Download Community-wide body size differences between nocturnal and

February 2013
NOTES
537
Ecology, 94(2), 2013, pp. 537–543
Ó 2013 by the Ecological Society of America
Community-wide body size differences between nocturnal and
diurnal insects
JENNIFER GUEVARA1
AND
LETICIA AVILÉS
Department of Zoology, University of British Columbia, Vancouver, British Columbia V6T 1Z4 Canada
Abstract. Examining community-wide patterns for the most diverse animal group,
insects, is fundamental to our understanding of the ecological and evolutionary factors that
maintain tropical diversity. Using several sampling techniques (malaise traps, pitfall traps,
visual searches, and social spider nest captures), we investigated the day–night community
composition of active insects to reveal differences in body size at three elevations in eastern
Ecuador. We show that insects active at night are, on average, larger than those active during
the day. Even though insect size decreased with increasing elevation, the observed diel pattern
was consistent across elevations, and for most insect orders. All sampling techniques
consistently detected day–night differences in insect size, except for social spider captures at
the two higher elevations, probably due to the reduced range of colony sizes at the higher
elevations and possibly lower spider activity at night. We suggest that the observed diel
patterns in insect size may be driven by a combination of factors, including increased risk
imposed on large insects by diurnal visual predators, mainly insectivorous birds, and
physiological responses to diel changes in abiotic conditions.
Key words: Anelosimus; body size; diel patterns; Ecuador; elevational gradients; insects; inventory;
social spiders; tropical community.
INTRODUCTION
Insects, through their unparalleled diversity and
numerical dominance, play key ecological roles shaping
and maintaining local and global biodiversity (Stork
2007). As such, ecologists have long sought to identify
and understand patterns in insect diversity, biomass,
abundance, activity, and body size (Janzen et al. 1976,
Basset and Kitching 1991, Stork and Blackburn 1993).
Community-wide insect size distributions are perhaps
some of the less explored patterns despite their critical
role in many aspects of ecological communities. Most of
our knowledge of insect size distributions involves
broad-scale geographical clines, including a seemingly
general trend for decreasing insect size with increasing
elevation and latitude (Janzen et al. 1976, Gaston and
Blackburn 2000, Blanckenhorn and Demont 2004,
Guevara and Avilés 2007, Powers and Avilés 2007),
but there are instances of the reverse trend within certain
insect taxa (Gaston and Blackburn 2000, Blanckenhorn
and Demont 2004). Insect size distributions at more
local scales have been less explored despite the
implications in local biotic processes (i.e., predator–prey
Manuscript received 6 January 2012; revised 27 August 2012;
accepted 10 September 2012. Corresponding Editor: J. A.
Rosenheim.
1
E-mail: [email protected]
and other multi-trophic interactions), community structure, and local diversity (Gaston and Blackburn 2000).
Existing studies already have revealed vertical stratification in forest insect size. At an Indonesian rain forest
site, Stork and Blackburn (1993) found that insects were
smaller in the understory, possibly as a result of higher
desiccation risk in the canopy. At a Bornean rain forest
site, Schulze et al. (2001) found, in contrast, that
Lepidopterans were larger in the understory, a pattern
presumably due to higher predation risk by birds in the
canopy. Local body size shifts may also arise from
habitat disturbance, with larger insects being found in
less disturbed areas (Shahabuddin et al. 2010). Similar
studies revealing local variation in insect size across time
and space remain notably scarce. In particular, none has
explored day–night patterns.
Day–night variation in local ecological conditions
may drive various community-wide insect patterns,
including diversity, abundance, and activity (Basset
and Springate 1992, Springate and Basset 1996). Similar
effects on insect size can be predicted, as body size
distributions are largely driven by their interaction with
the environment. Environmental differences between
day and night may either exert a selective pressure on the
size of insects active at different time periods or induce
behavioral or physiological responses leading to the
ecological sorting of day- and night-active insects.
538
NOTES
Relevant environmental factors include differences in
predation risk between day and night or diel changes in
abiotic factors such as temperature and solar radiation.
Larger insects would be expected at night, for instance,
under pressure from diurnal visual predators, such as
birds, the dominant size-dependent predators of insects
(Holmes et al. 1979, Berger and Gotthard 2008). Since
large insects are easier to see than smaller ones, birds
may be more likely to target them during the day (Curio
1977, Berger et al. 2006). This pressure would be relaxed
during the night, allowing increased activity of larger
insects. Nightly drops in temperature, particularly at
higher elevations, may also favor larger insects at night
as larger insects cool more slowly due to their smaller
surface area to volume ratio. In such scenarios, insect
activity may shift in unison with predation levels (Lima
and Dill 1990) or physiological responses to temperature
(Somme 2008), thus predicting that nocturnal active
insects would be larger than diurnal ones. If desiccation
risk, in contrast, is the predominant factor affecting diel
patterns in insect size, one might expect larger insects to
be active during the day because of their decreased
surface-area-to-volume ratio.
With these ideas in mind, we used a combination of
sampling techniques that allow comparison of day–night
active insects at three sites of varying elevation in eastern
Ecuador. Under the first two scenarios described above,
one would predict that night active insects would be
larger than those active in the daytime, while the third
scenario would predict the opposite pattern. We further
explored whether diel variation in insect size is
consistent across elevations and among insect taxa.
METHODS
Study sites
We collected our data at three nearby sites along an
elevation gradient on the eastern slope of the Andes in
the Napo Province, Ecuador. All sites occurred along or
within a few kilometers of the Baeza-Tena road. Our
low-elevation site was the Estación Biológica Jatun
Sacha (1.0728 S, 77.7318 W; elevation 385–410 m), a
lowland tropical rain forest consisting mainly of primary
forest with some areas of regenerated forest. We
conducted insect surveys during July and August
2005–2008. Our mid-elevation site was near the Hollin
River (0.6958 S, 77.7318 W; elevation 950–1100 m), a
montane rain forest site consisting of shrubby vegetation
and disturbed habitat along the Jondachi-Loreto road.
Here, we sampled insects in June and July 2006. Our
high-elevation site was in the Cocodrilos area of the
Reserva Ecologica Antisana (0.648 S, 77.88 W; elevation
1740–1940 m), a cloud forest site consisting of primary
forest, shrubby vegetation and disturbed habitat along
the Baeza-Tena road. Here, we performed insect surveys
in June and July 2008.
Ecology, Vol. 94, No. 2
Insect size surveys
We surveyed insects using four sampling techniques
that target active insects, thus permitting comparison of
day–night activity: no-kill malaise traps, pitfall traps,
visual searches, and social spider nest captures. The
prey-capture webs of social spider nests of the genus
Anelosimus (Araneae: Theridiidae) serve as a convenient
sampling method as they are three-dimensional structures that effectively intercept active flying insects
(Avilés 1997). Depending on the species and number
of individuals they contain, the nests range from a few
centimeters to sometimes a few meters in length for one
of the species in the lowland rain forest, hence capturing
prey that reflect the local range of insect sizes (Guevara
and Avilés 2007) as well as their activity patterns. The
four techniques used may target different insect types
and size ranges, thus they are complementary and
together can provide a better assessment of the insect
size composition than a single method. For the purpose
of our study, we did not include other popular methods
because these may either target sedentary taxa that may
not be representative of the taxa active at a given time
period (i.e., sweeping and beating) or do not provide a
means of comparison between periods (i.e., black lights
only sample night-active insects).
We conducted surveys with malaise traps and visual
searches at random points in the vicinity of existing
spider nests. At each site, we set up malaise traps
(height, 1.2 m; base, 1.2 m) in separate areas at
approximately 5–10 m away from spider nests (lowelevation N ¼ 8 traps in 2005, mid-elevation N ¼ 7 traps
in 2006, high-elevation N ¼ 5 traps in 2008). The traps
contained an insect-retention bottle at the top that
allowed entry of insects but not their exit. We checked
the traps daily in the evening (day sample between 17:00
and 18:00) and early in the morning (night sample
between 07:00 and 08:00). We removed the bottles with
the insects each time and replaced them with empty
bottles of the same size for the next sampling period. We
performed visual searches one to six times per day
(mode twice a day) at different haphazardly chosen
locations within each site. For this, we visually searched
for five consecutive minutes active and/or flying insects
in the surroundings. We sampled insects using pitfall
traps only at our low-elevation site.
We surveyed spider nests every 1–1.5 h in the day
(07:15–17:30) and night (18:30–01:00). Each time, we
noted the presence of new insects caught in the web or
being eaten by the spiders. At the low-elevation site, we
surveyed nests of two social spiders belonging to the
Family Theridiidae, Anelosimus eximius Keyserling and
Anelosimus domingo Levi (Avilés 1997), along a ;2 km
trail in the forest interior and in areas surrounding the
field station where they are typically found. At the midelevation site, we studied the subsocial Anelosimus
elegans Agnarsson and a second population of the
February 2013
NOTES
social A. eximius. We found nests of both species along
the road edge and nearby pastures. At our highelevation site, we studied the social A. guacamayos
Agnarsson (Avilés et al. 2007) and a second population
of the subsocial A. elegans. We found nests of both
species along an ;2-km transect along the main road.
We surveyed a total of 50 A. eximius and 26 A. domingo
nests in 2005, 2006, and 2009 at the low-elevation site, 32
A. eximius and 35 A. elegans nests in 2006 at the midelevation site, and 73 A. guacamayos and 141 A. elegans
nests in 2008 at the high-elevation site.
We measured insect size as the length (to the nearest
millimeter) between the most anterior point of the head
to the tip of the abdomen. For malaise, pitfall, and most
spider nest samples, we used a ruler for length
measurements. For visual searches, and sometimes for
insects trapped in spider nests, we estimated length by
eye to the nearest 0.5 cm, always by the same observer.
We classified the specimens to order. In total, our data
set included 6427 insects at the three sites.
Data analysis
We conducted three separate analyses where we used
as the response variable the log-transformed mean insect
size collected in a sampling bout (one day or night
malaise, pitfall, visual, or spider catches), weighted by
the number of insects in each bout. Number of insects
for daytime bouts (across all sites) ranged from 2 to 57
for malaise, 3 to 17 for pitfall, 6 to 74 for visual, and 0 to
27 for spider catches. Number of insects for night bouts
ranged from 1 to 41 for malaise, 4 to 12 for pitfall, 6 to
39 for visual, and 0 to 12 for spider catches.
In the first analysis, we tested for overall diel
differences in insect size across sites using a linear
mixed-effects model where the response variable was the
log-transformed mean body size per bout and the fixed
effects were period (day, night), site (low-, mid-, highelevation), and the period 3 site interaction. We treated
spider nest/trap ID as a random effect.
In the second analysis, we tested for differences
among techniques within each site using a similar model
where period, technique (malaise, pitfall, visual, or
spiders), and the period 3 technique interaction were
the fixed effects, and nest/trap ID the random effect. We
used independent contrasts to compare day vs. night for
each technique at each site, and for comparisons of
overall day–night insect size across sites.
In the third analysis, we tested for diel differences in
the size of different insect groups. We divided the
specimens in each bout into the seven most common
taxa plus an ‘‘other’’ category for less common taxa. Our
response variable was the log-transformed mean insect
size (per taxa) per bout, the fixed effects were period,
taxon, and period 3 taxon interaction, and the random
effect was nest/trap ID. All analyses were performed in
539
FIG. 1. Insect sizes (log-transformed body length, mean and
95% CI, measured in mm) from day and night samples obtained
by a combination of sampling techniques at three sites of
varying elevation in eastern Ecuador. Open circles are day
samples; solid circles are night samples.
R 2.14.2 (R Development Core Team 2012), using the
lme function in the nlme package (available online).2
RESULTS
At our low-elevation site, nocturnal insects were, on
average, larger than diurnal ones (F2, 199 ¼ 1733.5, P ,
0.0001; Figs. 1 and 2). This pattern was consistent across
sampling techniques as shown by a nonsignificant
interaction between technique and period (F4, 199 ¼
0.70, P ¼ 0.60; independent contrasts, day vs. night, A.
domingo F1, 199 ¼ 13.2, P ¼ 0.0004; A. eximius F1, 199 ¼
21.5, P , 0.0001; malaise F1, 199 ¼ 4.6, P ¼ 0.034; pitfall
F1, 199 ¼ 3.1, P ¼ 0.008; visual F1, 199 ¼ 5.7, P ¼ 0.01; Fig.
2).
At the two other sites, insects were also larger at night
than during the day (mid-elevation F2, 205 ¼ 847.4, P ,
0.0001; high-elevation F2, 352 ¼ 1624.1, P , 0.0001; Fig.
1) for all sampling techniques used, except the spider
catches (Fig. 2). This was evidenced by significant
interactions between technique and period at the two
higher elevations (mid-elevation F3, 205 ¼ 2.3, P ¼ 0.04;
high-elevation F3, 352 ¼ 7.6, P ¼ 0.0001). Thus, differences in insect size between day and night were
significant for samples from malaise traps (independent
contrasts, day vs. night, mid-elevation F1, 205 ¼ 24.3, P ,
0.0001; high-elevation F1, 352 ¼ 9.8, P ¼ 0.002; Fig. 2) and
visual searches (day vs. night, mid-elevation F1, 205 ¼
15.5, P ¼ 0.0001; high-elevation F1, 352 ¼ 20.9, P ,
0.0001), but not for spider samples at the mid-elevation
(A. elegans and A. eximius P . 0.05) and high-elevation
sites (A. elegans and A. guacamayos P . 0.05; Fig. 2).
Although overall insect size decreased with elevation
(F2, 428 ¼ 120.8, P , 0.0001; Fig. 1), insects were
consistently larger at night than during the day at all
2
http://cran.r-project.org/web/packages/nlme/index.html
540
NOTES
Ecology, Vol. 94, No. 2
FIG. 2. Insect sizes (log-transformed body length, mean and 95% CI, measured in mm) for day and night samples obtained by
different sampling techniques at three sites of varying elevation in eastern Ecuador (open circles are day samples; solid circles are
night samples) and the overall distribution of insect body sizes for each site (open bars are day samples day; black bars are night
samples). Abbreviations are, for spider nests samples: Anelosimus domingo (dom), A. eximius (exi ), A. guacamayos (gua), and A.
elegans (ele); malaise trap samples (mal); pitfall trap samples (pit); visual search samples (vis).
TABLE 1. Day and night mean insect body size for the most common insect orders captured with a combination of sampling
techniques at three different sites of varying elevations in eastern Ecuador.
Low-elevation (385–410 m)
Day
Order
Coleoptera
Diptera
Hymenoptera
Hemiptera/Homoptera
Lepidoptera
Orthoptera
Blattaria
Other
Body size
7.3
5.8
7.5
6.5
11.9
11.4
16.1
9.5
(5.6–9.4)
(4.5–7.6)
(5.8–9.7)
(5.0–8.4)
(9.1–15.7)
(8.7–15.0)
(11.7–22.3)
(7.1–12.8)
Mid-elevation (950–1100 m)
Night
n
175
103
176
99
77
89
37
43
Body size
14.3
6.7
8.4
10.9
19.6
21.5
19.4
14.1
(10.3–19.9)
(4.8–9.2)
(6.2–11.5)
(7.6–15.5)
(13.9–27.5)
(15.7–29.6)
(13.2–28.4)
(10.3–19.4)
Day
n
26
33
32
18
28
32
13
31
Body size
6.2
3.9
6.1
5.5
10.2
10.0
8.4
5.8
(4.0–9.6)
(2.5–6.0)
(4.0–9.4)
(3.6–8.4)
(6.7–16.1)
(6.3–15.8)
(5.1–14.0)
(3.7–9.0)
Night
n
40
55
55
87
25
23
11
25
Body size
7.6
4.7
6.7
7.2
12.2
15.4
11.6
8.0
(4.8–11.8)
(3.0–7.2)
(4.3–10.4)
(4.7–11.1)
(7.6–19.4)
(9.6–24.8)
(6.8–19.9)
(5.1–12.6)
n
32
61
37
52
34
23
7
25
Notes: Values represent back-transformed fitted mean body length (mm) with 95% CI. Sample size for the number of bouts in the
mean estimate is given by n.
February 2013
NOTES
sites (independent contrasts P , 0.05; Fig. 1). The lowelevation site had the largest insects for both time
periods (back-transformed values, means 6 95% CI, day
6.6 6 1.9 mm; night 12.1 6 2.8 mm) followed by the
mid- (day 4.4 6 1.2 mm; night 8.6 6 1.9 mm) and highelevation sites (day 3.5 6 0.98 mm; night 5.8 6 1.6 mm).
Although nocturnal insects were on average larger than
diurnal ones for the most common orders (lowelevation, F2, 401 ¼ 1283.6, P , 0.0001; mid-elevation,
F2, 283 ¼ 645.6, P , 0.0001; and high-elevation, F2, 497 ¼
1414.7, P , 0.0001; Table 1), this day–night difference
did not always hold across insect orders (taxon 3 period,
low-elevation, F7, 401 ¼ 3.2, P ¼ 0.003; mid-elevation,
F7, 283 ¼ 0.6, P ¼ 0.8; and high-elevation, F7, 497 ¼ 3.9, P
, 0.0001).
DISCUSSION
We used a combination of techniques to sample the
size range of night- and daytime-active insects at three
sites of varying elevation in eastern Ecuador. Our data
showed that night-active insects were on average larger
than insects active in the daytime, a pattern that held
across an elevational range, even though insect size
decreased with elevation (for the elevational pattern, see
also Guevara and Avilés [2007] and Powers and Avilés
[2007]). The pattern was also consistent across sampling
techniques, with the exception of the social spider
captures, which showed day–night insect size differences
at the low-elevation rain forest site, but not at higher
elevations.
There are several potential ecological, physiological,
and evolutionary explanations for the observed day–
night insect size pattern that need not be mutually
exclusive. Strong diurnal predation pressure, for instance, could drive day–night shifts in insect activity
(Lima and Dill 1990), thus potentially influencing shifts
in body size if size-dependent predation occurs. Just as
palatable caterpillars and other insects may have
evolved nocturnal lifestyles to avoid visual diurnal
predators (Janzen 1983, Berger and Gotthard 2008),
TABLE 1. Extended.
High-elevation (1740–1940 m)
Day
Body size
3.8
3.7
5.0
4.7
9.5
10.0
10.1
6.4
(2.9–5.2)
(2.8–5.0)
(3.7–6.7)
(3.5–6.3)
(6.8–13.1)
(7.1–14.2)
(6.5–15.9)
(4.6–8.9)
Night
n
141
282
177
250
33
7
38
Body size
6.5
4.1
6.2
4.7
15.2
23.0
12.4
8.9
(4.6–9.1)
(3.0–5.6)
(4.2–9.0)
(3.4–6.5)
(10.2–22.4)
(13.8–38.4)
(7.4–20.8)
(6.0–13.2)
n
28
70
21
44
26
5
4
14
541
day–night body size differences in active insects may
also arise if similar predation pressures exist on
conspicuously large insects, particularly at tropical
latitudes where biotic interactions are notably important
(Schemske et al. 2009). Insectivorous birds, which are
dominant visual insect predators (Holmes et al. 1979,
Cornell and Hawkins 1995), are diurnal foragers known
to cause high size-dependent insect mortality. Similarly,
ants, which mostly forage during the day, may also exert
strong predation pressure due to their high abundance in
the tropics (Floren et al. 2002). Some Neotropical ants,
for example, use ambush group foraging to specifically
target large mobile insects (Morais 1994). Size-dependent predation may influence insect size patterns either
via ecological sorting (i.e., more favorable nocturnal
conditions for large insects allow for their increased
activity at this time) or via natural selection either
favoring small insects during the day or allowing insects
to become larger at night under relaxed predation. Large
nocturnal insects, however, are not completely exempted
from nocturnal predation as they may be, for example,
more obvious ‘‘radar’’ targets for echo-locating predators (i.e., bats). Thus, the observed patterns should arise
only if the relative risk in size-dependent predation is
greater in the daytime.
Abiotic factors have also been shown to influence
physiological and behavioral responses of insects (Hodkinson 2005). Temperature and solar radiation, for
instance, may influence activity regimes associated with
body size given the larger ratio of surface area to volume
of small insects. Our patterns are consistent with the
prediction that larger insects should be active at night,
when temperatures are lower, but argue against desiccation risk, which should be greater during the day for
small insects, being the predominant factor driving diel
patterns in insect activity and size at our study sites.
Abiotic factors have been shown to influence physiological and behavioral responses of insects (Hodkinson
2005). Tropical rain forests, in general, exhibit higher
nocturnal insect activity compared to temperate forests
or higher elevations because night air temperatures
remain in warm ranges (Janzen 1967, 1973). As
minimum temperatures and relative humidity decrease
with elevation, however, diel temperature changes
become more pronounced, often alternating from high
daytime to very low night temperatures, particularly at
the equator (Sarmiento 1986 ). At mid to upper
elevations in the Andes, daily temperature fluctuations
often involve 208C or more. In the lowlands, the daily
temperature fluctuation is much lower, closer to about
108C (Neil and Jorgensen 1999). Smaller-sized insects
may be less active at night because they may need to seek
shelter to be able to withstand low night temperatures
(i.e., they cool faster due to their larger surface area to
volume ratios [Somme 2008]). Small-sized insects,
particularly tropical ones, may also face great desicca-
542
NOTES
tion challenges due to strong daytime insolation that
would favor higher daytime activity of larger insects.
Although this prediction is not supported by our overall
pattern of night-active insects being larger than diurnal
ones, the less pronounced difference in day–night insect
size at higher elevations (Fig. 1) may be partly explained
by a greater daytime desiccation risk of smaller insects at
higher elevations. This, in turn, is consistent with abiotic
factors playing a greater role at our higher elevation sites
and predation potentially playing a greater role at lower
elevations. That predation, rather than temperature,
may be the primary factor driving the larger size of
nocturnal insects is further supported by our finding (see
also Guevara and Avilés 2007, Powers and Avilés 2007)
that insects are smaller at the higher elevations where
temperatures are lower, which is the opposite of what is
expected under Bergmann’s rule (i.e., reverse Bergmann’s rule [Lomolino et al. 2010]). A reassuring finding
is that diel patterns were for the most part consistent
across elevations and across most insect orders.
Diel patterns in insect size obtained by the combination of techniques were also paralleled by spider catches
at the low-elevation site. Several reasons may account
for the lack of a difference at higher elevations. One
possibility is that the spiders at higher elevations may be
less active at night than their counterparts at the lowelevation site, as changes in ambient temperature and
wind speed may affect spider activity (Crouch and Lubin
2000). High elevation spider populations may thus not
be able to take advantage of the larger insects available
to them at night. Additionally, in Anelosimus spiders,
level of sociality and colony size decrease with elevation,
both within (Purcell and Avilés 2007) and among (Avilés
et al. 2007) species. The capture of the largest insects by
the subsocial A. elegans at the mid- and high-elevation
sites, in particular, may thus be constrained by its
smaller nests, as colonies consist only of the offspring of
a single female (Avilés 1997). A similar explanation may
apply to some extent to the social species, A. eximius at
the mid-elevation and A. guacamayos at the highelevation sites, as their populations consist of smaller
nests and a greater fraction of solitary females as
elevation increases (Avilés et al. 2007, Purcell and Avilés
2007). At the low-elevation site, on the other hand,
successful capture of large nocturnal prey is possible due
to the large nests and high cooperation levels of the
species found at this site (Yip et al. 2008).
Finally, we must mention that each sampling method
has limitations. Visual searches, for example, are biased
toward larger insects (Guevara and Avilés 2007),
particularly at night when small insects may be more
difficult to see. We aimed to minimize night–day
efficiency biases by performing searches under sufficient
night light and by the same observers at each site. We,
however, caution interpretation of visual search data as,
unlike the other sampling techniques used, the potential
Ecology, Vol. 94, No. 2
for human error is considerable. In addition, visual
searches and spider nest surveys were not done during
the entire night period, thus possibly excluding variation
in insect size due to crepuscular changes in abiotic or
biotic factors. Although malaise traps tend to miss large
as well as flightless active insects, they should be less
prone to night–day efficiency biases. The potential
shortcomings of individual techniques, however, should
be alleviated by the use of a combination of complementary techniques, including those that target small
and large insects or both active sedentary and flying
ones. Lastly, we note that although a phylogenetically
corrected analysis would be desirable for a strong
evolutionary inference of the processes behind the
patterns observed, such an analysis is beyond the scope
of this paper.
Studies reporting insect size distributions in the
tropics are generally scarce. Data on diel insect patterns
are, in particular, notably biased toward quantification
of species richness, abundance, and activity. To our
knowledge, community-wide diel patterns in insect size
have not been previously reported for the tropics or
other latitudes, thus there are no relevant studies
providing means of comparison to our results. We
conclude that, given the sampling techniques we used at
three elevations in eastern Ecuador, nocturnal insects
are larger than diurnal ones, at least at tropical latitudes.
Our study gives new insight into day–night shifts in
insect body size at local scales and provides testable
explanations of the processes creating such pattern.
ACKNOWLEDGMENTS
We thank the Ecuadorian Government for research permits,
the Corporación Sociedad para la Investigación y Monitoreo de
la Biodiversidad Ecuatoriana SIMBIOE and Pontifica Universidad Católica del Ecuador for sponsoring our work in
Ecuador, the Guardianı́a of Parque Nacional Sumaco and the
Estación Biológica Jatun Sacha for accommodation, field
assistants, and the Avilés lab members and two anonymous
reviewers for comments on the manuscript. Funding was
provided by an NSERC Discovery Grant and a James S.
McDonnell Foundation Grant to L. Avilés and an NSERC
Canada Graduate Scholarship to J. Guevara.
LITERATURE CITED
Avilés, L. 1997. Causes and consequences of cooperation and
permanent sociality in spiders. Pages 476–498 in J. C. Choe
and B. Crespi, editors. The evolution of social behaviour and
cooperation in insects and arachnids. Cambridge University
Press, Cambridge, UK.
Avilés, L., I. Agnarsson, P. A. Salazar, J. Purcell, G. Iturralde,
E. C. Yip, K. S. Powers, and T. C. Bukowski. 2007.
Altitudinal patterns of spider sociality and the biology of a
new mid-elevation social Anelosimus species in Ecuador.
American Naturalist 170:783–792.
Basset, Y., and R. Kitching. 1991. Species number, species
abundances and body length of arboreal arthropods associated with an Australian rainforest tree. Ecological Entomology 16:391–402.
Basset, Y., and N. D. Springate. 1992. Diel activity of arboreal
arthropods associated with a rainforest tree. Journal of
Natural History 26:947–952.
February 2013
NOTES
Berger, D., and K. Gotthard. 2008. Time, stress, predation risk
and diurnal-nocturnal foraging trade-offs in larval prey.
Behavioural Ecology and Sociobiology 62:1655–1663.
Berger, D., R. Walters, and K. Gotthard. 2006. What keeps
insects small? Size-dependent predation of two species of
butterfly larvae. Journal of Evolutionary Ecology 20:575–
589.
Blanckenhorn, W. U., and M. Demont. 2004. Bergmann and
converse Bergmann latitudinal clines in arthropods: two ends
of a continuum. Integrative and Comparative Biology
44:413–424.
Cornell, H. V., and B. A. Hawkins. 1995. Survival patterns and
mortality sources of herbivorous insects—some demographic
trends. American Naturalist 145:563–593.
Crouch, T. E., and Y. Lubin. 2000. Effects of climate and prey
availability of foraging in a social spider Stegodyphus
mimosarum (Eresidae). Journal of Arachnology 28:158–168.
Curio, E. 1977. How predators select their prey. Naturwissenschaften 64:575–578.
Floren, A., A. Biun, and K. E. Linsenmair. 2002. Arboreal ants
as key predators in tropical rainforest trees. Oecologia
131:137–144.
Gaston, K., and T. M. Blackburn. 2000. Patterns and processes
in macroecology. Blackwell, Oxford, UK.
Guevara, J., and L. Avilés. 2007. Multiple techniques confirm
elevations differences in insect size that may influence spider
sociality. Ecology 88:2015–2023.
Hodkinson, I. D. 2005. Terrestrial insects along elevation
gradients: species and community responses to altitude.
Biological Reviews 80:489–513.
Holmes, R. T., J. C. Schultz, and P. Nothnagle. 1979. Bird
predation on forest insect-exclosure experiment. Science
206:462–463.
Janzen, D. H. 1967. Why mountain passes are higher in the
tropics. American Naturalist 101:233–249.
Janzen, D. H. 1973. Sweep samples of tropical foliage insects:
effects of seasons, vegetation types, elevation, time of day,
and insularity. Ecology 54:687–708.
Janzen, D. H. 1983. Food webs: who eats what, why, how, and
with what effects in a tropical forest. Pages 167–182 in F. B.
Golley, editor. Tropical rain forest ecosystems: structure and
function. Elsevier Scientific, Amsterdam, The Netherlands.
Janzen, D. H., M. Ataroff, M. Farinas, S. Reyes, N. Rincon, A.
Soler, P. Soriano, and M. Vera. 1976. Changes in the
arthropod community along an elevational transect in the
Venezuelan Andes. Biotropica 8:193–203.
Lima, S. L., and L. M. Dill. 1990. Behavioural decisions made
under the risk of predation: a review and prospectus.
Canadian Journal of Zoology 68:619–640.
543
Lomolino, M. V., B. Riddle, and R. J. Whittaker. 2010.
Biogeography. Sinauer Associates, Sunderland, Massachusetts, USA.
Morais, H. C. 1994. Coordinated group ambush: a new
predatory behavior in Azteca ants (Dolichoderinae). Insectes
Sociaux 41:339–342.
Neil, A. D., and P. M. Jorgensen. 1999. Climate. Pages 8–13 in
P. M. Jorgensen and S. Leon-Yanez, editors. Catalogue of
the vascular plants of Ecuador, Monographs in Systematic
Botany from the Missouri Botanical Gardens 75. Missouri
Botanical Gardens, St. Louis, Missouri, USA.
Powers, K. S., and L. Avilés. 2007. The role of prey size and
abundance in the geographic distribution of spider sociality.
Journal of Animal Ecology 76:995–1003.
Purcell, J., and L. Avilés. 2007. Smaller colonies and more
solitary living mark higher elevation populations of a social
spider. Journal of Animal Ecology 76:590–597.
R Development Core Team. 2012. R: a language and
environment for statistical computing. R foundation for
Statistical Computing, Vienna, Austria. www.r-project.org
Sarmiento, G. 1986. Ecological features of climate in high
tropical mountains. Pages 11–45 in F. Vuilleumier and M.
Monasterio, editors. High altitude tropical biogeography.
Oxford University Press, Oxford, UK.
Schemske, D. W., G. G. Mittlebach, H. V. Cornell, J. M. Sobel,
and K. Roy. 2009. Is there a latitudinal gradient in the
importance of biotic interactions? Annual Review of Ecology
and Systematics 40:245–269.
Schulze, C. H., K. E. Linsenmair, and K. Fiedler. 2001.
Understorey versus canopy: patterns of vertical stratification
and diversity among Lepidoptera in a Bornean rainforest.
Plant Ecology 153:133–152.
Shahabuddin, P. Hidayat, S. Manuwoto, W. Noerdjito, T.
Tscharntke, and C. H. Schulze. 2010. Diversity and body size
of dung beetles attracted to different dung types along a
tropical land-use gradient in Sulawesi, Indonesia. Journal of
Tropical Ecology 26:53–65.
Somme, L. 2008. Adaptations of terrestrial arthropods to the
alpine environment. Biological Reviews 64:367–407.
Springate, N. D., and Y. Basset. 1996. Diel activity of arboreal
arthropods associated with Papua New Guinean trees.
Journal of Natural History 30:101–112.
Stork, N. E. 2007. Biodiversity: world of insects. Nature
448:657–658.
Stork, N. E., and T. M. Blackburn. 1993. Abundance, body size
and biomass of arthropods in tropical forests. Oikos 67:483–
489.
Yip, E. C., K. S. Powers, and L. Avilés. 2008. Cooperative
capture of large prey solves scaling challenge faced by spider
societies. Proceedings of the Natural Academy of Science
USA 105:11818–11822.