Like most herpetologists of my generation, I spent much of my boyhood, especially my teenage years, looking for snakes and frogs around local ponds and learning about their natural history. Graduate school extended my boyhood for another few years; earning a faculty position made it permanent. I have been fortunate to have spent the last 40+ years at the University of Victoria indulging in my major passion in life—finding and observing amphibians and reptiles in the wild—and combining that pursuit with my enthusiasm for science, asking questions that I would never have thought of in my youth (and being paid to do so).

Although I have made occasional research forays into “lizards,” turtles, anurans, and salamanders, snakes have always been my main obsession and my bread-and-butter study organisms. With my students, I have studied various aspects of the ecology of snakes and used assorted types of modern technology such as radiotelemetry (Charland and Gregory, 1995) and spectrophotometry (Isaac and Gregory, 2013), not to mention elaborate statistical analyses that would have been unimaginable in my pre-computer early days. All of this is richly rewarding, but what I find truly satisfying is time spent in the field and the little events that collectively open up windows into the lives (and deaths) of snakes. I particularly enjoy recapturing snakes that I have marked previously and seeing how much they have grown and wondering what they have been doing since I last saw them (sometimes 10 or more years). I can estimate survivorship from these mark–recapture data, but only rarely do I learn the ultimate fate of individual snakes that I mark and release. Nonetheless, every once in awhile, the hard evidence turns up in my hands. For example, I once found a marked juvenile Thamnophis sirtalis in the stomach of an adult Thamnophis elegans—no big mystery about what happened there.

This little observation underlines the fact that snakes, like most other animals, play multiple roles in the ecosystems they occupy. At one level, snakes are predators, eating a wide variety of prey. We know this partly because we sometimes see snakes catching or eating prey in the wild but mainly because we can sample stomach contents of snakes, either by dissection (Rodríguez-Robles, 2002) or by gently encouraging live snakes to regurgitate their prey (Tuttle and Gregory, 2009). As a result, we actually know quite a bit about the diets, if not the foraging habits, of many species of snakes.

Snakes are not only predators themselves but also prey for other species of predators. In this case, though, we usually know much less about the particular players involved (but see Shine et al., 2001; Ajtić et al., 2013). Predators are occasionally seen attacking or attempting to attack snakes, but in my experience, such observations are generally few and far between (and my camera never seems to be handy, either). We know that snakes are attacked because unsuccessful attacks often leave evidence in the form of injuries (Gregory and Isaac, 2005; Fig. 1), and also we sometimes see the remains of successful attacks (Fig. 2), but we usually do not know who the guilty party is. This is an important gap in our knowledge as even the mere threat of predation can have potentially significant population-level consequences for prey.

Fig. 1. 

Examples of injuries in Common Watersnakes (Nerodia sipedon). (A) Injury to body. (B) Loss of end of tail.

Fig. 2. 

A recently killed Natrix natrix found in the field.

Predation on snakes, and as a consequence, antipredation defenses of snakes, is an enigmatic subject that is ripe for further study. My experiences in the field have left me with many questions. Why do some snakes bite vigorously when captured, whereas others of the same species react more placidly? Why do snakes crossing open spaces, such as roads, often pause and rest immobile and exposed? Why do some snakes flee when approached, whereas others allow close approach and even contact? Why do some snakes play dead when handled? I cannot hope to answer these kinds of questions definitively here, but perhaps I can provide some sort of framework for tackling them.

Throughout my career, I have mostly studied natricine snakes, especially gartersnakes (Thamnophis). Why? Partly because they were the first snakes I encountered as a youth and I have a soft spot for them but mainly because they are diverse, widespread, and abundant and usually easily maintained in captivity, making them attractive for all sorts of studies, including comparative work within a restricted phylogenetic framework. Although some natricines bite readily when caught, their bites are usually innocuous; hence, snakes can be caught by hand without the need for tongs, tubes, and other paraphernalia that are needed for handling seriously venomous snakes. True, some natricines are mildly venomous, and people sometimes react to their bites (Hayes and Hayes, 1985; Gomez et al., 1994), but I am not aware of any fatalities or long-term effects from such bites. In fact, Rhabdophis is the only natricine genus that has any dangerously venomous species (Greene, 1988, 1997), but when molested, these snakes generally resort to other antipredator measures before biting (Mori and Burghardt, 2001).

Therefore, natricines represent an interesting class of prey. They are all small to medium-sized, even as adults, and thereby susceptible to capture by all manner of predators throughout their lives. Their physical defensive weaponry is weak to nonexistent in most cases (although biting might have an element of surprise). Nonetheless, they are highly successful and, as my mark–recapture records show, often live long lives. How do they do it? How do they avoid being eaten?

To a large extent, my interest in antipredator defenses has been fuelled by my experience with Grass Snakes (Natrix natrix), which often feign death when captured. This is curious behavior. I well remember the first grass snake I ever caught. As the snake hung limply from my hand, mouth agape, I was initially convinced that I somehow had accidentally killed it. It was a thoroughly persuasive display, but playing dead really seems like an odd thing to do when one's life is potentially at stake. Several colleagues, especially behavioral ecologists, have expressed their skepticism that this could be effective antipredator behavior. I have argued elsewhere why and under what circumstances it might be effective (Gregory, 2008a), but this remains an open question. On the one hand, questions about death-feigning underscore the fact that we know little or nothing about how (or even whether) most presumed antipredator behaviors of natricines are used against natural predators and their adaptive value; this is a general, if difficult, problem that demands attention (Lind and Cresswell, 2005). On the other hand, though, it makes me wonder whether there are any actions of natricines that do not have elements of predator avoidance or other antipredator behavior. From the perspective of a small animal, it must indeed be a scary world.

My first aim in this review is to summarize what we know about defensive armaments and behavior in natricine snakes. Because this is a “Perspectives” paper, my review is selective (perhaps idiosyncratic), rather than exhaustive but relies on both field and laboratory studies, with occasional reference, as appropriate, to studies of other animal taxa. Much of what I present about natricines also applies to other groups of snakes, but restricting my review to a single diverse clade is a practical constraint. Second, I discuss some current issues about predatory threat and defensive behavior that have broader ecological implications and merit further study. Collectively, they comprise a potential agenda for future research that we might profitably address using natricines as “model organisms.” At my advancing age, it is unlikely that I will tackle even a small fraction of this agenda, but I hope to inspire others to do so.

Because my focus here is on snakes as prey, further use of words like “predation” or “predators“ herein refers to predation on snakes rather than by snakes.

Antipredator Behaviors of Natricine Snakes

The wide variety of antipredator behaviors exhibited by snakes, including natricines, has been cataloged elsewhere (Greene, 1988), and it is not my intention to repeat this list in detail here but to cover antipredator defenses under broad categories. There are various ways that one could categorize such behaviors. Perhaps the most comprehensive is by Mori and Burghardt (2004), who categorized antipredator behaviors in three dimensions: 1) whether prey move toward or away from predators; 2) how much movement is involved; and 3) apparent function. More simply, Arnold and Bennett (1984) placed antipredator behaviors of neonate Plains Gartersnakes (Thamnophis radix) on a spectrum from defensive to offensive. Another convenient approach, which I adopt here, is to classify antipredator behaviors according to when they occur during a predator–prey interaction; different tactics may be called for at different points in an encounter with a predator. I pool predator avoidance and antipredator mechanisms (Brodie et al. 1991), but there is potential utility in making the distinction between them in other contexts.

Any interaction between predator and prey consists of sequential stages. The interaction can be terminated at any stage, by the prey escaping or being consumed, or by the predator giving up (Lima and Dill, 1990; Toledo et al., 2011; Sherbrooke, 2013). Presumably, it is in the prey's interest to end the encounter, in its favor, as soon as possible (Bowers et al., 1993; but see Bateman et al., 2014). Generally speaking, the stages of predation include: localization (detection), approach, attack, handling (including subjugation), and consumption. Approach and attack may be combined as one single action but can also be separate, as in the case of a predator approaching a prey but failing to attack because of aggressive defense posturing by the prey (Arnold and Bennett, 1984; Greene, 1997) or because of “flash” or “startle” coloration that is exhibited when the prey's body is flattened defensively (Shine et al., 2000; Westphal, 2007). This sequence can be simplified even further as a continuum from avoiding detection through avoiding capture to avoiding consumption.

By “avoiding detection,” I refer to detection of the prey by the predator, but the detection of the predator by the prey is equally important. Avoiding detection by the predator essentially means hiding (e.g., under cover; Gregory, 2009) or otherwise not stimulating the predator's senses (e.g., immobility coupled with visual crypsis; Isaac and Gregory, 2013). Hiding under cover may be more important for smaller snakes (Fig. 3) because they are presumably more vulnerable to predators than are large snakes (Bittner, 2003); small juvenile snakes also are slower than adults (Pough, 1977). Gartersnakes (Thamnophis) hiding under cover during the day are smaller than those basking or moving in the open (Gregory et al., 1983; Gregory, 1984), as are N. natrix (Fig. 4).

Fig. 3. 

An adult female Storeria occiptiomaculata. Although individuals of this small species are sometimes found in the open, they are typically found hiding under cover such as rocks.

Fig. 4. 

Boxplots of body sizes (snout–vent length, SVL) of Grass Snakes (Natrix natrix) found under cover and in the open at Fordwich, Kent, U.K., from 1999–2012 (unpubl. data; N = 302, of which 36 were under cover). Median is shown by horizontal line. Top and bottom of box are 75th and 25th quartiles, respectively; ends of vertical line are the furthest points not exceeding 1.5 times the distance between the quartiles; and dots are more extreme values. Lack of overlap of notches of boxes indicates significant difference between medians at P = 0.05.

If a predator detects a prey, but not vice versa, the prey may be doomed unless it detects the predator as the latter approaches. However, prey sometimes detect predators before the predator detects them. Once a prey animal has detected a predator, one option is to flee, but this entails the risk of drawing the predator's attention. If unsure about whether it has been detected by the predator, the prey's other option is to remain immobile (and perhaps thereby remain undetected), especially if cryptically colored, until the predator comes within some critical distance, usually called approach distance or flight initiation distance (FID), at which point the prey should flee.

Flight initiation distance has been well studied in diverse taxa (Stankowich and Blumstein, 2005), including lizards (Cooper, 2011; Cooper et al., 2012; Martín et al., 2009), but has been studied in only a few natricines (Layne and Ford, 1984; Weatherhead and Robertson, 1992; Burger, 2001; Shine et al., 2000, 2003a; Brown and Shine, 2004; Cooper et al., 2008). The point at which FID occurs is determined by the relative costs of fleeing versus remaining in place (Lima and Dill, 1990). More recent models emphasize the “decision” to remain still or flee as one that maximizes fitness (Cooper and Frederick, 2007, 2010). Flight initiation distance is influenced by multiple factors (Stankowich and Blumstein, 2005) and might even be reduced to zero in cases in which the prey has effective post-capture defensive behaviors (e.g., death-feigning; Cooper and Frederick, 2010) or the prey stands its ground with intimidating (even if bluffing) aggression toward the predator. Also, proximity to cover should allow closer approach of a predator (Stankowich and Blumstein, 2005); many natricines favor edge habitats when in the open (Wisler et al., 2008), presumably because they allow ready escape while basking or engaging in other activities. Finally, walking past snakes without stopping may be less likely to cause flight than approaching them directly (Burger, 2001; pers. obs.). One puzzle is why some snakes do not flee at all as they are approached and fail to move even when the human predator is lunging at them (Gregory, 2013).

A prey animal cannot be vigilant in all ways and in all directions all the time and I've caught my fair share of snakes that probably simply did not see me because their heads were hidden. But even snakes that are aware that a predator is approaching or close by might sometimes take the chance that the predator has not detected them and continue to try to evade detection by remaining immobile. I think I have a pretty good search image for snakes, but it's not perfect, and every once in awhile I put my foot down right beside a snake without seeing it. I am usually alerted to this if somebody else notices the snake or if the snake decides to move away soon after I've nearly stepped on it, but if the snake does not move, I easily might not see it at all and simply move on. In any case, the snake faces a dilemma—whether to move or to remain frozen in place. Either action could be the correct one, but either also could be an error. A decision has to be made especially quickly if the predator surprises the snake (and perhaps itself) by coming upon the snake suddenly and at close quarters, an experience I have had as a snake hunter many times.

This issue is similar to one I raised earlier: Why do snakes crossing open spaces, such as roads, often pause and rest immobile and exposed? Are they simply being wary? Snakes that have to cross open spaces have two main options: 1) cross the space and reach safety as quickly as possible but risk drawing the attention of predators through conspicuous movement; or 2) cross slowly with barely detectable movement but risk longer exposure. I have seen snakes do both of these, but what factors determine the direction of this apparent trade-off in any given case is an open question. I suspect that temperature plays a role—warmer snakes are faster than cooler snakes; furthermore, lingering on a dangerously hot substrate could be fatal.

However, temperature, which has a profound effect on the general ecology of snakes (Peterson et al., 1993), shows no consistent influence on FID—colder snakes flee earlier (Shine et al., 2000) or later (Layne and Ford, 1984) than warmer snakes or there is no difference in FID between warm and cold snakes (Weatherhead and Robertson, 1992; Brown and Shine, 2004; Cooper et al., 2008).

Trade-offs between putative antipredator behaviors should be common. Individual hatchling N. natrix that move quickly rely on flight to counter a perceived threat, whereas slower snakes rely more on aggression; locomotory speed, in turn, varies among clutches and is influenced by developmental temperature (Hagman et al., 2015). In DeKay's Brownsnakes (Storeria dekayi), the duration of death-feigning behavior is negatively related to maximum swimming speed, again indicating a trade-off between these two traits (Gerald, 2008).

Color and pattern play a role in avoiding detection through background matching and crypsis (King and Lawson, 1997; Isaac and Gregory, 2013), but also they are linked to flight behavior. For example, striped snakes should use flight readily because uniformity of body landmarks makes it difficult to follow the snake's movement (Jackson et al., 1976; Allen et al., 2013; but see von Helversen et al., 2013). In contrast, blotched snakes are cryptic when immobile and should be more likely to defend themselves than flee once detected by a predator (Jackson et al., 1976). Brodie (1992) tested the escape response of naïve newborn Northwestern Gartersnakes (Thamnophis ordinoides) from litters containing both striped and blotched individuals and found that the two morphs indeed had different responses. Laboratory work, coupled with mark–recapture studies in the field, showed that fitness was maximized in young snakes for two combinations of traits: snakes with stripes that fled in one direction when chased versus snakes without stripes that tended to reverse their movement and stop when chased instead of fleeing steadily in a single direction (Brodie, 1992). The sudden disappearance of the snake (and its blotches) when the snake stops is likely to confuse a predator and make it difficult to locate the prey. Color pattern and behavior are genetically correlated in this species (Brodie, 1989b). This correlational selection helps explain the color polymorphism. That is, neither pattern, by itself, has an advantage over the other—the advantage emerges only through correlation with behavior.

Also, sex and reproductive state should affect how snakes react to predatory situations. Male snakes that are preoccupied with searching for mates allow much closer approach than snakes not so engaged (Shine et al. 2003a; Brown and Shine, 2004). Because of their increased mass, gravid females are typically limited in their mobility (Seigel et al., 1987), which affects their antipredator options. Female T. ordinoides change their antipredator behavior during pregnancy, exhibiting a significantly higher frequency of reversals in movement direction than when not gravid and, thus, a higher reliance on sudden crypsis than on flight (Brodie, 1989a). By changing their behavior while pregnant, gravid females reduce their dependency on locomotion at a time when they are disadvantaged in this respect, thereby also reducing the cost of reproduction. Also, presumably because of their reduced locomotory ability, gravid Keelback Snakes (Tropidonophis mairii) flee at greater FID than do nongravid snakes (Brown and Shine, 2004). In contrast, when confronted with an immediately adjacent threat in the confines of the laboratory, gravid T. sirtalis are less likely to flee and more likely to be aggressive than males or nongravid females (Maillet et al., 2015).

Other characteristics of pregnant snakes also might be seen as part of their antipredator repertoire. For example, pregnant Thamnophis are typically anorexic, especially late in pregnancy (Gregory et al., 1999). I interpret this as a physiological suppression of appetite at a time when foraging might conflict with the more important function of thermoregulation. But not foraging also means that snakes can bask near shelter to which they can readily escape if a predator turns up; snakes that cannot move fast should ensure that they do not have to move far.

Gartersnakes often aggregate in the open when pregnant, sometimes touching, sometimes not (Gregory et al., 1987). Aggregation might be social behavior or merely common attraction to particular physical conditions, but it also might be antipredator behavior. Although an aggregation of prey might be somewhat more conspicuous to a predator than an isolated individual, there might be safety in numbers (the “selfish herd”; Hamilton, 1971) if the predator can focus on and catch only one or two prey while the others make good their escape, especially if they all move in different directions. A few years ago, I came across four pregnant T. elegans, all piled in a heap on a patch of open ground. The snakes all started to flee before I could touch them. I managed to catch them all, but then I am a specialized predator with grasping hands and a cloth bag in which to store the snakes while I process them one by one at my leisure. Other predators do not have this suite of attributes and probably would not have been able to catch, hold, and consume all four snakes very easily.

Snakes whose locomotion is affected by a full stomach (Garland and Arnold, 1983) might be expected to respond to threats in similar ways to gravid females. A large meal obviously represents a large payload of energy for future use, but also it can be costly. On at least two occasions, I have encountered snakes (one T. sirtalis and one N. natrix) that had eaten such large meals that they could hardly move. Their bodies were so distended in the stomach region that they mainly rocked back and forth when they tried to flee—easy pickings for me and no doubt for other predators.

Snakes that do not flee as the predator approaches may be confronted with exploratory examination by the predator before attack or a direct attack. Many studies mimic this situation of close approach and/or initiation of attack by approaching the snake with a finger, either held stationary or moving (Herzog and Burghardt, 1986), or touching the snake directly on some part of the body (Mori and Burghardt, 2000; Langkilde et al., 2004). In other cases, a model of an actual predator may be used to simulate approach and attack of a predator (Shine et al., 2000). Finally, attack may be simulated simply by catching the snake by hand and handling it (Gregory, 2008a). Collectively, these approaches have yielded a rich diversity of responses of snakes to approach and capture by simulated predators.

Studies in which snakes are exposed to escalating levels of threat (e.g., from approach to actual contact) show that defensive behaviors change as the approach of a predator becomes more threatening (Schieffelin and de Queiroz, 1991; Bowers et al., 1993). Although the sequence of behaviors employed as threat increases varies among species, the general pattern is to shift from more aggressive defense (e.g., striking) toward either more passive displays (e.g., head hiding, Fig. 5) or flight (Schieffelin and de Queiroz, 1991; Bowers et al. 1993). Aggressive behaviors are presumably attempts to intimidate the predator, buying time to escape, especially if directed at vulnerable parts of the body (e.g., eyes; Herzog and Bern, 1992), whereas behaviors such as head hiding (typically under or within the coiled body) protect the more vulnerable parts of the body from attack. Similarly, tail-waving, which also is a common behavior in such situations (Bowers et al., 1993), may serve to direct the predator's attention toward a less vulnerable part of the body (Langkilde et al., 2004). Tail injuries, including loss of parts of the tail, are common in natricines (Gregory and Isaac, 2005; Placyk and Burghardt, 2005, references therein). Given this variation in vulnerability of different parts of the body, where the body is touched by the “predator” may influence what behavior is shown in response (Arnold and Bennett, 1984; Langkilde et al., 2004). Finally, even when caught and handled, natricine snakes may continue to exhibit antipredator behaviors, such as struggling and twisting, biting, or expelling the contents of their cloacal glands, as well as feces and urine (“musking” behavior; cloacal smearing); the latter presumably makes the snake malodorous and unattractive to predators (Gangloff et al., 2014; Gray, 2015). Some behaviors, such as death-feigning, may be exhibited only when the snake is handled (Gregory et al., 2007; Gregory, 2008a). Studies of Tiger Keelbacks (Rhabdophis tigrinus), for example, often fail to reveal death-feigning behavior (Mori and Burghardt, 2001) unless significant handling is involved (Mutoh, 1983). Death-feigning is much easier to induce in N. natrix, but degree of handling and where on the body the snake is held still influence occurrence of this behavior (Gregory, 2008a; Fig. 6). Among other natricines, at least four species of Thamnophis (Gregory and Gregory, 2006; Schield and Waye, 2012; Enge, 2015), two species of Storeria (Jordan, 1970; Gerald, 2008), and one species of Regina (Oldham et al., 2015) may exhibit elements of death-feigning behavior when handled.

Fig. 5. 

Head-hiding behavior in a Thamnophis radix.

Fig. 6. 

Logistic regressions of probability of death-feigning vs. SVL for Grass Snakes (Natrix natrix) subjected to one of two treatments. (A) “Fully handled” snakes that were passed back and forth in the hands for 30 sec immediately upon capture and then subjected to routine handling and measurement plus being grasped at back of head for measurement of head size. (B) “Less-handled” snakes subjected to routine handling and measurement only. Solid line is predicted logistic regression, and dashed lines are 95% confidence limits around that prediction; vertical lines at top and bottom are data (snakes that feigned death and those that did not, respectively). Regression statistics: (A) χ²₁ = 0.72, P = 0.40, N = 97; (B) χ²₁ = 3.69, P = 0.055, N = 43. Data from Gregory (2013), but figure not shown there. For further details, see that reference.

Once an animal is caught by a predator, its options for escape are much more limited, but until the prey is dead, all is not necessarily lost. Escape from consumption might be possible via some of the behaviors discussed above. For example, struggling while held by the predator may cause the predator to lose its grip (L. A. Isaac, pers. comm.); many natricines also spin their bodies when held by the tail, apparently attempting to break the tail and escape (Cooper and Alfieri, 1993). Biting the predator while in its clutches also might deter the predator or perhaps simply irritate it further (Fig. 7). Some species, such as Watersnakes (Nerodia) are renowned for reacting to capture by biting vigorously. Not only do Nerodia bites sting, but they often bleed profusely. Of course, the hardened herpetologist ignores all this and hangs onto the snake, but biting might be an effective deterrent against other predators (including humans) in this case.

Fig. 7. 

A newly captured Thamnophis elegans vigorously bites the author's wrist. Photograph by G. Gregory.

Prey with chemical defenses have other options for deterring predators. Rhadophis tigrinus have dorsal nuchal glands, shared by only a few other species, just behind the head. Snakes sequester steroid toxins in these glands, obtained from Japanese Common Toads (Bufo japonicus) that they eat, and present the glands to oncoming predators; snakes from sites where toads are not present do not exhibit this behavior (Mori and Burghardt, 2000; Hutchinson et al., 2007; Hutchinson et al., 2013). Such behavior is clearly associated with nuchal glands because it is absent in close relatives that eat toads but lack nuchal glands (Takeuchi and Mori, 2012). Young R. tigrinus also obtain toxins as maternal provisions during embryonic development (Hutchinson et al., 2008); this provision presumably is enhanced by selective foraging of gravid females in habitats where they are likely to encounter toads (Kojima and Mori, 2014). Other natricines that eat poisonous prey also might use ingested toxins in their own defense (Williams et al., 2012), especially if paired with aposematic coloration (Williams et al., 2004).

Death-feigning, discussed above, is an example of a last-ditch attempt, perhaps a desperate one, to avoid death. Predators that relax their grip when a prey is apparently dead may momentarily look away, allowing the prey to escape. Alternatively, predators that cache their prey may mistakenly cache a death-feigning prey, which could escape when left alone. Clearly, however, this strategy will not work against a predator that kills and consumes its prey immediately. Unfortunately, we do not know how often death-feigning snakes use this behavior against real predators or whether they use it only against certain predators (see quotation at beginning of this article). This is true of virtually all the antipredator behaviors that we have described for snakes—we simply do not know whether they are used against real predators and in what context. This leads to the first of several overlapping areas in which we need to increase our knowledge about how snakes interact with their predators.

Some Areas for Further Research

Conclusion

A colleague once described my approach to research as “analytic natural history.” Whether I am sufficiently analytic or not is a question I leave for others to answer, but the term “natural history” is certainly one I embrace, including (perhaps inadvisably) in grant applications. By natural history, I mean the study of organisms in their natural habitats. In my view, natural-history research is not necessarily purely observational, as some would argue, but can readily accommodate tests of hypotheses. Controlled experiments in both field and laboratory obviously play a central role in ecology and behavior, but studies of the unrestrained, free-ranging animal bring another kind of truth, a different window into the animal's daily life. My bias toward this kind of field study underlies much of what I have written in this review.

Studying behavior of snakes can be tricky, especially in the field. Snakes are built low to the ground and generally cryptic, even when active, which makes them difficult to observe directly. Also, they can sometimes be difficult to watch simply because the observer might fall asleep waiting for something to happen. I once found the same individual gartersnake, preparing to shed its skin, under the same cover object for several days in a row. I have occasionally seen snakes feeding in the field but not often, and, in most of my study sites, I probably could spend another career waiting for another observation. However, encounters with predators, or predator-like stimuli, including humans, always elicit measurable behavioral responses from snakes, even in the form of immobility. These responses and their correlates can be measured in the field.

Reptiles often are excellent “model organisms” for studies of various general phenomena in biology (Luttershmidt, 2013). Like any other organism, reptiles have their own unique combinations of traits and cannot be perfect models for all other organisms, but broad principles should apply. Natricine snakes are smallish animals that presumably encounter risks similar to those faced by other vertebrates in the same size range. Although details of their antipredator tactics differ from those of other animals, they are faced with similar trade-offs in their behavioral “decisions.” Thus, studies of antipredator behavior of natricines should bear extrapolation not only to other snakes but, with a wider view, to other animals.

In this review, I have sketched out the broad patterns of defensive behavior in natricine snakes and suggested areas in which we might profitably extend our knowledge. My review has necessarily been selective, and I have ignored important areas such as phylogenetic distribution of traits and what that might tell us about evolutionary patterns and adaptive convergences. Similarly, although size and age must have profound effects on antipredator strategies, I have only hinted here and there at the complexities of ontogenetic change. Finally, body temperature, especially in interaction with other factors, is sure to play an important role in the expression of antipredator defenses, but again I have only touched upon it; it deserves a separate treatment.

Natricine snakes exhibit a diversity of antipredator behaviors that they deploy at various points in an encounter with a predator. However, we know little about use of these behaviors against predators other than humans nor about their adaptive value. Furthermore, antipredator behavior presumably has links to other phenomena such as acquisition of injury and stress, which in turn have potential population-level consequences. We need to integrate the study of antipredator behavior in these animals more closely with population and physiological ecology.