Content uploaded by Philip J Devries
Author content
All content in this area was uploaded by Philip J Devries on Feb 02, 2018
Content may be subject to copyright.
Biological Journal of the Linnean Society (1999), 68: 333–353. With 6 figures
Article ID: bijl.1999.0319, available online at http://www.idealibrary.com on
Species diversity in spatial and temporal
dimensions of fruit-feeding butterflies from two
Ecuadorian rainforests
PHILIP J. DVRIES∗AND THOMAS R. WALLA
Department of Biology, University of Oregon, Eugene, Oregon 97403-1210, U.S.A.
HAROLD F. GREENEY
Department of Entomology, University of Arizona, Tucson, Arizona 85721, U.S.A.
Received 6 July 1998; accepted for publication 4 January 1999
To test the hypotheses that butterflies in an intact lowland rainforest are randomly distributed
in space and time, a guild of nymphalid butterflies was sampled at monthly intervals for one
year by trapping 883 individuals of 91 species in the canopy and understory of four
contiguous, intact forest plots and one naturally occurring lake edge. The overall species
abundance distribution was well described by a log-normal distribution. Total species
diversity (c-diversity) was partitioned into additive components within and among community
subdivisions (a-diversity and b-diversity) in vertical, horizontal and temporal dimensions.
Although community subdivisions showed high similarity (1-b-diversity/c-diversity), sig-
nificant b-diversity existed in each dimension. Individual abundance and observed species
richness were lower in the canopy than in the understory, but rarefaction analysis suggested
that the underlying species richness was similar in both canopy and understory. Observed
species richness varied among four contiguous forest plots, and was lowest in the lake edge
plot. Rarefaction and species accumulation curves showed that one forest plot and the lake
edge had significantly lower species richness than other forest plots. Within any given month,
only a small fraction of total sample species richness was represented by a single plot and
height (canopy or understory). Comparison of this study to a similar one done in disturbed
forest showed that butterfly diversity at a naturally occurring lake edge differed strongly from
a pasture-forest edge. Further comparison showed that species abundance distributions from
intact and disturbed forest areas had variances that differed significantly, suggesting that in
addition to extrapolation, rarefaction and species accumulation techniques, the shapes of
species abundance distributions are fundamental to assessing diversity among sites. This
study shows the necessity for long-term sampling of diverse communities in space and time
to assess tropical insect diversity among different areas, and the need of such studies is
discussed in relation to tropical ecology and quick surveys in conservation biology.
1999 The Linnean Society of London
ADDITIONAL KEY WORDS:—Nymphalidae – species abundance distributions – species
diversity – rarefaction – vertical stratification – habitat disturbance – edges – conservation.
∗Corresponding author. Email: pdevries@darkwing.uoregon.edu. Address from April 2000: Center
for Tropical Diversity, Milwaukee Public Museum, 800 Wells Street, Milwaukee, WI 53233, U.S.A.
333
0024–4066/99/110333+21 $30.00/0 1999 The Linnean Society of London
P. J. DVRIES ET AL.334
CONTENTS
Introduction ....................... 334
Material and methods ................... 335
Study site ...................... 335
Study community ................... 337
Field methods ..................... 337
Statistical analyses ................... 338
Results ........................ 339
Discussion ....................... 346
Acknowledgements .................... 350
References ....................... 350
INTRODUCTION
Once upon a time lowland tropical forests were extensive wilderness areas teeming
with unknown life forms and buffered from the ravages of human civilization. In
the span of two generations, however, human commercial activities have largely
reduced these forests to degraded habitat remnants. Since tropical forests are being
destroyed faster than ever (Bowles et al., 1998), and the organisms within them remain
largely unknown, modern studies of species diversity are crucial for understanding
remaining tropical communities and their conservation (Heywood, 1995; Gaston,
1996; Wilson & Sandoval, 1996). Although documenting variation in species abund-
ance distributions of organisms through space and time can help identify general
ecological properties of tropical diversity, relatively few studies have done so (e.g.
Wolda, 1978, 1992; Hubbell & Foster, 1986; Morse, Stork & Lawton, 1988; Terborgh
et al., 1990; Gill, 1991; Kato et al., 1995; Condit et al., 1996; DeVries, Murray &
Lande, 1997; Novotny & Basset, 1998). Rather, the urgency of habitat destruction
has forced most biologists to perform rapid inventories (e.g. Roberts, 1991; Anon,
1993), or develop extrapolation techniques to estimate species richness in a variety
of habitat types (Noss, 1990; Ryti, 1992; Colwell & Coddington, 1994; Hammond,
1994; Pearson, 1994; Keister et al., 1996; Longino & Colwell, 1997). For example,
although recent books are devoted to measurement of diversity and its application
to community ecology and conservation biology (Magurran, 1988; Groombridge,
1992; Ricklefs & Schluter, 1993; Edwards, May & Web, 1994; Forey, Humphries
& Vane-Wright, 1994; Huston, 1994; Heywood, 1995; Gaston, 1996; Hayek &
Buzas, 1996), much focus in conservation biology is on inventories rather than
detailed studies of factors contributing to biological diversity. Regardless of their
widespread use, the validity of quick assessments or strict inventories must ultimately
be tested against long-term studies documenting variation of many species through
space and time.
Many plant and animal species in tropical forests exhibit stratified distributions
between canopy and understory (Allee, 1926; Bates, 1944; Richards, 1952; Pearson,
1977; Sutton & Hudson, 1980; DeVries, 1988; Stork, 1988; Longino & Nadkarni,
1990; Gill, 1991; Wolda, 1992; Malcolm, 1994; Erwin, 1995; Mallet & Gilbert,
1995; DeVries et al., 1997), and such vertical stratification is of obvious importance
to estimating diversity. Although vertical stratification is a significant component of
diversity, it is seldom addressed or measured directly. Given the recent surge of
interest in documenting canopy biotas (reviewed in Lowman & Nadkarni, 1995),
SPECIES DIVERSITY IN ECUADORIAN RAINFOREST BUTTERFLIES 335
surprisingly few investigations have measured species diversity simultaneously in
both canopy and understory environments through time (e.g. DeVries, 1988; Basset,
Aberlenc & Delvare, 1992; Wolda, 1992; Malcolm, 1994; Kato et al., 1995; DeVries
et al., 1997).
As the majority of all described species on earth are insects (Groombridge, 1992)
this group has great promise for illuminating patterns and processes of biological
diversity. Due to their relatively large size, colourful appearance and ease of sampling,
butterflies have broad appeal as models for understanding tropical insect diversity
and conservation biology (see Gilbert, 1984; DeVries, 1987, 1997; Brown, 1991;
Kremen, 1992, 1994; Beccaloni & Gaston, 1995; Daily & Ehrlich, 1995; Robbins
et al., 1996; Brown & Hutchings, 1997; DeVries et al., 1997 and references therein).
Although spatial and temporal factors are important components of ecological
diversity (e.g. MacArthur, 1972; Cody & Diamond, 1975; Ricklefs & Schluter, 1993;
Rosenzweig, 1995), their effects on diversity in tropical butterfly communities have
seldom been addressed directly. As a result many studies concerned with butterflies
and other insects are frequently limited by short sampling periods, use of non-
comparable sampling methods, presence-absence data only, small sample sizes, and
lack of data on spatial and temporal distributions within communities (e.g. Basset et
al., 1992; Daily & Ehrlich, 1995; Robbins et al., 1996). It is therefore often difficult
or impossible to compare diversity studies from different areas.
A recent Ecuadorian field study measured the diversity of fruit-feeding nymphalids
in spatial and temporal dimensions, and provided an assessment of habitat disturbance
on these butterflies (DeVries et al., 1997). Recognizing that, like most studies of
tropical diversity, their investigation was conducted in a forest with considerable
human disturbance, DeVries et al. (1997) concluded that the generality of their
findings required testing them against data sets gathered from forests with less
disturbance.
Accordingly this study was designed to characterize the diversity of an Ecuadorian
fruit-feeding nymphalid community from an intact forest and compare it to the
community studied by DeVries et al. (1997). To achieve our goals we first test the
hypothesis that fruit-feeding nymphalid butterflies are randomly distributed in
space and time among areas within continuous forest. After describing the species
abundance distribution of our total sample, we partition the measures of diversity
among subsets of the community in multiple dimensions, and analyse these partitions
statistically. Secondly, we compare our samples to those of DeVries et al. (1997) and
ask how the butterfly diversity of an intact forest compares to that from a similar forest
that has experienced greater disturbance. By gathering and analysing standardized
samples in different dimensions this study provides a unique comparison that accents
differences between two tropical butterfly communities, and points to patterns that
warrant comparative investigations from other areas.
MATERIAL AND METHODS
Study site
This research was conducted from 6 August 1993 to 19 July 1994 at the La Selva
Lodge (hereafter abbreviated LSL), Sucumbios Province, eastern Ecuador in the
P. J. DVRIES ET AL.336
~800 m
1 Km
1 Km
Garza Cocha
Forest 1 La Selva Lodge
Forest 3
Forest 2
Mandi Cocha
Forest 4
Rio Napo
N
= trap locations
= main trail
Figure 1. Schematic map of the La Selva Lodge study area showing approximate locations of the five
plots (Lake Edge, Forest 1 through 4) and five replicate sampling sites nested within each plot.
upper Amazon Basin 75 km E.S.E. of Coca in an area bounded by the Rio Napo,
and the oxbow lakes Garza Cocha and Mandi Cocha (0 29′50.3″S; 76 22′28.9″W)
near the settlement of Anyan
˜gu. Sampling sites were located within an approximately
1000 hectare section of forest around and between two oxbow lakes (Garza Cocha
and Mandi Cocha) with a further sample site situated about 800 m north of Mandi
Cocha. Rainfall data from 1995–1997 indicates that this area receives between 3.5
and 4.0 m of precipitation per year, with a dry season from January to March. The
forest surrounding the LSL study site includes at least 30 000 hectares of floodplain
forest that harbors an intact vertebrate fauna and flora, including some of the most
species rich forest known from Ecuador. All available evidence suggests that the
LSL study site represents continuous old growth forest that has escaped severe
disturbance of modern human civilization; although it is noteworthy that much of
the surrounding area is currently under threat of destruction by the petroleum
industry (Olson et al., 1996). Here we provide a schematic map of the LSL area
and trapping design in Figure 1 that is pertinent to this study. A detailed map and
analysis of the LSL trap study will appear elsewhere (DeVries & Walla, in prep.).
SPECIES DIVERSITY IN ECUADORIAN RAINFOREST BUTTERFLIES 337
Study community
Neotropical butterfly communities divide quite readily into two general adult
trophic guilds (see DeVries, 1987, 1988; DeVries et al., 1997): (1) flower feeders—those
species obtaining most nutritional requirements from flower nectar (most
Papilionidae, Pieridae, Lycaenidae, Riodinidae, and some Nymphalidae), and (2)
fruit-feeding nymphalids—whose adults gain their nutritional requirements by
feeding on juices of rotting fruits or plant sap. The fruit-feeding guild is generally
understood to include the nymphalid subfamilies Charaxinae, Morphinae
(Morphinae+Brassolinae of some authors, e.g. De Jong et al., 1996), Brassolinae,
Amathusiinae, Satyrinae, and particular genera of Nymphalinae (Limenitinae of
some authors). Fruit-feeding nymphalids are easily sampled in spatial and temporal
dimensions using traps baited with rotting fruits (e.g. DeVries, 1988; Brakefield &
Reitsma, 1991; Pinheiro & Ortiz, 1992; DeVries et al., 1997), and may comprise
between 40–55% of the total nymphalid richness in tropical forests (DeVries, 1987,
and unpublished). For completeness we note that at some neotropical sites a small
proportion of species in the subfamily Ithomiinae are found occasionally in fruit-
traps (DeVries, Lande & Murray, in press). Ithomiines, however, typically feed on
flower nectar, and are not strictly part of the fruit-feeding guild as defined here.
The few individual ithomiines trapped during this study were excluded from our
analyses.
Field methods
At LSL we established five sampling plots within intact, contiguous forest, each
containing five replicate sampling sites (Fig. 1). The five plots include: (1) Lake Edge:
along the oxbow lake, Garza Cocha, where traps were located at the interface of
forest edge and open water. This plot represented a naturally occurring forest edge
distinct from the other four plots that were all located within closed canopy forest;
(2) Forest 1: located approximately 400 m WNW offthe main trail between the two
lakes; (3) Forest 2: located along the northernmost 500 m of the central trail to
within 50 m S of Mandi Cocha; (4) Forest 3: located approximately 1 km due east
of the central trail and ranging from 500–700 m north of Garza Cocha; and (5)
Forest 4: located approximately 800 m north of Mandi Cocha.
In all plots each replicate sampling site was fitted with one understory trap and
one canopy trap for a total of ten traps in each area—five canopy, and five understory
(see DeVries, 1987; DeVries et al., 1997 for trap design and methods). Excluding
rare emergent trees, the average height of the forest canopy at La Selva ranged
between 18 and 29 m above the ground. In all cases our traps were positioned to
sample from within the canopy; that is, within the crown of the trap tree. Canopy
traps were suspended from thin ropes run over branches of an emergent tree, such
that all traps could be raised and lowered from the ground. Understory traps were
suspended from low branches such that the bases hung between 1 and 1.5 m above
ground and could be serviced directly. In the case of the Lake Edge plot all traps
(canopy and understory) were serviced from a dugout canoe.
Baited traps were maintained continuously for a 5-day sampling period within
the first week of every month. As in a similar study (DeVries et al., 1997) traps were
baited with locally-obtained bananas which were mashed, mixed well, and fermented
P. J. DVRIES ET AL.338
for 48 h in one large container prior to use, and on the day prior to the sampling
interval, bait was placed in a small plastic cup fixed inside each trap, and a small
amount of new bait from the common reservoir was added to each trap on the
third sampling day. During trap months all 50 baited traps were sampled daily for
5 days. On the last day of the sampling period baits were removed from all traps,
and traps remained unbaited for 3 weeks. New bait was made prior to the subsequent
sampling interval, and the protocol repeated throughout the study. As shown
previously (DeVries, 1988; DeVries et al., 1997) butterflies were not attracted to
unbaited traps.
All butterflies were identified to species, and depending on the species, individual
butterflies were treated in one of two ways. In most cases each individual was
collected and placed in a glassine envelope with all pertinent data written on the
envelope, and used for subsequent identification and ecological measurements. For
a few abundant species, individuals were marked with a unique number, released,
and the information recorded in a notebook. Only data for the first date of capture
of any individual were included in the analyses reported here. Results of the mark-
recapture study will be reported elsewhere (DeVries & Walla, in prep.).
Excepting a few refinements of Ehrlich’s (1958) higher classification of butterflies,
all subsequent systematic studies indicate that the phylogeny of nymphalid subfamily
relationships are unresolved (see summary in De Jong et al., 1996). In the absence
of a well resolved phylogeny it therefore seems almost arbitrary as to which higher
level classification is used, provided that the one chosen is unambiguous and well
known. As in a previous study (DeVries et al., 1997) we follow the conservative
synthesis of Ackery (1984) which is based upon Ehrlich (1958), and represents a
widely known, functional classification of nymphalid subfamilies.
Statistical analyses
Species abundance distributions were graphed following Williams (1964) who
noted that, in contrast to log base 2 (or any even number) interval widths, log base
3 interval widths with interval edges at 3
n
/2 do not overestimate rare species, or
violate the independence of data points. The species abundance distribution in
Figure 3 was plotted using log base 3 interval widths, and the goodness-of-fit to the
observed distribution was assessed for log series and log-normal distributions (Fisher,
Corbet & Williams, 1943; Williams, 1964; May, 1975). The position of the lowest
observed relative abundance (the ‘veil line’ of Preston (1948)) provided an estimate
of how completely the community had been sampled (Fig. 3).
We measure b-diversity as the component of total diversity among subdivisions
of the community in the dimensions of height (canopy and understory), area (Lake
Edge, Forest plots 1–4), or time (month). Specifically, the total, or c-diversity is
estimated by the diversity of the pooled data set for the entire sample; a-diversity
is the weighted average diversity within subdivisions (weighted by sample size); and
b-diversity equals c-diversity minus a-diversity. Thus, we use an additive partition
of diversity such that a-diversity plus b-diversity equals c-diversity. The proportion
of total diversity within subdivisions in a given dimension therefore provides a
natural measure of similarity among the subdivisions (Lande, 1996).
The hypothesis that total individual abundance for the entire community was
identical among areas was evaluated using Chi-squared tests.
SPECIES DIVERSITY IN ECUADORIAN RAINFOREST BUTTERFLIES 339
Chi-squared tests for homogeneity of observed species abundance distributions at
taxonomic levels of the total community and subfamilies were used to assess the
significance of b-diversity among sample subsets in dimensions of height, area, and
time. A sequential Bonferoni test (Rice, 1989) was then used to assess potential
table-wide type I errors at the a=0.05 level.
Species diversity was calculated using three measures: species richness, Shannon–
Wiener information and Simpson diversity (Magurran, 1988), and community
similarity indices corresponding to each of these measures were calculated as 1-b-
diversity/c-diversity (Lande, 1996).
Species accumulation curves for horizontal subdivisions of the LSL sample were
compared to assess the influence of sample size on species richness estimates (Colwell
& Coddington, 1994; Longino & Colwell, 1997). However, as estimates of species
richness in diverse communities are highly sensitive to sample size, direct comparisons
between subdivisions require a method that corrects for differences in sample sizes.
We calibrated species richness in vertical and horizontal subsets against the rarefaction
curve for the total sample (Sanders, 1968; Hurlbert, 1971; Gotelli & Graves, 1996)
which gives the expected species richness in a random subset of any particular size.
The statistical significance of such comparisons was evaluated using the approximate
95% confidence limits for the rarefaction curve, calculated as ±2 standard deviations
around expected values (Heck, van Bell & Simberloff, 1975).
To compare fruit-feeding nymphalid diversity from an intact forest (LSL) to that
of a disturbed forest ( Jatun Sacha) the larger sample size of the Jatun Sacha data
set (see DeVries et al., 1997) was rarefied to that of the present study and plotted
with 95% confidence intervals onto the rarefaction curve of the total LSL sample.
To test the hypothesis that the difference between variances of the log-normal
species abundance distributions fitted to the LSL and Jatun Sacha samples was an
artifact of the difference in the two sample sizes, the following random sampling
test was written using Mathematicaversion 3.0 (Wolfram, 1996). Individuals were
sampled at random, without replacement, from the larger Jatun Sacha sample to
form 10 000 simulated communities, each with the same number of individuals as
the smaller LSL data set. A log-normal distribution was fitted to the observed species
abundance distribution for each simulated community, and the variance of each
fitted distribution was calculated following Pielou (1975). The variance of the fitted
log-normal distribution for the LSL sample was compared to the distribution of
variances obtained for the fitted distributions from our 10 000 simulated communities,
and the proportion of the simulated variances greater than or equal to the observed
LSL variance was determined.
RESULTS
A total of 883 individual butterflies belonging to 91 species in five subfamilies
were captured during the 12 sampling periods (Table 1). The rank abundance
distribution of the entire sample showed that a large proportion of trapped butterflies
were accounted for by relatively uncommon species (Fig. 2): over 75% of species
were represented by 10 or fewer individuals. The species abundance distribution
ranged from 22 species represented by single individuals to one species, Nessaea
hewitsoni, represented by 104 individuals (Figs2&3).
P. J. DVRIES ET AL.340
T 1. A, species richness of the La Selva Lodge sample partitioned by vertical position. Rare species
are those represented by Ζ4 individuals, and common species are those represented by [5 individuals.
B, individual abundance of the La Selva Lodge sample partitioned by vertical position
Canopy Understory Both Total
(A) Species richness
Rare species 22 26 3 51
Common species 6 18 16 40
Total species 28 44 19 91
(B) Individual abundance
Total individuals 303 580 883
100
1
0
Species rank
Species abundance
5010 20 30 40 60 70 80 90
100
10
Figure 2. Rank-abundance distribution for total sample of fruit-feeding nymphalids.
Summary data showed that species richness and abundance were distributed
unequally between canopy and understory. Thirty-one per cent of the species were
found in canopy only, 48% were found in understory only, and the remaining 21%
of species were found in both strata (Table 1). When only the 51 rare species
(represented by Ζ4 individuals) were considered, the frequency of rare species was
distributed evenly with respect to vertical dimension; 22 species were found in
canopy only, 26 in understory only, and 3 were found in both. In contrast, a larger
proportion of common species (represented by [5 individuals) were found in the
understory in comparison to the canopy, and 66% of the total individual abundance
was found in the understory (Table 1).
Partitioning the entire sample into five sample areas (Lake Edge, Forest 1, Forest
2, Forest 3 and Forest 4) showed that observed species richness was unequal among
areas. Forest 3 had highest species richness and most unique species, the Lake Edge
had lowest species richness, and other areas had approximately equivalent species
richness and numbers of unique species (Table 2). The hypothesis of equal numbers
of species among areas, however, is not subject to standard statistical tests since
SPECIES DIVERSITY IN ECUADORIAN RAINFOREST BUTTERFLIES 341
8
35
00
Log
3
number of individuals
Number of species
10
30
25
20
15
5
246
B
8
35
00
Number of species
10
30
25
20
15
5
246
A
Figure 3. A, Species abundance distribution for total sample of fruit-feeding nymphalids (histogram)
from intact forest at the La Selva Lodge (91 species and 883 individuals). Parameters of the fitted log-
normal distribution (solid curve) using the method of Pielou (1975) on a log base 3 scale are: mean
1.053, variance 1.674, and estimated total number of species 100.73. The log-normal distribution
(v
2
=0.429, P=0.934) gives a better fit than the log-series distribution (v
2
=1.293, P=0.862). The
parameters of the log-series distribution (not illustrated) are a=25.456 and x=0.972. B, species
abundance distribution from disturbed forest (130 species and 6690 individuals) at Jatun Sacha (after
DeVries et al., 1997). Parameters of the fitted log-normal distribution (solid curve) are: mean 1.824
and variance 3.279. Note that variances of species abundance distributions of Jatun Sacha and La
Selva Lodge differ significantly (P<0.01) in a random sampling.
T 2. Distribution of total species richness (91 species), unique species, and total abundance among
sampling areas
Area Unique species Total species Abundance
Lake Edge 2 25 94
Forest 1 3 47 145
Forest 2 4 42 125
Forest 3 16 70 295
Forest 4 5 42 224
species are not independent (due to phylogenetic relationships), nor could they be
identically distributed (due to differences in abundance).
Species richness provided a measure of how rare and common species were
P. J. DVRIES ET AL.342
T 3. Overlap of 91 species among areas. Area ab-
breviations: LE=Lake Edge, F=Forest. Numbers in bold face
are species unique to particular areas. A, overlap of rare species
(nΖ4 individuals) occurring in area pairs. B, overlap of common
species (n[5 individuals) occurring in area pairs. C, overlap
of rare and common species occurring in 3–5 areas. Note that
most shared species are among the forest plots
(A) Rare species
LE F1 F3 F2 F4
F4 04505
F2 2374
F3 5 8 16
F1 2 3
LE 1
(B) Common species
LE F1 F3 F2 F4
F4 10 29 30 23 0
F2 13 26 27 0
F3 13 30 0
F1 13 0
LE 1
(C) Overlap of 3–5 areas
Area combinations Rare species Common species
LE-F1-F3 1 10
LE-F1-F2 0 12
LE-F1-F4 0 10
LE-F3-F2 1 11
LE-F3-F4 0 9
LE-F2-F4 0 9
LE-F1-F2-F3 0 10
LE-F1-F2-F4 0 9
LE-F1-F3-F4 0 9
LE-F2-F3-F4 0 9
F1-F3-F2 2 24
F1-F3-F4 3 28
F1-F2-F4 0 22
F3-F2-F4 0 23
F1-F3-F2-F4 0 22
All 5 areas 0 9
shared among sites (Table 3). A comparison using the 51 rare species (nΖ4
individuals) showed that the Forest 3 sample contained over three times as many
unique species as other sites, samples representing area triads of Forest 1, 3, 4 and
Forest 1, 2, 3 shared three and two species respectively, but no rare species were
shared by combinations of four areas (Table 3). For the remaining 40 common
species only the Lake Edge sample contained one unique species (Cissia erigone), but
no other areas contained unique common species. The greatest overlap of common
species (from 55 to 70%) occurred between those three and four way area com-
binations that did not include the Lake Edge sample, and 22% of the common
species were shared by all five areas (Table 3). In summary, the greatest number of
SPECIES DIVERSITY IN ECUADORIAN RAINFOREST BUTTERFLIES 343
Aug
94
100
0
Aug
93
Individual abundance
75
50
25
Sep
93
Oct
93
Nov
93
Dec
93
Jan
94
Feb
94
Mar
94
Apr
94
May
94
Jun
94
Jul
94
Canopy
Understory
Aug
94
35
0
Aug
93
Number of species
25
15
10
Sep
93
Oct
93
Nov
93
Dec
93
Jan
94
Feb
94
Mar
94
Apr
94
May
94
Jun
94
Jul
94
Canopy
Understory
B
30
20
A
Figure 4. Temporal variation of the total La Selva Lodge sample by vertical position. A, species
richness. B, individual abundance.
species were shared among combinations of plots that did not include the Lake
Edge.
Total abundances among the five areas differed significantly (v
2
=25.21, df=4,
P<0.0001), where the greatest abundance was found in the Forest 3 sample, and
the least in the Lake Edge sample (Table 2).
Temporal variation was evident in species richness and abundance in both canopy
and understory. Both species richness and abundance showed regular increases and
decreases through time (Fig. 4). Even these simple plots show the important role
temporal variation plays in the measurement of tropical species diversity.
Measures of community diversity (Magurran, 1988) and corresponding measures
of similarity among subdivisions of the community in space and time (Lande, 1996)
are provided in Table 4. The Shannon–Wiener and Simpson indices showed a
relatively high similarity among subdivisions in vertical, horizontal and temporal
dimensions in our sample, whereas species richness showed less similarity among
these dimensions.
P. J. DVRIES ET AL.344
T 4. Measures of community diversity and similarity for the total community of fruit-feeding
nymphalid butterflies at La Selva Lodge
Measure Community similarity among
heights areas months
Species richness 91 0.631 0.553 0.354
Shannon–Wiener 3.645 0.861 0.891 0.840
Simpson 0.954 0.961 0.980 0.974
∗Community similarity=1-b/c, where bis beta-diversity among subdivisions in a given dimension and cis
total community diversity (Lande, 1996).
T 5. Chi-squared tests for homogeneity of species abundance distributions among heights, areas,
and months for the total sample and subfamilies. Significance levels are: ns=not significant, ∗∗=
P<0.01, ∗∗∗=P<0.001. Application of the sequential Bonferroni test (Rice, 1989) did not affect the
significance of our results. Note: as all Morphinae were found in understory no statistical test was
performed for heights
Taxon Species richness Abundance Heights Areas Months
Total sample 91 883 ∗∗∗ ∗∗∗ ∗∗∗
Subfamily
Charaxinae 17 75 ∗∗∗ ns ∗∗∗
Nymphalinae 26 398 ∗∗∗ ∗∗∗ ∗∗∗
Morphinae 4 95 — ns ns
Brassolinae 12 96 ∗∗∗ ∗∗∗ ∗∗
Satyrinae 32 219 ∗∗∗ ∗∗∗ ns
Chi-squared tests for homogeneity of species abundance distributions demonstrated
that our sample was distributed non-randomly in all dimensions. The total sample
showed significant differences in species composition among subdivisions of vertical
position (canopy or understory), area, and sampling period (Table 5). Relative
frequencies of species in each subfamily differed significantly between canopy and
understory, among five areas (excepting Charaxinae and Morphinae), and with the
exception of Morphinae and Satyrinae, among sampling periods (Table 5).
Species accumulation curves showed that, among the plots, Forest 3 had the most
species, and Lake Edge and Forest 4 had the least species (Fig. 5). Whether Forest
1 and Forest 2 would continue to accumulate species more slowly than Forest 3
could only be determined by additional sampling.
Rarefaction of the entire LSL sample (Fig. 6) indicates that, in contrast to raw
data uncorrected for sample size (e.g. Table 2), Forest 1, 2 and 3 contained similar
numbers of species, while the Lake Edge and Forest 4 areas had significantly fewer
species than expected from a random sample of the entire community. Both canopy
and understory samples fell below the rarefaction curve for the total sample (Fig. 6)
indicating strong heterogeneity in the vertical dimension. The individual rarefaction
curves for canopy and understory showed no indication of crossing, and rarefaction
of the understory to the sample size of the canopy (with 95% confidence intervals)
suggests that the understory is expected to have approximately the same number of
species as the canopy (Fig. 6).
SPECIES DIVERSITY IN ECUADORIAN RAINFOREST BUTTERFLIES 345
350
75
Cumulative abundance
Species richness
100
30
50
45
40
35
25
20
15
10
50 150 200 250 3000
70
65
60
55
5
Lake Edge
Forest-1
Forest-2
Forest-3
Forest-4
Figure 5. Species accumulation curves showing total species versus cumulative individual abundance
through time in five sampling areas.
900
100
Sample size
Number of species
500
50
90
80
70
60
40
30
20
10
100 200 300 400 600 700 800
0
LE
F1
F4
C
U
T
JS
F2
F3
Figure 6. Rarefaction curve (solid curve) and approximate 95% confidence interval (dashed curves)
for the total La Selva Lodge sample of fruit-feeding nymphalids compared to observed species richness
in subdivisions of the sample along dimensions of height and area. Abbreviations: T=total community;
C=canopy; U=understory; LE=Lake Edge, F1=Forest 1, F2=Forest 2, F3=Forest 3, F4=Forest
4 (individual points=Β); (Χ) months; (Η) understory rarefied to sample size of canopy (with 95%
confidence intervals); (B) total Jatun Sacha sample rarefied to sample size of La Selva Lodge (with
95% confidence intervals).
P. J. DVRIES ET AL.346
When the total sample of Jatun Sacha (6690 individuals) was rarefied to that of
LSL (883 individuals) it fell below the LSL rarefaction curve, but within the 95%
confidence intervals (Fig. 6). Thus, rarefaction suggests that, despite the disparity
between species richness at these sites (130 species at Jatun Sacha and 91 species at
LSL), when standardized to the sample size of LSL species richness in these two
communities does not differ significantly.
Our random sampling test suggested that the difference between variances of
fitted species abundance distributions from LSL and Jatun Sacha was not an artifact
of sample size (Fig. 3). Ninety-nine per cent of the communities simulated from the
Jatun Sacha sample had variances greater than the observed variance of the LSL
sample, indicating that the shapes of the species abundance distributions of LSL
and Jatun Sacha samples differed significantly (P<0.01).
DISCUSSION
The scale and momentum of habitat destruction requires ecologists to accept
the practical need for quick surveys of biodiversity in ecological monitoring and
conservation planning (e.g. see Heywood, 1995; Laurence & Bierregaard, 1997).
Ultimately, however, the accuracy of quick surveys in tropical areas can be justified
only by testing their accuracy against long-term studies with intensive sampling that
partition diversity into spatial and temporal components. We show that fruit-feeding
nymphalid butterflies provide a model system for testing the accuracy of quick
biodiversity surveys, and performing detailed comparisons among sites. In contrast
to inventory or estimation techniques that rely on sampling with hand nets, sight
records or sampling at irregular intervals by different collectors (e.g. Daily & Ehrlich,
1995; Robbins et al., 1996; Brown & Hutchings, 1997), the system described here
allows more rigorous comparisons of butterfly diversity among samples in space and
time.
Trap studies have clearly advanced our understanding of tropical insect ecology
and diversity (e.g. Wolda, 1978, 1992; Hanski & Cambefort, 1991; Muirhead-
Thomson, 1991). However, it is well known that sampling bias might arise from
variance among trap positions, and variance among species in attraction to baits
(Williams, 1964; Muirhead-Thomson, 1991). Pooling replicate traps within plots (as
done here) can reduce individual trap variance, but species attraction to baits can
only be addressed by intensive mark-recapture studies (Seber, 1982) and/or detailed
observations on diet preference. As noted elsewhere (DeVries, 1988; DeVries et al.,
1997; DeVries, Lande & Murray, 1999) our methods estimate species abundance
of where adult butterflies were trapped, but not the distribution of host plants,
courtship sites, or other life history components. Nevertheless, our methods can
reduce or avoid the sampling biases in all hand net techniques that pool the efforts
of multiple persons, and trapping is clearly superior to sight records. Although
susceptibility of fruit-feeding nymphalids to traps has not been established for all
species, our methods are important for assessing tropical species diversity because
they permit more accurate comparisons and statistical analysis to be made among
samples in space and time than has been done previously.
This study provided unique estimates of spatial and temporal components of
species diversity and individual abundance for fruit-feeding nymphalid butterflies
SPECIES DIVERSITY IN ECUADORIAN RAINFOREST BUTTERFLIES 347
T 6. Comparison of species diversity between a natural lake edge at La Selva Lodge (this study)
and forest-pasture edge at Jatun Sacha (data from DeVries et al., 1997)
Species Total edge Cissia penelope Cissia penelope
richness abundance total abundance edge abundance
La Selva Lodge 25 94 2 0
Jatun Sacha 86 3686 1618 1556
within an intact tropical forest. Despite the fact that 78% of the species were
represented by fewer than 10 individuals, and 56% of the species were represented
by Ζ4 individuals (Fig. 2), the species abundance patterns in our sample fit the log-
normal distribution (Fig. 3). The position of the veil line (Preston, 1948) provided
an estimate of total species richness of 100.7 species using Pielou’s (1975) method,
and the method of Chao (1984) provided an estimate of 108.3±9.6 species.
Subsequent sampling has shown these to be underestimates of the fruit-feeding
nymphalid species richness occurring at LSL (DeVries & Walla, unpublished), and
this issue will be addressed elsewhere.
Chi-square tests for homogeneity of species abundance distributions in the total
community showed significant b-diversity in vertical, horizontal and temporal
dimensions. Significant b-diversity existed among heights for all subfamilies, and
most subfamilies showed significant b-diversity among areas and months (Table 5,
Fig. 4). Significant temporal and spatial variation in abundance and species richness
as seen in this and other studies (Kato et al., 1995; DeVries et al., 1997) suggests
that such heterogeneity is a fundamental feature of tropical forest insect communities.
Continuous observations over 5 years indicate that many species comprising Table
5 (in addition to others not sampled in this study) manifest consistent vertical
distributions at LSL (DeVries & Walla, unpublished), emphasizing the importance
of accounting for vertical dimension in tropical insect diversity estimates.
In many communities vegetation structure may have profound effects on species
richness (e.g. MacArthur, Recher & Cody, 1966; Southwood, Brown & Reader,
1979), and responses to edge effects may depend upon the focal group of organisms
(Murcia, 1995). Comparing LSL and Jatun Sacha serves to illustrate how Amazonian
butterfly species diversity may differ between forest edges. The edge at LSL comprised
a natural interface between a body of water and forest where butterfly diversity
showed a low individual abundance and species richness (Tables 2 & 6, Fig. 6). In
contrast, the edge at Jatun Sacha represents the interface between cattle pasture
and forest where butterfly diversity showed a relatively high individual abundance
and species richness (DeVries et al., 1997). Further, in a spectacular case of ecological
release (Table 6), over 1600 individuals of Cissia penelope were sampled at the Jatun
Sacha edge (DeVries et al., 1997), yet this species was represented by only two
specimens in the total LSL sample (neither from the Lake Edge). This simple
comparison highlights how different types of edges (one natural, the other due to
human disturbance) can affect two butterfly communities that share a large number
of species.
Although accumulation curves suggest a ranking of the five areas with respect to
species richness, ultimately the differences implied between them could only be
verified through continued sampling (Fig. 5). Given that intermediate levels of
P. J. DVRIES ET AL.348
disturbance may increase species richness (e.g. Connell, 1978; Huston, 1979, 1994;
Denslow, 1987), the greater richness of Forest 3 could suggest that plot has
experienced greater disturbance in the recent past than the other plots. This may,
in fact, be the case as indigenous residents report that a localized storm blew down
a significant number of large trees near this plot in August 1975, but left other areas
unaffected (L. Chowamongo, J. Hualinga and S. Machoa, pers. comms.).
Rarefaction analysis verified spatial heterogeneity in both vertical and horizontal
dimensions. The canopy and understory both fell below the 95% confidence intervals
of the rarefaction curve (Fig. 6) showing that each sample had a distinct species
composition, and only when combined did they reflect the species richness of the
total sample. Rarefaction of the understory sample to that of the canopy suggests
that both strata have similar species richness, a contrast to the disturbed Jatun Sacha
site where the canopy was shown to have a higher species richness (DeVries et al.,
1997). Rarefaction of the total community (Fig. 6) also showed that Lake Edge and
Forest 4 both fell below the rarefaction curve, confirming significant differences in
species richness among sampling plots.
The most suitable measures of species diversity for quick surveys have the desirable
statistical property of small bias when sample size is small. Of the three most
commonly used measures of species diversity, only Simpson diversity, 1-k, and to a
lesser extent Shannon–Wiener information, H, satisfy this criterion (Lande, 1996).
Although it is the least reliable statistic, species richness is often employed in
conservation applications because it is the only commonly used measure that is
sensitive to rare species (Peet, 1974). However, since species richness is highly
sensitive to the small sample sizes that are typical of quick surveys, comparison of
species richness among samples requires correction for differences in sample size
using rarefaction, species accumulation curves, and extrapolation techniques which
are best performed on large samples.
Comparisons of species richness are often used to discriminate among natural
areas to be conserved. For example, during a 12 month period we sampled 91
species of fruit-feeding nymphalids from an intact forest (Tables 1 & 2), but a
previous 12 month study (DeVries et al., 1997) sampled 130 species from a markedly
disturbed forest ( Jatun Sacha). In other words, using the same sampling regime
43% more species were found at Jatun Sacha than in the present study. Extrapolation
techniques also indicate that Jatun Sacha has a higher species richness than LSL
(see DeVries et al., 1997, and below), and under conservation practices favouring
species richness one might argue that LSL is less worthy of protection than Jatun
Sacha. Rarefaction, however, suggests that the apparent difference may be due to
disparity in the sizes of the two samples (Fig. 6).
Although rarefaction is a robust statistical tool for comparing samples of different
size, the shapes of the species abundance distributions are important for interpreting
comparisons among different communities. Estimates of total species richness using
the methods of Pielou (1975) and Chao (1984) suggest that Jatun Sacha has a higher
species richness than LSL. Alternatively, rarefaction suggests that at the sample size
of LSL the species richness of these communities do not differ significantly (Fig. 6).
Although the rarefaction curves for the two communities do not cross, the shapes
of the species abundance distributions (Fig. 3) indicate that further sampling from
LSL would yield rarefaction curves that intersect, and at large sample sizes the LSL
curve would fall below that of Jatun Sacha. This is because, at the LSL sample size,
rarefaction of the Jatun Sacha community considers a smaller proportion of the
SPECIES DIVERSITY IN ECUADORIAN RAINFOREST BUTTERFLIES 349
area beneath the log-normal species abundance distribution, and accounts for only
a small proportion of the total Jatun Sacha community (Fig. 3). This emphasizes
the importance of estimating the shape of the species abundance distribution when
comparing diversity of two sites.
Even though species abundance distributions are frequently used to describe
communities (Williams, 1964; Engen, 1978; Preston, 1980; Gray, 1987; Tokeshi,
1993), few field investigations consider the effect of disturbance on the shape of the
species abundance distribution. Both LSL and Jatun Sacha species abundance
distributions are well described by the log-normal distribution, but the LSL sample
has a significantly smaller variance than the more disturbed Jatun Sacha habitats
(Fig. 3). Results of our random sampling tests support the hypothesis that differences
in variance are not an artifact of sampling error, but reflect real differences in the
underlying distributions of our samples. We note that compared to LSL the higher
extrapolated total species richness at Jatun Sacha likely reflects a greater diversity
of habitats that have been created by recent disturbance at this site.
Species abundance patterns are known to vary depending on the community
sampled (Williams, 1964; May, 1981). For example, the classic study by Patrick,
Hohn & Wallace (1954) noted explicitly that diatom communities in polluted streams
had fewer species with greater abundances than communities in unpolluted streams,
resulting in differences in shapes of species abundance distributions. Studies of other
communities have also suggested that pollution causes similar changes in the shape
of species abundance distributions (Gray & Mirza, 1979; Ugland & Gray, 1982; but
see Lambshead, Platt & Shaw, 1983). These studies, in combination with our findings
based on two butterfly communities (Fig. 3), suggest that the generality of differences
in variances of species abundance distributions between disturbed and undisturbed
tropical forests deserves further critical investigation.
The vertical and horizontal distribution of organisms within tropical forests,
including insectivores, is well established (see Munn, 1985; Lowman & Nadkarni,
1995). Vertical and horizontal habitat partitioning by neotropical avian communities
(Munn, 1985; Cannaday, 1997), and their high degree of prey specialization (Snow,
1976; Rosenberg, 1990) reinforces the view that bird community structure plays a
significant role in the evolution and spatial association of tropical insects, and
particularly butterflies (Papageorgis, 1975; Ackery & Vane-Wright, 1984; DeVries,
1988; Turner & Mallet, 1996; Beccaloni, 1997; DeVries et al., 1999). Although the
strength of selection by predators on the structure of forest butterfly communities
remains obscure, given that small selective effects maintained over long periods of
time can produce major evolutionary changes (Wright, 1931; Haldane, 1932; Fisher,
1958; Lande, 1976), documenting diverse insect communities in space and time (as
done here) can reveal ecological patterns relevant to elucidating the evolution of
community structure in complex tropical habitats.
In concert with a previous investigation (DeVries et al., 1997) this study confirms
both the utility of long-term, standardized sampling in diverse tropical butterfly
communities and application of statistical techniques that allow comparisons among
spatial and temporal components of diversity. Our results reinforce the prevalence
of variation in diversity along vertical, horizontal and temporal dimensions, ac-
centuating the significance of considering tropical butterfly diversity in space and
time. The variation of species diversity observed here in fruit-feeding nymphalid
butterflies is likely due to underlying ecological and evolutionary factors, and implies
that such variation is inherent in other tropical insect communities. Our comparison
P. J. DVRIES ET AL.350
of two butterfly communities revealed intriguing differences in the shapes of species
abundance distributions from intact and disturbed forest that, when considered in
combination with previous work on diatom communities, invites further investigation.
Finally, studies of tropical butterfly communities have rarely, if ever, employed
standardized methods to compare diversity among different sites as reported here.
This study therefore encourages comparisons of fruit-feeding nymphalid butterfly
communities among other intact and disturbed tropical sites, and points to a means
of advancing our understanding of how spatial and temporal factors may influence
tropical insect diversity and community structure.
ACKNOWLEDGEMENTS
We thank Eric Schwartz for encouraging and supporting our field work, C.
Licuey, M. Lysinger and O. Tapuey for field assistance, L. Chowamongo, J. Hualinga
and S. Machoa for local natural history of the La Selva Lodge area, and G. Onore
(Universidad Catolica, Quito) with help obtaining research permits. For discussion
of tropical diversity and species abundance distributions we thank G. Austin, J.
Cadle, L. Emmons, S. Engen, C. Funk, R. Lande, E. Leigh, A. Magurran, R. May,
E. Pianka and M. Wood. For critical comments on drafts of this manuscript we
gratefully acknowledge R. Lande, C. Penz, E. Pianka, S. Ratner and two anonymous
reviewers. Portions of this study were supported by a National Geographic Society
grant (no. 5792-96). DeVries thanks the MacArthur Foundation and the Guggenheim
Foundation for broadly supporting his research. We dedicate this paper as an elegy
to the memory of Bert Kay, and to the rapidly dwindling biodiversity of Sucumbios
Province, Ecuador.
REFERENCES
Ackery PR. 1984. Systematic and faunistic studies on butterflies. Symposium of the Royal Entomological
Society of London 11: 9–21.
Ackery PR, Vane-Wright RI. 1984. Milkweed butterflies: their cladistics and biology. London: British
Museum (Nat. Hist.), Entomology.
Allee WC. 1926. Measurement of environmental factors in the tropical rainforest of Panama. Ecology
7: 273–302.
Anon. 1993. Rapid biodiversity assessment. Proceedings of the Biodiversity Assessment Workshop, Macquarie
University 1993. Sydney: Research Unit for Biodiversity and Bioresources, Macquarie University.
Basset Y, Aberlenc H, Delvare G. 1992. Abundance and stratification of foliage arthropods in a
lowland rain forest of Cameroon. Ecological Entomology 17: 310–318.
Bates M. 1944. Observations on the distribution of mosquitoes in a tropical forest. Ecology 25:
159–170.
Beccaloni GW. 1997. Vertical stratification of ithomiine butterfly (Nymphalidae: Ithomiinae) mimicry
complexes: the relationship between adult flight height and larval host-plant height. Biological Journal
of the Linnean Society 62: 313–341.
Beccaloni GW, Gaston KJ. 1995. Predicting the species richness of neotropical forest butterflies:
Ithomiinae (Lepidoptera: Nymphalidae) as indicators. Biological Conservation 71: 77–86.
Bowles IA, Rice RE, Mittermeir RA, daFonseca GAB. 1998. Logging and tropical forest
conservation. Science 280: 1899–1900.
Brakefield PM, Reitsma F. 1991. Phenotypic plasticity, seasonal climate and population biology
for Bicyclus butterflies (Satyridae) in Malawi. Ecological Entomology 16: 291–303.
SPECIES DIVERSITY IN ECUADORIAN RAINFOREST BUTTERFLIES 351
Brown KS Jr. 1991. Conservation of neotropical environments: insects as indicators. In: Collins NM,
Thomas JA, eds. The conservation of insects and their habitats. London: Academic Press, 449–504.
Brown KS Jr, Hutchings RW. 1997. Disturbance, fragmentation, and the dynamics of diversity in
Amazonian butterflies. In: Laurence WF, Bierregaard Jr, eds. Tropical forest remnants. Chicago:
University of Chicago Press, 91–110.
Cannaday C. 1997. Loss of insectivorous birds along a gradient of human impact in Amazonia.
Biological Conservation 77: 63–77.
Chao A. 1984. Non-parametric estimation of the number of classes in a population. Scandanavian
Journal of Statistics 11: 265–270.
Colwell RK, Coddington JA. 1994. Estimating terrestrial biodiversity through extrapolation.
Philosophical Transactions of the Royal Society of London B345: 101–118.
Cody ML, Diamond J, eds. 1975. Ecology and evolution of communities. Cambridge: Harvard University
Press.
Condit R, Hubbell SP, Lafrankie JV, Sukumar R, Manokaran N, Foster RB, Ashton PS.
1996. Species-area and species-individual relationships for tropical trees: a comparison of three 50-
ha plots. Journal of Ecology 84: 549–562.
Connell JH. 1978. Diversity in rain forests and coral reefs. Science 199: 1302–1310.
Daily GC, Ehrlich PR. 1995. Preservation of biodiversity in small rainforest patches: rapid evaluations
using butterfly trapping. Biodiversity and Conservation 4: 35–55.
De Jong R, Vane-Wright RI, Ackery PR. 1996. The higher classification of butterflies (Lepidoptera):
problems and prospects. Entomologica Scandanavica 27: 65–101.
Denslow JS. 1987. Tropical rainforest gaps and tree species diversity. Annual Review of Ecology and
Systematics 18: 431–451.
DeVries PJ. 1987. The Butterflies of Costa Rica and their Natural History. I: Papilionidae, Pieridae and
Nymphalidae. Princeton, NJ: Princeton University Press.
DeVries PJ. 1988. Stratification of fruit-feeding nymphalid butterflies in a Costa Rican rainforest.
Journal of Research on the Lepidoptera 26: 98–108.
DeVries PJ. 1997. The Butterflies of Costa Rica and their Natural History. II: Riodinidae. Princeton, NJ:
Princeton University Press.
DeVries PJ, Murray D, Lande R. 1997. Species diversity in vertical, horizontal, and temporal
dimensions of a fruit-feeding butterfly community in an Ecuadorian rainforest. Biological Journal of
the Linnean Society 62: 343–364.
DeVries PJ, Lande R, Murray D. 1999. Associations of co-mimetic ithomiine butterflies on small
spatial and temporal scales in a neotropical rainforest. Biological Journal of the Linnean Society 67:
73–85.
Edwards PJ, May RM, Web NR, eds. 1994. Large scale ecology and conservation biology. London:
Blackwell Scientific.
Ehrlich PR. 1958. The comparative morphology, phylogeny, and higher classification of the butterflies.
University of Kansas Science Bulletin 39: 305–370.
Engen S. 1978. Stochastic abundance models with emphasis on biological communities and species. London:
Chapman & Hall.
Erwin TL. 1995. Measuring arthropod biodiversity in the tropical forest canopy. In: Lowman MD,
Nadkarni N, eds. Forest Canopies. San Diego: Academic Press, 109–127.
Fisher RA. 1958. The Genetical Theory of Natural Selection. 2nd edition. New York: Dover.
Fisher RA, Corbet AS, Williams CB. 1943. The relation between the number of species and the
number of individuals in a random sample of an animal population. Journal of Animal Ecology 12:
42–58.
Forey PL, Humphries CJ, Vane-Wright RI, eds. 1994. Systematics and Conservation Evaluation.
Systematics Assoc. Special, Vol. 50. Oxford: Clarendon Press.
Gaston KJ. 1996. Biodiversity. Oxford: Blackwell Science.
Gilbert LE. 1984. The biology of butterfly communities. Symposium of the Royal Entomological Society of
London 11: 41–54.
Gill BD. 1991. Dung beetles in tropical American forests. In: Hanski I, Cambefort Y, eds. Dung beetle
ecology. Princeton, NJ: Princeton University Press, 211–229.
Gotelli NJ, Graves GR. 1996. Null models in ecology. Washington, D.C.: Smithsonian Institution Press.
Gray JS. 1987. Species-abundance patterns. In: Gee JHR, Giller PS, eds. Organization of communities
past and present. Oxford: Blackwell Scientific, 53–67.
Gray JS, Mirza FB. 1979. A possible method for detecting pollution-induced disturbance on marine
benthic communities. Marine Pollution Bulletin 10: 142–146.
Groombridge B, ed. 1992. Global biodiversity. London: Chapman & Hall.
P. J. DVRIES ET AL.352
Haldane JBS. 1932. The causes of evolution. London: Longmans, Green & Co.
Hammond PM. 1994. Practical approaches to the estimation of the extent of biodiversity in speciose
groups. Philosophical Transactions of the Royal Society of London B345: 119–136.
Hanski I, Cambefort Y. 1991. Spatial processes. In: Hanski I, Cambefort Y, eds. Dung beetle ecology.
Princeton, NJ: Princeton University Press, 283–304.
Hayek LC, Buzas MA. 1996. Surveying natural populations. New York: Columbia University Press.
Heck KL, van Belle G, SimberloffD. 1975. Explicit calculation of the rarefaction diversity
measurement and the determination of sufficient sample size. Ecology 56: 1459–1461.
Heywood VH, ed. 1995. Global biodiversity assessment. Cambridge: Cambridge University Press.
Hubbell SP, Foster RB. 1986. Commonness and rarity in a neotropical forest: implications for
tropical tree conservation. In: Soule
´ME, ed. Conservation biology. The science of scarcity and diversity.
Sunderland: Mass.: Sinauer Associates, 205–231.
Hurlbert SH. 1971. The non-concept of species diversity: a critique and alternative parameters.
Ecology 52: 577–586.
Huston M. 1979. A general hypothesis of species diversity. American Naturalist 113: 81–101.
Huston MA. 1994. Biological diversity, the coexistence of species on changing landscapes. Cambridge: Cambridge
University Press.
Kato M, Inoue T, Hamid AA, Nagamitsu T, Merdek MB, Nona AR, Itino T, Yamane S,
Yumoto T. 1995. Seasonality and vertical structure of light-attracted insect communities in a
dipterocarp forest in Sarawak. Researches on Population Ecology 37: 59–79.
Keister AR, Scott JM, Csuti B, Noss RF, Butterfield B, Sahr K, White D. 1996. Conservation
prioritization using GAP data. Conservation Biology 10: 1332–1324.
Kremen C. 1992. Assessing the indicator properties of species assemblages for natural areas monitoring.
Ecological Applications 2: 203–217.
Kremen C. 1994. Biological inventory using target taxa: a case study of butterflies of Madagascar.
Ecological Applications 4: 407–422.
Lambshead PJD, Platt HM, Shaw KM. 1983. Detection of differences among assemblages of
benthic species based on an assessment of dominance and diversity. Journal of Natural History, London
17: 859–974.
Lande R. 1976. Natural selection and random genetic drift in phenotypic evolution. Evolution 30:
314–334.
Lande R. 1996. Statistics and partitioning of species diversity, and similarity among multiple
communities. Oikos 76: 5–13.
Laurence WF, Bierregaard RO, eds. 1997. Tropical forest remnants. Chicago: University of Chicago
Press.
Longino JT, Nadkarni N. 1990. A comparison of ground and canopy leaf litter ants (Hymenoptera:
Formicidae) in a tropical montane forest. Psyche 97: 81–94.
Longino JT, Colwell RK. 1997. Biodiversity assessment using structured inventory: capturing the
ant fauna of a tropical rain forest. Ecological Applications 7: 1263–1277.
Lowman MD, Nadkarni N, eds. 1995. Forest canopies. San Diego, CA: Academic Press.
MacArthur RH. 1972. Geographical ecology. Princeton, NJ: Princeton University Press.
MacArthur RH, Recher H, Cody ML. 1966. On the relation between habitat-selection and species
diversity. American Naturalist 100: 319–325.
Magurran AE. 1988. Ecological diversity and its measurement. Princeton, NJ: Princeton University Press.
Malcolm JR. 1994. Edge effects in central Amazonian forest fragments. Ecology 75: 2438–2445.
Mallet J, Gilbert LE. 1995. Why are there so many mimicry rings? Correlations between habitat,
behavior and mimicry in Heliconius butterflies. Biological Journal of the Linnean Society 55: 159–180.
May RM. 1975. Patterns of species diversity and abundance. In: Cody ML, Diamond JM, eds. Ecology
and evolution of communities. Cambridge, MA: Harvard University Press, 81–120.
May RM. 1981. Patterns in multi-species communities. In: May RM, ed. Theoretical ecology: principles
and applications. Oxford: Blackwell, 197–227.
Morse DR, Stork NE, Lawton JH. 1988. Species number, species abundance and body length
relationships of arboreal beetles in Bornean lowland rain forest trees. Ecological Entomology 13: 25–37.
Munn CA. 1985. Permanent canopy and understory flocks in Amazonian species composition and
population density. Ornithological Monographs 36: 683–712.
Murcia C. 1995. Edge effects in fragmented forests: implications for conservation. Trends in Ecology &
Evolution 10: 58–62.
Muirhead-Thomson RC. 1991. Trap responses of flying insects. London: Academic Press.
Noss RF. 1990. Indicators for monitoring biodiversity: a hierarchical approach. Conservation Biology 4:
355–364.
SPECIES DIVERSITY IN ECUADORIAN RAINFOREST BUTTERFLIES 353
Novotny V, Basset Y. 1998. Seasonality of sap-sucking insects (Auchenorrhyncha, Hemiptera) feeding
on Ficus (Moraceae) in a lowland rain forest in New Guinea. Oecologia 115: 514–522.
Olson D, Dinnerstein E, Castro G, Maravi E, eds. 1996. Identifying gaps in botanical information
for biodiversity conservation in Latin America and the Caribbean. Washington, D.C.: World
Wildlife Fund.
Papageorgis CA. 1975. Mimicry in Neotropical butterflies. American Scientist 63: 522–532.
Patrick R, Hohn M, Wallace J. 1954. A new method for determining the pattern of the diatom
flora. Noctulae Natura Academy of Natural Sciences, Philadelphia 259: 1–12.
Pearson DL. 1977. A pantropical comparison of bird community structure on six lowland forest
sites. Condor 79: 232–244.
Pearson DL. 1994. Selecting indicator taxa for the quantitative assessment of biodiversity. Philosophical
Transactions of the Royal Society of London B345: 75–79.
Peet RK. 1974. The measurement of species diversity. Annual Review of Ecology and Systematics 5:
285–307.
Pinheiro CEG, Ortiz JVC. 1992. Communities of fruit-feeding butterflies along a vegetation gradient
in central Brazil. Journal of Biogeography 19: 505–511.
Pielou FC. 1975. Ecological diversity. New York: Wiley.
Preston FW. 1948. The commonness and rarity of species. Ecology 29: 254–283.
Preston FW. 1980. Noncannonical distribution of commonness and rarity. Ecology 61: 88–97.
Richards PW. 1952. The tropical rain forest, an ecological study. Cambridge: Cambridge University Press.
Rice WR. 1989. Analyzing tables of statistical tests. Evolution 43: 223–225.
Ricklefs R, Schluter D, eds. 1993. Species diversity in ecological communities. Chicago: University of
Chicago Press.
Robbins RK, Lamas G, Mielke OH, Harvey DJ, Casagrande MM. 1996. Taxonomic com-
position and ecological structure of the species-rich butterfly community at Pakitza, Parque Nacional
del Manu, Peru. In: Wilson DE, Sandoval A, eds. La Biodiversidad del Sureste del Peru. Lima, Peru:
Editorial Horizonte, 201–236.
Roberts L. 1991. Ranking the rain forests. Science 251: 1559–1560.
Rosenberg KV. 1990. Dead-leaf foraging specialization in tropical forest birds: measuring resource
availability and use. In: Morrison ML, Ralph CJ, Jehl JR Jr, eds. Avian foraging: theory, methodology,
and applications. Studies in Avian Biology 13: 360–368.
Rosenzweig ML. 1995. Species Diversity in Space and Time. Cambridge: Cambridge University Press.
Ryti RT. 1992. Effect of the focal taxon on the selection of nature reserves. Ecological Application 2:
404–410.
Sanders HL. 1968. Marine benthic diversity: a comparative study. American Naturalist 102: 243–282.
Seber GAF. 1982. The estimation of animal abundance and related parameters. 2nd ed. London: Charles
Griffin.
Snow DW. 1976. The web of adaptation: bird studies in the American tropics. New York: Quadrangle.
Southwood TRE, Brown VK, Reader PM. 1979. The relationship of plant and insect diversities
in succession. Biological Journal of the Linnean Society 12: 327–348.
Stork NE. 1988. Insect diversity: facts, fiction and speculation. Biological Journal of the Linnean Society
35: 321–337.
Sutton SL, Hudson PJ. 1980. The vertical distribution of small flying insects in the lowland rainforest
of Zaire. Zoological Journal of the Linnean Society 68: 111–123.
Terborgh J, Robinson SK, Parker TA, Munn CA, Pierpont N. 1990. Structure and organization
of an Amazonian forest bird community. Ecological Monographs 60: 213–238.
Tokeshi M. 1993. Species abundance patterns and community structure. Advances in Ecological Research
24: 111–186.
Turner JRG, Mallet JB. 1996. Did forest islands drive the diversity of warningly coloured butterflies?
Biotic drift and the shifting balance. Philosophical Transactions of the Royal Society B351: 835–845.
Ugland KI, Gray JS. 1982. Lognormal distributions and the concept of community equilibrium.
Oikos 39: 171–178.
Williams CB. 1964. Patterns in the balance of nature and related problems in quantitative ecology. London:
Academic Press.
Wilson DE, Sandoval A, eds. 1996. La Biodiversidad del Sureste del Peru. Lima, Peru: Editorial
Horizonte.
Wolda H. 1978. Fluctuations in abundance of tropical insects. American Naturalist 112: 1017–1045.
Wolda H. 1992. Trends in abundance of tropical forest insects. Oecologia 89: 47–52.
Wolfram S. 1996. Mathematica 3.0. Wolfram Industries.
Wright S. 1931. Evolution in Mendelian populations. Genetics 16: 97–159.
