Behavioral Ecology Vol. 11 No. 3: 239-245
© 2000 International Society for Behavioral Ecology
Effects of sex, body size, temperature, and location on the antipredator tactics of free-ranging gartersnakes (Thamnophis sirtalis, Colubridae)
a School of Biological Sciences A08, University of Sydney, NSW 2006 Australia b Zoology Department, Oregon State University, Cordley Hall 3029, Corvallis, OR 97331-2914, USA
Address correspondence to R. Shine. E-mail: rics{at}bio.usyd.edu.au .
Received 11 August 1998; revised 7 April 1999; accepted 24 June 1999.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Gartersnakes (Thamnophis sirtalis parietalis) in southern Manitoba are subject to intense predation (primarily by crows) during their spring breeding season. The huge numbers of snakes provide a unique opportunity to quantify behavioral traits. We simulated predator attacks by "pecking" more than 500 free-ranging snakes, to explore the determinants of snake response. Snakes responded to a human finger in the same way as they did to a more realistic stimulus (a model crow). A snake's response to attack depended on several factors, which interacted in complex ways. The primary influences on response were body temperature (warmer snakes tended to flee, whereas colder snakes remained cryptic or flattened and/or gaped and struck) and sex (males were more likely to flee). Responses also depended on microhabitat (i.e., inside the winter den versus in adjacent grassland) and on the snake's prior activity (e.g., courting snakes often ignored our close approach). These factors interacted in significant ways; for example, snakes outside the den were smaller and warmer than those inside, male snakes were smaller and warmer than females, and mean body temperatures were higher in larger snakes within each sex. Thus, a snake's body size and its location affected its defensive response indirectly (via their influence on body temperature). Our results differ from those of previous studies and suggest that antipredator responses in these animals depend in a flexible and complex way upon biotic and abiotic variables. Interactions among these variables also must be considered before we can identify underlying causal processes.
Key words: display, predation, snakes, temperature, Thamnophis sirtalis parietalis.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Predation is an important influence on survival rates for many kinds of animals, and a diverse array of morphological and behavioral traits have evolved in response to this selection pressure (e.g., Lima and Dill, 1990
;
Vermeij, 1982). Indeed, the
interplay between prey and predator is a central theme of behavioral ecology.
The ways in which potential prey items respond to a perceived threat of
predation can have substantial implications not only for the vulnerability of
individuals, but also for overall patterns of habitat use and demography
(e.g., Lima, 1998;
Sinclair and Arcese, 1995).
Unfortunately, there are formidable logistical obstacles to quantifying
organismal responses to predator attack for many species in their natural
habitats. Not only does the prey animal's response depend on factors such as
its sex, size, locomotor ability, and habitat (e.g.,
Lima, 1998), but these latter
factors are intercorrelated in complex ways. For example, sex differences in
response to predator approach might be secondary consequences of sex
differences in body size or location, rather than intrinsically
gender-specific responses. To disentangle such complexities, many authors have
taken the study system into the laboratory, where it is feasible to control
such confounding influences. Nonetheless, much is lost in this process of
simplification. If we are to understand why animals behave as they do in the
field, we need to appreciate not only the ways in which multiple factors
influence the organism's response to a predator, but also the ways in which
those factors interact with each other.
Snakes are ideal models to investigate these questions because they have
the most elaborate antipredator mechanisms yet described among reptiles
(Greene, 1988). Snakes have
been used to investigate the ways in which antipredator responses are
influenced by genetic factors (e.g.,
Brodie, 1989) as well as
environmental cues (e.g., temperature and feeding:
Brodie and Russell, 1999;
Herzog and Bailey, 1987). For
ectothermic species, body temperature may be one of the most important
influences on antipredator behavior because temperature will determine the
animal's ability to detect, repel, or escape the predator. Body size is
important for the same reasons, and habitat (especially proximity to a secure
retreat) may also influence an animal's response to a predatory attack (e.g.,
Duvall et al., 1985). Finally,
sex differences in antipredator behavior are widespread in animals (e.g.,
Clutton-Brock, 1991;
Fitch, 1999). We investigated
the role of these four variables as potential determinants of the response of
free-ranging gartersnakes to simulated predator attack. We also explored the
relationship of these variables to each other.
The massive aggregations of red-sided gartersnakes (Thamnophis sirtalis
parietalis) that occur in springtime in southern Manitoba, Canada, as the
snakes emerge from their overwinter dens, constitute an ideal study system.
This study system confers three major advantages. First, the extraordinary
abundance of the snakes enabled us to obtain independent measures on a large
number of animals in a short time. We could thus avoid the significant
artifacts that are likely to arise if the same study animals are used more
than once (because repeatabilities for antipredator behavior are typically
high; Arnold and Bennett, 1984;
Brodie, 1989;
Brodie and Russell, 1999;
Garland, 1988). Second, the
snakes concentrate on courtship and dispersal at this time of year. They do
not feed, and the females are not pregnant
(Gregory and Stewart, 1975),
thereby eliminating two sources of variation that might otherwise influence
antipredator tactics (Duvall et al.,
1985; Herzog and Bailey,
1987). Third, both direct and indirect evidence suggests that
these snakes experience strong selection for effective antipredator responses.
The direct evidence comes from the high levels of predation in our study
population (see below). The indirect evidence comes from intraspecific
geographic variation in color patterns across the range of this species. The
snakes in our study population display bright red lateral patches, unlike most
other subspecies of Thamnophis sirtalis
(Rossman et al., 1996). These
blotches are normally hidden by overlapping scales and thus are evident only
during antipredator displays (see Figure
1). The evolution of this conspicuous display coloration within
T. sirtalis suggests that antipredator displays may have been a
target of significant selection in this species.
|
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Study species and study area
Red-sided gartersnakes are slender, surface-active, nonvenomous natricine colubrids. In our population, adult females attain much larger sizes [mean 55-60 cm snout-vent length (SVL), 70-80 g) than do adult males (mean 45-50 cm SVL, 30-40 g). We studied this species near the northern limit of its range in Manitoba at three large communal dens within the Chatfield Community Pasture 100 km north of Winnipeg. At least 50,000 snakes overwinter in these dens every year, dispersing into the surrounding muskeg swampland during summer (Gregory, 1977
;
Gregory and Stewart, 1975;
Mason and Crews, 1985). We
studied the snakes during spring (May 1997 and 1998), at which time the snakes
were in the process of emerging from the den and mating. Thus, we encountered
snakes either inside the dens (rock-lined sinkholes approximately 20 m long, 3
m wide, and 2 m deep), as they emerged from underground cavities, or in the
mosaic grassland—aspen woodland habitat surrounding the den. Courting
and mating activity occurred primarily outside the den, probably because of
low temperatures in the deeply shaded den area (see below). The snakes were exposed to heavy predation during the course of our study. We recorded 590 predator-killed snakes at or near the dens over a 10-day period (8-17 May 1997), mostly due to attacks by American crows (Corvus brachyrhynchos). The crows inflicted injuries of several kinds, usually involving removal of the snake's liver. Crow attacks occurred at all times of the day, but the temporal distribution of fresh kills indicates that predation was concentrated in the early morning and late evening, perhaps because humans were absent from the den area at those times. Crows fled at our approach, so we were not able to observe predation as it occurred. The predator-killed snakes included both males (n = 433) and females (n = 157), comprising a wide size range in each case (males, 30.2-55.0 cm SVL; females, 31.0-73.3 cm). Hence, all size classes of both sexes were vulnerable to predation at this time.
Experimental techniques
Our aim was to simulate crow attack using humans rather than crows as the
actual stimulus. In using a human finger as the predatory stimulus, we follow
most previous analyses of defensive responses by snakes. Experimental work
provides a justification for this logistically convenient technique: the exact
form of the predatory stimulus appears to be unimportant in determining the
snakes' response; the human hand elicits the same kinds of responses as a
realistic model of the predator (Herzog et
al., 1989; Scudder and
Chiszer, 1977). Movement and elevation, rather than the exact
shape of the stimulus object, are the prime determinants of the snake's
response (Herzog et al.,
1989). Our technique has the additional advantage of avoiding
ethical difficulties involved with the use of live nonhuman predators.
However, to test whether responses to humans were similar to those to crows,
we also quantified snake responses to a model crow (see below).
We carried out three types of trials. In the first, the observer approached
the snake to a distance of approximately 1 m, then "pecked" the
snake in the midbody region with an outstretched finger. The peck was repeated
five times, with a 1-s pause between pecks. The snake was then picked up by
the midbody, and its response at the time it was seized was scored into one of
four categories: (1) flee, without any defensive display; (2) remain still,
without any overt reaction to the stimulus; (3) flatten the body (i.e.,
obvious dorsoventral compression); and (4) flatten the body and strike
open-mouthed. The strike display [equivalent to Hailey and Davies's
(1986) "viperine
display") was highly stereotyped and performed slowly. The strike was
launched toward the stimulus on some occasions, but the snakes often simply
waved their heads in the air with the mouth gaped widely open (see
Figure 1). These trials were
carried out on 18 and 19 May 1997, in cloudy conditions with air temperatures
around 2°-10°C. After the antipredator response was described, we
measured the snake's body temperature with a quick-registering cloacal
thermometer, and then recorded its sex, SVL, and mass. All snakes were
released at the site of capture after completion of the trials.
The second group of trials was conducted over a longer period (12-18 May
1997) and over a wider range of weather conditions (air temperatures
2-15°C). The range of body sizes of snakes used was almost identical to
that in the first trial (36-60 cm SVL), but the snakes in courting groups were
warmer than the solitary animals used in the first trials (means of 12.9°
versus 9.3°C; maximum temperature approximately 19°C in both trials,
but minima lower in the solitary-snake trials). In this second set of trials,
we targeted groups of snakes that were actively engaged in courtship behavior
(i.e., "mating balls"; Mason
and Crews, 1985). Two observers approached the group to within a
meter, and then suddenly moved their hands 20 cm above the snakes to simulate
a crow's wings. This brief (<2 s) stimulus generally induced a proportion
of the group to flee, whereas others remained, apparently unaware of the
observers' presence (i.e., they continued courtship activity). We then
captured both sets of snakes (i.e., those that fled and those that remained)
and recorded the body temperature, sex, and SVL of snakes in each of these two
groups.
The third set of trials (in May 1998) was designed to compare the snakes' response to a human and to a more accurate simulation of a natural predator. We used a commercially produced crow decoy (as used by hunters to attract crows). The model crow closely simulated a real crow in size (41 cm total length), shape, and color. We attached the plastic crow to a 2-m wooden dowel and pecked snakes with either the crow's bill or with a finger. Half of the snakes were first pecked with the crow and then (after a 30-s delay) with the finger. The other half were pecked by finger first, and then by the crow. As in the first set of trials, we recorded the snakes' response, sex, body size, and body temperature.
We tested data sets for normality and equality of variances before analysis. We relied on nonparametric contingency-table tests for most analyses. In these cases the chi-square calculations were based on raw numbers, although we report proportions to facilitate explanation of trends.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Sex, body temperature, body size, and location
Because sex, body temperature, body size, and location are plausibly associated with variation in antipredator behavior by the snakes, we first describe patterns of intercorrelation between these four factors. These patterns proved to be complex.
Contingency table analysis showed that the sex ratio was not significantly
different between samples of snakes collected inside versus outside the den
(
2 = 2.16, 1 df, p =.14). However, a two-factor ANOVA
with sex and location (inside versus outside the den) as factors and ln SVL as
the dependent variable showed that females were larger than males
(F1,234 = 157.49, p<.0001) and that snakes
outside the den were larger than those inside (F1,234 =
5.21, p <.025; see Figure
2). The interaction term between sex and location was not
significant (F1,234 = 0.42, p =.52), indicating
that the magnitude of sexual size dimorphism was similar inside versus outside
the den (despite the difference in mean body sizes in the two locations).
|
To clarify the ways in which sex, size, and location might influence a snake's body temperature, we carried out a two-factor analysis of covariance. The factors were sex and location (inside versus outside den), the covariate was body size (ln SVL), and the dependent variable was body temperature. This analysis revealed no significant heterogeneity in slopes (i.e., all interaction terms had p >.10), but body temperatures increased significantly with body size (covariate F1,234 = 17.95, p <.0001), and the intercepts of these regressions differed significantly as a function of both sex (males were warmer than females: F1,234 = 15.50, p <.0001) and location (snakes outside the den were warmer: F1,234 = 152.55, p <.0001).
The end result of these relationships is that a snake's body temperature depends on its size, its sex, and its location. In turn, size depends on both sex and location. Thus, the distribution of four traits (sex, size, temperature, location) that might potentially affect a gartersnake's antipredator behavior are themselves intercorrelated in complex ways. We now turn to the ways in which these four traits modify snake responses.
Antipredator responses
We obtained data on the antipredator responses of 238 snakes (159 males, 79
females) from the first set of trials, 293 snakes (263 males, 30 females) from
the second set, and 55 snakes (33 males, 22 females) from the third. We
analyzed these three data sets separately because of the different stimuli we
provided to the snakes in each case.
Solitary snake trials
First we consider overall responses from the first set of trials (those in
which the snakes were pecked by a finger). Logistic regression offers a
powerful approach to detecting influences on categorical response variables
such as those exhibited by the snakes. Our first analysis used "flee
versus remain" as the dichotomous dependent variable. The snake's sex,
body temperature, size (ln SVL), and location (inside versus outside den) were
entered as independent variables. Log-likelihood ratio tests showed that the
snake's decision (i.e., whether or not to flee from the predatory stimulus)
depended on its body temperature (2 = 35.48, 1 df, p
<.0001) and sex (2 = 8.04, 1 df, p <.005), but
not its body size (2 = 0.21, 1 df, p =.65) or
location (2 = 0.04, 1 df, p =.85). Male snakes were
more likely to flee than females, especially if they were warm.
Given that a snake does not flee, what factors determine whether it remains
still, flattens, or strikes? We used logistic regression with the same
independent variables as above, but restricted to the subset of animals that
did not flee from our approach and with the dependent variable being
"remain still, flatten, or strike". The responses of these animals
were affected by location (2 = 7.28, 2 df, p <.03;
62% of snakes inside the den struck at the stimulus, versus 40% of snakes
encountered outside the den), but not any of the other factors (all p
>.50).
In combination with these analyses, a more detailed inspection of patterns within the data set indicates that a snake's response to a predatory attack depends on the following factors:
- Body temperature. At body temperatures >12°C, the snakes generally
fled from our attack. Cooler snakes usually either remained still or launched
a defensive display (Figure
3).
- Sex. The above pattern was evident in both males and females, but the
threshold level for fleeing was lower in males. Many males attempted to flee
even at very low body temperatures, whereas females did not
(Figure 3). Even when analysis
was restricted to snakes at body temperatures close to the threshold for
flight (8°-12°C), contingency table analysis confirmed that the sex
difference in the propensity for flight was statistically significant
(2 = 11.25, 1 df, p <.003; 10% of 30 females fled
versus 59% of 92 males).
Although a snake's sex and its body temperature influenced the probability that it stayed rather than attempted to flee, these factors did not affect its response if it stayed. Approximately 23% of the sedentary snakes did not show any overt response to the stimulus, whereas 27% flattened without striking, and 50% showed both flattening and striking. These proportions were similar in males versus females (24 versus 22%; 27 versus 27%; 49 versus 51%;2 = 0.06, 2 df, p =.97), whereas the sex difference
in the proportion of snakes that fled from our approach was highly significant
(58 versus 25%; 2 = 21.15, 1 df, p <.0001).
Similarly, mean body temperatures did not differ significantly among the
response categories that did not involve attempts to flee (means = 7.1°C
for "remained still," 7.3 for "flattened," 6.8 for
"struck"; two-factor ANOVA with sex and response category as
factors, body temperature as dependent variable: effect of sex,
F1,120 = 0.14, p =.71; effect of response
category, F2,122 = 0.29, p =.75; interaction,
F2,120 = 0.15, p =.86).
- Location. Whether snakes were inside or outside the den had a substantial
effect on their antipredator behavior, but this was largely an indirect
consequence of the thermal environment. The proportion of snakes that fled
from our approach was much higher for snakes outside the den than for those
inside (59% versus 21%; 2 = 26.43, 1 df, p
<.0001). However, the logistic regression showed that once this thermal
effect was factored out of the analysis, location exerted little effect on
snake response. The only significant effect that we detected was that snakes
inside the den were more likely to strike at us than were those outside.
- Body size. A snake's size might affect its body temperature (due to thermal
inertia) or its ability to flee (i.e., locomotor speed) or display effectively
(i.e., be sufficiently intimidating). However, we did not detect any
relationship between size and behavior
(Figure 4; note nonsignificant
size effect in the logistic regressions). Larger snakes had higher body
temperatures within each sex, but males (the smaller sex) were consistently
warmer than females (Figure 4).
To determine whether the sex difference in the proportions of snakes fleeing
from our approach was simply due to a body size difference between the sexes,
we compared the responses of males with those of females over the same range
of body sizes (i.e., <61 cm SVL). Over this common size range, males were
more likely to flee than were females (2 = 7.15, 1 df,
p <.008). Thus, the behavioral difference between the sexes is not
simply an indirect effect of sexual size dimorphism.
|
|
Because the size-related shift in mean body temperature was relatively subtle (Figure 4), it had only a minor effect on defensive responses. For example, snakes inside the den were significantly larger than those outside the den (Figure 2), and hence would tend to be warmer for this reason (Figure 4). However, the magnitude of the thermal difference inside versus outside the den was so great that it overwhelmed the size effect. Thus, the proportion of snakes fleeing from our approach was more than twice as high outside the den as inside (see above and Figure 3), with the spatial effect on body temperatures completely overriding the more subtle size-related effect.
Response to model of a crow
We now consider the set of trials in which we compared the snakes' response
to a finger with their response to a plastic crow (a more realistic simulation
of the actual predator). The clear result was that the snakes responded in
similar ways to the two stimuli. In 43 of the 55 cases (78%), the snake
responded identically to the finger versus the plastic crow (9 snakes gaped to
both stimuli, 8 flattened, 3 remained still, 22 fled). Interest thus focuses
on the remaining 13 cases. Analysis reveals no consistent difference in
response to the two kinds of stimuli. For example, the number of times a snake
fled from the crow but flattened in response to the finger was exactly the
same as the number of times that the snake fled from the finger but flattened
in response to the crow (two in each case). To test the statistical
significance of minor differences between these two data sets, we compared the
observed frequencies of responses to those expected under two null models. The
data deviated markedly from the patterns expected if a snake's response to one
type of stimulus was completely independent of its response to the other
stimulus (2 = 26.69, 9 df, p <.01), but did not
differ significantly from the patterns expected if responses to the two
stimuli were identical (2 = 4.75, 9 df, p =.85).
Also, if analysis is restricted to responses to the plastic crow only, we see
the same patterns as for the trials using the human finger as the stimulus
(above). For example, the proportion of snakes that fled in response to the
crow was much higher for snakes with body temperatures > 15°C than for
snakes below that temperature (2 = 54.5, 1 df, p
<.0001).
Courting group trials
For the trials in which we frightened (but did not "peck")
courting groups, we never recorded any cases of flattening or striking;
instead, an average of 28% of the snakes in the group attempted to flee,
whereas the remainder continued courting. There was a strong sex bias in the
results, with only 1 of 27 females fleeing (4%) compared to 81 of 263 males
(31%; 2 = 7.58, 1 df, p <.006). The proportion of
males that fled from our approach was strongly related to mean body
temperature of snakes in the courting group: more snakes fled from warmer
groups (mean body temperature versus proportion that fled: r = -.48,
n = 36, p <.004). Analysis did not reveal any
determinants of "stay versus flee" within the males of each group.
Because both mean body sizes and mean body temperatures differed among
courting groups, we used a two-factor nested ANOVA (with group number and
"stay versus flee" nested within group number as the factors) to
investigate such effects. Male snakes that fled from us did not differ from
those that stayed in terms of body sizes (F25,35 = 0.58,
p =.92) or body temperatures (for males only,
F25,35 = 0.16, p >.99).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our data paint a complex picture of the determinants of antipredator responses in free-ranging gartersnakes. The tactic used by a snake in response to our simulated attack depended on a series of factors (such as body temperature, sex, location, and prior activity) and, just as importantly, these factors showed strong relationships with each other (for example, body temperatures were a function of sex, body size, and location). Although the stimulus we used was an artificial one, the similar responses to the human hand versus a model crow suggest that we are measuring a biologically meaningful response. The end result is that antipredator responses of these small snakes provide a useful model system in which to clarify the ways an ectotherm's behavioral decisions are influenced by abiotic and biotic variables.
The published literature on antipredator responses of snakes contains many
apparent inconsistencies, most notably in terms of the effect of body
temperature on defensive responses. Some of these inconsistencies undoubtedly
result from real interspecific differences
(Scudder and Burghardt, 1983);
for example, May et al.'s
(1996) extensive field study
clearly showed that pigmy rattlesnakes (Sistrurus miliarius) are more
likely to strike defensively when they are hot rather than when they are cold.
This pattern is the opposite to that shown by our gartersnakes and may reflect
different defensive capabilities of the two taxa (cf.
McLean and Godin, 1989; but
see also Duvall et al., 1985).
However, much of the published literature on antipredator tactics in snakes
has been based on natricine colubrids, especially gartersnakes.
Early reports noted (without quantification) that colder gartersnakes were
more likely to display rather than to flee
(Costanzo, 1986;
Fitch, 1965;
Heckrotte, 1967). Some of the
extensive laboratory studies on captive gartersnakes supported these early
reports (Arnold and Bennett,
1984), but others found the reverse pattern, whereby warmer snakes
display more often than cold snakes
(Schieffelin and de Quieroz,
1991; and see Keogh and
DeSerto, 1994, for a similar result on other small colubrid
species). Our study strongly corroborates the early reports: colder
gartersnakes were more likely to display rather than to flee, as has been
reported also in a variety of lizard species (e.g.,
Hertz et al., 1982). A recent
field study on Thamnophis sirtalis, similar to our own, also
concluded that colder snakes are less likely to flee
(Passek and Gillingham, 1997).
Remarkably, however, the threshold temperatures for flight recorded by these
authors (working on a Michigan population during May and June) were much
higher than those we measured in the Manitoba snakes; for example, they
reported mean body temperatures of 30.7°C for fleeing snakes and
27.4°C for snakes that did not flee. Thus, although the general pattern is
the same, the thermal ranges at which these responses were recorded were
dramatically different (see Figure
3).
Other aspects of our results also provide a strong contrast to previous
studies. Part of the reason may relate to interspecific (or even intraspecific
or ontogenetic) differences in responses, but we suspect that results from
laboratory studies may sometimes give an unreliable indication of the
behavioral responses of free-ranging snakes. Even brief captivity can
substantially affect defensive responses of snakes (e.g.,
Fitch, 1975;
Platt, 1969). Such
captivity-induced modifications to defensive responses accord well with one of
the main conclusions from our own study: these behaviors show considerable
flexibility in response to several variables.
What is the biological significance of the diversity we documented in
antipredator responses of red-sided gartersnakes? Optimality models suggest
that an organism should adopt the antipredator tactic most likely to maximize
its chances of surviving the encounter without injury
(Vermeij, 1982). Some of the
patterns we observed are easily interpretable in this light, but the adaptive
advantage of others (if any) remains obscure. The most straightforward case is
the consistent trend for warmer snakes to flee from our approach, whereas
colder snakes remained sedentary (e.g.,
Figure 3). Higher body
temperatures enhance locomotor ability in gartersnakes
(Heckrotte, 1967;
Scribner and Weatherhead,
1995), as in other ectotherms
(Huey and Slatkin, 1976), and
a snake is presumably more likely to escape if it can move rapidly. Because
gartersnakes are nonvenomous and relatively small, even the largest snake may
pose little threat to a predator such as a crow. This is especially true when
the snake is cold (and hence, less able to launch accurate strikes), as is
generally the case with displaying snakes. Perhaps for this reason, body size
had little effect on the snake's probability of launching an offensive display
(Figure 4).
One of the most interesting results from our study is the clear sex
difference in defensive tactics. Previous studies have often reported sex
differences in antipredator responses, but high levels of aggressive
antipredator display by females are typically associated with protection of
offspring (e.g., in several taxa of mammalian carnivores and ungulates). This
factor is irrelevant to gartersnakes, which show no postpartum care of
offspring (Rossman et al.,
1996). Higher levels of antipredator displays in males than in
females have been reported in two snake species
(Scudder and Burghardt, 1983;
Zinner, 1985), but a recent
report noted (without quantification) that female red-sided gartersnakes are
more likely than conspecific males to display and to bite when handled
(Fitch, 1999). This latter
report is consistent with our own data. An earlier study of neonatal
Thamnophis sirtalis reported no sex difference in levels of display
(Scudder and Burghardt, 1983).
Aggression is more closely linked to body temperature in gravid female
rattlesnakes than in nongravid females or males but is no more common overall
in one sex than the other (Goode and
Duvall, 1988). Similarly, Scribner and Weatherhead
(1995) reported that
antipredator responses did not differ between the sexes in three species of
natricine colubrids (including T. sirtalis). In their field study of
T. sirtalis, Passek and Gillingham
(1997) did not examine sex
differences in antipredator behavior.
In our study, female snakes engaged in a higher proportion of aggressive displays than did males, but this sex difference disappeared if attention was restricted to the subset of snakes that did not flee from the observer. That is, males and females that did not flee were equally likely to remain still, to flatten, or to flatten and strike. The sex difference lies in the fact that a high proportion of males fled from the observer, whereas fewer females adopted this tactic. Why were males more likely to flee? Male and female gartersnakes in our population differ in at least three factors—body size, thermal ecology, and emergence patterns—that might favor a sex divergence in antipredator tactics. We evaluate these possibilities below.
First, sex differences in body size per se cannot explain the behavioral difference, because size had little effect on response patterns within each sex (Figure 4).
Second, body temperature clearly plays a role. One reason that males fled so often is that they were warmer than females, despite a trend for mean body temperatures to increase rather than decrease with body size within each sex (Figures 3 and 4). Nonetheless, thermal factors cannot explain all of the sex difference in antipredator behavior. Even when snake response was plotted as a function of body temperature, the sexes followed different trajectories (Figure 3).
Third, males remain near the dens for some weeks during the mating period,
whereas females tend to disperse toward summer habitats soon after emergence
(Gregory and Stewart, 1975).
Many of the females that we tested were thus newly emerged and had not
attained full locomotor ability since emerging from their long overwinter
dormancy. Trials of locomotor ability show that females are substantially
slower (and thus less well-suited to relying on speed to escape predation;
Shine et al., unpublished data). This sex bias in locomotor ability provides a
plausible adaptive basis for the sex difference we observed in defensive
behavior.
More generally, antipredator responses in free-ranging snakes may offer an
unusually powerful opportunity to test adaptationist (optimality) models on
behavioral decisions. In combination with previous studies, our work suggests
that gartersnakes modify their antipredator behavior in response to a complex
suite of traits, with a strong focus on factors that influence the snake's
ability to successfully flee from a predator. Thus, snakes are more likely to
display rather than to flee if their locomotor ability is compromised by
factors such as low body temperature (see above), fatigue
(Arnold and Bennett, 1984),
poor condition (Andren, 1982),
pregnancy (Goode and Duvall,
1988), or a recent meal
(Herzog and Bailey, 1987), or
if the effectiveness of escape is influenced by the proximity of shelter or
conspecifics (Duvall et al.,
1985; and see above). This hypothesis is potentially testable by
experimental manipulation and by taking advantage of the wide range of
locomotor abilities often present within a single population (e.g.,
Arnold and Bennett, 1984;
Garland, 1988). This range may
be further increased by experimental manipulations early in ontogeny (e.g.,
Shine, 1995;
Sinervo and Huey, 1990).
Correlational evidence is consistent with the hypothesis that antipredator
responses in gartersnakes are fine-tuned by natural selection in such a way as
to maximize an individual's probability of surviving such an encounter, but we
will need additional data, both on predator behavior and prey responses,
before we can fully understand the dynamics of that interaction.
|
ACKNOWLEDGEMENTS |
|---|
We thank Dave Roberts (Manitoba Department of Natural Resources) for logistical support and the residents of Chatfield (especially A1 and Gerry Johnson) for help and encouragement. Keith Vanning provided the initial stimulus for our study and assisted throughout the work. Several Earthwatch volunteers also assisted with data collection. Financial support was provided by the Australian Research Council (to R.S.) and by a National Science Foundation National Young Investigator Award (IBN-9357245) and the Whitehall Foundation (W95-04) to R.T.M. Research was conducted under the authority of Oregon State University Institutional Animal Care and Use Committee Protocol no. LAR-1848B. All research was conducted in accord with the U.S. Public Health Service's Policy on Humane Care and Use of Laboratory Animals and the National Institutes of Health's Guide to the Care and Use of Laboratory Animals.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Andren C, 1982. Effect of prey density on reproduction, foraging and other activities in the adder, Vipera berus. Amphib Reptil 3: 81-96.
Arnold SJ, Bennett AF, 1984. Behavioral variation in natural populations. III. Antipredator displays in the garter snake Thamnophis radix. Anim Behav 32: 1108-1118.
Brodie ED III, 1989. Genetic correlations between morphology and antipredator behavior in natural populations of the garter snake Thamnophis ordinoides. Nature 342: 542-543.[Medline]
Brodie ED III, Russell NH, 1999. The consistency of individual differences in behaviour: temperature effects on antipredator behaviour in garter snakes. Anim Behav 57: 445-451.[ISI][Medline]
Clutton-Brock TH, 1991. The evolution of parental care. Princeton, New Jersey: Princeton University Press.
Costanzo JP, 1986. Influences of hibernaculum microenvironment on the winter life history of the garter snake (Thamnophis sirtalis). Ohio J Sci 86: 199-204.
Duvall D, King MB, Gutzweiler KJ, 1985. Behavioral ecology and ethology of the prairie rattlesnake. Nat Geo Res 1: 80-111.
Fitch HS, 1965. An ecological study of the garter snake, Thamnophis sirtalis. Univ Kans Publ Mus Nat Hist 15: 493-564.
Fitch HS, 1975. A demographic study of the ringneck snake (Diadophis punctatus) in Kansas. Univ Kans Mus Nat Hist Misc Publ 62: 1-53.
Fitch HS, 1999. A Kansas snake community: composition and changes over 50 years. Malabar, Florida: Kreiger.
Garland TJ, 1988. Genetic basis of activity metabolism. I. Inheritance of speed, stamina and antipredator displays in the garter snake Thamnophis sirtalis. Evolution 42: 335-350.
Goode MJ, Duvall D, 1988. Body temperature and defensive behavior of free-ranging prairie rattlesnakes, Crotalus viridis viridis. Anim Behav 38: 360-362.
Greene HW, 1988. Antipredator mechanisms in reptiles. In: Biology of the reptilia, vol. 16 (Gans, C, Huey RB, eds). New York: Alan R. Liss, 1-152.
Gregory PT, 1977. Life-history parameters of the red-sided garter snake (Thamnophis sirtalis parietalis) in an extreme environment, the Interlake region of Manitoba. Nat Mus Can Publ Zool 13: 1-44.
Gregory PT, Stewart KW, 1975. Long-distance dispersal and feeding strategy of the red-sided garter snake (Thamnophis sirtalis parietalis) in the Interlake of Manitoba. Can J Zool 53: 238-245.
Hailey A, Davies PMC, 1986. Effects of size, sex, temperature and condition on activity metabolism and defence behavior of the viperine snake, Natrix maura. J Zool 208: 541-558.
Heckrotte C, 1967. Relations of body temperature, size, and crawling speed of the common garter snake, Thamnophis sirtalis. Copeia 1967: 759-763.
Hertz PE, Huey RB, Nevo E, 1982. Fight versus flight: body temperature influences defensive responses in lizards. Anim Behav 30: 676-679.
Herzog HA, Bailey BB, 1987. Development of antipredator responses in snakes. II. Effects of recent feeding on defensive behaviors of juvenile garter snakes (Thamnophis sirtalis). J Comp Psychol 101: 387-389.
Herzog HA, Bowers BB, Burghardt GM, 1989. Stimulus control of antipredator behavior in newborn and juvenile garter snakes (Thamnophis). J Comp Psychol 103: 233-242.
Huey R, Slatkin M, 1976. Costs and benefits of lizard thermoregulation. Q Rev Biol 51: 363-384.[Medline]
Keogh JS, DeSerto FP, 1994. Temperature dependent defensive behavior in three species of North American colubrid snakes. J Herpetol 28: 258-261.
Lima SL, 1998. Nonlethal effects in the ecology of predator-prey interactions. Bioscience 48: 25-34.[ISI]
Lima SL, Dill LM, 1990. Behavioral decisions made under the risk of predation: a review and prospectus. Can J Zool 68: 619-640.
Mason RT, Crews D, 1985. Female mimicry in garter snakes. Nature 316: 59-60.[Medline]
May PG, Farrell TM, Heulett ST, Pilgrim MA, Bishop LA, Spence DJ, Rabatsky AM, Campbell MG, Aycrigg AD, Richardson WEI, 1996. Seasonal abundance and activity of a rattlesnake (Sistrurus miliarius barbouri) in central Florida. Copeia 1996: 389-401.
McLean EB, Godin JGJ, 1989. Distance to cover and fleeing from predators in fish with different amounts of defensive armour. Oikos 55: 281-290.
Passek KM, Gillingham JG, 1997. Thermal influence on defensive behaviors of the eastern garter snake, Thamnophis sirtalis. Anim Behav 54: 629-633.[ISI][Medline]
Platt DR, 1969. Natural history of the hognose snakes Heterodon platyrhinos and Heterodon nasicus. Univ Kans Publ Mus Nat Hist 18: 253-420.
Rossman DA, Ford NB, Seigel RA, 1996. The garter snakes. Evolution and ecology. Norman, Oklahoma: University of Oklahoma Press.
Schieffelin CD, de Quieroz Am, 1991. Temperature and defense in the common garter snake: warm snakes are more aggressive than cold snakes. Herpetologica 47: 230-237.
Scribner SJ, Weatherhead PJ, 1995. Locomotion and antipredator behavior in three species of semi-aquatic snakes. Can J Zool 73: 321-329.
Scudder KM, Chiszar D, 1977. Effects of six visual stimulus conditions on defensive and exploratory behavior in two species of rattlesnakes. Psychol Rec. 27: 519-526.
Scudder RM, Burghardt GM, 1983. A comparative study of defensive behavior in three sympatric species of water snakes (Nerodia). Z Tierpsychol 63: 17-26.
Shine R, 1995. A new hypothesis for the evolution of viviparity in reptiles. Am Nat 145: 809-823.
Sinclair ARE, Arcese P, 1995. Population consequences of predation-sensitive foraging: the Serengeti wildebeest. Ecology 76: 882-891.
Sinervo B, Huey RB, 1990. Allometric engineering: an
experimental test of the causes of interpopulational differences in locomotor
performance. Science 248:
1106-1109.
Vermeij GJ, 1982. Unsuccessful predation and evolution. Am Nat 120: 701-720.
Zinner H, 1985. On behavioral and sexual dimorphism of Telescopus dhara Forscal 1776 (Reptilia: Serpentes, Colubridae). J Herpetol Assoc Africa 31: 5-6.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  E. Rhoades and D. T. Blumstein Predicted fitness consequences of threat-sensitive hiding behavior Behav. Ecol., September 1, 2007; 18(5): 937 - 943. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Vanhooydonck, A. Herrel, and D. J. Irschick Determinants of sexual differences in escape behavior in lizards of the genus Anolis: a comparative approach Integr. Comp. Biol., August 1, 2007; 47(2): 200 - 210. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. D. Roth and J. A. Johnson Size-based variation in antipredator behavior within a snake (Agkistrodon piscivorus) population Behav. Ecol., March 1, 2004; 15(2): 365 - 370. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Shine, B. Phillips, H. Waye, and R. T. Mason Behavioral shifts associated with reproduction in garter snakes Behav. Ecol., March 1, 2003; 14(2): 251 - 256. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||