See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/330061964

Sensory basis of navigation in snakes: the relative importance of eyes and pit
organs

Article in Animal Behaviour · January 2019

DOI: 10.1016/j.anbehav.2018.11.004

CITATION READS

1 382

2 authors, including:

Rulon W Clark
San Diego State University
80 PUBLICATIONS 1,189 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Science Education View project

Ground Squirrel Antisnake Defenses View project

All content following this page was uploaded by Rulon W Clark on 27 March 2019.

The user has requested enhancement of the downloaded file.

 Animal Behaviour 147 (2019) 77e82

Contents lists available at ScienceDirect

Animal Behaviour
journal homepage: www.elsevier.com/locate/anbehav

Sensory basis of navigation in snakes: the relative importance of eyes

and pit organs
Hannes A. Schraft a, b, *, Rulon W. Clark a
a
Department of Biology, San Diego State University, San Diego, CA, U.S.A.
b
Graduate Group in Ecology, University of California, Davis, CA, U.S.A.

a r t i c l e i n f o
Animal movements govern most ecological interactions, from predation to reproduction and survival.
Article history: How animals move through the environment depends on available sensory information. Some snakes are
Received 23 August 2018 able to perceive infrared (IR) radiation in addition to visible light. Research on this sensory system has
Initial acceptance 1 October 2018 been almost exclusively focused on predation, and researchers have largely found that vision and IR
Final acceptance 24 October 2018 compensate for each other when one or the other is absent. However, IR sensing likely has much broader
functions, including navigation in the environment. Many features in the environment of pit vipers are
MS. number: A18-00599 both visually and thermally salient and could be used for orientation. Here, we tested how vision and IR
sensing interact in sidewinder rattlesnakes, Crotalus cerastes, in a simple navigation task in the field.
Keywords: Unlike in a predatory context, IR sensing did not compensate for the lack of vision. Snake movement
infrared sensing
paths were more tortuous, and snakes were less likely to encounter landmarks when eyes where
navigation
occluded but were unaffected when pit organs were occluded. These findings suggest that the interaction
perceptual range
thermography
between visual and IR cues may depend on context, and have important implications for our under-
vision standing of the evolution of IR sensing.
© 2018 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.

Nearly all animals move in a nonrandom way (Gould, 2004; duplex retinas (rods and cones; Davies et al., 2009), and the rod-
Åkesson, Bostro € m, Liedvogel, & Muheim, 2014). Directed move- packing density of boas and pythons is similar to that of cats
ments allow animals to find prey, mates and thermoregulation sites (280 000e460 000 rods/mm2; Sillman, Johnson, & Loew, 2001).
or avoid predators. How animals move through the environment Visual resolution, based on photoreceptor densities, is also similar
and where they go is determined by the available sensory infor- to that of cats (Sillman et al., 1999, 2001). Note, however, that other
mation (Lima & Zollner, 1996). The sensory channels involved in factors such as ganglion density and absolute eye size also influence
orientation depend on the species: many vertebrate species, resolution, and we still only have an approximate idea of snake
including humans, rely primarily on vision (Gould, 1998). However, visual acuity. No data are available on rattlesnake vision, but there
natural selection has equipped some species with sensory systems is reason to believe that they have sensitive night vision: pit vipers
that detect stimuli that are invisible to most other organisms. have elliptical pupils that must be contracted during daylight to
Certain lineages of snakes, including rattlesnakes (Viperidae: Cro- protect sensitive retinas (Goris, 2011). Rattlesnakes likely use vision
talinae), use both visible and infrared (IR) light to sense the envi- to orient; for example, western diamondback rattlesnakes, Crotalus
ronment (Goris, 2011). Here, we tested how these sensory channels atrox, use a solar compass to direct movement during the day
interact during navigation at night, when visual cues are reduced (Landreth, 1973). Presumably, visual cues play an important role
but IR contrast is high. during nocturnal movement as well.
Vision is likely a principal sensory channel for snakes, and most In addition to wavelengths of light typically detected by verte-
snakes have large, well-developed eyes (Ford & Burghardt, 1993). brate photoreceptive pigments (hereafter visible light), rattle-
The few existing studies on snake vision have found that nocturnal snakes can perceive thermal radiation (Goris, 2011). In rattlesnakes,
snakes can probably see relatively well. For example, snakes have facial pit organs house heat-sensitive membranes that detect IR
radiation of approximately 5e30 mm wavelength (Bakken &
Krochmal, 2007; de Cock Buning, 1983; Moiseenkova, Bell,
* Correspondence: H.A. Schraft, Department of Biology, San Diego State Univer-
Motamedi, Wozniak, & Christensen, 2003). IR sensing has
sity, San Diego, CA 92182, U.S.A. received most attention in a predatory context (e.g. Ebert &
E-mail address: [email protected] (H. A. Schraft).

https://doi.org/10.1016/j.anbehav.2018.11.004
0003-3472/© 2018 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
78 H. A. Schraft, R. W. Clark / Animal Behaviour 147 (2019) 77e82

Westhoff, 2006; Kardong, 1992; Theodoratus, 1997). However, In this study, we experimentally tested the hypothesis that IR and
several authors have pointed out that thermal vision is likely to be visible light detection interact in a simple navigation task involving
useful in other contexts, including general navigation to find suit- targets that are salient both visually and thermally. We quantified the
able microhabitats for thermoregulation (Goris, 2011; Greene, movement of rattlesnakes with different sensory occlusions (no oc-
1992; Krochmal & Bakken, 2003). This may even have been the clusions, only eyes occluded, only pits occluded, both eyes and pits
ancestral function of the pit organ, since environmental features occluded) placed in a seminatural arena with several bushes. If
provide strong thermal signals (Bakken & Krochmal, 2007; snakes find warm bushes using IR cues, then pit occlusions should
Krochmal, Bakken, & LaDuc, 2004; Fig. 1). lead to snakes being less likely to locate bushes and have more
The interaction of IR and visible light detection in rattlesnakes tortuous movement paths. Eye occlusions should lead to an analogue
has been an area of active research (e.g. Chen, Deng, Brauth, Ding, & result if snakes rely on visual cues even at night. This simple test
Tang, 2012; Hartline, Kass, & Loop, 1978; Kardong & Berkhoudt, allowed us to examine the possibility that IR sensing could be
1999). Laboratory-based behavioural studies generally find that extended to navigation using thermally contrasting objects at night.
eyes and pit organs can compensate for each other during preda-
tion, but compensation is not perfect (e.g. Chen et al., 2012; METHODS
Kardong & Berkhoudt, 1999). For example, rattlesnakes are able
to capture prey successfully even in complete darkness (Cadena, Study Site and Animals
Andrade, Bovo, & Tattersall, 2013). No such compensatory studies
have been performed for other ecological contexts or outside of a We performed 129 experimental trials on sidewinder rattle-
laboratory environment. There is reason to believe that both IR and snakes collected from the wild on the Barry M. Goldwater range
visible light sensing could be used for navigation in wild snakes. (managed by U.S. Marine Corps) near Yuma, Arizona, U.S.A. be-
Nocturnal rattlesnakes (such as sidewinders, Crotalus cerastes) tween 28 May and 9 August 2017. Habitat consists of small (~1 m)
move extensively throughout the night (Clark, Dorr, Whitford, sand dunes interspersed with bushes (Malusa & Sundt, 2015;
Freymiller, & Putman, 2016; Secor, 1994) and appear to move Fig. 1b). A system of larger dunes (up to 10 m high) runs
nonrandomly from bush to bush (H. A. Schraft, personal observa- northesouth adjacent to the study site. We measured all snakes
tion). Interestingly, bushes should stand out not only visually from (N ¼ 77), counted rattle segments to estimate age (Reiserer, 2001),
the ground-level view of a small animal like a rattlesnake, but also used nail polish to mark individuals with unique dorsal colour
through IR; bushes are typically 5e6
C warmer than the sur- combinations and inserted a radiofrequency identification (RFID)
rounding substrate at night (Fig. 1c), a temperature difference that tag (dorsally, subcutaneously; 8 mm passive integrated tran-
snakes should be able to detect easily (Goris, 2011). sponder (PIT) tags, Biomark Inc., Boise, ID, U.S.A.). We tested snakes

(a) (b)

°C
20.0

(c)

(d)
Temperature (°C)

18 (e)
12
6
–7.5
0° 90° 180° 270° 360°

Figure 1. (a) Schematic diagram of the arena, (b) photograph of the arena facing approximately northeast, (c) infrared panorama of the arena taken from the centre of the arena at
ground level, (d) same panorama as it may appear to a snake (see text) and (e) graph of mean temperature around the arena. The warm vertical bars at approximately 60
, 180
and
300
are the legs of the inverted tripod used to hold the camera. Image taken with a FLIR T-420 thermal imaging camera (FLIR Systems, Wilsonville, OR, U.S.A.; emissivity ¼ 0.95,
reflected temperature ¼ 20
C).
 H. A. Schraft, R. W. Clark / Animal Behaviour 147 (2019) 77e82 79

of all sizes (mean ± SD snoutevent length: 35.6 ± 5.6 cm), but our measure the following: total track length, straight-line distance to
sample was skewed towards smaller snakes because they were end of track and compass bearing to where the snake first touched
more abundant at our field site. the arena wall or where the track ended (measured from centre of
arena; Supplementary Fig. S1). We calculated each track's
Sensory Occlusions straightness index by dividing total track length by the straight-line
distance and arcsine-transformed the values (arcsine of the square
We performed trials on snakes with four different temporary root of the value; Maritz & Alexander, 2016). We used optical and
sensory occlusions: none (control), only eyes, only pit organs, and heat transfer analysis implemented in MaTLab R2017a (The Math-
both eyes and pit organs. All occlusion procedures took place under Works Inc., Natick, MA, U.S.A.) to simulate what the thermal fea-
isoflurane-induced anaesthesia. We blindfolded snakes with a tures of the arena might look like to a snake, depicted in Fig. 1d
small (~5  8 mm) strip of black tape (Ebert & Westhoff, 2006). To (elliptical spread function, major diameter ¼ 15
, minor diame-
block the pit organ, we placed a small sphere of beeswax directly in ter ¼ 7.5
). For details see Bakken and Krochmal (2007).
the pit with tweezers and secured it in place with a small drop of
Vetbond tissue adhesive (3M, Maplewood, MN, U.S.A.; see also Data Analysis
Chen et al., 2012; Haverly & Kardong, 1996; Krochmal & Bakken,
2003). Complete lack of response to a warm, moving object We used generalized linear mixed effects models to evaluate
assured us of the efficacy of sensory occlusions. We allowed snakes both straightness index and whether snakes reached a bush
to recover from anaesthesia and adapt to the sensory occlusions for (yes ¼ 1; no ¼ 0) as a function of occlusion status, light level (log-
at least 24 h before testing. After trials, we removed sensory oc- transformed because of logarithmic brightness perception; Bruce,
clusions under anaesthesia and released snakes at the point of Green, & Georgeson, 2003), and the interaction between occlu-
collection. Snakes in the treatment with no sensory occlusions sion status and light level. Light level could interact with occlusion
received sham procedures, including anaesthesia. status since the effect of blindfolding is presumably smaller during
dark, moonless nights than during bright nights. The model for
Arena whether a snake reached a bush was based on a binomial distri-
bution. All models included a random effect for snake identity since
We constructed a 10 m diameter circular arena at a flat, sandy some snakes were used in multiple trials. For both dependent
location on our field site (Fig. 1, Supplementary Fig. S1) using variables, we fitted four models with different sets of predictors
wooden stakes and black tarp. The arena included four bladdersage (intercept only, occlusion status as the only main effect; occlusion
bushes, Salazaria mexicana. During our night-time trials, bushes status þ light level; occlusion status þ light level þ occlusion sta-
were typically 5e6
C warmer than surrounding sand (Fig. 1c). The tus  light level) and used Akaike's information criterion corrected
arena barrier was the same temperature as the surrounding sand, for small samples (AICc) to identify the most informative model.
as verified by images taken with a thermal imaging camera (Fig. 1c). We then used Tukey tests to determine the contrasts between
This arena design did not allow us to manipulate the location or different occlusion statuses. We excluded trials from all analyses
temperature of bushes directly, but it did permit trials in a setting where snakes did not move at all (N ¼ 6 out of 129). In those trials,
closely mimicking the natural environment while still allowing for occlusion statuses were eyes and pits (N ¼ 3), eyes only (N ¼ 2) and
experimental control. pits only (N ¼ 1).
Several variables other than the location of bushes could influ-
Trials ence snake movement. First, the direction of snake movement
could be influenced by the compass bearing from which the
We placed snakes at the centre of the arena in an inverted 7.5- observer released the snake, and the compass bearing of a snake's
litre bucket. After 2 min, the observer released the snake by lift- collection location (if snakes orient towards ‘home’). To explore
ing the bucket with a 3.6 m extension pole. The observer then these possibilities, we fitted circular regression models, which
quietly left the arena and remained out of visual and thermal sight predict the sine and cosine of the response variable by fitting two
behind a sand dune. We randomly determined the compass bearing trigonometric polynomials given a predictor variable (Batschelet,
from which the observer released the snake and then hid during 1981; Pewsey, Neuha €user, & Ruxton, 2013). We are not aware of
each trial. After 10 min, the observer collected the snake and any published approach for modelling circular data with random
mapped the tracks left by the snake's movement until first contact effects, so we randomly deleted all but one repeated trial from the
with the arena wall or the end of the tracks. We raked the entire same individual for these circular regression models, resulting in a
arena before every trial to obliterate previous tracks and scatter reduced sample size of 74 trials. Second, not all bushes were the
potential scent trails. During all trials, we recorded air temperature same distance from the arena centre. If snakes have depth
(
C) and light level (lx). Sidewinders are primarily nocturnal during perception (vision and IR), closer bushes may be more attractive
the summer months (Clark et al., 2016), so all trials took place than distant ones, even for the same subtended angle. Snakes may
between 2000 hours and 0130 hours. Final sample sizes for each thus simply move to the nearest bush (Maritz & Alexander, 2016;
treatment group were: no occlusions, 31 trials; only pit organs Supplementary Table S1). Visual inspection of the data allowed us
occluded, 31 trials; only eyes occluded, 34 trials; both eyes and pits to dismiss this possibility (Supplementary Fig. S2). Third, seasonal
occluded, 33 trials. Some individuals were tested multiple times large-scale movement patterns (i.e. migration; Secor, 1994) or
within and between treatment groups. Analyses showed that there movements along some environmental gradient could lead to
was no significant order effect; we accounted for repeated mea- overall directedness of movement. We tested this with a Rao's
sures on the same individual by including a random intercept for spacing test for angular uniformity (Batschelet, 1981; Pewsey et al.,
snake identity in the models (see below). 2013).
All data manipulation and analyses were performed in R v.3.2.3
Data Extraction (R Core Team, 2015); we used the packages ‘dplyr’ (Wickham,
Francois, Henry, & Müller, 2017; data manipulation), ‘lme4’
We used ImageJ v.1.48i (National Institutes of Health, Bethesda, (Bates, Ma €chler, Bolker, & Walker, 2015; mixed model fitting),
MD, U.S.A., http://rsbweb.nih.gov/ij/) on mapped snake tracks to ‘MuMIn’ (Barton  2016; model comparison), ‘multcomp’ (Hothorn,
80 H. A. Schraft, R. W. Clark / Animal Behaviour 147 (2019) 77e82

Bretz, & Westfall, 2008; Tukey tests), ‘effects’ (Fox, 2003; extracting Table 2
effect sizes) and ‘circular’ (Agostinelli & Lund, 2017; circular sta- Tukey pairwise comparisons between different occlusion statuses for the best-
fitting model of straightness index
tistics). Original data and R script are included in the
Supplementary Material. Contrast Estimate SE Z P

None e Pits 0.054 0.068 0.794 0.857

Ethical Note None e Eyes 0.294 0.067 4.404 <0.001
None e Both 0.422 0.068 6.227 <0.001
Pits e Eyes 0.240 0.067 3.569 0.002
All procedures adhered to the ASAB/ABS Guidelines for the use Pits e Both 0.368 0.068 5.384 <0.001
of animals in research. The San Diego State University Institutional Eyes e Both 0.128 0.067 1.904 0.226
Animal Care and Use Committee approved all methods (APF 16-08-
014C). We minimized stress to snakes by releasing them within
72 h of initial capture. Experiments were unlikely to have affected
1
snakes negatively, as all snakes displayed natural behaviour during a
a
trials and upon release. We recaptured snakes tested previously in
52 cases, implying that trials did not affect survival.

Straightness index (fited)

0.8
b
RESULTS
0.6 b
Snakes moved from the point of release in all but six trials,
confirming that no combination of sensory occlusions dissuaded
0.4
snakes completely from undertaking movements.
Straightness index was best predicted by a model with occlusion
status as the only predictor (Table 1). We therefore focus on the 0.2
model without light level (range <0.01 lxe0.15 lx) and the inter-
action between occlusion status and light level. Movement was
significantly straighter with no occlusions at all or only pit organs 0
occluded, compared to only eyes occluded or both pit organs and None Pits Eyes Both
eyes occluded (Table 2, Fig. 2).
Sensory occlusion status
Whether a snake reached one of the bushes was best predicted by
a model including occlusion status as the only predictor (Table 1). We Figure 2. Fitted estimates of straightness index for each sensory occlusion status ac-
therefore again focus on the model without light level and the cording to the top model. Higher values indicate straighter movements. Shaded areas
interaction between occlusion status and light level. The model represent 95% confidence intervals.

including the interaction term produced a near-singular fit

(q < 0.001), likely because of marginal power to estimate random Table 3
intercepts. We nevertheless retained the model for model selection Tukey pairwise comparisons between different occlusion statuses for the best fitting
because (1) a model excluding repeated observations from the same model of whether snakes reached one of the bushes in the arena
individual produced qualitatively identical results and (2) near- Contrast Estimate SE Z P
singular fits increase the probability of false positives, not false neg-
None e Pits 0.182 0.668 0.273 0.993
atives. Tukey pairwise comparisons on the top model showed that
None e Eyes 2.216 0.593 3.734 0.001
snakes were more likely to reach a bush with no sensory occlusions or None e Both 2.274 0.604 3.763 0.001
only pits occluded (Table 3). Snakes reached bushes at lower levels Pits e Eyes 2.398 0.621 3.862 <0.001
with only eyes occluded, or both eyes and pits occluded (Fig. 3). Pits e Both 2.457 0.631 3.891 <0.001
Neither observer bearing (circular regression: r ¼ 0.47, N ¼ 74, Eyes e Both 0.059 0.551 0.107 0.999

P ¼ 0.85 and 0.75) nor origin bearing (r ¼ 0.46, N ¼ 74, P ¼ 0.96 and
0.32) predicted the bearing of a snake's location at the end of a trial.
Snakes also did not simply go to the nearest bush. In fact, snakes DISCUSSION
visited the nearest bush least often (3.3 m from arena centre; 13
trials), and the farthest bush most often (4.3 m; 25 trials). This Rattlesnakes relied on visual cues for nocturnal navigation.
slight preference for one of the bushes manifested itself in signifi- Indeed, we found no evidence for the hypothesis that vision and IR
cant overall directedness of movement (Rao's spacing test: detection compensate for each other in a simple navigation task.
U ¼ 175.4, N ¼ 74, P < 0.05). With eyes occluded, movement paths were more tortuous, and

Table 1
Model comparison results for models of straightness index and whether a bush was reached as a function of various predictors

Predictors df Log likelihood AICc DAICc Weight

Straightness index
Occlusion status 6 17.68 48.1 0.00 0.943
Occlusion status þ light level 7 19.36 53.7 5.61 0.057
Occlusion status þ light level þ occlusion status  light level 10 21.41 64.8 16.70 0.000
Intercept only 3 33.10 72.4 24.32 0.000
Reached bush
Occlusion status 5 66.95 144.4 0.00 0.732
Occlusion status þ light level 6 66.95 146.6 2.21 0.243
Occlusion status þ light level þ occlusion status  light level 9 65.78 151.2 6.75 0.025
Intercept only 2 84.34 172.8 28.37 0.000

All models include a random intercept for snake identity.

 H. A. Schraft, R. W. Clark / Animal Behaviour 147 (2019) 77e82 81

1 2017; Grace, Woodward, Church, & Calisch, 2001). Sensory

Probability of reaching bush (fitted)

a a
compensation may be imperfect because the resolution of IR
detection is likely crude compared to vision. Optical and heat
0.8 transfer analyses show that the image on the pit membrane is
poorly focused and has a low temperature contrast (Bakken &
0.6 Krochmal, 2007), although neural amplification and sharpening
b b could alleviate this (Stanford & Hartline, 1980; Westhoff, Morsch, &
Ebert, 2006). Fig. 1d, generated with Bakken and Krochmal's (2007)
0.4
optical and heat transfer routine, shows that the thermal features of
the arena should be detectable by a snake using IR radiation alone.
0.2 However, IR cues appear to be overridden by visual cues in a nav-
igation context.
0 Our results suggest that the usefulness of IR detection in pit
vipers could be limited to certain ecological contexts by the details
None Pits Eyes Both
of receptor physiology (see also Bakken & Krochmal, 2007). The pit
Sensory occlusion status organ membrane receptor response is phasic (Bullock & Cowles,
1952; Goris & Terashima, 1973), and so IR detection may be most
Figure 3. Estimated probability of reaching one of the bushes for each sensory oc-
useful to detect small moving objects, such as endothermic prey
clusion status according to the top model. Shaded areas represent 95% confidence
intervals. animals (Ebert & Westhoff, 2006). Given our results, it seems that
snakes are not able to make stationary objects equally conspicuous
through scanning head movements (Bakken & Krochmal, 2007;
snakes had a lower probability of reaching a bush. On the other Goris & Terashima, 1973). Previous authors suggested that the pit
hand, occluded pit organs did not affect snake behaviour. These organ likely evolved to find suitable microhabitats, for example for
results were robust even though we observed overall directedness thermoregulation (Krochmal et al., 2004). We did not find (indirect)
to snake movement directions. This suggests that sidewinder rat- support for this hypothesis here, although further carefully
tlesnakes do not use IR detection for basic orientation in an unfa- designed field studies are needed to explore additional proposed
miliar area, even when thermally contrasting features are present. functions (e.g. predator avoidance, burrow searching for prey) of
The compensation between these two sensory channels reported in infrared sensing in snakes.
previous studies on predatory behaviour may not extend to navi-
gation behaviour. Instead, the degree to which snakes use visual Declaration of Interest
and IR cues may depend on context or spatial scale.
Our results suggest that visual cues play a primary role in We have no conflicts of interest to declare.
directing movement even at night. Although the visual sensitivity
of vipers has not received any experimental attention, our study Data Availability
shows that nocturnal sidewinder rattlesnakes can see well enough
to detect large environmental features such as bushes and use them Original data and R script for analysis and figures are included as
for basic orientation. Five of the six snakes that never moved at all Supplementary Material.
during trials had either only eyes or both eyes and pits occluded,
further underlining the apparent importance of visual cues for
Acknowledgments
movement. The overall directedness of snake movement could be
due to large-scale seasonal movements such as migration to or
We thank Abigail Rosenberg and the Marine Corps Air Station
from overwintering sites (Secor, 1994), but this seems unlikely
Yuma for providing access to the Barry M. Goldwater Range. Cha-
considering that we performed trials in midsummer. Alternatively,
nont Alvord, Elizabeth Signore and Tom Ganter collected data. We
snakes may have a preference to travel along the northesouth
are grateful to George Bakken for access to and help with the optical
running dune system located next to the study site. The natural
and heat transfer simulation. Members of the Clark lab, Shannon
history and migratory patterns of this particular population of
Whelan, George Bakken, Gail Patricelli, John Phillips and an anon-
sidewinder rattlesnakes is poorly known.
ymous referee offered valuable comments on earlier versions of
Contrary to our results, previous studies found that IR-sensing
this manuscript. This study was funded by San Diego State Uni-
snakes use IR cues in the absence of visual cues (e.g. Chen, Liu,
versity, the San Diego Chapter of the Explorer's Club, the Chiricahua
Brauth, Fang, & Tang, 2017; Ebert & Westhoff, 2006; Kardong &
Desert Museum, the Society for Integrative and Comparative
Berkhoudt, 1999; Van Dyke & Grace, 2010). For example, snakes
Biology and the Herpetologists' League.
can track moving thermal targets even when blindfolded (Ebert &
Westhoff, 2006; Van Dyke & Grace, 2010), and the absence of vi-
Supplementary Material
sual cues does not impair successful prey capture (Cadena et al.,
2013; Chen et al., 2012, 2017). Arboreal pit vipers (Gloydius she-
Supplementary Material associated with this article is available, in
doaensis) select ambush sites based on both visual and thermal
the online version, at https://doi.org/10.1016/j.anbehav.2018.11.004.
backgrounds (Shine & Li-Xin, 2002). There is substantial overlap
between the visual and IR fields of view (Hartline et al., 1978), and
eye size is inversely correlated with pit organ size (Liu, Chen, References
Papenfuss, Lu, & Tang, 2016), indicating that the two systems
Agostinelli, C., & Lund, U. (2017). Circular: Circular statistics (R package version 0.4-
perform a similar function. However, compensation between eyes 93). https://r-forge.r-project.org/projects/circular/.
and pit organs is imperfect even in a predatory context. Kardong Åkesson, S., Bostro€m, J., Liedvogel, M., & Muheim, R. (2014). Animal navigation. In
and Berkhoudt (1999) found that both eye occlusions and pit or- L. A. Hansson, & S. Åkesson (Eds.), Animal movement across scales (pp. 151e178).
Oxford, U.K.: Oxford University Press.
gan occlusions make snakes more reluctant to strike, and blind- Bakken, G. S., & Krochmal, A. R. (2007). The imaging properties and sensitivity of the
folding reduced strike accuracy in several studies (Chen et al., 2012, facial pits of pitvipers as determined by optical and heat-transfer analysis.
82 H. A. Schraft, R. W. Clark / Animal Behaviour 147 (2019) 77e82

Journal of Experimental Biology, 210, 2801e2810. http://doi.org/10.1242/jeb. Kardong, K. V., & Berkhoudt, H. (1999). Rattlesnake hunting behavior: Correlations
006965. between plasticity of predatory performance and neuroanatomy. Brain,
Barton, K. (2016). MuMIn: Multi-model inference (R package version 1.15.6). https:// Behavior and Evolution, 53, 20e28. http://doi.org/10.1159/000006579.
CRAN.R-project.org/package¼MuMIn. Krochmal, A. R., & Bakken, G. S. (2003). Thermoregulation is the pits: Use of thermal
Bates, D., M€ achler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects radiation for retreat site selection by rattlesnakes. Journal of Experimental
models using lme4. Journal of Statistical Software, 67, 1e48. Biology, 206, 2539e2545. http://doi.org/10.1242/jeb.00471.
Batschelet, E. (1981). Circular statistics in biology. New York, NY: Academic Press. Krochmal, A. R., Bakken, G. S., & LaDuc, T. J. (2004). Heat in evolution's kitchen:
Bruce, V., Green, P. R., & Georgeson, M. A. (2003). Visual perception: Physiology, Evolutionary perspectives on the functions and origin of the facial pit of pit-
psychology, & ecology. New York, NY: Psychology Press. vipers (Viperidae: Crotalinae). Journal of Experimental Biology, 207, 4231e4238.
Bullock, T. H., & Cowles, R. B. (1952). Physiology of an infrared receptor: The facial http://doi.org/10.1242/jeb.01278.
pit of pit vipers. Science, 115, 541e543. Landreth, H. F. (1973). Orientation and behavior of the rattlesnake, Crotalus atrox.
Cadena, V., Andrade, D. V., Bovo, R. P., & Tattersall, G. J. (2013). Evaporative respi- Copeia, 1973 (pp. 26e31).
ratory cooling augments pit organ thermal detection in rattlesnakes. Journal of Lima, S. L., & Zollner, P. A. (1996). Towards a behavioral ecology of ecological
Comparative Physiology A, 199, 1093e1104. http://doi.org/10.1007/s00359-013- landscapes. Trends in Ecology & Evolution, 11, 131e135.
0852-4. Liu, Y., Chen, Q., Papenfuss, T. J., Lu, F., & Tang, Y. (2016). Eye and pit size are
Chen, Q., Deng, H., Brauth, S. E., Ding, L., & Tang, Y. (2012). Reduced performance of inversely correlated in Crotalinae: Implications for selection pressure relaxa-
prey targeting in pit vipers with contralaterally occluded infrared and visual tion. Journal of Morphology, 277, 107e117. http://doi.org/10.1002/jmor.20483.
senses. PLoS One, 7. e34989 http://doi.org/10.1371/journal.pone.0034989. Malusa, J., & Sundt, P. (2015). Vegetation of the Barry M. Goldwater Range West,
Chen, Q., Liu, Y., Brauth, S. E., Fang, G., & Tang, Y. (2017). The thermal background Marine corps air station e Yuma, Arizona. https://doi.org/10.13140/
determines how the infrared and visual systems interact in pit vipers. Journal of RG.2.1.2058.7681 [Report submitted to USMC under Cooperative Agreement
Experimental Biology, 220, 3103e3109. http://doi.org/10.1242/jeb.155382. DACA87-05-H-0018].
Clark, R. W., Dorr, S. W., Whitford, M. D., Freymiller, G. A., & Putman, B. J. (2016). Maritz, B., & Alexander, G. J. (2016). Scale-dependent orientation in movement
Activity cycles and foraging behaviors of free-ranging sidewinder rattlesnakes paths: A case study of an African viper. Ethology, 122, 207e214. http://doi.org/
(Crotalus cerastes): The ontogeny of hunting in a precocial vertebrate. Zoology, 10.1111/eth.12459.
119, 196e206. http://doi.org/10.1016/j.zool.2016.02.005. Moiseenkova, V., Bell, B., Motamedi, M., Wozniak, E., & Christensen, B. (2003).
de Cock Buning, T. (1983). Thermal sensitivity as a specialization for prey capture Wide-band spectral tuning of heat receptors in the pit organ of the copperhead
and feeding in snakes. American Zoologist, 23, 363e375. snake (Crotalinae). American Journal of Physiology - Regulatory, Integrative and
Davies, W. L., Cowing, J. A., Bowmaker, J. K., Carvalho, L. S., Gower, D. J., & Hunt, D. M. Comparative Physiology, 284, R598eR606.
́
(2009). Shedding light on serpent sight: The visual pigments of henophidian Pewsey, A., Neuh€ auser, M., & Ruxton, G. D. (2013). Circular statistics in R. Oxford,
snakes. Journal of Neuroscience, 29, 7519e7525. http://doi.org/10.1523/ U.K.: Oxford University Press.
jneurosci.0517-09.2009. R Core Team. (2015). R: A language and environment for statistical computing. Vienna,
Ebert, J., & Westhoff, G. (2006). Behavioural examination of the infrared sensitivity Austria: R Foundation for Statistical Computing. http://www.R-project.org/.
of rattlesnakes (Crotalus atrox). Journal of Comparative Physiology A, 192, Reiserer, R. S. (2001). Evolution of life histories in rattlesnakes (Dissertation). Berke-
941e947. http://doi.org/10.1007/s00359-006-0131-8. ley, CA: University of California.
Ford, N., & Burghardt, G. (1993). Perceptual mechanisms and the behavioral ecology Secor, S. M. (1994). Ecological significance of movements and activity range for the
of snakes. In R. A. Seigel, & J. T. Collins (Eds.), Snakes: Ecology and behavior (pp. sidewinder, Crotalus cerastes. Copeia, 1994, 631e645.
117e164). New York, NY: McGraw-Hill. Shine, R., & Li-Xin, S. (2002). Arboreal ambush site selection by pit-vipers Gloydius
Fox, J. (2003). Effect displays in R for generalised linear models. Journal of Statistical shedaoensis. Animal Behaviour, 63, 565e576. http://doi.org/10.1006/anbe.2001.
Software, 8, 1e27. 1928.
Goris, R. C. (2011). Infrared organs of snakes: An integral part of vision. Journal of Sillman, A., Carver, J., & Loew, E. (1999). The photoreceptors and visual pigments in
Herpetology, 45, 2e14. http://doi.org/10.1670/10-238.1. the retina of a boid snake, the ball python (Python regius). Journal of Experi-
Goris, R. C., & Terashima, S. (1973). Central response to infra-red stimulation of the mental Biology, 202, 1931e1938.
pit receptors in a crotaline snake, Trimeresurus flavoviridis. Journal of Experi- Sillman, A. J., Johnson, J. L., & Loew, E. R. (2001). Retinal photoreceptors and visual
mental Biology, 58, 59e76. pigments in Boa constrictor imperator. Journal of Experimental Zoology, 290,
Gould, J. L. (1998). Sensory bases of navigation. Current Biology, 8, 731e738. http:// 359e365. http://doi.org/10.1002/jez.1076.
doi.org/10.1016/S0960-9822(98)70461-0. Stanford, L., & Hartline, P. (1980). Spatial sharpening by second-order trigeminal
Gould, J. L. (2004). Animal navigation. Current Biology, 14, 221e224. http://doi.org/ neurons in crotaline infrared system. Brain Research, 185, 115e123.
10.1093/acprof. Theodoratus, D. H., Chiszar, D., & Smith, H. M. (1997). Rattlesnake orientation to
Grace, M. S., Woodward, O. M., Church, D. R., & Calisch, G. (2001). Prey targeting by prey as function of thermal backgrounds and edges. Psychological Record, 47(3),
the infrared-imaging snake Python: Effects of experimental and congenital 461e472.
visual deprivation. Behavioural Brain Research, 119, 23e31. Van Dyke, J. U., & Grace, M. S. (2010). The role of thermal contrast in infrared-based
Greene, H. W. (1992). The ecological and behavioral context for pit viper evolution. defensive targeting by the copperhead, Agkistrodon contortrix. Animal Behaviour,
In J. A. Campbell, & E. D. Brodie (Eds.), Biology of the pit vipers (pp. 107e117). 79, 993e999. http://doi.org/10.1016/j.anbehav.2010.01.012.
Tyler, TX: Selva. Westhoff, G., Morsch, M., & Ebert, J. (2006). Infrared detection in the rattlesnake
Hartline, P. H., Kass, L., & Loop, M. S. (1978). Merging of modalities in the optic Crotalus atrox: From behavioural studies to midbrain recordings. In M. Vences,
tectum: Infrared and visual integration in rattlesnakes. Science, 199, 1225e1229. € hler, T. Ziegler, & W. B. Bonn (Eds.), Herpetologia Bonnensis II Proceedings of
J. Ko
http://doi.org/10.1126/science.628839. the 13th Congress of the Societas Europaea Herpetologica, 27 September e 2
Haverly, J. E., & Kardong, K. V. (1996). Sensory deprivation effects on the predatory October 2005 (pp. 225e228). Bonn, Germany http://www.seh-herpetology.org/
behavior of the rattlesnake, Crotalus viridis oreganus. Copeia, 1996, 419e428. publications/proceedings.
http://doi.org/10.2307/1446858. Wickham, H., Francois, R., Henry, L., & Müller, K. (2017). dplyr: A grammar of data
Hothorn, T., Bretz, F., & Westfall, P. (2008). Simultaneous inference in general manipulation (R package version 0.7.4). https://CRAN.R-project.org/
parametric models. Biometrical Journal, 50, 346e363. package¼dplyr.
Kardong, K. V. (1992). Proximate factors affecting guidance of the rattlesnake strike.
Zoologische Jahrbuecher Anatomie, 122, 233e244.

View publication stats