To hold or not to hold? The effects of prey type ... - Wiley Online Library

Journal of Zoology. Print ISSN 0952-8369

To hold or not to hold? The effects of prey type and size on the predatory strategy of a venomous snake X. Glaudas1

, T. C. Kearney1,2 & G. J. Alexander1

1 School of Animal, Plant and Environmental Sciences, University of the Witwatersrand, Johannesburg, South Africa 2 Vertebrate Department, Ditsong National Museum of Natural History, Pretoria, South Africa

Keywords Bitis arietans; predatory strategy; foraging; predator–prey relationship; puff adder; prey size; venomous snakes. Correspondence Xavier Glaudas, School of Animal, Plant and Environmental Sciences, University of the Witwatersrand, Johannesburg, P.O. Wits, 2050, South Africa. Tel: +27 (76) 652 8018; Fax: +27 (11) 717 6494 Email: [email protected] €del Editor: Mark-Oliver Ro Received 16 September 2016; revised 30 January 2017; accepted 1 February 2017 doi:10.1111/jzo.12450

Abstract The dangerous prey hypothesis predicts that when predators can discriminate between harmless and dangerous prey, they should alter their predatory behavior according to the risk. Venomous snakes, which rely on an envenomating strike to kill prey, often feed on potentially dangerous prey such as rodents, and have the choice between two alternative strategies following a bite: they either hold onto prey until it is incapacitated, or release it immediately, relying on scent trailing to find the bitten prey. In this study, we combined observational and experimental data collected in the field in South Africa to test the hypothesis that prey type and size affected the predatory strategy of a venomous snake, the puff adder (Bitis arietans; Viperidae). Relative prey size, but not prey type, affected the snake’s strategy: puff adders typically struck and held prey of small size (e.g., toads, small adult mice and shrews), but often struck and released larger prey (e.g., large adult mice and rats), presumably because retaliatory bites from the latter ones are more dangerous. Hence, puff adders discriminated according to prey size and adjusted their predatory strategy consequently. Contrary to the common belief that viperid snakes typically strike and release prey, our findings agree with empirical evidence found in the literature, that is, puff adders generally alter their predatory tactics based on prey size. Although a few laboratory studies have examined the theme of the present paper in other venomous snakes, our research was conducted in a natural setting and thus provides a better understanding of the factors used by predators to choose between alternative hunting strategies.

The dangerous prey hypothesis predicts that when predators can discriminate between harmless and dangerous prey, they should alter their predatory behavior according to the risk (Forbes, 1989). These adaptations to minimizing risk from dangerous prey have been reported for a variety of organisms (see Mukherjee & Heithaus, 2013 for a review) and may involve a wide variety of tactics. For example, Burton’s legless lizard (Lialis burtonis) strike large lizard prey, but not small ones, on the head or neck to prevent them from biting back (Wall & Shine, 2007); and herons and grebes increase prey handling time when feeding on ictalurid catfish, which possess locking spines that could injure these gape-limited predators (Forbes, 1989). These adaptations of predators represent part of their evolutionary response to the arms races between predator and prey. The evolution of snakes has been strongly shaped by their diet (Gans, 1961; Greene, 1983). Many snakes feed on potentially dangerous and sometimes large prey, and they have evolved efficient mechanisms (e.g., constriction, venom) to quickly incapacitate them (Greene, 1997; Wall & Shine, 2007).

Accumulating evidence also indicates that snakes often make context-dependent predatory decisions (Reiserer, 2002; Glaudas & Alexander, 2017a). For instance, constrictors frequently seize and swallow small and harmless prey alive, while larger and/or more dangerous prey are killed by constriction prior to ingestion (de Queiroz, 1984; Mori, 1994), indicating both prey discrimination and behavioral adjustment. Venomous snakes are particularly interesting in this regard, because they generally exhibit one of two stereotyped predatory strategies following an envenomating strike: either they hold onto prey until it is incapacitated or dead (‘strike-andhold’) or release it immediately (‘strike-and-release’; hereafter, the hold and release strategies, respectively). The injection of venom insures that prey will soon succumb, and if released, the bitten prey can be recovered through scent trailing, a mechanism known as strike-induced chemosensory searching (SICS; Chiszar et al., 1982). Hence, venomous snakes are ideal study organisms to examine some of the factors that affect their predatory decisions. The predatory strategies of venomous snakes has received considerable attention, from research investigating SICS

Journal of Zoology 302 (2017) 211–218 ª 2017 The Zoological Society of London

211

Introduction

Venomous snake predatory strategy

(Chiszar et al., 1982) and the factors involved in the amount of venom injected by snakes into prey (Hayes, 2008), to studies examining the determinants of the hold and release strategies. For the latter, viperid snakes, by far the most commonly studied taxa, are generally considered to release prey after a strike (Deufel & Cundall, 2006), despite evidence demonstrating that some species adjust their strategies based on prey type and/or size (Allon & Kochva, 1974; Kardong, 1986; Hayes, 1992), while others do not (Chiszar et al., 1982; Hayes, 1995). In this field study conducted in South Africa, we tested the effect of prey type and size on the predatory strategy of a venomous snake, the puff adder (Bitis arietans; Viperidae). Puff adders are ambush-foraging snakes which, in our population, feed mostly on rodents and amphibians, occasionally on birds and lizards, and are reported to hold onto prey (Chiszar et al., 1982). We quantified the relationship between predatory tactic and prey type and size, using two methods: first, we collected observational data using remote video-cameras focused on ambush-hunting snakes, and secondly, we experimentally manipulated rodent prey size by offering pre-killed rodents to free-ranging radio-tagged puff adders. We specifically predicted that snakes would release potentially dangerous rodents more often than relatively harmless amphibians; and that the probability of snakes holding onto prey would decrease with increasing prey size, because larger prey are more dangerous (Radcliffe, Chiszar & O’connell, 1980). All prior experimental studies on this topic have been conducted in the laboratory (but see Putman, Barbour & Clark, 2016 for field observational data on Crotalus oreganus). Hence, our research conducted in a biologically relevant context provides a better understanding of the factors used by predators to choose between alternative hunting strategies.

Materials and methods Study site and species The study took place in the Dinokeng Game Reserve, a ca. 18 500 ha area in the Gauteng Province of South Africa ( 25.38S, 28.31E; ca. 1100 m. a.s.l.). The site, which is composed of a mosaic of savannas and open woodlands, falls within the Savanna Biome, and is seasonal with hot, wet summers and mild, dry winters (Shulze, 1997). The puff adder is a heavy-bodied medium-sized viperid snake (ca. 700–900 mm snout-to-vent length [SVL]), which occurs in savannas and open woodlands throughout most of sub-Saharan Africa and parts of the Arabian Peninsula. In the area of study, puff adders are most active from the onset of the rain season, typically in late October–November, to the beginning of the dry season in June-July. During the coldest months of the austral winter (July–August), puff adders do not hibernate but generally remain inactive.

X. Glaudas, T. C. Kearney and G. J. Alexander

males, 44 females) in accordance with established procedures (Reinert, 1992). Transmitter mass never exceeded 5% of the snake body mass. We released snakes at their capture locations, typically 3–4 days following surgery, and located snakes every 2–3 days, using a R1000 radio receiver (Communications Specialists, Inc., Orange, CA, USA) and a Yagi antenna (Africa Wildlife Tracking, Pretoria, Gauteng, South Africa). At the end of the study, we removed all radio transmitters and released the snakes where they were last caught. From 23-Sep-2013 to 23-Dec-2015, we used fixed videography to monitor the foraging behavior of radio-tagged puff adders in the field. This method consisted of setting up continuously recording video-cameras focused on ambushed predators and is a powerful approach to recording precise feeding data on ambush-hunting snakes (Clark, 2006). Each video-camera unit consisted of a closed-circuit television surveillance camera (model PC177IRHR-8, Supercircuit Inc., Austin, TX, USA) connected to a recording mini digital video recorder (model MDVR 14-4, Supercircuit Inc.; 30 frames s 1), and powered by a 12-V sealed lead-acid battery. The video-cameras recorded in color during the day and automatically switched to night-time vision, using the built-in infrared LEDs at low light levels. We located radio-tagged snakes, using radiotelemetry and set up the video-camera units ca. 70 cm in front of the snakes. The following day, we collected the memory cards and reviewed recordings to identify successful strike at prey, prey identity to the lowest possible taxonomic level, and latency to prey death (time elapsed from the strike to the last muscular twitch or motion of prey, s). We also recorded the post-strike predatory strategy of the snakes: we considered that a snake released prey when snake let go of it within 2-s of a strike (with the exception of a case when a snake got its teeth entangled in the fur of a rat for ca. 3 s, while the snake attempted to let go; see Video S18). Further, we estimated prey length (cm), after the feeding event happened, by placing a ruler at an angle that matched the prey trajectory in the camera’s field of view, and estimated size-specific prey mass by comparing images of prey with specimen information for intact congeners or conspecifics of similar size from the nearest locality in the Ditsong National Museum of Natural History collections.

Experimental manipulation of rodent prey size

As part of a radiotelemetric study on puff adders, we implanted radio transmitters (model SI-2, 13 g; Holohil Ltd, Carp, ON, Canada) into the body cavity of 86 puff adders (42

As part of a food supplementation study on adult male puff adders from 2013 to 2015 (Glaudas & Alexander, 2017b), we offered pre-killed striped mice (Rhabdomys pumilio), a natural prey species, for a 2-month period to six individuals each year (18 individuals overall). We used radiotelemetry to locate the snakes, and fed them in situ from 8-Jan to 13-Mar in 2013, 15-Jan to 7-Mar in 2014, and 13-Jan to 5-Mar in 2015. During these periods, each snake was offered 1–4 R. pumilio per week in a single feeding event, up to a maximum of ca. 15% of its body mass (recorded at the start of the food supplementation period), using a pair of 1-m snake tongs. Prior to feeding the snakes, we recorded their behavior. We defined snakes as resting when we found them in a coiled position with their head

212


Radiotelemetry and fixed videography


unraised, typically at the center of the body coil, or in ambush when the chin rested on the ground in a striking position with the body loosely coiled, the typical foraging posture for puff adders. We held the mice, presented headfirst, in the tongs, and slowly approached it toward the snake location from ca. 0.5 m away at a speed of ca. 0.3 m/s 1. We held mice as close to the ground as possible, and selected the angle of approach that presumably maximized prey detection by snakes. The mice were randomly selected from a container and weighed using a portable scale (accuracy 1 g) prior to being presented to the snakes. We offered mice to the snakes, and whenever snakes struck, immediately released them and recorded the snake’s post-strike predatory strategy (released ≤ 2 s; monitored with a stopwatch). For logistic reasons, we were unable to videotape the experimental part of this study. The behaviors observed were recorded in the field, and the strategies used by snakes were in all cases clearly distinguishable.

Statistical analysis We used generalized estimating equations to model the binary logistic responses of snakes (holding prey or not; treating releasing prey as the reference category) to variation in prey type and size. We used generalized estimating equations to account for multiple observations on the same individuals. We used the observational data collected through videography to characterize the relationship between prey type and size and predatory strategy. We grouped the prey captured on videos to their respective group (mammals, amphibians, avian and non-avian reptiles). Because our sample size for avian and non-avian reptiles was small (3 and 2, respectively), we only analyzed data for the most common prey types, amphibians and mammals. We used prey type as the categorical variable and ln-transformed relative prey mass (prey mass/snake mass) as the covariate, allowing us to control for snake size. We further examined the link between rodent prey size and predatory strategy, using our experimental manipulation of prey size. We accounted for the fact that we often fed snakes several mice per day (1–4) over several weeks (up to 10) in the following way. First, snakes rarely struck at prey after having eaten two mice (n = 3), and therefore we only included the first two mice in our analysis. We accounted for a possible order effect on snake strategy by including mouse order in the model. Second, we tested for a week effect by initially running the model with week number as a covariate. The lack of statistical significance (P = 0.44) indicated that measurements on the same snakes across weeks were relatively independent, and we removed this variable from the analysis. Similarly, we included snake behavior (resting vs. ambush) in the initial analysis, but removed this factor due to a lack of significance (P = 0.37). Hence, our model included relative prey mass as the covariate, and mouse order as the categorical variable. For each model, we used the quasi-likelihood under the independence model criterion (QIC) to select the most appropriate working correlation matrix structure for the data, that is, the one with the smallest QIC value (Pan, 2001). We conducted all statistical analyses using STATISTICA, version 12.5 Journal of Zoology 302 (2017) 211–218 ª 2017 The Zoological Society of London


(StatSoft Inc., 2014), and SPSS, version 23.0 (IBM Corp., 2015). Values given are means SD unless otherwise mentioned, and all reported P-values are two-tailed. Significance level for all tests was a = 0.05.

Results Fixed videography data We reviewed 4634 h of video recordings of snakes foraging (18 females and 18 males). Fourteen snakes (seven males, seven females; SVL: 772 85 mm; mass: 630 191 g), which were all in ambush, successfully struck at 26 prey (11 mammals, 10 amphibians, three avian and two non-avian reptiles; see Data S1 for a list of prey and Video S2–S26 for 25 videos): six snakes were successful once, and four individuals successfully struck at prey twice and thrice. (Only one snake was experimentally fed prior to collecting the videography data, but this snake was last offered food 20 months before being filmed in the field. Hence, we are confident that snakes exhibited natural behaviors.) We collected reliable mass estimates for 25 prey. Relative prey mass averaged 24 83.6% (range: 0.45–422%), and absolute prey mass averaged 124 382 g (2.5–1933 g). The (relatively) smallest prey item was a bushveld rain frog (Breviceps adspersus; see Video S26) eaten by a 558 g snake, while the largest was a Lepus hare struck by a 458 g puff adder (see Video S17), which the snake was unable to swallow. The only significant effect in the model was caused by relative prey mass (Table 1): snakes were more likely to hold onto smaller prey than larger ones (Fig. 1). Anecdotally, the three birds captured by snakes were held, and of the two lizards caught, the arboreal Acanthocercus atricollis was held and the terrestrial Gerrhosaurus flavigularis was released. Whenever possible, we recorded latency to prey death: eight cases involved prey that were held, and we added an additional case of a young Pronolagus rabbit, which was released but died in the camera’s field of view. Our analysis included four mammals and five amphibians that did not differ in (ln-transformed) relative prey mass (one-way ANOVA; F1,7 = 0.08, P = 0.77). Relative prey mass and prey type did not affect latency to death (general linear model; ln-transformed mass: F1,6 = 0.5, P = 0.5; type: F1,6 = 2.4, P = 0.17), but our sample size was small. Further, considering only prey

Table 1 Generalized estimating equation of the ‘strike-and-hold’ predatory strategy in Bitis arietans using the fixed videography data

Variable (intercept) Prey class (amp) Relative prey mass (ln)

Estimate SE (95% CI) 3.05 1.6 ( 0.38 to 6.1) 0.83 1.35 ( 1.8 to 3.5) 1.71 0.72 ( 3.1 to 0.31)

Wald v² (d.f.)

P

3.75 (1) 0.38 (1) 5.70 (1)

0.053 0.54 0.017*

Estimates (SE; and their 95% confidence intervals) and tests of model effects are shown. For prey class, amp refers to amphibian prey and mammals are treated as the reference category. The asterisk denotes significant effect of a variable in the model.

213



Figure 1 The relationship between relative prey mass (%) and predatory strategy (R for release and H for hold) in puff adders. The horizontal dashed line distinguishes observational from experimental data. Note the scale break on the x-axis between 40 and 420.

that were held, amphibians (mass: 27.6 21 g) died in 221 148 s (n = 5), and mammals, which were small (9.8 5.6 g), died in 44 23 s (n = 3). The rabbit died in 228 s.

Experimental manipulation of rodent prey size We offered a total of 192 Rhabdomys pumilio to 18 different snakes, which resulted in 87 strikes from 17 snakes (SVL: 790 100 mm; mass: 827 239 g). The number of mice struck per individual averaged 5.1 3.3 (range: 1–13). Relative prey mass of struck prey averaged 7 2.6% (1.5–14.4%), and absolute prey mass averaged 56 23.6 g (14–131 g). We included 84 strikes at prey in the model: snakes released and held onto prey on 54 (relative prey mass: 7.6 2.6%), and 30 (6 2.4%) occasions, respectively. Relative prey mass and mouse order significantly affected the probability that a snake would hold onto prey (Table 2): relative prey mass negatively co-varied with the hold strategy (Fig. 1), and snakes were more likely to release the first mouse, which were relatively heavier (one-way ANOVA; F1,82 = 3.9, P = 0.05), compared to the second. The odds ratio (e 0.23; the exponent being

the parameter estimate for relative prey mass in the model; see Table 2) indicated that the probability of holding onto prey decreased by a factor of 0.792 (95% confidence interval: 0.63– 0.99) or by 20.8% per 1% increase in relative prey mass. Therefore, a 10% increase in relative prey mass decreased the probability that snakes held onto prey by a factor of 0.1 (0.79210) or by 90%.

Discussion

Estimates (SE; and their 95% confidence intervals) and tests of model effects are shown. For the mouse order variable, 1 refers to the first mouse offered, and hence the second mouse offered is treated as the reference category. Asterisks denote significant effects of variables in the model.

We combined observational and experimental data collected in the field to test the hypothesis that the predatory strategy of a venomous snake is affected by prey type and size. Our results showed that prey size, but not prey type, affected the odds of snakes holding onto prey, with increased probability of puff adders releasing relatively larger prey. This indicates that puff adders discriminated between small and large prey, and made risk-sensitive decisions when capturing them. Although a few studies have examined this theme in venomous snakes, our research is unique because it is the first conducted in the field. Laboratory experiments are extremely valuable, partly because they allow control of many variables, but captivity can alter the natural behavior of animals (Warwick, 1990) and may not allow observations of the actual breadth and complexity of snake behavior (Glaudas & Alexander, 2017a). We used radiotelemetry and fixed videography to examine the relationship between predatory strategy and prey type. We predicted that snakes would hold onto harmless amphibian prey, but release potentially dangerous mammals. Our videography analysis did not support this idea. However, we cannot conclude that prey type does not affect puff adder predatory strategy: relative prey mass was much higher for mammals (mean SE; 54.7 41%) compared to amphibians (2.5 0.7%; t-test with unequal variance, P = 0.01), and snakes held onto amphibians in 9/10 cases. The single instance where a toad was released was possibly due to it escaping the snake’s jaws, which we conservatively scored as a release (see Video S8). Inclusion of prey type only in the model revealed a significant effect in that amphibians were significantly more often held than mammals (P = 0.03), but this effect

214


Table 2 Generalized estimating equation of the ‘strike-and-hold’ predatory strategy in Bitis arietans when offered pre-killed Rhabdomys pumilio

Variable (intercept) Mouse order (1) Relative prey mass (%)

Wald v² (d.f.)

P

1.89 0.93 (0.6 to 3.7) 1.5 0.6 ( 2.6 to 0.4)

4.10 (1) 7.10 (1)

0.04* 0.01*

0.23 0.11 ( 0.46 to

3.96 (1)

0.046*

Estimate SE (95% CI)

0.004)



disappeared when relative prey mass was included. Hence, due to the important difference in relative prey mass between amphibians and mammals, we unfortunately cannot separate the effect of prey size from prey type. Many snakes exhibit prey-specific variation in prey-handling behavior (Rodrıguez-Robles & Leal, 1993; Mehta & Burghardt, 2009). However, experimental evidence for an effect of prey type on the hold/release strategy of venomous snakes is limited. Banded rock rattlesnakes (Crotalus lepidus) presumably held onto lizards but not rodents (Chiszar, Radcliffe & Feiler, 1986), although no data were presented to support the former claim. The only two quantitative experimental studies that we are aware of reported that cottonmouths (Agkistrodon piscivorus) held onto fish but typically released mice (O’Connell, Chiszar & Smith, 1981), and that prairie rattlesnakes (C. viridis) held onto birds more often than mice (Hayes, 1992). Presumably, releasing fishes and birds would make their recovery by snakes difficult, if fishes/birds swam away/flew off leaving no chemical trail (O’Connell et al., 1981; Hayes, 1992). Our anecdotal observations that puff adders held onto birds and an arboreal lizard agree with these hypothesis, suggesting that prey type may indeed affect their predatory strategy. Our findings for prey size were more straightforward, because relative prey size negatively co-varied with the hold strategy in both our observational and experimental studies. Therefore, as we predicted, puff adders discriminated prey based on size and adjusted their predatory tactic accordingly. Interestingly, an earlier laboratory study anecdotally reported that puff adders held onto adult rodent prey (Chiszar et al., 1982). Our findings do not support this, but these contrasting results are likely due, in part, to the difference in (relative) prey mass between studies: all mice offered to puff adders in Chiszar et al. (1982)’s laboratory study were small (ca. 20 g), while we observed/offered prey that varied substantially in (relative) prey mass. Altogether, these results demonstrate that the generalization that puff adders hold onto adult rodent prey is incorrect (e.g., an adult laboratory rat often weights 10 times the mass of an adult laboratory mouse), and emphasize the fact that variable methodologies can lead to different conclusions. Holding onto prey increases the risk of a retaliatory bite (Kardong, 1982), which is supported by two of our video recordings: a large puff adder that held onto the arboreal lizard Acanthocercus atricollis was repeatedly bitten on the body, without any obvious detrimental consequences, and the snake did not let go (see Video S2). The other instance involved a snake that held onto a small mammal, but seemingly released it following a bite to the head (see Video S20). However, the snake seemed unaffected and started SICS within minutes following release. In all other cases, hanging onto small mammal prey apparently did not come at a cost to puff adders, because mammals that were held quickly died, and none struggled sufficiently to pose a threat to the snake. Puff adders strike hard and fast (Young, 2010), and have long fangs relative to most vipers (Cundall, 2009). These characteristics may partly explain the tendency for puff adders to hold onto small prey as the strike is likely to knock prey out and/or the long fangs hit vital organs, which would quickly incapacitate prey and decrease the odds of retaliation.

On the other hand, holding onto prey has the benefit that the snake is less likely to lose its meals, and would not need to invest time and effort in trailing prey. Trailing prey, a consequence of the release strategy, could expose snakes to predators, reveal the snake’s location to other potential prey, and also represent an opportunity cost, if time involved trailing prey significantly decreases foraging time. A female puff adder that struck and released a black rat (Rattus rattus) did not recover the bitten prey, despite investing more than 10 h in prey trailing. This indicates that losing prey is indeed a potential cost associated with the release strategy, and that tracking prey may require a significant time investment. Factors other than prey size likely affected the tendency of snakes to hold or release prey: for example, some vipers are known to target specific prey body parts, which decreases time to death and the risk of being bitten (Kardong, 1986; Barr, Wieburg & Kardong, 1988), and/or the fangs may in some cases have contacted hard-to-penetrate body parts (Cundall, 2009), which may all contribute to the predatory strategy used by snakes. Unfortunately, the resolution limitations of our field video recordings, coupled with the facts that prey was often small enough to be entirely swallowed and that we did not videotape the experimental part of the study preclude us from examining the importance of these variables. Therefore, although we acknowledge that other factors probably contributed to the use of a specific strategy, our data show that some of this variation is significantly explained by prey size. Viperid snakes are generally considered to release prey following an envenomating strike (Deufel & Cundall, 2006), yet empirical data show that the hold response often depends on prey size, as supported by our findings. In fact, a tabulation of published species-specific strategies clearly shows that most viperid snakes studied to date adjust their predatory strategy based on prey size (Table 3), although the strength of this relationship varies by species (Radcliffe et al., 1980). Thus, a more appropriate generalization would be that vipers typically adjust predatory strategies based on prey size. A similar conclusion was reached in a study investigating the hold/release strategy by snake families and lifestyles (Deufel & Cundall, 2006). The latter study also found a link between habitat preference and predatory strategy, in that arboreal snakes held onto prey comparatively more than terrestrial species. A prey item released in an arboreal environment would likely drop to the ground and leave no scent trail, making prey recovery by the snake unlikely (Deufel & Cundall, 2006). A variety of factors have been proposed as cues allowing snakes to assess prey size. These include thermal, chemical, proprioceptive, and visual cues (Gans, 1966; Radcliffe et al., 1980). Unlike pitvipers, the most studied viperid clade, puff adders are not known to be able to detect the infrared radiation signature of their prey, and thus thermal cues are unlikely in this instance. Furthermore, we never saw a puff adder using chemosensory tongue-flicks just prior to striking at prey (but see Glaudas & Alexander, 2017a; for an alternative tongueflick function). Therefore, the most parsimonious explanation is that puff adders used proprioceptive and/or visual cues to assess prey size and decide on predatory tactic. Manipulation of prey-like models in a laboratory study demonstrated that


215



Table 3 Tabulation of viperid snake predatory strategy when capturing rodents (H for hold, R for release; asterisks denote the main strategy observed in the study), with reference to prey size effect (Y for yes, N for no)

Species Agkistrodon piscivorus Bitis arietans Calloselasma rhodostoma Crotalus durissus Crotalus enyo Crotalus oreganus Crotalus viridis Crotalus viridis Daboia palaestinae Gloydius blomhoffii Lachesis muta Porthidium nummifer

Acknowledgments

Strategy (H/R)

Prey size effect (Y/N)

Reference

H, R*

Y

Kardong (1982)

H, R H*, R

Y Y

This study Barr et al. (1988)

H, H, H, H, R H, H, H, H

Y Y Y Y N Y Y Y ?

Radcliffe et al. (1980) Radcliffe et al. (1980) Kardong (1986) Radcliffe et al. (1980) Hayes (1995) Allon & Kochva (1974) Barr et al. (1988) Boyer et al. (1995) Chiszar & Radcliffe (1989)

R* R R* R* R* R* R

to important discoveries about the foraging decisions made by predators in nature.

All prey size effect indicates that the probability of holding onto prey decreases with increasing prey size.

We thank K. Erlwanger, M. A. Costello, and K. Thambu and the central animal service staff at the University of Witwatersrand for assisting with surgical procedures, G. Sauthier and H. Van Der Vyver for field assistance, and the landowners of the Dinokeng Game Reserve that granted us access to their properties to track snakes, specifically the Graf, Anderson, Engelbrecht, Keith, Leroux, Pretorius families and F. Erasmus. Last but not least, Gerd and Tienie Graf at Ikhaya lamaDube Game Lodge greatly facilitated this study by providing free accommodation and good company for more than 3 years. This project was conducted under animal protocol #2012-4204. Specimens were collected under scientific research permits (CPF6-0167, CPF6-0024) issued by the Gauteng Department of Agriculture and Rural Development. This work was partly supported by a research grant from the Committee for Research and Exploration at the National Geographic Society (#9443-14) to XG and GJA, and by postdoctoral fellowships from the Claude Leon Foundation and the University of the Witwatersrand’s research office to XG.

several distantly related viper species (including Bitis caudalis, a close relative of B. arietans) relied on visual cues to distinguish between prey types (Reiserer, 2002), and we speculate that this sensory modality is partly used by puff adders to assess prey size. Finally, the presumed ultimate factor for the effect of prey size on the hold/release strategy used by snakes is typically linked to prey dangerousness. Although very plausible, it may not be the sole explanation for the observed pattern. That is, released small prey may be more likely to be lost than larger ones (Radcliffe et al., 1980): for instance, small prey may flee into refuges that snakes may not be able to access (e.g., small burrows), thereby preventing prey recovery, an important selective pressure driving the evolution of snake foraging strategy (Radcliffe et al., 1980). This idea could easily be tested in the laboratory, and we encourage researchers to examine alternative, but not necessarily mutually exclusive, ultimate factors for the prey size dependency of venomous snake predatory strategy. In summary, we demonstrated that the predatory strategy of a venomous snake is affected by prey size. Puff adders held onto small prey items, but released large ones, which contrary to common beliefs, is the typical pattern found in viperid snakes. This indicates that vipers often assess the costs and benefits of using one strategy over the other. Hence, ambush-hunting viperid snakes are ideal organisms with which to examine context-dependent predatory decisions. The technologies available, such as the combination of radiotelemetry and fixed videography, now allow recording of predatory events in the field (Cundall, 2002). Combining these techniques with an experimental procedure manipulating various aspects of prey characteristics could ultimately lead

References

216


Allon, N. & Kochva, E. (1974). The quantities of venom injected into prey of different size by Vipera palaestinae in a single bite. J. Exp. Biol. 188, 71–75. Barr, A.D., Wieburg, S.A. & Kardong, K.V. (1988). The predatory strike behavior of the mamushi, Agkistrodon blomhoffi blomhoffi, and the Malay pit viper, Calloselasma rhodostoma. Jpn J. Herpetol. 12, 135–138. Boyer, D.M., Garrett, C.M., Murphy, J.B., Smith, H.M. & Chiszar, D. (1995). In the Footsteps of Charles C. Carpenter: facultative strike-induced chemosensory searching and trailfollowing behavior of bushmasters (Lachesis muta) at Dallas zoo. Herpetol. Monogr. 9, 161–168. Chiszar, D. & Radcliffe, C.W. (1989). The predatory strike of the jumping viper (Porthidium nummifer). Copeia 1989, 1037–1039. Chiszar, D., Andren, C., Nilson, G., O’Connell, B., Mestas, J.S., Smith, H.M. & Radcliffe, C.W. (1982). Strike-induced chemosensory searching in Old World vipers and New World pit vipers. Anim. Learn. Behav. 10, 121–125. Chiszar, D., Radcliffe, C.W. & Feiler, F. (1986). Trailing behavior in banded rock rattlesnakes (Crotalus lepidus klauberi) and prairie rattlesnakes (Crotalus viridis viridis). J. Comp. Psychol. 100, 368–371. Clark, R.W. (2006). Fixed videography to study predation behavior of an ambush foraging snake, Crotalus horridus. Copeia 2006, 181–187. Cundall, D. (2002). Envenomation strategies, head form and feeding ecology in vipers. In Biology of the vipers: 149–161. Schuett, G., H€oggren, M., Douglas, M.E. & Greene, H.W.



(Eds). Eagle Mountain: Eagle Mountain Publishing. [Correction added on 22 April 2017, after first online publication: This reference was wrongly named and this has been corrected in this current version.] Cundall, D. (2009). Viper fangs: functional limitations of extreme teeth. Physiol. Biochem. Zool. 82, 63–79. Deufel, A. & Cundall, D. (2006). Functional plasticity of the venom delivery system in snakes with a focus on the poststrike prey release behavior. Zool. Anz. 245, 249–267. Forbes, L.S. (1989). Prey defences and predator handling behaviour: the dangerous prey hypothesis. Oikos 55, 155– 158. Gans, C. (1961). The feeding mechanism of snakes and its possible evolution. Am. Zool. 1, 217–227. Gans, C. (1966). The biting behavior of solenoglyph snakes? Its bearing on the pattern of envenomation. In Proceedings of the International Symposium on Venomous Animals: xii–xiii. Sao Paulo: Instituto Butantan Publications. Glaudas, X. & Alexander, G.J. (2017a). A lure at both ends: aggressive visual mimicry signals and prey-specific luring behaviour in an ambush-foraging snake. Behav. Ecol. Sociobiol. 71, 2. Glaudas, X. & Alexander, G.J. (2017b). Food supplementation affects the foraging ecology of a low-energy, ambush-foraging snake. Behav. Ecol. Sociobiol. 71, 5. Greene, H.W. (1983). Dietary correlates of the origin and radiation of snakes. Am. Zool. 23, 431–441. Greene, H.W. (1997). Snakes: evolution of mystery in nature. Berkeley: University of California Press. Hayes, W.K. (1992). Prey-handling and envenomation strategies of prairie rattlesnakes (Crotalus v. viridis) feeding on mice and sparrows. J. Herpetol. 26, 496–499. Hayes, W.K. (1995). Venom metering by juvenile prairie rattlesnakes, Crotalus v. viridis: effects of prey size and experience. Anim. Behav. 50, 33–40. Hayes, W.K. (2008). The snake venom-metering controversy: levels of analysis, assumptions, and evidence. In Biology of rattlesnakes: 191–220. Hayes, W.K., Beaman, K.R., Cardwell, M.D. & Bush, S.P. (Eds). Loma Linda: Loma Linda University Press. [Correction added on 22 April 2017, after first online publication: The year in this reference was wrongly stated and this, together with its in-text citation, has been corrected in this current version.] IBM Corp. (2015). IBM SPSS Statistics for Windows, Version 23.0. www.ibm.com/analytics/us/en/technology/spss/ Kardong, K.V. (1982). Comparative study of changes in prey capture behavior of the cottonmouth (Agkistrodon piscivorus) and Egyptian cobra (Naja haje). Copeia 1982, 337–343. Kardong, K.V. (1986). Predatory strike behavior of the rattlesnake, Crotalus viridis oreganus. J. Comp. Psychol. 100, 304–314. Mehta, R.S. & Burghardt, G.M. (2009). Contextual flexibility: reassessing the effects of prey size and status on prey restraint behaviour on macrostomate snakes. Ethology 114, 133–145.

Mori, A. (1994). Prey-handling behaviour of newly hatched snakes in two species of the genus Elaphe with comparison to adult behaviour. Ethology 97, 198–214. Mukherjee, S. & Heithaus, M.R. (2013). Dangerous prey and daring predators: a review. Biol. Rev. 88, 550–563. O’Connell, B., Chiszar, D. & Smith, H.M. (1981). Poststrike behavior in cottonmouths (Agkistrodon piscivorus) feeding on fish and mice. Bull. Phil. Herpetol. Soc. 29, 3–7. Pan, W. (2001). Akaike’s information criterion in generalized estimating equations. Biometrics 57, 120–125. Putman, B.J., Barbour, M.A. & Clark, R.W. (2016). The foraging behavior of free-ranging rattlesnakes (Crotalus oreganus) in California ground squirrel (Otospermophilus beecheyi) colonies. Herpetologica 72, 55–63. de Queiroz, A. (1984). Effects of prey type on the prey-handling behavior of the bullsnake, Pituophis melanoleucus. J. Herpetol. 18, 333–336. Radcliffe, C.W., Chiszar, D. and O’connell, B. 1980.Effects of prey size on poststrike behavior in rattlesnakes (Crotalus durissus, C. enyo, and C. viridis). Bull. Psychon. Soc. 16: 449–450. Reinert, H.K. (1992). Radiotelemetric field studies of pitvipers: data acquisition and analysis. In Biology of the pitvipers: 185– 198. Campbell, J.A. and Brodie. E.D. Jr (Eds.). Tyler: Selva Press. Reiserer, R.S. (2002). Stimulus control of caudal luring and other feeding responses: a program for research on visual perception in vipers. In Biology of the vipers: 361–383. Schuett, G.W., H€oggren, M., Douglas, M.E. & Greene, H.W. (Eds). Eagle Mountain: Eagle Mountain Publishing. Rodrıguez-Robles, J.A. & Leal, M. (1993). Effects of prey type on the feeding behavior of Alsophis portoricensis (Serpentes: Colubridae). J. Herpetol. 27, 163–168. Shulze, R.E. (1997). South African atlas of agrohydrology and climatology. Pretoria: Water Research Commission. StatSoft Inc. (2014). STATISTICA (data analysis software system), version 12.5. www.statsoft.com Wall, M. & Shine, R. (2007). Dangerous food: lacking venom and constriction, how do snake-like lizards (Lialis burtonis, Pygopodidae) subdue their lizard prey? Biol. J. Linnean Soc. 91, 719–727. Warwick, C. (1990). Reptilian ethology in captivity: observations of some problems and an evaluation of their aetiology. Appl. Anim. Behav. Sci. 26, 1–13. Young, B.A. (2010). How a heavy-bodied snake strikes quickly: high-power axial musculature in the puff adder (Bitis arietans). J. Exp. Zool. A Ecol. Genet. Physiol. 313, 114–121.


217

Supporting Information Additional Supporting Information may be found in the online version of this article: Data S1. List of prey items captured by videotaped puff adders (Bitis arietans) in the field. Thirteen of these prey were reported in an appendix in Glaudas & Alexander (2017b).



Video S2. A puff adder striking and holding onto an arboreal lizard (Acanthocercus atricollis). Video S3. A puff adder striking and holding onto a red toad (Schismaderma carens). Video S4. A puff adder striking and holding onto a soricid shrew. Video S5. A puff adder striking and holding onto a red toad (Schismaderma carens; Glaudas & Alexander, 2017a). Video S6. A puff adder striking and holding onto an arrowmarked babbler (Turdoides jardineii). Video S7. A puff adder and striking holding onto a Sclerophrys toad (Glaudas & Alexander, 2017a). Video S8. A puff adder successfully striking at a bufonid toad. The toad seemingly escaped the snake’s jaws, but was conservatively scored as a release. Video S9. A puff adder striking and holding onto a bufonid toad. Video S10. A puff adder striking and holding onto a red toad (Schismaderma carens; Glaudas & Alexander, 2017a). Video S11. A puff adder striking and releasing an unknown rodent. Video S12. A puff adder striking and releasing a black rat (Rattus rattus ‘sensu lato’). Video S13. A puff adder striking and releasing a Pronolagus rabbit. Video S14. A puff adder striking and holding onto a bufonid toad (Glaudas & Alexander, 2017a).

Video S15. A puff adder striking and holding onto a willow warbler (Phylloscopus trochilus). Video S16. A puff adder striking and holding onto a southern gray-headed sparrow (Passer diffusus). Video S17. A puff adder striking and releasing a Lepus hare. Video S18. A puff adder striking and releasing a black rat (Rattus rattus ‘sensu lato’). Video S19. A puff adder striking and releasing a yellowthroated plated lizard (Gerrhosaurus flavigularis). Video S20. A puff adder seemingly holding onto a small mammal. The video suggests that the snake was bitten as a result, which caused the snake to release prey. Video S21. A puff adder striking and releasing a rodent (cf. Mastomys sp.). Video S22. A puff adder striking and holding onto an unknown amphibian. Video S23. A puff adder striking and holding onto an African common toad (Sclerophrys gutturalis). Video S24. A puff adder striking and holding onto a rodent (cf. Mastomys sp.). Video S25. A puff adder striking and holding onto a rodent (cf. Mastomys sp.). Video S26. A puff adder striking and holding onto a bushveld rain frog (Breviceps adspersus).

218
