Behavioral Ecology Advance Access originally published online on February 24, 2008
Behavioral Ecology 2008 19(3):621-626; doi:10.1093/beheco/arn012
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The effects of perceived mortality risk on habitat selection in a terrestrial salamander
a Princeton University, Box 845, Princeton, NJ 08544, USA b Department of Biology, University of Virginia, Charlottesville, VA, USA
Address correspondence to A.M. Roberts. E-mail: amrobert{at}princeton.edu.
Received 22 August 2007; revised 4 January 2008; accepted 7 January 2008.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONSFUNDINGREFERENCES |
|---|
Animals select microhabitats based on food availability, physiological cost, and mortality risk relative to other available habitats. The best habitat selection models take into account a trade-off between predation risk and foraging success. We investigated the effect of simulated predation on microhabitat use in light of differences in prey availability in Plethodon cinereus, a small terrestrial salamander which displays nighttime plant-climbing behavior. We tested the hypothesis that plant climbing is a predator avoidance behavior in P. cinereus by causing autotomization of salamanders’ tails to simulate attempted predation and subsequently tracking them with fluorescent powder. We found that, on average, simulated predation increased the maximum height climbed. To ensure that salamanders were not climbing to access greater numbers of prey, we measured prey abundance on and above the ground and found more prey on the ground than on plants. Finally, we conducted observations of unmanipulated individuals of P. cinereus and found that weather variables affected the height climbed and males climbed higher than females, perhaps due to lower energetic costs in males. We conclude that salamanders use the plant habitat as a refuge from predation despite reduced foraging potential and increased physiological cost.
Key words: microhabitat, Plethodon cinereus, predation risk, tail autotomy.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONSFUNDINGREFERENCES |
|---|
Animals select microhabitats for a variety of reasons. Three primary factors are the animal's metabolic gain from successful foraging in a habitat, the animal's metabolic loss from physiological cost, and the animal's mortality risk in a habitat. In general, predators are expected to affect an animal's habitat selection by altering the mortality risk in one or more available habitats. Indeed, predators have been shown in many empirical studies to influence the habitat use of foragers (e.g., Milinski and Heller 1978
; Werner et al. 1983; Gilliam and Fraser 1987; Lima and Dill 1990). Nevertheless, quantification of an animal's perception of and reaction to predation risk can be elusive, especially in terrestrial habitats; terrestrial animals overall show a smaller response to the presence of predators than aquatic animals (Preisser et al. 2005). Many studies in terrestrial systems use "giving-up densities" (GUDs; the resource density remaining when an animal has finished foraging in a habitat) as a measure of perceived risk (reviewed in Verdolin 2006; see, e.g., Hochman and Kotler 2007), but such measurements are probably inaccurate (Lima and Bednekoff 1999).
Models that take into account a trade-off between predation risk and foraging success tend to have the most explanatory power (Gilliam and Fraser 1987; Railsback and Harvey 2002). Evidence for such a trade-off has been found in sophisticated laboratory studies of aquatic species (e.g., Milinski and Heller 1978; Werner et al. 1983; Fraser and Gilliam 1987). Terrestrial field studies, which have the advantage of addressing habitat selection under natural conditions, usually either describe risk–foraging trade-offs in more basic terms (e.g., Kullberg 1998) or detail reactions to predation risk without assessing foraging criteria (e.g., Martín and López 2000; Amo et al. 2003; Cooper 2003; Stankowich and Coss 2007). Terrestrial studies have rarely attempted to quantify risk–foraging trade-off, with some notable exceptions (e.g., Abramsky et al. 2002). In the present study, we investigated the effect of simulated predation on microhabitat use in light of differences in prey availability, examining this phenomenon in a terrestrial salamander, a taxon not typically studied for these purposes.
Plethodon cinereus, the red-backed salamander, is potentially a good model species for studying predation risk in terrestrial species because it occupies distinct nocturnal habitats: underground burrows, ground surface, and vegetation. At night, if P. cinereus emerge from their underground retreats, which they inhabit during the day (reviewed in Petranka 1998), they often forage on the open forest floor (Taub 1961; Burton and Likens 1975; Dakin 1978; Mathis 1991; Liebgold and Jaeger 2007). However, they are also frequently found on small plants and larger shrubs or trees (Burton and Likens 1975; Jaeger 1978) and have been observed to remain on vegetation for over 2 h (Wise S, personal communication).
These 3 habitats vary in prey availability, physiological costs, and risk of predation. Plethodon cinereus are euryphagic (reviewed in Petranka 1998), so we can assess food abundance in a particular microhabitat by collecting small invertebrates in that habitat, without regard to taxon. Jaeger (1978) found that stomachs of P. cinereus captured at night on plants had a greater food volume than those captured at night on the forest floor. On the other hand, like all plethodontids, P. cinereus is a lungless salamander, with gas exchange occurring only through its skin (reviewed in Petranka 1998), such that it may rapidly lose moisture under dry conditions, both in the lab (Grover 2000) and in the forest (Jaeger 1978). This suggests that dehydration may physiologically constrain their microhabitat selection during nighttime foraging.
Jaeger (1978) concluded, on the basis of high stomach content for individuals on plants combined with these physiological facts, that P. cinereus individuals climb plants in search of better prey. In his words, "plants appear to contain a richer, but only periodically exploitable, pool of food items than the normal forest litter habitat of P. cinereus" (Jaeger 1978, p. 690). However, an alternative explanation of his result is that the forest floor has better prey that is exploited prior to satiation, after which salamanders climb plants, where predation risk may be lower. Hungry three-spined sticklebacks, for example, feed in the denser portions of swarms of prey where predation risk is higher, whereas partially satiated sticklebacks feed in the safer, less dense edges of prey swarms (Milinski and Heller 1978; Heller and Milinski 1979). Plethodon cinereus may behave similarly.
We predicted that red-backed salamanders climb plants, despite the physiological constraint of dehydration, to avoid predation but not to increase foraging. Nighttime predators of P. cinereus include larger plethodontids like Pseudotriton ruber (Bock and Fauth 1992) and potentially Desmognathus fuscus (Jaeger et al. 1998; but see Ransom and Jaeger 2006), as well as snakes, such as the eastern garter snake Thamnophis sirtalis (Hamilton 1951), and shrews, such as Blarina brevicauda (Brodie et al. 1979). To our knowledge, none of these predators or potential predators typically climb plants (shrews reviewed in Churchfield 1990; plethodontids reviewed in Petranka 1998), but no direct assessment of actual or perceived mortality risk to P. cinereus has been conducted to our knowledge.
In this study, we simulated predation on P. cinereus and used a fluorescent powder-tracking method to determine whether simulated predation increased climbing behavior (Test 1) in conjunction with a comparison between prey abundance on the ground and on plants using sticky traps (Test 2). It is unlikely that an animal should perceive its mortality risk to be zero even in the absence of predators or predator-simulating stimuli (Lima and Dill 1990). Therefore, we also captured salamanders on the ground and on plants on a separate survey plot to determine the prevalence of plant climbing in natural circumstances, as well as to assess the effects of age and sex on plant-climbing behavior in the absence of artificial stimuli (Test 3).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONSFUNDINGREFERENCES |
|---|
Study site
All data were collected in the forest at Mountain Lake Biological Station (University of Virginia), Salt Pond Mountain, Giles County, Virginia, within 1 km of 37°22'32''N, 80°31'20''W and at an elevation of approximately 1200 m. Temperature (°C), humidity (%), pressure (mB), wind speed (m/s), and rainfall (mm) were collected every 30 min at Mountain Lake Biological Station by a C10 datalogger (Campbell Scientific, Logan, UT).
Test 1: the effects of simulated predation on plant climbing
On 24 nights during the nonbreeding season, between 16 June and 25 July, we tracked 95 red-backed salamanders (2–9 per night) with fluorescent powder (Day Glo Corporation, Cleveland, OH). This method has been used to track other small vertebrates, including mammals, amphibians, and reptiles (Lemen and Freeman 1985; Fellers and Drost 1989; Blankenship et al. 1990; Eggert 2002). Rittenhouse et al. (2006) found that fluorescent powder was harmless to amphibians (Ambystoma maculatum and Rana sylvatica). Powdered individuals of P. cinereus do not lose more weight than control individuals over 1 week (Roberts A, unpublished data). Between 1845 and 2200, we located salamanders under logs and rocks, except on 6 occasions when salamanders were located after dark; we found these salamanders foraging on leaf litter. Each salamander we tracked was searched for (and found) at least 5 m from other salamanders we tracked in order to ensure that home ranges did not overlap. Liebgold and Jaeger (2007) found that the mean home range of P. cinereus in this area was 1.36 m2. Sampling was performed along forest transects systematically with no location sampled twice in order to avoid reusing the same salamander on multiple nights and to avoid confusing old and current powder trails.
When we located a salamander, we measured its snout-vent length (SVL). We assigned salamanders to 2 age classes using Sayler's (1966) criteria based on SVL: juvenile (<33.5 mm) or adult (>33.5 mm); additionally, we took into account an estimated SVL increase of approximately 2.5 mm for juveniles during the study period (0.5 weekly). If individuals had SVLs on the cusp of juvenile/adult determination, data from these individuals were excluded from analyses involving sex–age classes. We used the shape of the snout to determine sex (sensu Test and Bingham 1948): male snouts have a more angular shape than female and juvenile snouts. This method, as opposed to candling (sensu Gillette and Peterson 2001), was used to minimize the disturbance to the salamander, even though it is slightly less accurate (unpublished data). Then, we dipped the salamander in fluorescent powder such that it was dorsally covered along its tail and on the latter 10 mm of its torso. We induced autotomization of 5–10 mm of the tails of treatment individuals by squeezing those sections of tail with tweezers (henceforth referred to as tail clipping) to simulate a predation attempt. We then released the salamander where we had found it. Control individuals were measured, dipped in powder, and released with their tails intact.
Powder trail data were collected between 2300 and 0230 (2–6 h after capture). Individuals of P. cinereus begin foraging just after dusk, with surface activity peaking 2–4 h after dusk (Dakin 1978; see also Maerz et al. 2001); therefore, beyond 2 h after sunset, trail-forming activity probably did not strongly depend on time elapsed, which we confirmed based on our data (below). Using a deep-violet LED flashlight (Northwest Marine Technology, Shaw Island, WA), we followed the trails from their origin either until we located the salamander or until a thorough search of the area yielded no new dots of powder. For each plant with powder on it, we recorded the height off the ground of the highest fluorescent mark. Height of powder found on logs and dead trees (and in one case an elevated rock) was also recorded. Marks that may have been left on plants above 185 cm were not recorded because they were too high for the observer to find. This physical limitation is unlikely to have affected our results because very few observed marks even approached this height. For each trail, we also measured the maximum displacement (D), that is, the straight linear distance between the salamander capture location and the mark furthest from that location. We took note of the time between powdering the salamander and observation of its trail (T). From our data, we calculated 3 parameters to quantify the extent of climbing demonstrated by the powder trails: mean height climbed (
), maximum height climbed (H), and number of climbs (m).
We performed a logistic regression (SAS Institute Inc. 2004) to determine whether treatment (tail clipped or control), sex–age category (male, female, or juvenile), or the covariates, SVL and D (log transformed), had an effect on whether a salamander climbed. We then used general linear models (GLMs) (SAS Institute Inc. 2004) to determine the effect of tail clipping and sex–age category on , H, and m (all log transformed). Trails with values of zero for , H, and m (i.e., those of salamanders that did not climb) were excluded from the GLM analyses to obtain normal distributions after log transformation. In the 3 tests with a climbing parameter as the dependent variable, we used SVL, D, and T as covariates. (We included T to test whether variation in time powdered and time observed affected our results.) In the tests with mean and maximum heights ( and H, respectively) as the dependent variables, because of their relatedness, we used a Bonferroni correction, so that 
= 0.025. Full models included all interaction terms derived from the independent variables. Models were reduced using backward selection, removing parameters with P > 0.10 (SAS Institute Inc. 2004).
Test 2: prey availability on-ground and off-ground
On 15 of the nights on which we tracked salamanders, falling between 28 June and 23 July, we set up on average 7.1 ± 0.55 yellow 62 x 77–mm sticky traps after powdering salamanders. Sticky traps were placed within 100 m of the salamander captures (i.e., at many distinct locations in the forest) and picked up after measuring the fluorescent trails (3–6 h after sunset). On each night, one of the traps was on the ground, with the rest of the traps off the ground, because in preliminary studies we found that off-ground traps captured significantly lower numbers of insects than on-ground traps; more traps improved the accuracy of our estimate for off-ground prey abundance. On each night, ground and off-ground traps were left out for equal lengths of time (2–6 h). Adjacent to the ground trap, other traps were placed on the trunk of a tree (typically Quercus rubra or Acer rubra) at 3 possible heights (0, 110, or 185 cm), on the leaves of large understory shrubs (Acer pensylvanicum and Castanea dentata), on small shrubs (e.g., Vaccinium spp.), on ferns (Thelypteris noveboracensis and Osmunda cinnamomea), or on logs.
We surveyed each trap with a dissecting scope, and all organisms were recorded. Typical taxa included flies, collembolans, lepidopterans (adult and larval), beetles, orthopterans, as well as arachnids (mites and harvestmen). All taxa captured were listed in Petranka (1998) as prey of the euryphagic P. cinereus.
Each organism was placed into 1 of 6 size classes based on body length (see Table 1). We calculated a prey volume index (PVI) for each trap by the following method. First, we assumed, as did Fraser (1976) and Maerz et al. (2006) in a similar context, that organisms were roughly cylindrical, and we further assumed a body diameter equal to half of the body length, so that 
, where v, r, and l are the volume, radius, and length of the cylinder, respectively. To estimate the average volume of organisms in each size class, we calculated the mean of the maximum and minimum volumes, that is, 
is the average volume estimate for the ith size class. These 
are given in Table 1. Finally, to estimate the total prey volume found on a trap, we multiplied the average volume for a size class by the number of organisms in that size class, summing these products. That is, 
, where ni is the number of organisms in the ith size class found on a trap. Although this method does not offer an exact metric for prey volume available, we took it to be a good approximation.
|
Maerz et al. (2006)
found that the mean maximum prey volumes for individuals of P. cinereus were 12.3 ± 1.4 and 27.4 ± 3.4 mm3 for 2 habitat types. Because the approximate minimum volume of prey in v6 > 67.3 mm3 was much greater than this approximate upper limit on prey consumption, we were confident that organisms in size class 6 were too large for P. cinereus to eat. Consequently, we excluded them in the calculation of the PVI. Thus, the sum in the expression for PVI is taken to run from i = 1 to i = 5. For each night, the mean off-ground PVI was taken to be the mean of the PVIs of the off-ground traps. We used a paired t-test (SAS Institute Inc. 2004) to compare the log-transformed variables (ground PVI and mean off-ground PVI for each night).
Test 3: effects of age and sex on natural plant climbing
On 13 nights between 19 June and 27 July 2006, we conducted searches by headlamp in a 12 x 12–m plot in the forest near Mountain Lake Biological Station. Nocturnal surveys were conducted only if there had been precipitation within the last 24 h. We comprehensively surveyed the ground as well as all plants on the plot in search of P. cinereus. We recorded the height off the ground of all individuals located on plants. We measured SVL of each individual to 0.01 mm using Mitutoyo digital calipers (Mitutoyo America Corporation). We distinguished adults from juveniles by their size as above. We determined the sex of adults using the candling method of Gillette and Peterson (2001) to look for the presence (male) or absence (female) of black testes because in this test we were not concerned with their behavior after release. We injected each salamander with fluorescent elastomer (Northwest Marine Technologies) to ensure that recaptured salamanders were not included in our analyses. Data from recaptured individuals were not used in the analyses in order to maintain independence of the data points. After data collection, we released salamanders at their capture locations.
We performed 2 logistic regressions (SAS Institute Inc. 2004). The first excluded juveniles and was used to determine whether sex (male or female) or the covariates—SVL and 4 log-transformed weather parameters, wind, temperature, pressure, and humidity (each averaged over the day of the survey)—were correlated with whether an adult salamander was found off the ground; the interaction term, SVL x sex, was included as well. The second logistic regression was identical except that it included juveniles, and we used age (adult or juvenile) as the binary independent variable instead of sex (correspondingly, SVL x sex was replaced with SVL x age). Because we used the data for whether adults climbed in 2 tests, we applied a Bonferroni correction: = 0.025.
We used a GLM (SAS Institute Inc. 2004) to determine the effect, among salamanders found off the ground, of sex–age category and the covariates—SVL, the same 4 weather parameters used in the logistic regressions, log transformed, and the SVL x sex–age category interaction term—on the height (log transformed) at which they were found, with 2 a priori contrasts: juvenile versus adult and male versus female. We reduced our original model using backward selection, removing parameters with P > 0.10.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONSFUNDINGREFERENCES |
|---|
Test 1: the effects of simulated predation on plant climbing
Data from 9 individuals were excluded from the analyses because of powder spills, indiscernible trails, and ambiguous sex determination. Thirteen of the remaining 86 salamanders did not climb at all. Neither treatment nor sex–age category had a significant effect on whether or not the salamanders climbed (
= 1.2764, P = 0.2586 and = 6.7191, P = 0.8214, respectively).
The reduced model for mean height climbed () was significant (R2 = 0.27, F7,65 = 3.39, P = 0.0038, = 0.025). Increases in SVL significantly increased (F1,65 = 5.32, PSVL = 0.0243; Figure 2). The tail-clipping treatment alone did not have a significant effect (F1,65 = 2.93, Ptreatment = 0.0915), but the treatment x sex–age category interaction was a nonsignificant trend after the Bonferroni correction (F2,65 = 3.55, P = 0.0343), with treatment tending to increase in males but not in females or juveniles (Figure 2). Also included in the reduced model were sex–age category (F1,65 = 0.60, Psex–age = 0.5496) and treatment x SVL (F2,65 = 3.35, PtreatmentxSVL = 0.0719). D and T were nonsignificant, having been removed in the process of backward selection.
|
The reduced model for maximum height (H) was significant (R2 = 0.27, F4,68 = 6.24, P = 0.0002,
= 0.025). There was a significant effect of tail-clipping treatment on H (F1,68 = 5.46, P = 0.0224) with tail-clipped individuals (
) climbing higher than controls (
). Maximum height climbed also increased with increases in SVL (F1,68 = 9.94, P = 0.0024). Other variables remaining in the model were D (F1,68 = 4.54, P = 0.0367) and D x treatment (F2,68 = 4.69, P = 0.0338). Again, time between powdering and observation had no effect. The reduced model for m, the total number of plants climbed, was significant (R2 = 0.33, F1,70 = 17.03, P < 0.0001), including only D (F1,70 = 28.07, Pdistance < 0.0001) and T (F1,70 = 3.62, Ptime = 0.0613), the latter of which was not significant. Salamanders that traveled a greater distance (and thus may be considered more active) tended to climb more plants.
Test 2: prey availability on-ground and off-ground
More prey (in terms of volume, log transformed) were available on the ground (
) than on vegetation (
; t14 = 2.98, P = 0.0099).
Test 3: effects of age and sex on natural plant climbing
Sex was the only variable left in the first logistic regression (which included adults but not juveniles) after stepwise elimination (N = 82), and it was not significant (R2 = 0.0432, = 3.6131, P = 0.0573). In total, 19 individuals (23%) were found on plants and 63 (77%) were not. Males were 2.778 (95% confidence interval: 0.969–7.965) times as likely as females to be found on plants.
The weather parameters were eliminated from the second logistic regression, which included juveniles; remaining in the model were age, SVL, and the SVL x age interaction term. The model itself was significant (R2 = 0.0714, 
= 10.4506, P = 0.0151, = 0.025). Adults were more likely to be found on plants than juveniles ( = 5.7784, Page = 0.0162), and the interaction between SVL and age was significant ( = 5.4275, Page x SVL = 0.0198). There was also a tendency for SVL to be greater when individuals were found on plants, but this was not significant after the Bonferroni correction ( = 4.7328, PSVL = 0.0296, = 0.025).
The reduced GLM for height when found (log transformed) was significant (R2 = 0.52, F8,29 = 2.87, P = 0.0253) and included all variables from the full model except humidity—that is, sex–age category (F2,29 = 3.83, Psex–age = 0.0382), SVL (F1,29 = 0.39, PSVL = 0.5410), SVL x sex–age (F2,29 = 3.57, Psex–age = 0.0464), wind (log transformed; F1,29 = 5.33, Pwind = 0.0313), temperature (log transformed; F1,29 = 4.94, Ptemperature = 0.0374), and pressure (log transformed; F1,29 = 6.55, Pwind = 0.0183). Wind and pressure were negatively correlated with the height at which salamanders were found, whereas temperature was positively correlated with the height. A priori contrasts found that males were found higher than females (F1,29 = 7.60, P = 0.0118; Figure 3) but that there was no significant difference between heights of juveniles and adults (F1,29 = 1.14, P = 0.2969; Figure 3).
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONSFUNDINGREFERENCES |
|---|
Unlike in aquatic systems, risk–foraging trade-offs have rarely been quantified for terrestrial animals. Studies inferring such trade-offs in terrestrial animals usually quantify predation risk (e.g., Martín and López 2000
; Amo et al. 2003; Cooper 2003; Stankowich and Coss 2007) without assessing prey abundance or quality (but see Abramsky et al. 2002). In this study, we used simulated predation attempts and quantification of prey abundance to identify a risk–foraging trade-off for the red-backed salamander, P. cinereus. We found that individuals of P. cinereus that experience a nonlethal tail injury climb higher than those that are not injured. Prey abundance (in terms of volume) was far greater on the ground suggesting that individuals who have sustained such an injury have a heightened perception of risk near the ground and perceive lower risk higher on plants. Our results, then, are consistent with the hypothesis that plant climbing is a predator avoidance strategy in P. cinereus. Before concluding in favor of this hypothesis, however, we address 2 alternative hypotheses.
Jaeger (1978) hypothesized that individuals of P. cinereus climb plants because more prey is available on plants than on the ground. Plethodon cinereus stores important fat reserves in its tail (Fraser 1980), a portion of which we autotomized during the simulated predation experiment; it was therefore possible that tail loss caused salamanders to attempt to compensate for lost energy stores by increasing foraging behavior and/or seeking out prey-rich habitats. However, we found that plants actually provided lower prey availability than the ground. Furthermore, some common prey items (e.g., earthworms) are totally unavailable on plants. We conclude that plants are food poor as compared with the ground. Therefore, our results contradict the hypothesis that climbing in P. cinereus is a foraging strategy.
Second, intraspecific competition may cause salamanders to climb. Loss of dominance has been suggested as an explanation for changes in habitat selection in response to induced tail autotomy in the lizard Holbrookia propinqua (Cooper 2003). Plethodon cinereus exhibits intraspecific competition for cover object space during the day (Jaeger et al. 1995), with larger individuals usually outcompeting smaller conspecific individuals (Mathis 1990). However, if the salamanders climbing higher on plants were doing so because they were being competitively excluded from nocturnal ground foraging space by larger conspecifics, then we would expect to see a negative correlation between SVL and maximum and/or average height (H and/or ). The positive correlation in our results (Figure 2) contradicts this competition hypothesis. However, we note that there is some evidence that smaller species of salamanders tend to dehydrate faster than larger species (Grover 2000). This study did not test differential intraspecific physiological constraints of size within this species, which could have affected climbing, so further research may be warranted to definitively refute the competition hypothesis.
The results of this study suggest that plant climbing is a predator avoidance strategy in P. cinereus. Future studies quantifying actual mortality risk or predation pressure in the salamander's 3 main microhabitats—underground burrows, ground surface, and vegetation—would be useful for confirming that the ground surface is more dangerous than vegetation as well as determining predation risk for underground retreats, which may be low.
It has been proposed that prey availability is very low in underground burrows (Fraser 1976). However, prey availability in underground burrows has not been quantified. Adding data on underground prey availability and actual predation risk in all microhabitats to our data would complement our data in this terrestrial system.
Our study found variation in plant-climbing behavior between males and females of P. cinereus, that is, males were found higher than females (Test 3; Figure 3), and there was a trend for males to climb higher than females when tail clipped (Test 1; Figure 2). Foraging requirements may explain this variation. Maerz (2000) found sex differences in P. cinereus foraging behavior, with females capturing more prey (but see Maerz et al. 2001). This difference in climbing by different sexes is of interest because females’ ability to yolk a clutch is constrained by energy consumption (Petranka 1998), and at least one study in another plethodontid salamander found that spermatophore production has a low cost (Marks and Houck 1989). We found that the ground is more prey rich than vegetation, so females may remain on the ground to forage despite the increased risk of predation because they are more energetically stressed (sensu DeLaet 1985; Hegner 1985; Godin and Sproul 1988; Hogstad 1988). Future studies investigating this possible sex difference in habitat selection could shed light on sex-related behavior in P. cinereus, as well as providing opportunities for further evaluating the effects of sex on habitat selection.
In concert with other recent studies in terrestrial systems (e.g., Abramsky et al. 2002; Cooper 2003), this field study has helped establish experimentally that the relative costs of habitat selection can play a large role in terrestrial systems as well as in aquatic ones. This study complements studies of risk–foraging trade-offs in aquatic species and adds to the limited data on terrestrial taxa by establishing the existence of such a trade-off in a sedentary terrestrial poikiotherm.
|
FUNDING |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONSFUNDINGREFERENCES |
|---|
Mountain Lake Biological Station; University of Virginia IACUC Protocol Number 3283; National Science Foundation (DBI-0453380).
|
|
ACKNOWLEDGEMENTS |
|---|
We thank H. Wilbur and V. Rudolf for advice and constructive criticism as well as those who helped with fieldwork: J. Armstrong, M. Longest, A. Roe, and S. Schrock. We thank K. Grayson for providing the fluorescent powder. Comments from E. Brodie III, from the editor, H. Hofmann, and from 2 anonymous reviewers greatly improved the quality of the manuscript.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSCONCLUSIONSFUNDINGREFERENCES |
|---|
Abramsky Z, Rosenzweig ML, Subach A. The costs of apprehensive foraging. Ecology (2002) 83:1330–1340.[Web of Science]
Amo L, López P, Martín J. Risk level and thermal costs affect the choice of escape strategy and refuge use in the Wall Lizard, Podarcis muralis. Copeia (2003) 2003:899–905.[CrossRef]
Blankenship EL, Bryan TW, Jacobsen SP. A method for tracking tortoises using fluorescent powder. Herpetol Rev. (1990) 21:88–89.
Bock SF, Fauth JE. Pseudotriton (northern red salamander). Diet. Herpetol Rev. (1992) 23:58.
Brodie ED Jr, Nowak RT, Harvey WR. The effectiveness of antipredator secretions and behavior of selected salamanders against shrews. Copeia (1979) 1979:270–274.[CrossRef]
Burton TM, Likens GE. Salamander populations and biomass in the Hubbard Brook Experimental Forest, New Hampshire. Copeia (1975) 1975:541–546.[CrossRef]
Churchfield S. The natural history of shrews (1990) Ithaca (NY): Cornell University Press.
Cooper WE. Shifted balance of risk and cost after autotomy affects use of cover, escape, activity, and foraging in the keeled earless lizard (Holbrookia propinqua). Behav Ecol Sociobiol (2003) 54:179–187.[Web of Science]
Dakin SF. The influence of interspecific competition on the microhabitat distribution of terrestrial lungless salamanders in the southern Appalachian mountains [PhD dissertation] (1978) [Durham (NC)]: Duke University.
DeLaet JF. Dominance and anti-predatory behaviour of great tits Parus major: a field study. Ibis (1985) 127:372–377.[CrossRef]
Eggert C. Use of fluorescent pigments and implantable transmitters to track a fossorial toad (Pelobates fuscus). Herpetol J (2002) 12:69–74.
Fellers GM, Drost CA. Fluorescent powder: a method for tracking reptiles. Herpetol Rev (1989) 20:91–92.
Fraser DF. Empirical evaluation of hypothesis of food competition in salamanders of the genus Plethodon. Ecology (1976) 57:459–471.[CrossRef][Web of Science]
Fraser DF. On the environmental control of oocyte maturation in a plethodontid salamander. Oecologia (1980) 46:302–307.[Web of Science]
Fraser DF, Gilliam JF. Feeding under predation hazard: response of the guppy and Hart's rivulus from sites with contrasting predation hazard. Behav Ecol Sociobiol (1987) 21:203–209.[CrossRef][Web of Science]
Gillette JR, Peterson MG. The benefits of transparency: candling as a simple method for determining sex in red-backed salamanders (Plethodon cinereus). Herpetol Rev (2001) 32:233–235.
Gilliam JF, Fraser DF. Habitat selection under predation hazard: test of a model with foraging minnows. Ecology (1987) 68:1856–1862.[CrossRef][Web of Science]
Godin J-GJ, Sproul CD. Risk-taking in parasitized sticklebacks under threat of predation: effects of energetic need and food availability. Can J Zool (1988) 66:2360–2367.
Grover MC. Determinants of salamander distributions along moisture gradients. Copeia (2000) 2000:156–168.[CrossRef]
Hamilton WJ Jr. The food and feeding behavior of the garter snake in New York state. Am Midl Nat (1951) 46:385–390.[CrossRef]
Hegner RE. Dominance and anti-predatory behaviour in blue tits (Parus caeruleus). Anim Behav (1985) 33:762–768.[CrossRef][Web of Science]
Heller R, Milinski M. Optimal foraging of sticklebacks on swarming prey. Anim Behav (1979) 27:1127–1141.[CrossRef][Web of Science]
Hochman V, Kotler BP. Patch use, apprehension, and vigilance behavior of Nubian Ibex under perceived risk of predation. Behav Ecol (2007) 18:368–374.
Hogstad O. Social rank and antipredator behavior of willow tits Parus montanus in winter flocks. Ibis (1988) 130:45–56.[CrossRef]
Jaeger RG. Plant climbing by salamanders: periodic availability of plant-dwelling prey. Copeia (1978) 1978:686–691.[CrossRef]
Jaeger RG, Gabor CR, Wilbur HM. An assemblage of salamanders in the southern Appalachian Mountains: competitive and predatory behavior. Behaviour (1998) 135:795–821.
Jaeger RG, Wicknick JA, Griffis MR, Anthony CD. Socioecology of a terrestrial salamander: juveniles enter adult territories during stressful foraging periods. Ecology (1995) 76:533–543.[CrossRef][Web of Science]
Kullberg C. Spatial niche dynamics under predation risk in the Willow Tit Parus montanus. J Avian Biol (1998) 29:235–240.[CrossRef]
Lemen CA, Freeman PW. Tracking mammals with fluorescent pigments: a new technique. J Mammal (1985) 66:134–136.[CrossRef][Web of Science]
Liebgold EB, Jaeger RG. Juvenile movements and potential inter-age class associations of red-backed salamanders. Herpetologica (2007) 63:51–55.[CrossRef][Web of Science]
Lima SL, Bednekoff PA. Temporal variation in danger drives antipredator behavior: the predation risk allocation hypothesis. Am Nat (1999) 153:649–659.[CrossRef][Web of Science]
Lima SL, Dill LM. Behavioral decisions made under the risk of predation: a review and prospectus. Can J Zool (1990) 68:619–640.
Maerz JC. Prey availability and phenotypic differences between local terrestrial salamander populations [PhD dissertation]. (2000) [Binghamton (New York)]: State University of New York at Binghamton.
Maerz JC, Myers EM, Adams DC. Trophic polymorphism in a terrestrial salamander. Evol Ecol Res (2006) 8:23–35.
Maerz JC, Panebianco NL, Madison DM. Effects of predator chemical cues and behavioral biorhythms on foraging activity of terrestrial salamanders. J Chem Ecol (2001) 27:1333–1344.[CrossRef][Web of Science][Medline]
Marks SB, Houck LD. Partial cost of spermatophore production in the salamander, Desmognathus ochrophaeus. J Herpetol (1989) 23:81–84.[CrossRef]
Martín J, López P. Costs of refuge use affect escape decisions of Iberian rock lizards Lacerta monticola. Ethology (2000) 106:483–492.[CrossRef][Web of Science]
Mathis A. Territoriality in a terrestrial salamander: the influence of resource quality and body size. Behaviour (1990) 112:162–175.[CrossRef]
Mathis A. Territories of male and female terrestrial salamanders: costs, benefits, and intersexual spatial associations. Oecologia (1991) 86:433–440.[CrossRef][Web of Science]
Milinski M, Heller R. Influence of a predator on the optimal foraging behaviour of sticklebacks (Gasterosteus aculeatus L.). Nature (1978) 275:642–644.[CrossRef]
Petranka JW. Salamanders of the United States and Canada (1998) 1st ed. Washington: Smithsonian Institute Press.
Preisser EL, Bolnick DI, Benard MF. Scared to death? The effects of intimidation and consumption in predator-prey interactions. Ecology (2005) 86:501–509.[CrossRef][Web of Science]
Railsback SF, Harvey BC. Analysis of habitat-selection rules using an individual-based model. Ecology (2002) 83:1817–1830.[CrossRef][Web of Science]
Ransom TS, Jaeger RG. An assemblage of salamanders in the southern Appalachian Mountains revisited: competitive and predatory behavior? Behaviour (2006) 143:1357–1382.[CrossRef]
Rittenhouse TAG, Altnether TT, Semlitsch RD. Fluorescent powder pigments as a harmless tracking method for Ambystomatids and Ranids. Herpetol Rev (2006) 37:188–191.
Sayler A. The reproductive ecology of the red-backed salamander, Plethodon cinereus, in Maryland. Copeia (1966) 1966:183–193.[CrossRef]
Stankowich T, Coss RG. Effects of risk assessment, predator behavior, and habitat on escape behavior in Columbian black-tailed deer. Behav Ecol (2007) 18:358–367.
Taub FB. The distribution of the red-backed salamander, Plethodon c. cinereus, within the soil. Ecology (1961) 42:681–698.[CrossRef][Web of Science]
Test FH, Bingham BA. Census of a population of the red-backed salamander (Plethodon cinereus). Am Midl Nat (1948) 39:362–372.[CrossRef]
Verdolin JL. Meta-analysis of foraging and predation risk trade-offs in terrestrial systems. Behav Ecol Sociobiol (2006) 60:457–464.[CrossRef][Web of Science]
Werner EE, Gilliam JF, Hall DJ, Mittelbach GG. An experimental test of the effects of predation risk on habitat use in fish. Ecology (1983) 64:1540–1548.[CrossRef][Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||