hyla nana 10 0 0 4
Hyla raniceps 0 1 0 0-4
Hyla rondoniae 0 10 0 0
Hyla rubra 13 3 14
Hyla sp. (large rubra) 110 0 4
Hyla sp. ( n/b/fl-like ) 4 4 12 1
Ostcocephalus taurinus 0 0 0 0 1
Phrynohyas venulosa 10 113
Phyllomcdusa bicolor 0 2 0 0 2
Phyllomedu.m hypochondrialis 0 0 0 0 4
Phyllomedusa vaUhmti 0 10 0 0
Sphaenorhynchus eurhostiis ..._ 0 0 0 0 3
Bolitoglossa altamazonica 4 4 2 3 0
Gonatodes humeralis 4 4 3 3 0
Hemidactylus mabouia 0 0 0 0 4
Thccadactylus rapicaudus 10 0 0 0
Lepidoblephanis festae 0 110 0
Anolis fuscoauratus ...^_... __ 3 3 110
Anolis ortoni 10 0 0 0
Anolis punctatus 110 10
Iguana iguana 0 10 0 1
Plica umbra 3 112 0
Polyclirus inarmoratus 11110
0
0
1
4
2
1
1
1
4
4
3
2
1
0
0
0
0
0
3
4
0
0
0
0
0
0
0
3
1
0
0
0
0
0
0
1
4
4
2
0
2
0
0
1
2
0
0
0
1
0
1
2
20 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
Table 2. — (Concluded)
Terra Open &
Species Firme Varzea Igapo Capoeira Edge
Tropidunis torquatus 1
U ranoscodon siiperciliosa 1
Mabutja mahomja ._ 3
Alopoglossus carinicaiidatus ... 0
Ameiva ameiva 1
Arthrosaura kochii 1
Cneniidophorus lemniscatus ._._ 0
Crocodilurus lacertina 0
Dracaena guianensis 0
Kentropijx calcaratus 4
Leposoma percarinatum 0
Prionodaciijlus argulus 1
Ttipinambus nigropunctatus .... 1
altamazonica, is abundant in capoeira, terra firme, and varzea areas,
but few individuals are found in the terra firme-varzea transition
area. LeptocJactylus marmoratiis occurs predominantly in well-
drained areas, but Gonatodes humeralis tolerates wet and dry areas.
Resource Partitioning
Extensive field observations were carried out in an effort to
discern whether species do indeed partition environmental resources.
Resources examined were those aspects of the habitat which are im-
portant to the daily and seasonal activities of the species as follows:
frogs — standing bodies of water, calling sites, and vegetation and
ground area used for daily activities; salamanders — vegetation used
for nocturnal activities; lizards — vegetation and ground area used
for basking sites and other diurnal activities.
For purposes of analysis, the environment can be divided into
vertical and horizontal components such as arboreal (high and low),
terrestrial, aquatic margin, and aquatic. The distribution of species
in these subdivisions of each of the five major areas is shown in
table 3. Most species studied are either low arboreal or terrestrial.
Table 3. — Distribution of Species of Frogs and Lizards Within Subdivisions
of Five of the Major Areas. Numbers preceding hyphens are frogs, and num-
bers following hyphens are lizards.
Terra Open &
Subdivision Firme Varzea Igapo Capoeira Edge
Arboreal (high) 0-3 1-3 0-1 0-2 1-1
Arboreal (low) 9-6 11-6 9-5 4-4 16-2
Terrestrial 9-6 7-5 3-2 4-5 4-7
Aquatic Margin 1-0 2-0 1-0 0-0 3-0
Aquatic 0-0 1-1 0-0 0-0 0-0
ANALYSIS OF ECOLOGICAL DISTRIBUTION 21
Spatial overlap among some species does exist. Salamanders and
Hyla sp. (m/;ra-Iike) overlap greatly in their utilization of low
vegetation in terra firme and varzea areas at night, presumably for
obtaining food; also present, sharing the same vertical component,
are numerous sleeping lizards (Gonatodes humeralis). Although
these three species are the most abundant vertebrates using this
aspect of the environment at night, the population densities appear
to be so low that it is unlikely that significant interspecific competi-
tion exists.
There is evidence of breeding site partitioning in frogs, probably
indicative of differing requirements of various species. Some tree
frogs, such as Hyla haumgardneri, H. egleri, and H. goughi breed
in diverse types of ponds and swamps, large or small, deep or
shallow; apparently the frogs require only standing water and
emergent vegetation. On the other hand, Hyla mimita, H. raniceps,
and PlujUomeduso hypochondrialis are found in only some of the
same areas as H. haumgardneri, egleri, and goughi. Hyla raniceps
breeds only in larger bodies of water, at least 8 m by 15 m, usually
at least 0.6 m in depth. Fhyllomedusa hypochondrialis is restricted
to ponds bordered by dense vegetation. The distribution of H.
mimita is more difficult to interpret; the frogs occur in all types of
areas, but without any regular pattern. For instance, numerous
males call from one pond and not from a nearby pond having
similar size, water depth, and emergent vegetation. The population
density of this species appears to be lower than those of H. haum-
gardneri, H. egleri, and H. goughi. Perhaps male H. mimita attract
other males to an area for the purpose of forming breeding congrega-
tions. This formation would be of greater importance to a less
abundant species than to a more common one and would explain
the fact that usually these frogs call in groups of at least 15 in-
dividuals in contrast to H. haumgardneri, H. egleri, and H. goughi
which often call in groups of 10 or less.
Many species of frogs which breed sympatrically demonstrate
calling site segregation (Tables 4 and 5). Most species characteris-
tically call from a certain physiognomic type of vegetation, at a
relatively uniform height from the water. The type of vegetation
utilized is correlated with the body build and size of the animal.
Large, heavy frogs generally call from the ground, sturdy vegetation
near the ground, or from branches of trees; small frogs usually call
from grass stems or leaves and small branches from emergent and
edge vegetation. Some species have a broader range of calling sites
than do others. For example, Hyla goughi commonly calls from
both emergent and edge vegetation, 0.05-1.5 m above the water,
22
OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
<
>.
w
U
c
63
to
X IXXX iX iXXXXX
X IXXX IX xxxxx
X IXXX IX xxx^x
XiXXXiX XX XXX
X IXXX IX IXXXXX
X IXXX IX 'XXXXX
X IXXX IXXXXX 1 1
X iXXX iXXXXX 1 1
X IXXX :XXXXX i I
XXXX XXXXX II
XXXX XXXXX II
X IXXX IXXXXX 1 1
^ X X
X X X X X X X 1
'^ ^ X
X X X X X X X 1
XXIX
X X X X X X X 1
XXIX
X X X X X X X 1
XXIX
XX X X X X X 1
XXIX
X X X X X X X '
XXXXXXXX XXXXX
xxxxx^xxxxxxx
xxxxxxxxxxxxx
xxxxxxxxxxxxx
xxxxxxxxxxxxx
xxxxxxxxxxxxx
XXXX
XX X X X X 1 X
y y X X
X X X X X X IX
XXXX
X X X X X X IX
XXXX
X X X X X X IX
XXXX
XX X X X X IX
X X X X
X XXXX X IX
XX 1 X
X 1 1 XXX 1 1
X X 1 X
XI 1 X XX 1 1
X X 1 X
XI 1 XXX 1 1
X X i X
XI : XX X 1 1
XXIX
XI 1 XXX 1 1
X X 1 X 1
XI 1 XXX 1 1
X X 1 X 1
1 1 IXXX IX
X X 1 X 1
1 1 IXXX IX
X X 1 X 1
1 1 IXXX IX
X X 1 X 1
1 1 IXXX IX
X X 1 X 1
1 1 1 XXX 1 X
X X 1 X 1
1 1 1 XXX 1 X
tC tl "U
C^ O w
^ < < Hi 1 1 1 1
-^ :<: -^ ~ ~ ~ :, -^ :j :j
tJCiXttWtJIMtC
C/2
a
1-9
a
<
■~n
5:5:y;55c5i:^_S_2_=_2.2_2
h ^
^ «
■ S
^ g = 5 _
ANALYSIS OF ECOLOGICAL DISTRIBUTION 23
whereas PliyUomecIusa hypocJwndriaUs always calls from edge veg-
etation usually 0.6-1.5 ni abo\e the ground or water. Complete seg-
regation of calling sites does not exist for all species in all areas.
Segregation is partially dependent on species composition at the
site, relative abundance of the calling individuals, and on the size
of the breeding site relative to the population densities. Generally,
in large, mixed congregations segregation tends to break down, and
the frogs call from whatever sites are a\'ailable. Interspecific com-
petition for calling sites is probably significant during times of much
reproductive acti^'ity. Segregation is more pronounced in large
areas with distinct physiognomic vegetational di\'ersity than in
smaller areas with less calling site diversity. A commonly accepted
explanation for the evolution of partitioning of calling sites is the
resultant tendency to reduce the chances of interspecific mating.
However, because segregation breaks down in large, mixed congre-
gations at the time it is most needed, I propose that calling site
partitioning exists due to the structural and behavioral attributes of
each species rather than as a necessary reproductive isolating mech-
anism; advantages likely include improved mating efficiency and
reduced energy expenditure.
There is a definite replacement of several species of tree frogs
at breeding sites because of calling site overlap. Hyla hoesemani,
H. multifasciato, H. raniceps, and H. rubra all call from thick
clumps of emergent vegetation, usually within 20 cm of the water.
Individuals of all four species call from the same swampy areas,
but not all at the same time; the only two of these species ever
found calling sympatrically and synchronically are H. hoesemani
and H. rubra, the two smaller species. Every congregation of Hyla
sp. (large rubra) observed was found calling in association with
//. rubra. Male HyJa sp. (large rubra) call from the ground or
low, thick vegetation. They seem to be dominant over H. rubra as
indicated by calling site displacement of H. rubra when the two
species call sympatrically. Hyla rubra usually calls from low vegeta-
tion, but when Hyla sp. (large rubra) is also calling from the area,
the former calls from higher \ egetation.
Perhaps some syntopic species (species with similar habitats)
coexist with minimum interspecific competition as a result of
temporal partitioning of the environment, in terms of diel and sea-
sonal acti\ities. For example, the nocturnal gecko, Thecadaciyhis
rapicaudus, is likely the temporal replacement for diurnal lizards
feeding on similar species of insects and utilizing the same habitat.
The two species of dendrobatid frogs use the same forest floor by day
that several species of leptodactylids utilize at night. Frogs demon-
24 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
to
o
U5
0)
O
CO
C/3
bo
a
15
U
O
bc
C
•a
o
■■D
■■a
C8
ID
W
ca
<
O
Q
w
2
CM W)
13
C
3
p
O
c
c
•X3
03
O
M
X I X xx
I I I I I I X I I I
I I I IXXIXXXIXXXXXXXIXX
I I I I I I I M I I I I I I I I I I X
13
c
s
o
o
^ c
< o
O "-g
J— I _2
be
QJ
>
O
i-I
I I X I I I I I I I I I I I I I I I I I
IIIIIIXIIXIIXIIIIIXIIII
I I I I M I I I I I I I I I I I I
T3
B
P
o
<:
w
N C
rt .2
>^
Ml
b£
X I XX X I
I I X I I I
I I I I I I
I I I I I I ! IXXXXXI I IXIXI IXI
I I I I I I I I I I I I I I I I I I I I X I
C/5
'5
a
CO
3
?? 3 *- ^
s a 5= 5.
« jj a .s•
Jr ^ i) ijs
C Q s '^i
to
-d ^ ~s 2 '^ ~
to
25
to ■
to
-2 ^
"a -2
2
O "S ?n ?^ 5 .= = to •= =
■ o 2
to
O
.to ~-,
« c ii
JS-CtototjCtt'tx-SisSC
to
c a a ^ fe
►J ^ S w i:^':^';:^';^';^ i^ i:^'::^':^':^'!^':^'^'^'^ § c^
ANALYSIS OF ECOLOGICAL DISTRIBUTION
25
y:
§13 o
« y c/.
tn c c
OJ C8 ,1)
a> lu s-
M-. ^ o
C nj O
.» O rt
:^ i
g c« 3
o a; t«
Trt ^ ^
« r- o
to
o TS o
II > a
1— I CS OJ
o c ■'•
o 3 «
^ cS (U
H
5 !«
^ O
bC bC ,
2.S S
-o o
bC'O =«
m o
£
u
^
y
!>
C
O
CU
(—'
'TS
t2
'>
1
QJ
1
II
CD
ffl
W
. ^^
►4
•<
"«
H
s
U <J <; cq
oi ca (M
cq cq
CQ Cq cq (J (J <
m <:
CO
a <i< uu <:u
--I <: <:
!j (J <: c<i oq
<; <M
PQ <; fQ oq
c<)
<
C<1
!U
<N CM < OJ CO OI
U
CQ < CO
m<<
<
U
CO
I I
<
a
C/2
- a PQ -S- CO CO CO p3 «<;<<«< cQ < CO cq C5 (25
CO d Ol '^
CQ
CO
UP3UUUUUc^ifQ<!<p3U<P5<;m
Eq<lPQpQ<;cq pQcoco
CO
<: «
<
CO
u CO <: < CO
CO ca
uc^ip5<uuum
05 « <: 05
CO
(J CO <| < CO <| u
P5 CO CO CO
oicq<;o5<;coco i<l05.
05 <
CO
fO !
m i <:<^
i «
CO CO
cl
+ + +
o in o in lo
in oi o oi CO
■— I f— I 1— I r
I
in
oi
o o in in
in t>
o o
'st^ CO
+
S o -■
o m >— I
o
+ +
o o
OI in o in o
CO ■— I oi
in '
I
o
o
o
o
05
OOOOOt^OOOCOiC
inCD-^O5COCOC0C0t^
I I I I I vi. -J.
o o in o in in o
T}< in CO 00 CO .-I CO
45
2 ■S
S-2
a.
CO
CO
o o u
c a a
ns -cj -TS
o o o
1^ i-- t^
a. a. a,
!^ ca i;^
k4 ►J hj
=o 5 s;
S s o
so o
s s
o
^3
s s
"^ « „ ~
I = s I
t: ^ CO o
2 = «^ >i^ — ^
~J Q o a bjD 5d
g -a ~a o ^ t^l
'^ a Q a ^ ^
■3 ,1~
bc y
bc
<3
~ -W M-,
&5
S
CO
O
V dj a
a; 2 S
bc^ §
i? 5 "^
o •« ^
o
1^ ,J£
« S^
■■s «
a 5 1^ ■'^ -!
cc
= a
CO
a. cc
CO
-5
CO
ft
o
Q^-5-2-2-5-2-2^^-5
f.^ r-^ .-^ -^ -^ -.-^ --. -.-^ ->i» -*.
<3 a a ^ J3 si'
qq:^:!':!':^:^':^^^"^'^'^'^'^'^'^'^'^^^^
Co
.^
CJ
O
to
«
a.
en
26 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
strate seasonal partitioning of the environment in terms of breeding
activities (Table 6). Hijla muUifasciata and H. raniceps both call
from low vegetation in swampy, open areas; they are rather large
tree frogs with similar mating calls. The former is a wet season
breeder, whereas the latter is a dry season breeder. Most of the
frogs are nocturnal (83.8%) and most of the lizards are diurnal
(91.7%); the salamander is nocturnal. In the terra firme and varzea
forests, 50 percent of the frogs, salamanders, and Hzards considered
as a group are diurnal; the distribution in the igapo forest is similar,
with 45.5 percent diurnal and 54.5 percent nocturnal. Most of the
species in open and edge areas are nocturnal (73.5%); breeding
tree frogs account for most of this distribution. On the other hand,
most of the species in the capoeira area are diurnal (70%); over
half of these species are lizards, many of which are heliotherms
(Table 7).
Contingency Table Analysis
The contingency table analysis technique, developed by Wil-
liams ( 1952 ) as an extension of Fisher and Yates' ideas for dealing
with frequency counts in two-way tables, is employed here for two
reasons: 1) to measure the degree of association between species
and plots; and 2) to partition the species-plot association into in-
dependent components representative of environmental gradients.
The analysis was carried out on 20 species of frogs, salamanders,
and lizards from 44 sampling plots, each 20x30 m, from the capoeira-
terra firme transition, terra firme-varzea transition, varzea, and igapo
areas. All plots received approximately equal amounts of sampling
time from mid-January to the end of July. The 20 species were the
only species of frogs, salamanders, and lizards found within the
boundaries of the particular plots analyzed. The total sample in-
cludes 1218 individuals (Table 8). Most individuals were not re-
moved from the habitat, so the relative abundance indices are pos-
sibly inclusive of re-counted individuals; each observation was
treated as a unit indicative of species-habitat association.
Table 7. — Comparison of Activity Cycles of Amphibians and Lizards in Five
of the Major Areas. AbsoKite number of species and percentage of species
within each area are given.
Period of
Activity Tana Firme Varzea Igapo Capoeira Open & Edge
Diurnal 18 19 10 14 9
50% 50% 45.5% 70% 26.5%
Nocturnal -- 18 19 12 6 25
50% 50% 54.5% 30% 73.5%
ANALYSIS OF ECOLOGICAL DISTRIBUTION
27
c
<!
J2
o
c
C
o
u
■73
CS
o
s
X
o
o
3
00
H
ffl
<;
H
I— ( cq
03 =2
>05
N O
5 "a
C C
H
H c
4- °
2 x:
a 2
a
C/2
00 CM 05
CO 05
lO
05 cq
: ic ^
CO
(N 00
1 CO
in
(M
cq
«N
cq CO
CO
CO
.— < IC 1— I
00 CD Ol O
i ^
■— (
Tp
00 CO ^
oq
i-H
(M
O
cq CM
■-H -T — 1 t-
: CO 00 CD (M -r -p
CO
t^ (M (M ^
1 »o ^
CM
I-H
1 "—1
iO
CM
lO "* o>
CO r-l
1— I CO 00 00
<M CO
lO 05
O CM
CO rf
CM ■— I
CD O O CM
CO CO 05 -— I
>— lOOCDCMCOCM'^'^'^CO
CMi— ICOCMt— 1-^ UOi— I
CO rH
CJ
■"^ tr i J^ ti' ~
CQ
X S
3
•^ «i.^ r^
?3 = 3
a
CJ
c
o
ft
•i ii c
~ 2 ^
« i a
~ t o
cj si-i: ~ = o s o
P^ r". ^ ^^ rs , r^
■^ 2.
<3
O
^ ^ C^ ^
a ~ a
^ ^ ^ S-f
a;a:a;a;a;a:,jKja,ei,
5j ,C
~2 2 5
Cj
5 ~ 5 « :;
^ ^ »_ K.^ (^^ i^_. )^_ N»_ ^ ^ ■— » '?*- .V "^ ^ IB k J -» ^*^ '^" >-~-, r^^ r'^
o =: ^ 5u
CO
00 O
CO I-H
00 o
CM
•t-H
>
~ •- .a
o
a
C/3
C
c
Kj;5a,;D H H
28 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
Capoeira
Terra Firme
Transition
ndex I
Fig. 12. Scores on index I plotted against scores on index II for each of
44 plots. Each dot represents the position of a particular plot relative to the
X and y axes. Index I is a moisture gradient from dry (negative) to wet
(positive). Index II is a vegetation density gradient from dense ground cover
(negative) to grassy ground cover (positive).
The data were assembled into a species X plot table; the species
frequency counts represent the number of individuals of each
species which occurred in a particular plot. Williams (1952)
showed that when actual environmental measurements were un-
available, scores could be calculated from the data of the contin-
gency table by simply using those sets of scores for which there is
maximum correlation. The interpretation is feasible because the
scores are adjusted to have a mean of zero and a variance of one.
The computer print-out for the analysis consists of a series of
indices, each representing an environmental gradient or a composite
of such gradients. Each index maximizes the measure of association
between the two sets of variables — species and plots. Index scores
relative abundances are presented in table 9. The results are
presented by Cartesian (x, y) scattergrams of two sets of scores
(Figs. 12-17). In this way, two gradients (two indices) can be
studied simultaneously and their interaction examined. Species or
plots having similar index scores appear close together on the
diagram. Thus, ecologically similar plots and species with similar
distributions can be identified.
No actual environmental measurements were taken; resource
ANALYSIS OF ECOLOGICAL DISTRIBUTION
29
o
■£ CO
(t>
C
M
•I-*
s
t3
3
OJ
z
03
tn
0;
O
T
(J
y;
C/2
^
X
--r
0 4-*
CO O p TT CO C^l CO ^ t--_ 00 t--_ '^. »0 CO P QD p CO '* O
oiio^^'^coiOTiiOi-HCDcdi/Joooi^-^cdioco
1— I ^H CO (M 1-H
'S X
^°
pa oa
-S.2
-o a
•—I c«
Sx
c
0)
c
<A
H
>>
o
c
a>
M)
C
o
U
<a
XI
a
o
U
0) o
O -I-'
S a
< '^
en
CC ^
«H -^
c o
o 'z;
•J3 >>
03
05
H
OCOCOOSCO-^t'MCDOip
Tj^Tpinosoocooooo
05 CM CO 03 00 O-l ■*_ ■*.
coaJt-^oiioooJin-^'t-'odcDCDt---
^(M (M'-l ICOII I 1
II I I
— lOCOCOKtjlO'-Ht-COCDCDi— I— l^CD05COmcO^
t- o p M o 05 1> Tf in CO CO 00 p -<* p CO i> p p p
Qc^i^i^TiJc<ic4ioc<i-*'>-HO»ncop-^'poa50o
(M
I— I M I i-H C^ ^ (M
05 05 05 0 1^
'1^1
c^ t- CD csi 10 lo in i> CO p in p »n p c<) i> CO 00 p »n
ind^''N'--<«oi'-Hcdinc-i'*'-Ht7'Hcocoo^'-|
•coin— lO-ll'-H II r-<llll i'-j<
^^^^inini>i>t-(Min(M^'-Hcoc35053;cO'H
J/-I CO ^ t- O ^ t-; It-; p p p p ^. ^. "* P P °0 '~1 <^
c^^^-^o6ocd^ajp<^^o<^iopop<^iO'— iMcO'^
,— I r-H 01 05 CM
I
I I
I
a;
o
c
OS
c
(0001 X)
aAC}B[nuinoov
odB§i
B9ZJB^
UOfllSUBJX
BazjB^-9uuij BJiax
uopisuBJX Qui-iIjI
^cooa505^oa^(cc^l:^moco
Q ,—1 ,— ii— icO^lOOSt— {i— (COi— I
CO
Tf —I
c?
t- in 05
in 00 CO
CD 05 t~-
000
000
d o d
CO CO
o t-
—I o
o o
d d
CO
■—I
o
o
d
in M CO ^
1 C<1 00 O CO
i t^ O 00
1 1— I O ^H
1000
o
0000
in
1— I
in
'^
o
d
C5
o
O
' d
I— ( N
00 CO
CO Ir-
(M CD
o o
d d
i> 00
00 o
<N o
o o
d d
(McctcoTt^inrrfM i^
oocooMincoc-i i^i
O^Ot^fMOO 10
O O p P p p P ; P
'•d d d d d d d ' d
in
CO
00
o
d
CD CD
1 — I ^H
o o
o o
d d
O ^ 'i^ CD
tT t^ (M Oi
O 05 I> CO
Tjl 1— I 1— I ^H
.-H O O O
d> d) <z> d>
CO M CO TT 00 00
Tfi t^ — ^ CD (N C.1
(M 0> CO ^ CO CO
o c<i :::! 2 S 2
o ^ o p p p
d d d d d d
;^>j S CL :::3
30 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
Terra Firme-
Varzea
Transition
ndexl
Fig. 13. Scores on index I plotted against scores on index III for each of
44 plots. Each dot represents the position of a particular plot relative to the
X and y axes. Index I is a moisture gradient from dry (negati\'e) to wet
(positive). Index III is the vertical distribution of species found within the
plots from terrestrial (negative) to low vegetation (positive).
requirements for each species were analyzed indirectly by assuming
that a given sampling plot provides necessary resources for the
species found therein. For this reason interpretation of the indices
is inferential.
Etwironmental gradients. — The first four index scores from the
contingency table analysis were analyzed in an attempt to : 1 ) de-
termine the major limiting environmental parameters affecting the
ANALYSIS OF ECOLOGICAL DISTRIBUTION
31
20
10
H
^ 0
■D
c
10-
20
Capoeira-
Terra Firme
Transition
Varzea
•-^414
Terra Firnne-Varzea
Transition
0
Fig. 14. Scores on index II plotted against scores on index III for each of
44 plots. See figures 12 and 13 for explanation of dots and indices. The three
dots not included witliin forest boundaries all are igapo plots.
distribution of species; 2) characterize the four major forest areas
in terms of those limiting factors relevant to frogs, salamanders, and
lizards; and 3) identify the habitat of each of the 20 species in terms
of the environmental parameters represented by the indices.
The first index indicates a moisture gradient from dry (low
values ) to wet ( high values ) . Moisture probably is the most critical
factor affecting the ecological distribution of amphibians and rep-
tiles in the study area.
Probably the next most critical limiting factor is the physiognomy
of the vegetation. The second index is indicative of vegetation
32
OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
X
-SO
10
(p
23.6
©-
523
617
t ®
©-
-^-21. 2
©
®
@
0
®r
10
0
Indexl
10
20
Fig. 15. Scores on index I plotted against scores on index II for each of
20 species. See figin-e 12 for explanation of the indices. Each circle repre-
sents the position of a particular species relative to the x and y axes. The
species numbers are associated with species names in table 9, p. 29.
ground cover. The spectrum is from dense ground cover (low
values) to grassy areas (high values). The second index may also
indicate light intensity, resulting from the structure and density
of the vegetation. In general, areas with dense ground cover are
darker habitats than are grassy open areas.
The third index probably is a combination of factors affecting
vertical distribution. Terrestrial species have low values, and spe-
cies which inhabit low vegetation have high values. Lizards found
on tree trunks and along the boardwalks have intermediate scores.
The fourth index seems to be a composite of many factors.
Some of the following may be involved, but no one of them is
responsible for the separation of the plot or species scores: 1)
temporal activity (diel and seasonal); 2) organism size; 3) phylo-
genetic position of organisms; 4) heliophilous versus sciophilous or-
ganisms; 5) niche breadth of organisms; 6) abundance of animals
within plots; and 7) solitary organisms versus congregations. The
fourth index segregates the following species pairs, which are
similar on the basis of the first three indices: Hyh rubra and Htjia
sp. (rubra-\ike), Hyla geograpJiica and Uranoscodon superciliosa,
and Hyla rubra and Bolitoglossa altamazonica.
ANALYSIS OF ECOLOGICAL DISTRIBUTION
33
■
' ©
©
1 1
1 1 — :
20
-
@-28l_
©
©
©
10
@
@
"
1
©
X
1°
©
©
©
08)
© ©
©
©
10
-
©
•
®
-
20
1
, ©
1 1 1
1
10
0
10
20
Index I
Fig. 16. Scores on index I plotted against scores on index III for each of
20 species. See figures 13 and 15 for explanation of indices and circles/species
numbers respectively.
When the scores of the first three indices for the 44 plots ( Table
10) are plotted against each other, it is possible to characterize the
four major areas in terms of the environmental gradients analyzed
(Figs. 12-14). Likewise, when the species scores are plotted, it is
possible to get some understanding of the habitat of each species
in terms of the environmental gradients inferred from the indices
(Figs. 15-17).
Plot index scores. — Four capoeii"a-terra firme transition plots
34 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
Table 10. — Contingency table indices for each of 44 plots analyzed. Plot
numbers are located on figures 10 and 11. Indices are plotted on figures 12-14.
Plot No. Index I Inde.x II Index III
1
4
11
12
15
16
17
18
19
20
22
23
24
25
26
27
28
29
30
31
33
42
43
44
45
46
49
50
51
52
53
54
65
66
67
68
69
70
71
72
73
74
75
76
3.66
-6.77
10.62
-3.13
-7.83
20.81
-6.50
-2.36
8.60
-5.57
-4.80
14.94
-5.74
0.88
-1.96
-0.67
-1.32
-8.52
-8.55
2.79
-2.38
-7.60
6.06
-1.55
-10.78
4.77
-4.29
-12.62
5.93
-2.88
-12.39
6.03
-3.31
-10.17
3.26
-1.02
-12.35
7.34
-3.19
-1.16
-0.88
-11.55
3.26
-3.47
-13.31
-2.30
-1.51
-9.13
1.87
-6.29
-25.82
-5.26
-0.48
-10.31
3.02
-1.95
-8.83
-0.67
-1.45
-11.32
2.82
-2.87
-6.13
-12.05
5.54
-2.63
-12.58
6.11
-2.95
6.64
-9.18
5.09
0.55
-8.65
10.87
2.34
-11.48
22.55
5.89
-5.08
-4.86
1.74
-5.26
-0.18
8.56
-8.12
-13.45
6.33
-8.77
11.53
3.38
-4.85
1.81
3.44
-5.34
-0.96
22.25
41.36
13.27
14.31
-5.39
-14.28
15.00
4.45
-5.85
13.83
0.45
-5.74
13.92
-6.85
-2.69
15.77
-4.80
-1.69
12.38
-4.32
-8.75
15.28
-5.04
-7.98
9.76
-4.10
-7.00
8.98
-4.33
-5.59
9.82
-4.66
-9.33
12.64
-5.78
-15.93
were studied. In general, these plots are characterized by low to
mid-range \alues on the first index, fairly low on the second, and
high on the third (Figs. 12-14). The area is towards the drier end
of the moisture spectrum and near the denser end of the vegetation
ANALYSIS OF ECOLOGICAL DISTRIBUTION
35
20
10-
0
'^ 0
10-
20-
aiD
1
1 1
1 1 :
-
©*6l.7
(4)*52.3
© ®
©
^
©
©
© ©
©
r
o
©
-
-
©
*
®
-
- 1 1 ©
1 I
1 1
10
0
Index IT
10
20
Fig. 17. Scores on index II plotted against scores on index III for each of
20 species. See figures 12 and 13 for an explanation of indices and figure 15
for an explanation of circles.
density spectrum. The value on the thu-d index suggests that the
herpetofauna of this area is predominantly found on low \egetation
rather than on the ground.
Nineteen plots are terra firme-varzea transition areas, and can
be divided into tsvo groups, dry transition and wet transition (Fig.
10 ) . The two transition areas are clearly segregated when the index
values are plotted against each other (Figs. 12-14). The entire
transition zone is characterized by low to mid-range values on the
36 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
first index, high on the second, and low to middle on the third.
The area represents an intermediate zone with regard to the
physical environmental parameters, except on the second index,
indicating that the ground cover is relatixely grassy. One plot is
extremely low on the third and fourth indices and segregates from
the other transition plots. The low \'alue on the third index is ex-
plained by the many terrestrial leptodact\lids found calling from
temporary puddles.
Nine varzea plots were analyzed (Fig. 11). There is a very
small range of variation on the first and second indices, but a wide
range on the third and fourth. In general, most plots have a fairly
high value on the first index, low on the second, and from low to
high on the third. The xarzea is a wet em ironment \\'ith relatively
dense ground cover; the organisms are neither predominantly ter-
restrial nor inhabitants of low \ egetation.
Twelve igapo plots were studied. These plots have the highest
values on the first index, indicating that the igapo is the wettest area.
Most values on the second index range from low to middle and most
on the third are low. The igapo forest has a relati\ely dense to
intermediate vegetation ground cover. Most of the lizards are ter-
restrial or are found predominantly on the boardwalks. When the
indices are plotted against each other, one plot is segregated from
the other igapo plots by high xalues on the second and third indices
(Figs. 12-14). The second index score is explained by the presence
of large clumps of tall emergent grass in the plot. Two species of
tree frogs not found elsewhere in the igapo forest utilize the grass
for calling sites; this creates a higher third index score than those
values for igapo plots in which there are many terrestrial lizards.
Species index scores. — The Cartesian plots of index values ( Figs.
18-21) and the bar diagrams (Figs. 22-24) illustrate that each spe-
cies has requirements and tolerances with regard to the environ-
mental gradients. Several trends represented by correlations be-
tween species abundances and availabifity of a particular resource
are evident ( Figs. 22-24 ) . For example, index I represents a mois-
ture gradient; those species with the highest positive values are
those found in association with wet areas. Each of the seven species
with the highest scores (Leptodactylus wogneri, Kentropyx cakara-
tus, Mahmja mabouya, Dendrohates ventrimaculatus, Hyla egleri,
H. granosa, and H. boumgardneri) is most abundant in the igapo
forest, less abundant in the \arzea forest, still less common in the
terra firme-varzea transition area, and rare in the capoeira-terra
firme transition area, if found in the last two areas at all. The two
species with extremely low negative scores on the first index (Lepto-
ANALYSIS OF ECOLOGICAL DISTRIBUTION 37
dactijlus marmoratus and Bnfo typlwnius) are more abundant in the
terra firme-varzea transition area than in capoeira-terra firme transi-
tion plots, contrary to what one might expect. The distribution is
better understood when the second index scores are considered;
both species have positive scores, but neither one is extreme. Ap-
parently these terrestrial species inhabit relatively dry areas but
avoid the open areas characteristic of capoeira forest in preference
to the denser undergrowth of high ground, dry terra firme-varzea
transition areas.
In most instances, those species with scores closest to zero are
the most abundant. This is probably because those species requiring
neither extreme (considered generalized) are able to utilize more
of the en\'ironment. If more of the environment is potentialh'
available for exploitation by a species, it can be assumed that the
potential carrying capacity of the environment for that species is
greater than that for a specialized species restricted to a particular
habitat. Gonatodes humeralis is the most abundant of the twenty
species and has scores near zero on each of the four indices. The
three next most abundant species, Lepfodactyhis marmoratus,
Kentropyx calcaratus, and Bufo typJionius, have scores relatively
close to zero on all indices except the first. Species with extremely
high positive or low negative scores on index IV are relatively un-
common.
Most species of lizards do not have extreme values on any of the
environmental gradients. Gonatodes humeralis, Kentropyx cal-
caratus, and Mahuya mahoiiya are the only species found in all
four major areas; none has extreme index values. The combined
cumulative relative abundances (the three species from the four
areas) represents 509 individuals, or 41.8 percent of the total
sampled herpetofauna. Gonatodes humeralis is less abundant in
the igapo forest than in the other three areas, whereas K. calcaratus
and M. mahouya are most abundant in the igapo forest. The other
four species of lizards are relatively uncommon in all of the areas.
Anolis fuscoauratus, Leposoma percarinattim, and Plica umbra have
no extreme index scores; the first two species are near zero on the
moisture gradient, and P. umbra is near zero on the fourth index.
Uranoscodon superciliosa is relatively generalized with respect to
all the environmental gradients except vegetation density; the
score on the second index is low, indicative of its occurrence in
areas of dense vegetation.
In general, the amphibians demonstrate more extreme environ-
mental requirements than do the lizards. None of the thirteen
38 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
o
O
O
c\J
UOjijSUDJl 9UJJIJ DJjai-DJiaodDQ/ UOjilSUDJl D9ZJDA-8UJJ1-J
OOOIX souopunqv 8A!;D|8y
O
DJJ8J_
ANALYSIS OF ECOLOGICAL DISTRIBUTION 39
species is found in all four areas. The salamander, BoUto^Jossa
aUaniazonica, has a score near zero on the moisture gradient, but
exhibits extreme scores on the second and third indices, indicative of
the occurrence of individuals on low vegetation in relatively dense
areas. Pliysalaemus petersi is generalized with regard to all of the
environmental gradients. Bufo tijphonius and Leptodactylus mar-
moratus are specialized only with regard to the moisture gradient;
they inhabit relatively dry areas. Ihjla granosa is found in relatively
open, very wet areas. This species is more abundant in the igapo
forest than in the varzea forest; the fourth index score is almost
zero. Leptodoctylus tcagneri is terrestrial, as indicated by the ex-
tremely low third index score; the species is more common in the
varzea and igapo forests than in the terra firme-varzea transition
area, apparently due to the absence of permanent standing water
in the transition area. The fourth index yields extreme values for
several of the species of frogs. Plujsahemus ephippifer, Hijla
geographica, H. haumgardneri, Hijla sp. (rj//;/fl-like), and H. egkri
all have low scores; Dendrohates ventrimactdatus and Hyla rubra
have high values. Plujsahemus ephippifer is terrestrial, found only
in the terra firme-varzea transition area. Dendrohates ventriinacu-
hittis occurs in very wet areas of the varzea and igapo forests; the
species is relatively uncommon in both areas. Hyla geographica and
//. rubra are found in places of rather dense vegetation. The three
most specialized species seem to be Hyla haumgardneri, H. egleri,
and Hyla sp. (n//;ra-like). The first two species are found in very
wet, open grassy areas, whereas Hyla sp. (ruhra-Mke) is found in
plots having intermediate values on the moisture gradient, with
dense vegetation. All three species are found on low vegetation;
all have extremely low values on the fourth index.
Of the twelve species of frogs, the only abundant ones are Bufo
typhonius, Leptodactylus marmoratus, and L. icagneri, all of which
are terrestrial, and mainly forest inhabitants. Three species of tree
frogs, Hyla haumgardneri, H. egleri, and H. ruhra, are found
principally in open, non-forested areas, where they congregate at
ponds and swamps to breed, thus explaining their relative uncom-
monness in the forest plots.
transition (figure 18, upper) and terra firnie-\'arzea transition (figure 19,
lower). See figures 12 and 13 for an explanation of indices. Species numbers
are associated with species names in table 9, p. 29. Numbers enclosed in
squares indicate species that are habitat specialists; circles are habitat inter-
mediates; diamonds are relatively unconnnon generalists; triangles are moder-
ately common generalists, and hexagons are abundant generalists.
40
OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
o
1^
o o o o o
to lO ^ f^ CM ~
isajoj D8ZJDA ui OOOIX aouopunqy aAUD|8y
ANALYSIS OF ECOLOGICAL DISTRIBUTION
41
1 r
-^t^
fco
_L
-L.
_L
^
_L
O
o c o o
^tsajoj odoBi U! OOOIX aouopunqv aAjPiay
o
(NJ
0
s
©
_L
©
0
O
-©-
£L
HE
©
©
_L
O
o
(M
O
to
O
M
X
O <"
~ (\J "g
02
u
O
O
o
o
X
o
a
3
^d
■ >. -frt
=« 5
>— ■ TO
•- a
-w li
c ^-
«2
a; C
ii CS
C CO
.y =>
0)
O S
o
a
ID
o
o
CM
o
o
'o
C/5
- O
CO
fcn
42 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
Relative Abundance XIOOO
^10 35 ^60^10 35 ^60^10 35 ^60^10 35 ^60
1 1 1
•
— a
1 1 ^
1 1 1
1 I 1
Leplodactylus marmoratus
Bufo typhonius
Physalaemus petersi
^ Plica umbra
10
5
•
Gonutoaes humeralis
Physalaemus ephippifer
Hyla sp. (w/bw-likel
Anolis fuscoauratus
'Boliloglossa altamazonica
Leposoma percarinafum
Hyla geographica
uranoscodon superciliosa
Lepfodactylus wagneri
•^ Hyla rubra f
0
5
•
--•-
— 0
— A
"^
H
B A
•-
• -
"O
•
^^^^^■~"
_c
^Kenlropyx. calcarafus
15
— Dendrobafes ventrimaculafus
— Hyla egleri
Hyla granosa
^ Hyla baumgardneri
20
25
#
Capoeira-
Terra Firme
Transition
1 1 1 1
Terra Firme-
Varzeo
Transition
1 1 r 1
Varzea
Igapo
5 15 25
Niche Breadth
35
Fig. 22. Relationship of species scores on index I, relative abundance,
and niche breadth scores for 20 species in each of the four major areas. Dots
represent niche breadth scores. Horizontal bars indicate index scores and rela-
tive abundance.
ANALYSIS OF ECOLOGICAL DISTRIBUTION
43
Relative Abundance XIOOO
^10 35 ^60^10 35 ?605|0 35 ^60^10 35 ?60
n — r
10
— •
0
H
I oh
X
25
5or
• —
55
50-
•—
— I 1 1
^y/a sp.{rudra-\'\ke)
BolHoglossa aliamazonica
'^^::::!^yla geographica
I Uranoscodon superciliosa
^Hyla rubra
Anolis fuscoauratus
Lepfodacfy/us wagnen'*,
— Plica umbra
Leposoma percarinatum
Physalaemus ephippifer \.
Kent ropy X calcaratus-*-^
Gonatodes humeralis
'Dendrobates ventrimaculatus
Mabuya mabouya
Physalaemus peters!
Bufo typhonlus
Leptodactylus marmoralus
• Hyla granosa
• Hyla egleri
•Hyla baumgardrieri
Capoeira-
Terra Firme
Transition
Terra Firme-
Varzea
Transition
Varzea
Igapo
15 25
Niche Breadth
35
Fig. 23. Relationship of species scores on inde.x II, relative abundance,
and niche breadth scores for 20 species in each of the four major areas. See
figure 22 for explanation.
44 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
Relative Abundance XIOOO
^10 35 ^60^10 35 ^60^10 35 ^^60^10
35 ^60
1 1
•
— •
-•
1 1 1
1 1 1
1 1 ~i
Physalaemus ephippifer
20
15
10
5
>ft_ . .
1 1
Leptodactylus wagneriJ
Physalaemus petersi
Leposoma percarinafum
Kentropyx calcaratus-^
._»
•
0 . _... . .
Hyla geographica
Leptodactylus marmoratus
— Dendrobates ventrimaculatus
Bufo typhonius
•
0
a>
XI
c
10
_»
-•
Conatodes humeralis
^Mabuya mabouya
— Hyla granosa
Uranoscodon superciliosa
m
— Plica umbra
^Hyla rubra
Anolis fuscoauratus
— Hyla egleri
"^Hyla baumgardneri
Bolitoglossa altamazonica
Hyla %p\rubra-\\V.^)
15
20
9
25
•^-
Copoeira-
Terra Firme
Transition
— 1_ 1 1 1
Terra Firme-
Varzea
Transition
1 1 1 1
Varzea
Igapo
5 15 25
Niche Breadth
35
Fig. 24. Relationship of species scores on index III, relative abundance,
and niche breadth scores for 20 species in each of the four major areas. See
figure 22 for explanation.
ANALYSIS OF ECOLOGICAL DISTRIBUTION 45
Niche Breadth Analysis
Niche breadtli is used in this paper to refer to habitat niche
breadth and is presumed to be correlated with the range of en-
\'ironmental tolerances. Niche breadth scores were calculated from
the standard formula proposed by Levins (1967), where p,v, is the
proportion of occurrences of species / in plot i, niche breadth of
species / ( B; ) equals :
1/B, = i:p,r
No actual environmental measurements were taken; resource
requirements for each species were measured indirectly by assum-
ing that a given sampling plot provides the necessary resources for
the amphibians and lizards found therein. Although the niche
dimension is referred to as being habitat, there may be certain latent
biotic interactions influencing the distribution of species which are
included in the niche breadth measurement. The limitation of using
occurrence in sampling plots as an indirect method of measuring
recjuirements of species is acknowledged. However, the analysis is
the only one feasible due to the lack of direct physical environmen-
tal measurements. The data are from the matrix (plot X species)
used in the contingency table analysis. Niche breadth values are
included in table 9. The niche breadth analysis lised here was not
meant to describe the entire niche of each species, but rather to de-
limit the spectrum of the habitat dimension of the niche of each
species.
Three species of lizards (Gonatodes humeraUs, Kentropyx cal-
caratus, and Mahmja ma])omja) have much higher niche breadth
scores ( 16-32 ) than the next highest species, Leptodactijlus icagneri
(approximately 12). These three species of lizards are found in
all four of the major areas, account for 41.8 percent of the entire
sample of 1218 individuals, and do not have extreme scores on any
of the contingency table indices representing environmental gra-
dients. These lizards are considered to have wide niche breadths
with regard to habitat requirements and tolerances and are referred
to as habitat-generalists ( Fig. 25 ) . The relative abundances in each
major area are plotted in figure 26.
Five species can be considered habitat-specialists; all of them
have niche breadth scores in the range of 1-4, indicating that they
have very narrow tolerances and specialized requirements with
regard to the environmental parameters measured indirectly by
the analysis (Fig. 25). Each species is found in only one or two
of the four major areas and is relatively uncommon. The cumulative
46 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
SPECIALISTS
Hyla baumgardneri (|0)
Hyla egleri (9)
Uranoscodon superciliosa (2)
Hyla gran OS a (II)
Leposoma percarinatum (3)
Hyla geographica (19) INTERMEDIATES
Hyla z^Xrubra-\\V^) (49)
Dendrobafes ventrimaculalus (8)
Hyla rubra (30)
/^//'c^ umbra (II)
Bolifoglossa allamazonica (30)
Physalaemus peters! (15)
Physalaemus ephippifer (17)
Anolis fuscoauratus (18)
/S'y/b lyphonius (101)
Leptodactylus marmoratus (156)
Leptodactylus wagneri (9 2)
_L
10 15 20
Niche Breadth
Mabuya mabouya^ GENERALISTS
(44)
A6>A' fro pyx ca lea rat us (117)
Gona lodes hum era lis (257)
I I I
35
Fig. 25. Niche breadth scores. The bars represent iiiche breadth scores.
Nuniliers in parentheses indicate tlie accumulative relative abundance X 1000
in all of the four major areas. The dashed lines separate the species into
habitat specialists, intermediates, and generalists.
relative abundance of the five species in all of the areas is only 44
out of the total of 1218 individuals, or 3.6 percent. Two of the
habitat-specialists are lizards (Leposoma percarinatum and Uranos-
codon superciliosa), and three are frogs (Hyla baumgardneri, H.
egleri, and H. granosa). Leposoma percarinatum, a secretive ter-
restrial lizard found within the leaf litter by day, is probably more
ANALYSIS OF ECOLOGICAL DISTRIBUTION
47
>I00-
8 80
o
CD
o
c
o
■o
<
>
60-
40
Q: 20
1 1
r
1
ounuiuocs numc'ui iS •
#
A
Kentropyx calcaratus •—
Mabuya mabouya • —
— •
— •
/
•
-
: /
\ /
\ /
\/
/ \
/ \
/
•
•
-
^^ ^-
-
r - - - T
1
1
Capoeira-
Terra Firme
Transition
Terra Firme -
Varzea
Transition
Varzea
Igapo
Fig. 26. Relative abundance X 1000 of the three haliitat generalists in each
of the four major areas.
widely distributed and more abundant than the data indicate.
Uranoscodon siiperciliosa is found mainly near pools of standing
water in the varzea forest. Hyla granosa is predominantly an igapo
specialist, not found outside of the forest. The other two species of
tree frogs, H. haumgardneri and H. egleri, are not primarily forest
inhabitants, but are found abundantly in open areas; therefore, these
two species are not specialized for the particular forest areas, but
rather are dependent on standing water. For this reason the species
have low niche breadth scores relative to the forest analysis. If a
similar study were carried out in open areas, these species would
probably have wide habitat niche breadth scores, for they are
abundant and seem to have a wide range of environmental toler-
ances in open areas.
The remaining twelve species are considered to be habitat-
intermediates (Fig. 25). In general, these species demonstrate
intermediate niche breadth scores, corresponding to relatively few
extreme values on the environmental indices from the contingency
table analysis. They are generally more abundant and more widely
distributed than the habitat-specialists, but less so than the habitat-
48 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
generalists. This category includes the one species of salamander,
two lizards, and nine frogs. Several of the habitat-intermediates
have niche breadth scores similar to those of the habitat-specialists.
The artificial line separating the two groups is obviously based on
more than niche breadth values; representation and relative
abundance in the major areas were also considered.
There seems to be a definite relationship between cumulative
relative abundance and niche breadth scores (Fig. 27). In general,
those species with wide habitat tolerances (high niche breadth
values) are more abundant than those with narrow habitat toler-
ances. The abundant generalist has the highest niche breadth value,
the moderately common generahsts have lower niche breadth values,
and the five habitat-specialists have the lowest niche breadth values
and are extremely uncommon.
Another way of looking at the association is to plot index scores
against niche breadth values (Fig. 28). All of the habitat-specialists
have positive values on the first index, indicative of wet environ-
ments. Three of the habitat-specialists are restricted to open, grassy
areas. One of the specialists is terrestrial, and the other four are
found predominantly on low vegetation. Three of the specialists
have more extreme negative values on the fourth index than does
the generalist having a negative value. The relationship of niche
breadth values to both index scores and relative abundances within
each major area is presented ( Figs. 22-24 ) .
When index scores are plotted against relative abundance values
for each area, it is possible to characterize the areas with regard to
species composition in terms of habitat-generalists, intermediates,
and speciahsts (Figs. 18-21). The capoeira-terra firme transition
area provides suitable habitat for the three habitat-generalists ( one
is moderately common and the other two are relatively uncommon ) ,
but the five habitat-specialists are absent. The terra firme-varzea
transition area is composed of two habitat-specialists, the three
generalists, and numerous habitat-intermediates. One of the gen-
eralists is very abundant in this area, and the other two species are
relatively uncommon; the two specialists are rare. Both of the
specialists are found in one additional major area. The varzea area
is represented by the three generalists (one abundant, one moder-
ately common, and one relatively uncommon) and three habitat-
specialists ( all rare ) ; one of the specialists is restricted to the varzea
forest. Two of the generalists are moderately common in the igapo
forest, and the third is abundant; three habitat-speciahsts inhabit
the area, only one of which (Hijla baumgardneri) is restricted to the
ANALYSIS OF ECOLOGICAL DISTRIBUTION
49
o
o
c\j
y^
/v
©
o
a
a5
e
<u
en
V> c
<
— o
- E E
a> E
£ o
CU o <i>
en rz
_>^ .g
,
C O) T3
to
1 "5 p
T3 1- C
^
_ C <U i_
o
3 T3 OJ
cu
J3 O -tl
CL
< ^ ^
00
0<O
1 1 1
o
iai
o
c
o
■o
c
o<
>
a:
O)
>
5
c
u
en
o
<a
a
to
O
cq
3
e
o
CD
O
CM
o
O
CM
o
i3
o
o
o
^.s
(11 -^ IT
O U
o .2
c
X -^
o
OJ
1)
J2
o
o
<
C<]
mpD9jg aqoiN
50 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
©0
©
0
-<#-
<^
J©
0
©
©
o
1 N
X
c
B
©© @
^
Q
©r
.©
o
o
00
o
U3
o
in
o
o
ro
H ::
X
0)
c\j —
O
o
ro
o
c\j
o
ANALYSIS OF ECOLOGICAL DISTRIBUTION 51
igapo. All three specialists have the highest positive scores of any
of the igapo species on the first and second indices, indicating stand-
ing water and areas of grassy, emergent \'egetation; two species
have the highest positixe scores on the thtird index. The habitat-
specialists in the igapo forest exhibit more extreme index scores
and are more abundant than other specialists in other areas.
Niche Overlap Analysis
In a consideration of niche overlap, it is appropriate to ask:
Proportionately, how often do species / and / occur together? Niche
oxerlap can be crudely estimated by plot overlap if we assume that
species requirements are intiinsic properties of the plots. The mea-
sure does not indicate what the overlapping requirements of the
species are, but merely that oxerlap exists. Niche overlap scores
\\'ere obtained from a formula suggested by Horn (1966); p,, is the
proportion of occurrences of species / in plot i. Overlap of species
/' and k ( oc jk) is then estimated by the following:
cc jk = 2 SPo- P//,7 ( 2p,r+2p-A- )
The index is from 0.0 (no overlap) to 1.0 (complete overlap).
A high niche o\'erlap value for two species indicates they are found
together in the same plots. For example, Hyki hanmgardneri and
H. egleri have an overlap value of 0.971, the highest of any two
species associations; these frogs breed in the same plots in the igapo
forest. Other high correlations are Biifo typJioniiis and Leptodacty-
lus mavmoratus (0.928) and Uyla sp. {ruhra-Mke) and Bolitoghssa
altomazonica (0.913). Both species pairs usually occur sympatrically
and therefore probably overlap greatly with regard to certain en-
viionmental requirements.
The following species pairs frequently occur together and have
fairly high correlations, likely indicating similarities in environ-
mental re({uirements: 1) Kentropyx calcaratus and Mabiiya ina-
houya (0.718); 2) IlyU rubra and Anolis fuscoauratus (0.631); 3)
Hy]a egleri and H. granosa (0.589); 4) Leptodactylus marmoratus
and PJujsalaemus petersi (0.586); 5) Leptodactylus marmoratus
and Gonatodes humeralis (0.575); 6) Hyla haumgardneri and H.
granosa (0.547); 7) Leptodactylus wagneri and Mahuya mahouya
(0.539); 8) Hyla rubra and Uranoscodon superciliosa (0.5.33);
9) Bufo typhonius and Gonatodes humeralis (0.529); 10) Anolis
fuscoauratus and Plica umbra (0.513); 11) Hyla geographica and
Uranoscodon superciliosa (0.512); and 12) Bufo typhonius and
52 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
Phijsahemus pefersi (0.507). Physalaemus ephippifer has the least
association with any other species, the highest being with
Phijsahemus petersi (0.210). A complete, ordered tabulation of
niche overlap values for every species pair combination is given in
table 11.
0; 1
-c S
4-* ^
>. "
t-
r^
tfl
OJ
0
13 -G
'o
m M
dj
1^
CO
o c
^+-' G
r-"
^H
t«^
"^ "o
;d
^^ o
V)
^ a
'0
^ c
a
-2:s
CAl
H-^
>
w OJ
■ — ^
d
lO
>-T!
_«
y^
^5
is
O
a
O
C/i
V3
M q;
<u
m I-
t)
o 8
c/;
tTi
«
t!
;.^
Oj
o; ^
q;
■£ 2
4-'
c3
H
a;
a
« s
C/}
P <+H
-2 c
03
° >=
U
u o
^^
O
CO
(U 13
'o
a
C X
y
1— 1 ^
53
CO
wi T!
OJ CD
>- C
o -3
C/3 C
oq
erlap
> dete
a
> rt
C/3
o .
, ^
0) b
-c f5
o y
■S t«
^ a
CS
Oj
a
C/3
■-I >
o
J 0)
« _c
-< o
H-a
OOOiCOt^C-lOlOCOlO^OOOlCOOlCOCO^
OOOfM-riOcocj5(Mco*fOoc^Haccoa5coco
OOOOOOOO-Hi— ii— ii— II— lOlCMCOCOlCCO
o o o o o o o" o o" o o o o o o o o o o
Oi^^fM-HCO-^COt^
ocxDioa5if5cgocoooo-f
l-H I— ll — I i-H^Hi-HCnIi — I
oooooK:i':0't-rffCDOtco30fM--(t^a5
ooooo^Hcoioot^t^oooo-rcocc-roc
OOpopOOpOOOO'— JCMOICM^ICIO
ooo'oooooooo'ooooo'ooo
1— I C2 i-H
(Maj-*coooi>omt^ioocD(Nooco-ti
OOOOOlOM'Ol-^OiCI'-^COCOO'I-f'lCCOOl
OOOOOdOO'^lOO^^CvlOt^OCO-H
pOOOOp^_^'— ;rt_C<jC<10"j(MCOCO'^'1^lO
oooooooooooo'ooo'oooo
1— I CO Tt< CT C5 C<I CO
t-oc'^cooiooqoot^oo
1— ii-Ht— ll— ii— I 1— (I— (oq
ooooooir:iu:it^cX'C<i(Mt^o:)cc!iot:^o?'H
OOOOO— i^HCO-r-f<lClO»Ot^G5aCOOCt~-
P p O O O O O O O O O O p O O ^ (M in 05
o o o o o o o o o o o o o o o o o o o
lO 00
CO o
-H oq
^a5-Hl01>C75t^OfMCD(M00C0C0
OOOOOOOOOOOir:it^O>— ^t:^liOI>r-l
OOOOOOOOOOO(M-1''-r00CDt^^t-
ppppppoopooooooi-H— iiop
o o o o o o o o o o o" o o o o o o o o
^10Q003^C<ICO'*'t>a50lOt>OCD(NOOCDTt<
I— ll— Ir-Hi— ll— ll— (C<li— I 1— ll— I 1—1
OOOOOt^^l>t-lLC00in(MO^(M(MXO
oooooioa3'-icDoooo^o-imiococDt--c<i
p p p p o p p ^ ^ ^_ -H eq oi (N oi oi CO CO '^t*
oooooo'ooooooooooooo
1— I OJ ^^ (M
mcoco-<*oot-i:^oocDio;Dco
1— I ^^ 1— I d 1— I ^H 1— I
00000<MCOICCOC<1C<1— ^OOCOOO^t^OiOO
OOOOO(MCMC0iCO000J05OOC0O(M(M
ppoooooooi— i^Hi— i^c<i(Mc<iioioa5
o o o d o o d o o o o o o o o o o o o
(Mcoiocooot^Trcooot^oO'-HCO'tiaicqioa^
Cq^H rHi— <^H ^^1— ll— I^Hi— ll— I
ANALYSIS OF ECOLOGICAL DISTRIBUTION
53
^3
o
<
■*
C/5
CO
0)
'o
C/3
(M
O
a
CD
'o
OJ
a
CD
a
05
'o
Oj
a
C/2
00
CD
'o
a
C/2
OOOH^-^^-COCDCD(MOO^<Ot~--^OOOOCO'— I
OO^-H'tlOOICOOSOOCOt^COCO-^CD^CO
oooooo'oooooo'ooooooo
CO
'*COc<lt^(Ma5lO
00 CD X lo CO o a>
i-H 1 — I i-H .— I 05 I— (
O O O 05 CO CTj t^
O O O ^ CO O '-I
O O O O O -H — ;
d d d o d
Q0'-rCOC0^t~-^O1C<lt^00C0
^05G50--i^a505COC"I'ti^
-H ^ ^H O-l C-1 01 Ol 01 CO -^^ ■^_ 02
oooooooooooooo
CO
■rr^HC<lCDO<M0>t^CD^lCO0C10l>a>
i-Hi-H J— I i-Hi — I Csl^Hi— ( f— I
rt^ 00
OOOOOC505l0a5C0O(Mt^
oooooo^tMoi-rt^t^co
OOOOOOOOOOpO-H
ooodddoddoddd
(MCOCDOOOt^COlOOOCOCitO
CM I— I i-H r— I 1— I ^H
CD O 0> CO t^ CD
CD ^H ^H I— I O 00
^ oq CO -r Lo lo
d di d d di d
-t -H lO t^
05
o o o o o o
o o o o o o
o o o o o c
d d d d d d
OOOlOCOOitO
O O O Ol O O -H ol
o o o p ^_ ^ ^ — ;
d d d d d d d d
CD "^ G5 X) O
ic 00 t- C5 -H
^ -H -H ^ <>1
d d d d d
(MCO'^CDt^XCOTOOOOlOCDt^OiOlC'-HCl
. — ^1 — lO]<— (I — I I— li — ( I-H^^ 1 — I
05 oq 05 -t a> t^
•— I oq (M t" -f lO
o o o o o o
d d d o o o
a ^^ a
C^OiCDt^t^t-CD'-^t-CDCDOlO
d o CI oi CO lo o »o CD o 'CO CO CO
^ rt ^ --4 ^ ^ CM (M oi CO -t -r lo
ddddddddc>'Z>ddc:>
-fCO-t^-iCOt^COC-lOCDOlt^lOlOCDCO
0000000>--^CvlOCDCCCOt^CDOOlCCDCO
oooooooi-riot-^i-Hi— iioi>05cot^ooc<i
OOOOCOOOOCi-H-^-^'— I'— I^iciop
ddddodddddddddddodd
C-q CO K; CD t^ O O
(M -H
-rxooco-HCJi-^r^iocq^H
OOOOTCDt^'OOX.— i-HOTOCDCO'^r^CO
OOOOlO^OICOCDXCT^COn'CDt^t^X'— I
p q p p p '-I ^ -H —; ^ — ; oi oi CI CI CO CO CO p
ddddddddddddddddooo
coTr^cicDcjiOt^cDci-Ht-oxmcnm-^co
I-Hi-H I-H 1— I I— <C<1'— I'— li— I I— !■— I
Species Diversity and Equitability
The Shannon index was used in the present analysis as a means
of comparing the four major forest areas. The index is calculated
as follows :
H' «. -2 p; log pi ^ C/N (N logio N-S"< logio n,),
54 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
o
pa
<
o
en
'y
a
C/2
en
o
a
00
en
'o
a
t-
'a
OJ
a
C/2
o
'o
OJ
a
en
!U
'o
a
CO
.- O o 00 oi CO
CO lO 00 O — I CO
oooooco-tr^^ccot-
OOOOOOOt^LOCCOO^H'— 1-.".'~^ ■--
O O O O O O O O ^ ^. ^ Ol Ol fSj (M_ O-l -r lO uo
o" o d o
^H CO -t 05
dddoodooooooooo
^fMOJCOOlOOOOcOOOCMt flCt^
I — ii— (I— t 1 — If— ii — ii— ti — I I— If— (
o o o o o o
o o o o o o
o o o o o o
d d <6 d d d
lOroOQOOCM-fl^^OCOt^CO
■-H'— it^lOCCCCCCl^COI>l>(M^H
OOO^H'-i^H^H-HCMCMCO't'lO
ddddddddddddd
CM CO in
t- o
•-H (M
-fOoicDt^QO^ro
in CO CO
ITiOOfMG^G^OllOOOt^O
CM CM t- O -T iJO CD t^ 00 O
ddddo'ddddddo'dddddo'd
pH O CO CD r-i 05 00
CD05-ta^ (MCOCO^
CM CM CO CO T t in I>
^ CM a>
t^ in 05 CO
-fOO-tCOmt-CMCOOCD
ooc^-tmocoa^cMO-t-ti-HCMXt^ccincM
oo-rcorocMcM-rcxja^o^o^cMcococco^
o O p o O '-; ^ ^_ —; ^. —; CM cm_ cm_ cm cm_ oi -:f -t_
CO 05
t— ^H -f 00 1— t
cocoinoocMOiOomcM
I — ^ T— I ^H I — ^ CM '—I
_icoro^H<c-rc/:;oocDOxccocx)OcMoo
Tt<-r'ccoco5^-iincD05CJ5^Hi— i-rt^cct^mcc^^
o O O O O ^ —; ^ -H ^. oi c>i c>i oi oi CO -r -t t-^
dddddd ddddddddddddd
OJo-i^co-t-HO^oot^cooincc-tt^cMinooo
1— < I— (1— ll— li-HCM '-^ t-Hi— I^H
ini-CD^O^-HOO-tOCCD OCMCCC50-f005in
oi-^t-c:5t-oooccoicct-o50--;-i;coi2J>ir;;
o o o o ^ ^H ^ CM CM oi o] oi CO CO CO "t "t in in
dddddd ddddddddddddd
CO
CD CM -H O
^ (M
l^l>inGO 0>C0OCM00TC0^05
where H' = average diversity per species, p, = probability of en-
countering the jf" species, C = the constant 3.321928 when using
Base 2, N = total number of indixiduals, and n, = number of in-
dixiduals in the /"" species: The \alues are interpreted as follows.
If in Area I there are 20 individuals of species A and 20 individuals
of species B, then H' ^ 2.00; there are two equally common species.
If in Area II there are 40 individuals of species A and 10 individuals
of species B, then H' «= 1.65; there are 1.65 equally common species.
Area I is considered to have a higher species diversity than Area II.
The data used were those in the original data matrix which also
served as the basis for the contingency table analysis and niche
breadth and overlap analyses.
ANALYSIS OF ECOLOGICAL DISTRIBUTION 55
The calculated species diversity values for the four major forest
areas are as follows: capoeira-terra firme transition = 3.00; terra
firme-varzea transition = 2.45; varzea = 3.02; and igapo = 2.72.
Thus, in terms of the sampling plots analyzed, the varzea is the
most diverse area with regard to species richness and evenness;
the terra firme-varzea transition area is the least diverse. A deter-
mination of statistical and biological significance of these values
would require additional data, ideally with an equal number of
plots sampled per major area.
MacArthur ( 1957 ) proposed a "broken-stick model" dealing
with species equitability based on one million individuals of 200
species. The model is based on the equation
77 1-
l/s2 l/s-i+l),
i-i
where 77, is the theoretical proportion of individuals in the r"' most
abundant species (/• = 1, 2, . . . , s), each theoretical proportion itself
being obtained by summing over r terms (/ = 1, 2, . . . , r). By using
this formula, it is possible to obtain an apportionment of the in-
dividuals among the species in a sample in about as equitable a
manner as ever occurs in nature. An advantage of MacArthur's
model is that there is no set of parameters into which data must
conform; for each possible number of species (s), the equation
generates a complete set of s proportions 77,. ( r = 1, 2, . . . , s ) . The
model yields a curve whereby species abundances are graduated
from the rarest to the most common. A maximum equitability
curve, whereby for every sample size each species is equally abun-
dant, can also be calculated. Species diversity values (H') obtained
from the Shannon index can then be compared to the broken-stick
and maximum equitability curves. Any community falling between
the two curves is considered to be extremely diverse.
The species diversity values for the four areas were plotted in
relation to the curve expected from the broken-stick model and the
maximum equitabihty curve ( Fig. 29 ) . All areas fall to the left of
the broken-stick distribution with the exception of the capoeira-
terra firme transition area; the position of this area suggests that it
is highly diverse in terms of species equitability. The validity of
this model was questioned by Hairston (1969). He claimed that
the broken-stick model lacks ecological meaning, because conformity
to the model is largely a function of sample size. He demonstrated
that large samples tend to make rare species even more rare and
common species even more abundant; the reverse distortion is ap-
parent in small samples. Because of the small sample sizes in the
present analysis, the distribution of abundances may be distorted
56 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
CO
o
Q.
CA)|
0
^5
1
I
1
1
1. Capoeira-Te
rra
Firme Transition
2. Terra Firme-
Vc
rzea Transition
- 3. Varzea
•
3
•
/-
4. Igapo
2
•
4 A y
/
y\
•
•
Broken
_
Stick Distribution-
^^^^*'^^^
^«
1
-• '^Maximal Diversity
1 1
1
12 3 4
Species Diversity (h')
Fig. 29. Relationship of species diversity and numbers of species for each
of the four major areas to MacArthur's broken stick distribution and maximal
diversity.
such that rare species appear to be more common relative to abun-
dant species than actually is the case.
Lloyd and Ghelardi ( 1964 ) proposed an equitability equation
for the measure of fit of observed relative abundances of species to
those predicted by MacArthur's broken-stick model, as follows:
E = s'/s,
where s is the actual number of species and s' is the theoretical
number of species that should be present according to the broken-
stick model at the actual diversity {W), as calculated from the
Shannon index. Maximum conformation to the model is 1.00.
The following equitability indices were calculated from Lloyd and
Ghelardi's table: capoeira-terra firme transition := 1.10; terra firme-
varzea transition ^ 0.58; varzea ^ 0.79; and igapo = 0.90. Because
of inequitability in the distribution of individuals among the species,
the sample from the capoeira-terra firme transition forest has a
species diversity appropriate to a community with 10 percent more
species than actually occur in the particular sample. On the other
hand, the samples from the terra firme-xarzea transition, varzea,
and igapo areas have species diversities appropriate to communities
with only 58, 79, and 90 percent as many species as actually occur
in the respective areas. Therefore, only the capoeira-terra firme
transition area is more diverse than would be expected by the
ANALYSIS OF ECOLOGICAL DISTRIBUTION 57
Table 12. — Comparisons of Major Areas by Coefficients of Coniinuiiity. Num-
bers in Roman are the numl^er of shared species of amphil^ians and lizards
lietween two major areas; numbers in bold face are the actual number of
species in a given area; numliers in italics are the coefficient of community
values.
Major Terra Open &
Areas Firme Varzea Igapo Capoeira Edge
Terra Firme 36 24 16 20 16
\'arzea 0.480 38 19 16 18
Igapo 0.381 0.463 22 13 12
Capoeira 0.556 0.381 0.448 20 10
Open & Edge -. . 0.296 0.333 0.273 0.227 34
broken-stick model (Fig. 29). Again, because of small sample sizes,
the statistical and biological significance of this analysis is uncertain.
Coefficient of Community
The coefficient of community ( CC ) , used to measure the relative
similarity of samples from two communities (major areas), is
calculated,
CC = S,,;,/ ( ^(; + ^/< <i(ih),
where S„i, is the number of species shared by samples A and B, S„
is the total number of species present in sample A, and S/, is the
total number of species present in sample B.
Coefficients were calculated for every two area combinations
for five major areas: open and edge areas, capoeira, terra firme,
\'arzea, and igapo. The distribution data used are found in table 2,
consisting of 62 species of frogs, salamanders, and lizards. The
coefficients are presented in table 12, in addition to the actual num-
ber of species every two areas have in common. The varzea and
terra firme forests have the most species in common (24), but the
coefficient of community is the second highest (0.480). The capoeira
and terra firme forests have 20 species in common and have the
highest coefficient of community (0.556). The capoeira forest and
open and edge areas have the fewest species in common ( 10 ) and
have the lowest coefficient of community (0.227). Likewise, there
is low similarity between the igapo forest and open and edge areas
( 12 shared species, with a coefficient of 0.273) .
SUMMARY AND CONCLUSIONS
The ecological distribution of each of 62 species of frogs, sala-
manders, and lizards was determined by means of continuous
58 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
sampling throughout the environment from mid-January to the end
of July, 1969, two weeks in April 1970, and June-July, 1970. Each
species exhibits a characteristic distribution, according to its genetic,
morphological, and physiological make-up, its life cycle, its way of
relating to the physical environment, and its interactions with
other species.
The contingency table analysis was used to obtain a measure of
the association between species of amphibians and reptiles and
their habitats (plots) and to partition this association into inde-
pendent components (indices) which determine the distribution
of species within four of the major forest areas. The components
are interpreted as follows: the first is a moisture gradient; the sec-
ond, a vegetation density gradient; the third, a vertical distribution
gradient, and the fourth seems to be a composite of environmental
parameters. Each species can be characterized in terms of the
indices. Species with scores near zero are the most generalized
with regard to the environmental parameters studied and are gen-
erally the most abundant species; those species with extremely high
positive or low negative scores are restricted to a particular range
of the envii-onmental spectrum and are relatively uncommon. The
species of frogs exhibit more environmental extremes than do the
lizards, indicating that the particular species of frogs studied have
more narrow environmental tolerances than do the lizards included
in the analysis. The environment likely produces greater restric-
tions on frogs than on lizards in the carrying out of life processes
due to basic physiological differences between the animals, resulting
in more restricted distributions for frogs than for lizards.
Niche breadth scores, as calculated from Levins' formula, are
presumed to be correlated with the range of en\'ironmental toler-
ances. Three species of lizards have much higher habitat niche
breadth scores than the other 17 species of amphibians and lizards;
these three species are the only ones found in all of the major forest
areas. There is a definite relationship between cumulative relative
abundance and niche breadth values. In general, those species
with wide environmental tolerances (high niche breadth scores)
are more abundant than those with narro\\' tolerances (low niche
breadth scores ) .
When niche breadth scores, abundance indices, and scores on
the enx'ironmental gradients are analyzed together, three species
are referred to as habitat-generalists, five species as habitat-spe-
cialists, and the remaining twelve as habitat-intermediates. The
generalists occur in all major areas, have high niche breadth scores,
are abundant, and exhibit no extreme scores on the environmental
ANALYSIS OF ECOLOGICAL DISTRIBUTION 59
gradients; the specialists are found in only one or two major areas,
have low niche breadth scores, are relatively uncommon, and ex-
hibit exti'eme recjuirements on one or more environmental gradient.
Partitioning of resources is evident, in terms of both space and
time. One of the most striking examples is the calling site segrega-
tion among breeding male frogs in a given area. Habitat niche
overlap can be estimated by plot overlap. Many species pairs have
relatively high o\erlap \'alues, thereby indicating that they fre-
quently occur in association with each other. Assuming the plot is
indicative of the recjuirements and tolerances of the species found
therein, we can conclude that some niche overlap does exist.
Four of the major forest areas were compared and contrasted
by \'arious analytical technifjues. Each area was characterized by
the contingency table indices. The capoeira-terra firme transition
area is relati\'ely dry and has rather dense vegetation; the herpeto-
fauna predominantly inhabits low vegetation and tree trunks. The
terra firme-varzea transition area can be divided into high ground
transition and lo\\', wet transition. The entire transition zone is an
intermediate area with respect to the environmental gradients, ex-
cept that in many areas the ground cover is less dense than that of
the capoeira-teiTa firme transition area. The varzea plots are rela-
tixely wet and have fairly dense ground co\'er. The igapo forest is
the wettest area and has intermediate to relatively dense ground
co\'er; most of the lizards are either terrestrial or are found on
the boardwalks, and the frogs are found both on low vegetation
and on the ground.
Another way of comparing the areas is in terms of species rich-
ness and equitability. Species richness \'alues for five major areas
are: varzea — 38 species; terra firme — 36 species; open and edge
areas — 34 species; igapo — 22 species; and capoeira — 20 species.
Coefficients of community were calculated on these data to deter-
mine relative similarity between every tAvo areas. The highest
similarity is between capoeira and terra firme forests (0.556), and
the lowest is between capoeira and open-edge areas (0.227). Spe-
cies diversity (Shannon index) scores were calculated from the
contingency table data matrix; the results are: varzea — 3.02;
capoeira-terra firme transition — 3.00; igapo — 2.72; and terra firme-
varzea ti'ansition — 2.45. Equitability values were then calculated
from the species diversity estimates (H^) and compared to Mac-
Arthur's broken-stick model. The capoeira-terra firme transition
area has an equitability of 1.10, indicating that the area is more
diverse than would be expected by MacArthur's model. As dis-
60 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
cussed in the relevant section, the statistical and biological sig-
nificance of this analysis is uncertain due to small sample size.
The ecological requirements and tolerances characteristic of
species in a community vary widely. Every species exploits the
available resources in the most effective way possible for that par-
ticular species. Some species accomplish this by specializing in
part of the environment, whereas others are generalized and utilize
a greater spectrum of environmental parameters. It is concluded
that the niche segregation existent among the frogs, salamanders,
and lizards living in various habitats within the rainforest at Belem
permits many species to coexist with highly efficient utilization of
environmental resources.
ACKNOWLEDGMENTS
I am extremely grateful to the many persons who made the field
studies in Belem possible. Initial research was supported by A.R.O.
Grant 7184-EN to the Smithsonian Institution and by the Museum
of Natural History Director's Fund, University of Kansas. Field
work in 1970 was supported through a combination of monies from
the W. G. Saul Fund, the Museum of Natural History Director's
Fund, the Watkin's Museum of Natural History Fund, and a grant
from the National Science Foundation through the Committee on
Systematic and Evolutionary Biology, all at the University of Kan-
sas. I thank the officials of IPEAN (Instituto de Pesquisas e Ex-
perimentagao Agronomicas do Norte) for permission to carry out
my studies in the APEG reserves ( Area de Pesquisas Ecologicas do
Guama). John P. Woodall, Director of the Belem Virus Laboratory,
provided field and laboratory facilities.
Many persons made my field work enjoyable and profitable; my
thanks go to William E. Duellman, Philip S. Humphrey, Thomas
E. Lovejoy, and Alan D. Crump for their guidance, help, and com-
panionship in the field. I am also grateful to Carlos Cabe^a and
the many other Brasilians who assisted with field collections; with-
out their eff^orts my study would not have been possible.
I am indebted to the numerous persons who helped with the
synthesis of this paper. William E. Duellman aided with identifica-
tion of specimens; William H. Hatheway ran my data through com-
puter programs at the Unixersity of Washington. Philip S. Humph-
rey, William E. Duellman, and Or ley R. Taylor contributed nu-
merous suggestions and ideas with regard to the writing of this
paper. Also, I am grateful to Stephen R. Edwards, Paul B. Robert-
son, and Janalee P. Caldwell for their comments and suggestions.
ANALYSIS OF ECOLOGICAL DISTRIBUTION 61
Linda Tiueb assisted with formats for illustrations and Thomas H.
Swearingen executed most of the final drawings. Judy Macura
typed the manuscript.
Finally, I am grateful to all the many persons who provided
endless encouragement throughout the course of study; special
thanks go to Abraham Goldgewicht for his constant patience, under-
standing, and confidence.
LITERATURE CITED
Belem Virus Laboratory, Belem, Para, Brazil
1967. Annual report. Unpublished.
Caix, S. a., G. M. de O. Castro, ]. M. Pires, and N. T. da Silva
1956. Application of some phytosociological techniques to Brazilian rain
forest. Anier. J. Botany, 4.3:911-941.
Duellmax, W. E.
1967. Courtship isolating mechanisms in Costa Rican hylid frogs. Herpe-
tologica, 23:169-183.
Hairston, N. C, J. D. Allex, R. K. Colwell, D. J. FxiruYi^fA, J. Howell,
M. D. LxjBix, J. Mathias, and J. H. Vaxdermeer
1968. The relationship between species diversity and stability: an experi-
mental approach with Protozoa and bacteria. Ecology, 49(6):1091-
1101.
Hatheway, W. H.
1967. Physiognomic characterizations of three \egetational types at the
Guama Ecological Research Area, Belem, Brazil. UnpuljJished
report to the Smithsonian Institution, June 15, 1967:1-35.
HoRx, H. S.
1966. The measurement of "oxerlap" in comparati\e ecological studies.
Amer. Nat., 100:419-424.
IIUTCHIXSOX, G. E.
1957. Concluding remarks. Cold Spring Harbor Symposia on Quantita-
ti\e Biology, 22:415-427.
IxGER, R. F. and B. Greexberg
1966. Ecological and competitive relations among three .species of frogs
( genus Rana ) . Ecology, 47 ( 5 ) : 746-7.59.
Kexdeigh, S. C.
1961. Animal ecology. Prentice-Hall, Inc., Euglewood Cliffs, N.J., 468 pp.
Lloyd, M. and R. J. Ghelardi
1964. A table for calculating the "equitability" component of species
diversity. J. Animal Ecol., 33:217-225.
Lloyd M., R. F. Ixger, and F. W. Kixg
1968. On the di\ersity of reptile and amphibian species in a Bornean
rainforest. Amer. Nat, 102(928) :497-515.
Mac Arthur, R. H.
1957. On the relative abundance of bird species. Proc. Natl. Acad. Sci.
U.S., 43:293-295.
1964. Environmental factors affecting bird species divcLsity. Amer. Nat.,
98:387-398.
1965. Patterns of species diversity. Biol. Rev., 40:510-533.
62 OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY
MacArthur, R. H. and J. W. MacArthur
1961. On bird species diversity. Ecology, 42:594-598.
MacArthur, R. H., J. W. MacArthur, and J. Preer
1962. On bird species di\ersity: II. Prediction of bird census from habi-
tat measurements. Amer. Nat., 96:167-174.
MacArthur, R. H., H. Recher, and M. Cody
1966. On the relation between habitat selection and species diversity.
Amer. Nat., 100:319-332.
Pianka, E. R.
1967. Lizard species diversity. Ecology, 48(3) :333-351.
Pielou, E. C.
1966. Shannon's formula as a measure of specific di\'ersity: its use and
misuse. Amer. Nat., 100:463-465.
Rand, A. S.
1964. Ecological distribution in anoline lizards of Puerto Rico. Ecology,
45:745-752.
Rand, A. S. and S. S. Humphrey
1968. Interspecific competition in the tropical rain forest: ecological
distribution among lizards at Belem, Para. Proc. U.S. Natl. Mus.,
125(3658): 1-17.
Schoener, T. W.
1968. The Anolis lizards of Bimini: resource partitioning in a complex
fauna. Ecology, 49:704-726.
1970. Nonsynchronous spatial overlap of lizards in patchy habitats. Ecol-
ogy, 51(3) :408-418.
ScHOEXER, T. W. and G. C. Gorman
1968. Some niche differences in three Lesser Antillean lizards of the
genus Anolis: Ecology, 49( 5 ) : 819-830.
Shannon, C. E.
1948. The mathematical theory of communication. In C. E. Shannon
and W. Wea\'er, The mathematical theory of communication. Univ.
Illinois Press, Urbana, 117 pp.
Whittaker, R. H.
1970. Communities and ecosystems. The Macmillan Company, Collier-
Macmillan Limited, London, 162 pp.
Williams, E. J.
1952. Use of scores for the analysis of association in contingency tables.
Biometrika, 39:274-289.
University of Kansas Publications
MUSEUM OF NATURAL HISTORY
The University of Kansas Pulilications, Museum of Natiual
History, beginning with vokmie 1 in 1946, was discontinued
with volume 20 in 1971. Shorter research papers formerly pub-
lished in the above series are now published as Occasional
Papers, Museum of Natural History. The Miscellaneous Pub-
lications, Museum of Natural History, began with number 1 in
1946. Longer research papers are published in that series.
Monographs of the Museum of Natural History were initiated
in 1970.
Institutional libraiies interested in exchanging publications
may obtain the Occasional Papers and Miscellaneous Publica-
tions by addressing the Exchange Librarian, University of Kan-
sas Library, Lawrence, Kansas 66044. Individuals may pur-
chase separate numbers of all series. Prices may he obtained
upon request addressed to Publications Secretary, Museum of
Natiual History, Uni\er.sity of Kansas, Lawrence, Kansas 66044.