Flash behavior increases prey survival

Flash behavior increases prey survival Abstract Flash behavior, in which otherwise cryptic prey exhibit conspicuous coloration or noise when fleeing from potential predators, has been postulated to hinder location of prey once they become stationary. Here, using artificial computer-generated prey and humans as visual predators, we show that human subjects are more likely to abandon their search for prey that flash, compared to continuously cryptic fleeing controls. Survivorship of flashing prey was an additional 20% higher than the survivorship of continuously cryptic prey, depending on the background against which it was depicted. This survivorship advantage was consistent regardless of whether prey showed flash colors continuously or intermittently during flight. The advantage over continuously cryptic prey was highest when the flashing prey was presented first. Likewise, the more search areas containing no prey that the volunteers had initially viewed, the more likely they were to give up when there was a cryptic prey present. Collectively, these 3 findings indicate that volunteers inferred the flashing prey was absent from the search area when they failed to see a prey in the same form as they saw it move. Our results demonstrate first proof of concept: flash behavior, widely seen in taxa from insects to mammals, is an effective antipredator escape mechanism. INTRODUCTION Flash behavior describes the way in which an otherwise cryptic prey suddenly reveals a conspicuous patch of color or emits noise during movement after disturbance and then hides the patch or ceases to make a sound on stopping (Cott 1940; Edmunds 1974; Edmunds 2008). Examples of patches of color that are only displayed when fleeing include the bright hindwings (underwings) of noctuid and sphingid moths, and the proximal portions of the hindwings of cicadas (Hemiptera), grasshoppers (Orthoptera), and stick insects (Phasmatodea) (Cott 1940; Edmunds 1974). Such insects are cryptic at rest, but conspicuous in flight. Before flight and shortly after coming to rest following escape, these individuals fold their brightly colored hindwings beneath their cryptic forewings. Flash behavior exhibited during movement is also seen in many vertebrate taxa including frogs (Williams et al. 2000), some even-toed ungulates (Artiodactyla) (Caro et al. 2004) and rabbits and hares (Leporidae) (Stoner et al. 2003). At present, the adaptive significance of this putative antipredator defense is unknown. Cott (1940) noted that it seems “to confuse or misdirect an enemy in the pursuit of prey” (p. 376), stressing that it is the “sudden disappearance of color combined with the equally sudden suspension of movement which tends to mislead the eye and to render the animal’s exact whereabouts on alighting all the more difficult to detect”. Edmunds (1974) similarly suggested that the predator “may be caused to hesitate by the sudden movement and appearance of the bright color ... and it may follow this color and be deceived by its sudden disappearance into assuming the prey has vanished whereas in reality the prey has come to rest in its normal cryptic posture with the colored structures hidden.” As such, flash colors could serve to initially startle a would-be predator while the prey is on the move and/or hinder the predator’s search if it follows the color rather than the organism (Edmunds 2008). These hypotheses are entirely plausible, but, as yet, there is no proof that flash behavior has evolved for these reasons. Here, we conduct experiments to test the second component of the above explanations, namely that flash behavior hinders subsequent search for potential prey. Before describing the experiments, we first need to clarify precisely what we mean by flash behavior because it has been used rather broadly to represent almost any conspicuous display of color. Naturally, we focus here on the role of flash color as an anti-predator defense, rather than for mate attraction (see for example Lloyd 1971; Schultz and Fincke 2009; Ballantyne and Lambkin 2013). In that context, flash coloration is a term that has been used to describe startle (“deimatic”) displays produced by stationary organisms in response to the approach or contact by a predator, thereby slowing or halting attack (Umbers and Mappes 2015; Kang et al. 2016; Umbers et al. 2017). While the sudden appearance of something previously hidden could serve to intimidate or distract predators, whether the prey is stationary or mobile (Olofsson et al. 2012; Kang et al. 2017), the second hypothesized protective mechanism associated with flash behavior is different: it involves hindering subsequent search for previously mobile prey. By contrast, deimatic displays do not involve a search-hindering affect. Flash coloration is occasionally used to refer to conspicuous signals that are easily seen during flight, such as the black flank stripe of the Thomson’s gazelle Eudorcas thomsoniii, thought to amplify stotting, thus informing the predator that the prey is difficult to catch (Caro and Stankowich 2010). However, the signal is not hidden once the prey has stopped, so it is unlikely to have evolved to hinder search, and we do not consider this further. Flash coloration in mobile prey has 2 forms: a continuous display of a conspicuous patch during flight, or an intermittent display. Continuous flash behavior has been described in many taxa including cryptic Herennia ornatissima spiders showing their red ventrum while descending rapidly on a thread from their webs (Anonymous 1945), in Oedipoda grasshoppers showing bright red lateral surfaces as they fly away from a disturbance (Edmunds 2008), in dull colored shorebirds showing their white backs, rumps or tails when fleeing en masse (Del Brooke 1998), and in white-tailed deer Oedicolus virginianus tail-flagging (Caro et al. 1995). Intermittent flashing is seen in some lepidoptera as they beat their wings (Edmunds 2008) and in some leporids such as the black-tailed jackrabbit Lepus californicus showing its black tail during flight (Kamler and Ballard 2006). Whether continuous and intermittent flashes serve different functions is unclear, so our experiments investigated the properties of both forms of display. Flash colors could protect the prey by affecting predator psychology. As Cott (1940) and Edmunds (2008) suggested, flash colors displayed during movement may deceive visual predators about the subsequent location of the prey at rest. Additionally (or alternatively), flash colors may confuse predators about the resting colors of prey. Because fleeing prey allow predators to observe the prey momentarily, predators may formulate a transient “search image” based on flash colors (Tinbergen 1960; Langley 1996), although strictly speaking search image formation applies only to cryptic prey (Lawrence and Allen 1983). It is possible that flash colors are the most salient feature of prey during fleeing, thus making predators perceive prey to be of a completely different color than their resting color. If true, predators will be more likely to give up their search for prey that have exhibited flash behavior than those that have not, believing them to have fled the scene if they do not readily detect them. In this study, we used computer-generated artificial prey and humans as visual predators, which is a useful and convenient system to test the adaptive significance of animal coloration, and specifically to test the hypothesis that flash behavior causes predators to abandon search for cryptic prey. METHODS We recruited anonymous human volunteers and evaluated the responses of each participant to flash behavior using a custom-built computer-based game. Our volunteers were visitors (largely undergraduates) to the University Centre at Carleton University, Ottawa, Canada and both experiments were conducted from May to July 2017. Games were presented on a 23” LCD monitor (Dell ST-2310f). Prior to testing, we showed each participant a tutorial video about the game, but details of the hypothesis were not disclosed. Each game had 2 phases, namely “prey escape” and “search.” Once each participant understood the nature of the game and was ready to play, the game was started. At the start of the prey escape phase, a single square prey item (375 × 375 twips, with the number of pixels = twips/15) of given color was placed at a random position on a complex background image (grass or sand, dimensions 8172 × 16,764 twips, Figure 1). When a participant moved the mouse pointer on the background image within 20,000 twips of the prey (hence any movement of the mouse pointer into the background image), the prey moved to escape. Escape was to either the left or right side of the screen (whichever was the furthest to move). Prey movement was simulated as a directed random walk comprising of a rapid sequence of discrete steps in randomly drawn directions taken from a given distribution. The step size of the prey throughout was 600 twips, the step angle was chosen at random from an even distribution within ± 45° from the horizontal, and the duration of wait before new step was 30 ms. Figure 1 View largeDownload slide The prey (here shown together, although note that they were presented alone and in pseudo-random sequence to volunteers) displayed against the (a) grass and (b) sandy background. Individual prey were either cyan, magenta, blue, green, yellow, cryptic or red in color (left to right). Cryptic prey were presented only in the final 2 trials after presenting the prey with fixed conspicuous colors. These cryptic prey either kept their cryptic color when they moved, or they flashed red as they moved. Cryptic prey presented against the grass background had an red, green, blue values of each pixel (RGB) of (73, 151, 19) and an RGB of (192, 156, 111) when presented against a sand background. Figure 1 View largeDownload slide The prey (here shown together, although note that they were presented alone and in pseudo-random sequence to volunteers) displayed against the (a) grass and (b) sandy background. Individual prey were either cyan, magenta, blue, green, yellow, cryptic or red in color (left to right). Cryptic prey were presented only in the final 2 trials after presenting the prey with fixed conspicuous colors. These cryptic prey either kept their cryptic color when they moved, or they flashed red as they moved. Cryptic prey presented against the grass background had an red, green, blue values of each pixel (RGB) of (73, 151, 19) and an RGB of (192, 156, 111) when presented against a sand background. These parameter settings allowed the participant to observe the retreating prey for 1–2 s, depending on its starting position. Following the “escape” of prey, volunteers were asked to click a follow button, after which the search phase began. During search phase, the volunteers were shown an inverted version of the same background screen (to prevent prey being revealed through the contrasting otherwise identical images), with the same prey present at a new random location. The volunteers were asked to find the escaped prey and click on it (generating a pleasant sound and happy face when complete). In this phase, volunteers had 2 options: moving the mouse to the prey and clicking directly on it, or pressing a “give up” button to move on to the next prey if they could not find it (inferring that no prey was present in the screen). Training and general experimental procedure We presented 8 preys sequentially to each participant. The first 6 were considered as training prey and each of these was colored differently (magenta, cyan, red, green, yellow, and blue; see Figure 1), conspicuous against the grass or sand background and presented in a random order. The grass background was of lush green grass while the sandy background showed grains of light brown sand with tiny pebbles scattered on it. The grass background image was readily discriminable from the conspicuous green colored prey used in the training sessions. Crucially, all of these prey maintained the same color while moving (escape phase) and when sedentary (search phase). For the first 6 prey types, there was a 25% probability that the prey would not be present in the search phase (i.e., the second screen). We used these blank “duds” in order to get the volunteers used to the fact that there can sometimes be no prey in a search screen after it had escaped (mimicking a rabbit disappearing down a hole, for example). When the prey was absent from the search screen, volunteers would inevitably press the “give up” button and the same color trial was entered again into the list of prey that had not yet been attacked, so that it could be presented again (in a random sequence of prey not yet searched for). See Supplementary Video S1 for a sample video. Experiment 1: Flash prey versus cryptic prey After all 6 training prey were eventually attacked, volunteers were each presented with 2 treatment prey: a cryptic prey with continuous flash when it moved (continuous flash, CF) and a cryptic prey without flash when it moved (cryptic throughout, CR). The color of cryptic prey had mean R, G, and B values of the background image (Wyszecki and Stiles 1982), making it relatively hard to detect because of the background color matching between prey and background (Merilaita and Stevens 2011). CR prey displayed the cryptic color both when sedentary and when moving. CF prey showed the same cryptic color when sedentary, but exhibited a red color while moving; red is a common flash color found in insect prey thought to exhibit flash behavior (Edmunds 2008). CR prey were presented prior to CF prey for half of the volunteers and the order was reversed for the other half. The prey was always present in the search phase. To test whether the results could be generalized across different background types, the experiment was repeated twice, once with a grass and once with a sandy background. Following the experiment, a noncomprehensive red/green color-blindness test was performed. Data from volunteers who demonstrated signs of color blindness (n = 5) were excluded from analysis. For each training and test image shown to each volunteer that was present on the search screen, we recorded 1) a binary response whether the participant attacked the prey or clicked the “give up” button and so moved on to the next image without finding the prey, and 2) the time spent on each image until either they detected the prey (detection time) or gave up searching (giving up time). There was no maximum cut-off time. A total of 120 volunteers judged to be non-color blind were tested (60 for each background). Each volunteer experienced only one type of background. Experiment 2: Intermittently flashing prey versus cryptic prey In this experiment, we tested whether intermittently displayed flash (hereafter, IF) color could increase survivorship of the prey. All conditions remained the same as in Experiment 1, except that we changed the nature of the flash behavior. Whenever a prey with flash behavior moved on the screen its color flickered (flickering rates: mean of 16.7 changes/seconds) between red and cryptic colors by randomly selecting red or cryptic at each step in the rapid random walk rather than constantly showing red color. Again, we tested 120 volunteers for this experiment (60 for each background). Data analysis In both experiments, our primary goal was to test whether cryptic prey with flash behavior had a survival advantage over cryptic prey without flash behavior. To test this, we fitted generalized (and general) linear mixed models to the data and compared 1) prey survivorship (binary response) and 2) prey searching time of volunteers hunting for cryptic and flashing prey (continuous response). The search time has 2 different endpoints: it indicates the time taken to detect the prey when the prey was detected (detection time) and it indicates the time taken until the volunteer gave up looking for the prey if the prey was not detected (giving up time). Thus, we separately analyzed detection time and giving up time. In each case, the search time response was log transformed to meet the assumptions of the fitted general linear model. For Experiment 1, we treated prey type (CR or CF), background type (grass or sand), and the interaction between these 2 main effects as explanatory variables. We also included 1) the order of presentation of CR and CF as a fixed effect to control for any effect of presentation order, and 2) the interaction between presentation order and our color treatment as predictors, since the order that prey appeared first could have influenced the volunteers’ performance in finding prey. Since “duds” appeared stochastically in the training, the number of “duds” that occurred during a trial varied among volunteers ranging from 0 to 9 duds (on average, volunteers experienced 1.9 duds while 19.1% of the tested individuals did not experience any duds). To estimate and control for any effect of duds, we additionally included the number of duds that appeared for each participant in the training sequence as a covariate. To account for possible differences among human volunteers, and ensure independence, the volunteer was treated as a random effect. The analysis of Experiment 2 proceeded in the same way, with prey type treatments now being CR and IF (intermittent flash). To compare the effectiveness of CF prey with IF prey in reducing predation between experiments, we fitted 2 separate models using 1) survivorship (binomial), 2) giving up time (continuous), and 3) detection time (continuous) as response variables respectively (only one observation for each human was used in each analysis, so a random effect was not necessary to control for independence). Prey type (CF vs. IF), background type, the order of the presentation (whether flashing prey was presented prior to or after cryptic prey in each experiment), number of duds, and the interaction between background and prey type were included as explanatory variables. In all results, we present the model with the lowest Akaike Information Criterion (AIC) among competing candidate models. The analysis of deviance table for each analysis is shown in Supplementary Tables S1 and S2. All the analyses were conducted in R (R Core Team 2017) using “lme4” package (Bates et al. 2015). RESULTS Experiment 1: Continuously flashing prey versus cryptic prey Survivorship of the CF prey was on average 19% higher than CR prey (Figure 2a, χ21 = 10.55, P = 0.001), and the overall survivorship of prey was significantly higher on a grass background than on a sandy background (Figure 2a, χ21 = 6.78, P = 0.009). We also found a significant interaction effect between background and prey type (Figure 2a, χ21 = 25.63, P < 0.001) in that the survivorship advantage of the continuously flashing prey was stronger against grass compared to the sandy background. The overall trend of CF prey surviving better than CR prey was consistent between the 2 backgrounds. Interestingly, however, presentation order had a significant effect on prey survivorship (order effect, χ21 = 25.42, P < 0.001) and survivorship of CF prey higher when they were presented prior to CR prey rather than presented later (treatment × order interaction, χ21 = 5.09, P = 0.02). Volunteers were less likely to detect the prey as they experienced a greater number of duds (χ21 = 9.36, P = 0.002). Figure 2 View largeDownload slide Comparisons of the survivorship of cryptic prey and flashing prey when presented against 2 different backgrounds (grass and sand). (a) shows the comparison between cryptic prey and continuously flashing prey in Experiment 1 while (b) shows the comparison between cryptic prey and intermittently flashing prey in Experiment 2. Bars represent the mean survivorships and error bars represent Wilson binomial 95% confidence intervals. CR = cryptic prey; CF = continuously flashing prey; IF = intermittently flashing prey. Figure 2 View largeDownload slide Comparisons of the survivorship of cryptic prey and flashing prey when presented against 2 different backgrounds (grass and sand). (a) shows the comparison between cryptic prey and continuously flashing prey in Experiment 1 while (b) shows the comparison between cryptic prey and intermittently flashing prey in Experiment 2. Bars represent the mean survivorships and error bars represent Wilson binomial 95% confidence intervals. CR = cryptic prey; CF = continuously flashing prey; IF = intermittently flashing prey. For those prey that were eventually detected, we found no effect of prey type (CF or CR) on detection time (Figure 3, χ21 = 0.81, P = 0.37), but volunteers spent longer trying to find prey on grass than on a sandy background (χ21 = 5.88, P = 0.02). We found no interaction effect between prey type and background type on detection time (χ21 = 0.27, P = 0.60), and no effect of presentation order effect (χ21 = 0.25, P = 0.62) nor an interaction between order and treatment (χ21 = 0.54, P = 0.46). The number of duds volunteers experienced did not affect detection time (χ21 = 0.001, P = 0.97). Figure 3 View largeDownload slide The log transformed detection time (when the prey was found) and giving up time (when volunteers stopped looking, inferring the prey was not present in the scene) for cryptic prey and continuously flashing prey in Experiment 1. Bars and error bars represent mean and standard error of the mean. CR = cryptic prey; CF = continuously flashing prey. Figure 3 View largeDownload slide The log transformed detection time (when the prey was found) and giving up time (when volunteers stopped looking, inferring the prey was not present in the scene) for cryptic prey and continuously flashing prey in Experiment 1. Bars and error bars represent mean and standard error of the mean. CR = cryptic prey; CF = continuously flashing prey. On the other hand, the giving up time was shorter for CF prey in comparison to CR prey (Figure 3, χ21 = 9.92, P = 0.002) and shorter on grass than on the sandy background (χ21 = 13.15, P < 0.001). The giving up time was also shorter for whichever prey was presented later (order effect: χ21 = 4.76, P = 0.03). We found no effect of the interaction between background and prey type on giving up time (χ21 = 0.88, P = 0.35) nor evidence of an interaction between presentation order and treatment (χ21 = 0.30, P = 0.58). For volunteers who experienced a greater frequency of duds, they gave up finding the prey earlier (χ21 = 13.95, P < 0.001). Experiment 2: Intermittently flashing prey versus cryptic prey Survivorship of IF prey was on average 23% higher than cryptic (CR) prey (Figure 2b, χ21 = 13.44, P < 0.001), but here we found no effect of background type (χ21 = 1.69, P = 0.19) and there was no evidence of an interaction between background and prey type (χ21 = 0.99, P = 0.32) on prey survivorship. As before, the survivorship of both CR and IF prey were higher when they were presented prior to the other (χ21 = 5.38, P = 0.02), but we did not find significant effects of either the interaction between presentation order and treatment (χ21 = 1.74, P = 0.19) nor the number of duds experienced (χ21 = 1.83, P = 0.18). In terms of search time, we did not find any effects of our explanatory variables on both detection time (all P > 0.3) and giving up time (all P > 0.07) except that volunteers gave up finding the prey earlier when they experienced more duds (χ21 = 9.39, P = 0.002). Continuously flashing prey versus intermittently flashing prey We found no effect of flash type (χ21 = 3.27, P = 0.07) nor an interaction between flash type and background (χ21 = 2.19, P = 0.13) on prey survivorship, although there was a trend showing that prey survived better on grass background than on the sandy background overall (χ21 = 7.08, P = 0.007). In both CF and IF prey, survivorship was higher when prey were presented before cryptic prey (order effect: χ21 = 17.41, P < 0.001). Prey survivorship was higher when volunteers experienced a greater number of duds for both CF and IF prey (χ21 = 8.02, P = 0.005). In terms of searching time, we found no effects of background type, flashing type, the interaction between background and flashing type, and presentation order on either detection time or giving up time (all P > 0.06) except that volunteers gave up earlier (χ21 = 35.10, P < 0.001) and detected prey earlier (χ21 = 5.84, P = 0.02) as they experienced a greater number of duds. DISCUSSION Our results demonstrate that flash behavior gives artificial prey a survival advantage of approximately 20% (a ~0.2 increase in survivorship, depending on the background) compared to cryptic fleeing prey, whether the prey flashes continuously or intermittently. A related benefit of flash behavior is that human predators gave up searching for retreating prey that continually flashed earlier than for prey that remained cryptic throughout. This suggests that expectations of finding a prey that exhibited conspicuous colors previously were lower than if the prey had remained cryptic. However, we found no differences in the detection time between prey that were cryptic throughout and cryptic prey that flashed when they escaped. Our results collectively support the hypothesis that flash behavior protects the prey by sending “false information” about the prey’s resting color (see also our analyses of the effects of duds below). In relation to the background effect, grass backgrounds appeared to be difficult habitats in which to locate prey based on overall longer detection times and higher prey survivorship than on the sandy background. However, the background effect is only apparent for the continuously flashing prey, yet we found no differences in detection times or prey survivorship between the 2 backgrounds for intermittently flashing prey. This implies that visual background might have affected the evolution of different flashing types. Conspicuousness of flash colors might play a role in the observed decrease in giving up time in the experiment 1. Visual predators adaptively modify their search rate (i.e., the time spent looking for a given prey in a background) based on prey crypticity: if prey are more cryptic, predators spend more time focusing in a given area to find prey (Smith 1974; Gendron 1986). Our finding that human volunteers gave up earlier when searching for flashing prey compared to cryptic prey is consistent with the idea that flash behavior leads predators to anticipate conspicuous prey, and therefore spend less time searching for flashing prey. This effect consequently increases the flash-displaying prey’s probability of survival (but note here that we could not find the same effect for intermittent flash colors). Nevertheless, the observation that intermittently and continuously flashing prey derived similar survivorship benefits suggests that it is not the on-and-off nature of the signal per se but its suppression in conjunction with ceasing movement that is key to being effective. Our findings using human volunteers indicate that exhibiting conspicuous colors during flight either continuously or intermittently and then returning to crypsis on stopping or alighting provides clear antipredator benefits just as Cott (1940) and Edmunds (1974) anticipated. The mechanism(s) by which flash behavior increases survivorship seems to be that conspicuousness sets up an expectation that the prey will be conspicuous and therefore makes it more difficult to locate the now cryptic prey. There is some evidence that the expectations of our volunteers did indeed shape the survival of prey they search for. In particular, the survival advantage of both continuously and intermittently flashing prey was higher as human volunteers experienced more “duds” in their search phase. In effect, if the volunteers were used to finding prey all the time (an unlikely phenomenon in nature) then they would be more liable to continue until they found one. This interpretation is further supported by the evidence for an effect of presentation order on survivorship: when flashing prey were seen before cryptic prey then their survivorship was substantially higher because subjects were less aware that such prey could be highly cryptic. Therefore 3 sources of evidence, our 20% improved survival, the importance of expectations produced by duds, and presentation order, together suggest that the predator expectations of what to find once the prey had settled were influenced by flash behavior. Our findings suggest that flash displaying prey may have evolved in situations in which 1) predators do not spend a long time in one spot to find a prey, and 2) predators search in complex environments where prey are more difficult to find (Dimitrova and Merilaita 2009). An alternative or additional reason why flash behavior results in higher survivorship is that the predator may expect to see the animal continuing on its obvious trajectory and search ahead of where the prey has alighted. Given the discrete nature of our screen shots for fleeing and settling, this was not tested, however. Finally, as suggested by Edmunds (2008), it is possible that the color patch may be followed rather than the outline of the prey, making it difficult to discern its outline at rest. The observation that flash behavior has greater advantages on backgrounds where prey are overall hardest to detect is not surprising—if prey camouflage is poor, then such prey would be detected whether or not they flash during escape. As such, flash behavior may be more likely to evolve in particular habitats where crypsis is likely to be effective. Edmunds (1974) also proposed that flash behavior is more likely to evolve in species which lack the stamina for a long chase—after all if a prospective prey item could flee the vicinity entirely, then its appearance on retreat would be immaterial. Comparative analyses could explore these propositions. While our experiments have focused on elucidating the possible role of flash behavior in hindering subsequent detection, there may be other benefits. In particular, anecdotal evidence suggests that escape mimicry occurs in some insect taxa, for example, the flight of the grasshopper Arphia conspersa mimics that of the palatable but difficult to catch pierid butterfly Colias eurytheme (see Balgooyen 1997). Likewise, it is possible that flash behavior has intimidatory and confusion effects, not dissimilar to deimatic displays in stationary prey. At present, however, our data show for the first time, that flash behavior can hinder predator search in a manner suggested by early scholars of anti-predator defenses. Displays may well exploit a conspicuous illusion to lower the probability of detection. Despite this advance, flash behavior continues to constitute a widespread yet largely unexplored antipredator defense. SUPPLEMENTARY MATERIAL Supplementary data are available at Behavioral Ecology online. FUNDING We thank NSERC (T.S. and C.K.) for funding, and the Wissenschaftskolleg zu Berlin for the time to formulate ideas (T.C. and T.S.). Ethics: Ethical approval was obtained (protocol # 13385 14-0276) from the Carleton University Research Ethics Board-B (CUREB-B) following the Canadian Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans (TCPS2). Data accessibility: Analyses reported in this article can be reproduced using the data provided by Loeffler-Henry et al. (2018). REFERENCES Anonymous . 1945 . Animal concealment and flash coloration . Nature . 155 : 232 – 233 . Balgooyen TG . 1997 . Evasive mimicry involving a butterfly model and grasshopper mimic . Am Midl Nat . 137 : 183 – 187 . Google Scholar CrossRef Search ADS Ballantyne LA , Lambkin CL . 2013 . Systematics and phylogenetics of Indo-Pacific Luciolinae fireflies (Coleoptera: Lampyridae) and the description of new genera . Zootaxa . 3653 : 1 – 162 . Google Scholar CrossRef Search ADS PubMed Bates D , Mächler M , Bolker B , Walker S . 2015 . Fitting Linear Mixed-Effects Models Using {lme4} . J Stat Softw . 67 : 1 – 48 . Google Scholar CrossRef Search ADS Caro TM , Graham CM , Stoner CJ , Vargas JK . 2004 . Adaptive significance of antipredator behaviour in artiodactyls . Anim Behav . 67 : 205 – 228 . Google Scholar CrossRef Search ADS Caro TM , Lombardo L , Goldizen AW , Kelly M . 1995 . Tail-flagging and other antipredator signals in white-tailed deer: new data and synthesis . Behav Ecol . 6 : 442 – 450 . Google Scholar CrossRef Search ADS Caro TM , Stankowich T . 2010 . The function of contrasting pelage markings in artiodactyls . Behav Ecol . 21 : 78 – 84 . Google Scholar CrossRef Search ADS Cott HB . 1940 . Adaptive coloration in animals . London (UK) : Methuen . Del Brooke M . 1998 . Ecological factors influencing the occurrence of ‘flash marks’ in wading birds . Funct Ecol . 12 : 339 – 346 . Google Scholar CrossRef Search ADS Dimitrova M , Merilaita S . 2009 . Prey concealment: visual background complexity and prey contrast distribution . Behav Ecol . 1 : 176 – 181 . Edmunds M . 1974 . Defence in animals: a survey of anti-predator defences . New York (USA) : Longman Publishing Group . Edmunds M . 2008 . Flash colors . In: Capinera JL , editor. Encyclopedia of entomology . Dordrecht : Springer Netherlands . p. 1466 . Google Scholar CrossRef Search ADS Gendron RP . 1986 . Searching for cryptic prey: evidence for optimal search rates and the formation of search images in quail . Anim Behav . 34 : 898 – 912 . Google Scholar CrossRef Search ADS Kamler JANF , Ballard WB . 2006 . Ear flashing behavior of black-tailed jackrabbits (Lepus californicus) . Am Midl Nat . 155 : 402 – 403 . Google Scholar CrossRef Search ADS Kang C , Moon H , Sherratt TN , Lee S-I , Jablonski PG . 2016 . Multiple lines of anti-predator defence in the spotted lanternfly, Lycorma delicatula (Hemiptera: Fulgoridae) . Biol J Linn Soc . 120 : 115 – 124 . Kang C , Zahiri R , Sherratt TN . 2017 . Body size affects the evolution of hidden colour signals in moths . Proc R Soc Lond B . 284 : 20171287 . Google Scholar CrossRef Search ADS Langley CM . 1996 . Search images: selective attention to specific visual features of prey . J Exp Psychol Anim Behav Process . 22 : 152 – 163 . Google Scholar CrossRef Search ADS PubMed Lawrence ES , Allen JA . 1983 . On the term “search image” . Oikos . 40 : 313 – 314 . Google Scholar CrossRef Search ADS Lloyd JE . 1971 . Bioluminescent communication in insects . Annu Rev Entomol . 16 : 97 – 122 . Google Scholar CrossRef Search ADS Loeffler-Henry K , Kang C , Yip Y , Caro T , Sherratt TN . 2018 . Data from: flash behavior increases prey survival . Dryad Digital Repository. http://dx.doi.org/10.5061/dryad.g80d3j3 Merilaita S , Stevens M . 2011 . Crypsis through background matching . In: Merilaita S , Stevens M . editors. Animal camouflage: mechanisms and functions . Cambridge (UK) : Cambridge University Press . p. 17 – 33 . Google Scholar CrossRef Search ADS Olofsson M , Eriksson S , Jakobsson S , Wiklund C . 2012 . Deimatic display in the European swallowtail butterfly as a secondary defence against attacks from great tits . PLoS One . 7 : e47092 . Google Scholar CrossRef Search ADS PubMed R Core Team . 2017 . R: a language and environment for statistical computing . Vienna (Austria) : R foundation for Statistical Computing . http://www.R-project.org/. Schultz TD , Fincke OM . 2009 . Structural colours create a flashing cue for sexual recognition and male quality in a neotropical giant damselfly . Funct Ecol . 23 : 724 – 732 . Google Scholar CrossRef Search ADS Smith JNM . 1974 . The food searching behaviour of two European thrushes. II. The adaptiveness of the search patterns . Behaviour . 49 : 1 – 61 . Google Scholar CrossRef Search ADS Stoner CJ , Bininda-Emonds ORP , Caro T . 2003 . The adaptive significance of coloration in lagomorphs . Biol J Linn Soc.W . 79 : 309 – 328 . Google Scholar CrossRef Search ADS Tinbergen L . 1960 . The natural control of insects in pine woods. I. Factors influencing the intensity of predation by songbirds . Arch Nerl Zool . 13 : 265 – 343 . Google Scholar CrossRef Search ADS Umbers KDL , De Bona S , White TE , Lehtonen J , Mappes J , Endler JA . 2017 . Deimatism: a neglected component of antipredator defence . Biol Lett . 13 : 20160936 . Google Scholar CrossRef Search ADS PubMed Umbers KDL , Mappes J . 2015 . Postattack deimatic display in the mountain katydid, Acripeza reticulata . Anim Behav . 100 : 68 – 73 . Google Scholar CrossRef Search ADS Williams CR , Brodie Jr. ED , Tyler MJ , Walker SJ . 2000 . Antipredator mechanisms of Australian frogs . J Herpetol . 34 : 431 – 443 . Google Scholar CrossRef Search ADS Wyszecki G , Stiles WS . 1982 . Color science . New Jersey : Wiley . © The Author(s) 2018. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Behavioral Ecology Oxford University Press

Flash behavior increases prey survival

Loading next page...
 
/lp/ou_press/flash-behavior-increases-prey-survival-YusZcRn3nI
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1045-2249
eISSN
1465-7279
D.O.I.
10.1093/beheco/ary030
Publisher site
See Article on Publisher Site

Abstract

Abstract Flash behavior, in which otherwise cryptic prey exhibit conspicuous coloration or noise when fleeing from potential predators, has been postulated to hinder location of prey once they become stationary. Here, using artificial computer-generated prey and humans as visual predators, we show that human subjects are more likely to abandon their search for prey that flash, compared to continuously cryptic fleeing controls. Survivorship of flashing prey was an additional 20% higher than the survivorship of continuously cryptic prey, depending on the background against which it was depicted. This survivorship advantage was consistent regardless of whether prey showed flash colors continuously or intermittently during flight. The advantage over continuously cryptic prey was highest when the flashing prey was presented first. Likewise, the more search areas containing no prey that the volunteers had initially viewed, the more likely they were to give up when there was a cryptic prey present. Collectively, these 3 findings indicate that volunteers inferred the flashing prey was absent from the search area when they failed to see a prey in the same form as they saw it move. Our results demonstrate first proof of concept: flash behavior, widely seen in taxa from insects to mammals, is an effective antipredator escape mechanism. INTRODUCTION Flash behavior describes the way in which an otherwise cryptic prey suddenly reveals a conspicuous patch of color or emits noise during movement after disturbance and then hides the patch or ceases to make a sound on stopping (Cott 1940; Edmunds 1974; Edmunds 2008). Examples of patches of color that are only displayed when fleeing include the bright hindwings (underwings) of noctuid and sphingid moths, and the proximal portions of the hindwings of cicadas (Hemiptera), grasshoppers (Orthoptera), and stick insects (Phasmatodea) (Cott 1940; Edmunds 1974). Such insects are cryptic at rest, but conspicuous in flight. Before flight and shortly after coming to rest following escape, these individuals fold their brightly colored hindwings beneath their cryptic forewings. Flash behavior exhibited during movement is also seen in many vertebrate taxa including frogs (Williams et al. 2000), some even-toed ungulates (Artiodactyla) (Caro et al. 2004) and rabbits and hares (Leporidae) (Stoner et al. 2003). At present, the adaptive significance of this putative antipredator defense is unknown. Cott (1940) noted that it seems “to confuse or misdirect an enemy in the pursuit of prey” (p. 376), stressing that it is the “sudden disappearance of color combined with the equally sudden suspension of movement which tends to mislead the eye and to render the animal’s exact whereabouts on alighting all the more difficult to detect”. Edmunds (1974) similarly suggested that the predator “may be caused to hesitate by the sudden movement and appearance of the bright color ... and it may follow this color and be deceived by its sudden disappearance into assuming the prey has vanished whereas in reality the prey has come to rest in its normal cryptic posture with the colored structures hidden.” As such, flash colors could serve to initially startle a would-be predator while the prey is on the move and/or hinder the predator’s search if it follows the color rather than the organism (Edmunds 2008). These hypotheses are entirely plausible, but, as yet, there is no proof that flash behavior has evolved for these reasons. Here, we conduct experiments to test the second component of the above explanations, namely that flash behavior hinders subsequent search for potential prey. Before describing the experiments, we first need to clarify precisely what we mean by flash behavior because it has been used rather broadly to represent almost any conspicuous display of color. Naturally, we focus here on the role of flash color as an anti-predator defense, rather than for mate attraction (see for example Lloyd 1971; Schultz and Fincke 2009; Ballantyne and Lambkin 2013). In that context, flash coloration is a term that has been used to describe startle (“deimatic”) displays produced by stationary organisms in response to the approach or contact by a predator, thereby slowing or halting attack (Umbers and Mappes 2015; Kang et al. 2016; Umbers et al. 2017). While the sudden appearance of something previously hidden could serve to intimidate or distract predators, whether the prey is stationary or mobile (Olofsson et al. 2012; Kang et al. 2017), the second hypothesized protective mechanism associated with flash behavior is different: it involves hindering subsequent search for previously mobile prey. By contrast, deimatic displays do not involve a search-hindering affect. Flash coloration is occasionally used to refer to conspicuous signals that are easily seen during flight, such as the black flank stripe of the Thomson’s gazelle Eudorcas thomsoniii, thought to amplify stotting, thus informing the predator that the prey is difficult to catch (Caro and Stankowich 2010). However, the signal is not hidden once the prey has stopped, so it is unlikely to have evolved to hinder search, and we do not consider this further. Flash coloration in mobile prey has 2 forms: a continuous display of a conspicuous patch during flight, or an intermittent display. Continuous flash behavior has been described in many taxa including cryptic Herennia ornatissima spiders showing their red ventrum while descending rapidly on a thread from their webs (Anonymous 1945), in Oedipoda grasshoppers showing bright red lateral surfaces as they fly away from a disturbance (Edmunds 2008), in dull colored shorebirds showing their white backs, rumps or tails when fleeing en masse (Del Brooke 1998), and in white-tailed deer Oedicolus virginianus tail-flagging (Caro et al. 1995). Intermittent flashing is seen in some lepidoptera as they beat their wings (Edmunds 2008) and in some leporids such as the black-tailed jackrabbit Lepus californicus showing its black tail during flight (Kamler and Ballard 2006). Whether continuous and intermittent flashes serve different functions is unclear, so our experiments investigated the properties of both forms of display. Flash colors could protect the prey by affecting predator psychology. As Cott (1940) and Edmunds (2008) suggested, flash colors displayed during movement may deceive visual predators about the subsequent location of the prey at rest. Additionally (or alternatively), flash colors may confuse predators about the resting colors of prey. Because fleeing prey allow predators to observe the prey momentarily, predators may formulate a transient “search image” based on flash colors (Tinbergen 1960; Langley 1996), although strictly speaking search image formation applies only to cryptic prey (Lawrence and Allen 1983). It is possible that flash colors are the most salient feature of prey during fleeing, thus making predators perceive prey to be of a completely different color than their resting color. If true, predators will be more likely to give up their search for prey that have exhibited flash behavior than those that have not, believing them to have fled the scene if they do not readily detect them. In this study, we used computer-generated artificial prey and humans as visual predators, which is a useful and convenient system to test the adaptive significance of animal coloration, and specifically to test the hypothesis that flash behavior causes predators to abandon search for cryptic prey. METHODS We recruited anonymous human volunteers and evaluated the responses of each participant to flash behavior using a custom-built computer-based game. Our volunteers were visitors (largely undergraduates) to the University Centre at Carleton University, Ottawa, Canada and both experiments were conducted from May to July 2017. Games were presented on a 23” LCD monitor (Dell ST-2310f). Prior to testing, we showed each participant a tutorial video about the game, but details of the hypothesis were not disclosed. Each game had 2 phases, namely “prey escape” and “search.” Once each participant understood the nature of the game and was ready to play, the game was started. At the start of the prey escape phase, a single square prey item (375 × 375 twips, with the number of pixels = twips/15) of given color was placed at a random position on a complex background image (grass or sand, dimensions 8172 × 16,764 twips, Figure 1). When a participant moved the mouse pointer on the background image within 20,000 twips of the prey (hence any movement of the mouse pointer into the background image), the prey moved to escape. Escape was to either the left or right side of the screen (whichever was the furthest to move). Prey movement was simulated as a directed random walk comprising of a rapid sequence of discrete steps in randomly drawn directions taken from a given distribution. The step size of the prey throughout was 600 twips, the step angle was chosen at random from an even distribution within ± 45° from the horizontal, and the duration of wait before new step was 30 ms. Figure 1 View largeDownload slide The prey (here shown together, although note that they were presented alone and in pseudo-random sequence to volunteers) displayed against the (a) grass and (b) sandy background. Individual prey were either cyan, magenta, blue, green, yellow, cryptic or red in color (left to right). Cryptic prey were presented only in the final 2 trials after presenting the prey with fixed conspicuous colors. These cryptic prey either kept their cryptic color when they moved, or they flashed red as they moved. Cryptic prey presented against the grass background had an red, green, blue values of each pixel (RGB) of (73, 151, 19) and an RGB of (192, 156, 111) when presented against a sand background. Figure 1 View largeDownload slide The prey (here shown together, although note that they were presented alone and in pseudo-random sequence to volunteers) displayed against the (a) grass and (b) sandy background. Individual prey were either cyan, magenta, blue, green, yellow, cryptic or red in color (left to right). Cryptic prey were presented only in the final 2 trials after presenting the prey with fixed conspicuous colors. These cryptic prey either kept their cryptic color when they moved, or they flashed red as they moved. Cryptic prey presented against the grass background had an red, green, blue values of each pixel (RGB) of (73, 151, 19) and an RGB of (192, 156, 111) when presented against a sand background. These parameter settings allowed the participant to observe the retreating prey for 1–2 s, depending on its starting position. Following the “escape” of prey, volunteers were asked to click a follow button, after which the search phase began. During search phase, the volunteers were shown an inverted version of the same background screen (to prevent prey being revealed through the contrasting otherwise identical images), with the same prey present at a new random location. The volunteers were asked to find the escaped prey and click on it (generating a pleasant sound and happy face when complete). In this phase, volunteers had 2 options: moving the mouse to the prey and clicking directly on it, or pressing a “give up” button to move on to the next prey if they could not find it (inferring that no prey was present in the screen). Training and general experimental procedure We presented 8 preys sequentially to each participant. The first 6 were considered as training prey and each of these was colored differently (magenta, cyan, red, green, yellow, and blue; see Figure 1), conspicuous against the grass or sand background and presented in a random order. The grass background was of lush green grass while the sandy background showed grains of light brown sand with tiny pebbles scattered on it. The grass background image was readily discriminable from the conspicuous green colored prey used in the training sessions. Crucially, all of these prey maintained the same color while moving (escape phase) and when sedentary (search phase). For the first 6 prey types, there was a 25% probability that the prey would not be present in the search phase (i.e., the second screen). We used these blank “duds” in order to get the volunteers used to the fact that there can sometimes be no prey in a search screen after it had escaped (mimicking a rabbit disappearing down a hole, for example). When the prey was absent from the search screen, volunteers would inevitably press the “give up” button and the same color trial was entered again into the list of prey that had not yet been attacked, so that it could be presented again (in a random sequence of prey not yet searched for). See Supplementary Video S1 for a sample video. Experiment 1: Flash prey versus cryptic prey After all 6 training prey were eventually attacked, volunteers were each presented with 2 treatment prey: a cryptic prey with continuous flash when it moved (continuous flash, CF) and a cryptic prey without flash when it moved (cryptic throughout, CR). The color of cryptic prey had mean R, G, and B values of the background image (Wyszecki and Stiles 1982), making it relatively hard to detect because of the background color matching between prey and background (Merilaita and Stevens 2011). CR prey displayed the cryptic color both when sedentary and when moving. CF prey showed the same cryptic color when sedentary, but exhibited a red color while moving; red is a common flash color found in insect prey thought to exhibit flash behavior (Edmunds 2008). CR prey were presented prior to CF prey for half of the volunteers and the order was reversed for the other half. The prey was always present in the search phase. To test whether the results could be generalized across different background types, the experiment was repeated twice, once with a grass and once with a sandy background. Following the experiment, a noncomprehensive red/green color-blindness test was performed. Data from volunteers who demonstrated signs of color blindness (n = 5) were excluded from analysis. For each training and test image shown to each volunteer that was present on the search screen, we recorded 1) a binary response whether the participant attacked the prey or clicked the “give up” button and so moved on to the next image without finding the prey, and 2) the time spent on each image until either they detected the prey (detection time) or gave up searching (giving up time). There was no maximum cut-off time. A total of 120 volunteers judged to be non-color blind were tested (60 for each background). Each volunteer experienced only one type of background. Experiment 2: Intermittently flashing prey versus cryptic prey In this experiment, we tested whether intermittently displayed flash (hereafter, IF) color could increase survivorship of the prey. All conditions remained the same as in Experiment 1, except that we changed the nature of the flash behavior. Whenever a prey with flash behavior moved on the screen its color flickered (flickering rates: mean of 16.7 changes/seconds) between red and cryptic colors by randomly selecting red or cryptic at each step in the rapid random walk rather than constantly showing red color. Again, we tested 120 volunteers for this experiment (60 for each background). Data analysis In both experiments, our primary goal was to test whether cryptic prey with flash behavior had a survival advantage over cryptic prey without flash behavior. To test this, we fitted generalized (and general) linear mixed models to the data and compared 1) prey survivorship (binary response) and 2) prey searching time of volunteers hunting for cryptic and flashing prey (continuous response). The search time has 2 different endpoints: it indicates the time taken to detect the prey when the prey was detected (detection time) and it indicates the time taken until the volunteer gave up looking for the prey if the prey was not detected (giving up time). Thus, we separately analyzed detection time and giving up time. In each case, the search time response was log transformed to meet the assumptions of the fitted general linear model. For Experiment 1, we treated prey type (CR or CF), background type (grass or sand), and the interaction between these 2 main effects as explanatory variables. We also included 1) the order of presentation of CR and CF as a fixed effect to control for any effect of presentation order, and 2) the interaction between presentation order and our color treatment as predictors, since the order that prey appeared first could have influenced the volunteers’ performance in finding prey. Since “duds” appeared stochastically in the training, the number of “duds” that occurred during a trial varied among volunteers ranging from 0 to 9 duds (on average, volunteers experienced 1.9 duds while 19.1% of the tested individuals did not experience any duds). To estimate and control for any effect of duds, we additionally included the number of duds that appeared for each participant in the training sequence as a covariate. To account for possible differences among human volunteers, and ensure independence, the volunteer was treated as a random effect. The analysis of Experiment 2 proceeded in the same way, with prey type treatments now being CR and IF (intermittent flash). To compare the effectiveness of CF prey with IF prey in reducing predation between experiments, we fitted 2 separate models using 1) survivorship (binomial), 2) giving up time (continuous), and 3) detection time (continuous) as response variables respectively (only one observation for each human was used in each analysis, so a random effect was not necessary to control for independence). Prey type (CF vs. IF), background type, the order of the presentation (whether flashing prey was presented prior to or after cryptic prey in each experiment), number of duds, and the interaction between background and prey type were included as explanatory variables. In all results, we present the model with the lowest Akaike Information Criterion (AIC) among competing candidate models. The analysis of deviance table for each analysis is shown in Supplementary Tables S1 and S2. All the analyses were conducted in R (R Core Team 2017) using “lme4” package (Bates et al. 2015). RESULTS Experiment 1: Continuously flashing prey versus cryptic prey Survivorship of the CF prey was on average 19% higher than CR prey (Figure 2a, χ21 = 10.55, P = 0.001), and the overall survivorship of prey was significantly higher on a grass background than on a sandy background (Figure 2a, χ21 = 6.78, P = 0.009). We also found a significant interaction effect between background and prey type (Figure 2a, χ21 = 25.63, P < 0.001) in that the survivorship advantage of the continuously flashing prey was stronger against grass compared to the sandy background. The overall trend of CF prey surviving better than CR prey was consistent between the 2 backgrounds. Interestingly, however, presentation order had a significant effect on prey survivorship (order effect, χ21 = 25.42, P < 0.001) and survivorship of CF prey higher when they were presented prior to CR prey rather than presented later (treatment × order interaction, χ21 = 5.09, P = 0.02). Volunteers were less likely to detect the prey as they experienced a greater number of duds (χ21 = 9.36, P = 0.002). Figure 2 View largeDownload slide Comparisons of the survivorship of cryptic prey and flashing prey when presented against 2 different backgrounds (grass and sand). (a) shows the comparison between cryptic prey and continuously flashing prey in Experiment 1 while (b) shows the comparison between cryptic prey and intermittently flashing prey in Experiment 2. Bars represent the mean survivorships and error bars represent Wilson binomial 95% confidence intervals. CR = cryptic prey; CF = continuously flashing prey; IF = intermittently flashing prey. Figure 2 View largeDownload slide Comparisons of the survivorship of cryptic prey and flashing prey when presented against 2 different backgrounds (grass and sand). (a) shows the comparison between cryptic prey and continuously flashing prey in Experiment 1 while (b) shows the comparison between cryptic prey and intermittently flashing prey in Experiment 2. Bars represent the mean survivorships and error bars represent Wilson binomial 95% confidence intervals. CR = cryptic prey; CF = continuously flashing prey; IF = intermittently flashing prey. For those prey that were eventually detected, we found no effect of prey type (CF or CR) on detection time (Figure 3, χ21 = 0.81, P = 0.37), but volunteers spent longer trying to find prey on grass than on a sandy background (χ21 = 5.88, P = 0.02). We found no interaction effect between prey type and background type on detection time (χ21 = 0.27, P = 0.60), and no effect of presentation order effect (χ21 = 0.25, P = 0.62) nor an interaction between order and treatment (χ21 = 0.54, P = 0.46). The number of duds volunteers experienced did not affect detection time (χ21 = 0.001, P = 0.97). Figure 3 View largeDownload slide The log transformed detection time (when the prey was found) and giving up time (when volunteers stopped looking, inferring the prey was not present in the scene) for cryptic prey and continuously flashing prey in Experiment 1. Bars and error bars represent mean and standard error of the mean. CR = cryptic prey; CF = continuously flashing prey. Figure 3 View largeDownload slide The log transformed detection time (when the prey was found) and giving up time (when volunteers stopped looking, inferring the prey was not present in the scene) for cryptic prey and continuously flashing prey in Experiment 1. Bars and error bars represent mean and standard error of the mean. CR = cryptic prey; CF = continuously flashing prey. On the other hand, the giving up time was shorter for CF prey in comparison to CR prey (Figure 3, χ21 = 9.92, P = 0.002) and shorter on grass than on the sandy background (χ21 = 13.15, P < 0.001). The giving up time was also shorter for whichever prey was presented later (order effect: χ21 = 4.76, P = 0.03). We found no effect of the interaction between background and prey type on giving up time (χ21 = 0.88, P = 0.35) nor evidence of an interaction between presentation order and treatment (χ21 = 0.30, P = 0.58). For volunteers who experienced a greater frequency of duds, they gave up finding the prey earlier (χ21 = 13.95, P < 0.001). Experiment 2: Intermittently flashing prey versus cryptic prey Survivorship of IF prey was on average 23% higher than cryptic (CR) prey (Figure 2b, χ21 = 13.44, P < 0.001), but here we found no effect of background type (χ21 = 1.69, P = 0.19) and there was no evidence of an interaction between background and prey type (χ21 = 0.99, P = 0.32) on prey survivorship. As before, the survivorship of both CR and IF prey were higher when they were presented prior to the other (χ21 = 5.38, P = 0.02), but we did not find significant effects of either the interaction between presentation order and treatment (χ21 = 1.74, P = 0.19) nor the number of duds experienced (χ21 = 1.83, P = 0.18). In terms of search time, we did not find any effects of our explanatory variables on both detection time (all P > 0.3) and giving up time (all P > 0.07) except that volunteers gave up finding the prey earlier when they experienced more duds (χ21 = 9.39, P = 0.002). Continuously flashing prey versus intermittently flashing prey We found no effect of flash type (χ21 = 3.27, P = 0.07) nor an interaction between flash type and background (χ21 = 2.19, P = 0.13) on prey survivorship, although there was a trend showing that prey survived better on grass background than on the sandy background overall (χ21 = 7.08, P = 0.007). In both CF and IF prey, survivorship was higher when prey were presented before cryptic prey (order effect: χ21 = 17.41, P < 0.001). Prey survivorship was higher when volunteers experienced a greater number of duds for both CF and IF prey (χ21 = 8.02, P = 0.005). In terms of searching time, we found no effects of background type, flashing type, the interaction between background and flashing type, and presentation order on either detection time or giving up time (all P > 0.06) except that volunteers gave up earlier (χ21 = 35.10, P < 0.001) and detected prey earlier (χ21 = 5.84, P = 0.02) as they experienced a greater number of duds. DISCUSSION Our results demonstrate that flash behavior gives artificial prey a survival advantage of approximately 20% (a ~0.2 increase in survivorship, depending on the background) compared to cryptic fleeing prey, whether the prey flashes continuously or intermittently. A related benefit of flash behavior is that human predators gave up searching for retreating prey that continually flashed earlier than for prey that remained cryptic throughout. This suggests that expectations of finding a prey that exhibited conspicuous colors previously were lower than if the prey had remained cryptic. However, we found no differences in the detection time between prey that were cryptic throughout and cryptic prey that flashed when they escaped. Our results collectively support the hypothesis that flash behavior protects the prey by sending “false information” about the prey’s resting color (see also our analyses of the effects of duds below). In relation to the background effect, grass backgrounds appeared to be difficult habitats in which to locate prey based on overall longer detection times and higher prey survivorship than on the sandy background. However, the background effect is only apparent for the continuously flashing prey, yet we found no differences in detection times or prey survivorship between the 2 backgrounds for intermittently flashing prey. This implies that visual background might have affected the evolution of different flashing types. Conspicuousness of flash colors might play a role in the observed decrease in giving up time in the experiment 1. Visual predators adaptively modify their search rate (i.e., the time spent looking for a given prey in a background) based on prey crypticity: if prey are more cryptic, predators spend more time focusing in a given area to find prey (Smith 1974; Gendron 1986). Our finding that human volunteers gave up earlier when searching for flashing prey compared to cryptic prey is consistent with the idea that flash behavior leads predators to anticipate conspicuous prey, and therefore spend less time searching for flashing prey. This effect consequently increases the flash-displaying prey’s probability of survival (but note here that we could not find the same effect for intermittent flash colors). Nevertheless, the observation that intermittently and continuously flashing prey derived similar survivorship benefits suggests that it is not the on-and-off nature of the signal per se but its suppression in conjunction with ceasing movement that is key to being effective. Our findings using human volunteers indicate that exhibiting conspicuous colors during flight either continuously or intermittently and then returning to crypsis on stopping or alighting provides clear antipredator benefits just as Cott (1940) and Edmunds (1974) anticipated. The mechanism(s) by which flash behavior increases survivorship seems to be that conspicuousness sets up an expectation that the prey will be conspicuous and therefore makes it more difficult to locate the now cryptic prey. There is some evidence that the expectations of our volunteers did indeed shape the survival of prey they search for. In particular, the survival advantage of both continuously and intermittently flashing prey was higher as human volunteers experienced more “duds” in their search phase. In effect, if the volunteers were used to finding prey all the time (an unlikely phenomenon in nature) then they would be more liable to continue until they found one. This interpretation is further supported by the evidence for an effect of presentation order on survivorship: when flashing prey were seen before cryptic prey then their survivorship was substantially higher because subjects were less aware that such prey could be highly cryptic. Therefore 3 sources of evidence, our 20% improved survival, the importance of expectations produced by duds, and presentation order, together suggest that the predator expectations of what to find once the prey had settled were influenced by flash behavior. Our findings suggest that flash displaying prey may have evolved in situations in which 1) predators do not spend a long time in one spot to find a prey, and 2) predators search in complex environments where prey are more difficult to find (Dimitrova and Merilaita 2009). An alternative or additional reason why flash behavior results in higher survivorship is that the predator may expect to see the animal continuing on its obvious trajectory and search ahead of where the prey has alighted. Given the discrete nature of our screen shots for fleeing and settling, this was not tested, however. Finally, as suggested by Edmunds (2008), it is possible that the color patch may be followed rather than the outline of the prey, making it difficult to discern its outline at rest. The observation that flash behavior has greater advantages on backgrounds where prey are overall hardest to detect is not surprising—if prey camouflage is poor, then such prey would be detected whether or not they flash during escape. As such, flash behavior may be more likely to evolve in particular habitats where crypsis is likely to be effective. Edmunds (1974) also proposed that flash behavior is more likely to evolve in species which lack the stamina for a long chase—after all if a prospective prey item could flee the vicinity entirely, then its appearance on retreat would be immaterial. Comparative analyses could explore these propositions. While our experiments have focused on elucidating the possible role of flash behavior in hindering subsequent detection, there may be other benefits. In particular, anecdotal evidence suggests that escape mimicry occurs in some insect taxa, for example, the flight of the grasshopper Arphia conspersa mimics that of the palatable but difficult to catch pierid butterfly Colias eurytheme (see Balgooyen 1997). Likewise, it is possible that flash behavior has intimidatory and confusion effects, not dissimilar to deimatic displays in stationary prey. At present, however, our data show for the first time, that flash behavior can hinder predator search in a manner suggested by early scholars of anti-predator defenses. Displays may well exploit a conspicuous illusion to lower the probability of detection. Despite this advance, flash behavior continues to constitute a widespread yet largely unexplored antipredator defense. SUPPLEMENTARY MATERIAL Supplementary data are available at Behavioral Ecology online. FUNDING We thank NSERC (T.S. and C.K.) for funding, and the Wissenschaftskolleg zu Berlin for the time to formulate ideas (T.C. and T.S.). Ethics: Ethical approval was obtained (protocol # 13385 14-0276) from the Carleton University Research Ethics Board-B (CUREB-B) following the Canadian Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans (TCPS2). Data accessibility: Analyses reported in this article can be reproduced using the data provided by Loeffler-Henry et al. (2018). REFERENCES Anonymous . 1945 . Animal concealment and flash coloration . Nature . 155 : 232 – 233 . Balgooyen TG . 1997 . Evasive mimicry involving a butterfly model and grasshopper mimic . Am Midl Nat . 137 : 183 – 187 . Google Scholar CrossRef Search ADS Ballantyne LA , Lambkin CL . 2013 . Systematics and phylogenetics of Indo-Pacific Luciolinae fireflies (Coleoptera: Lampyridae) and the description of new genera . Zootaxa . 3653 : 1 – 162 . Google Scholar CrossRef Search ADS PubMed Bates D , Mächler M , Bolker B , Walker S . 2015 . Fitting Linear Mixed-Effects Models Using {lme4} . J Stat Softw . 67 : 1 – 48 . Google Scholar CrossRef Search ADS Caro TM , Graham CM , Stoner CJ , Vargas JK . 2004 . Adaptive significance of antipredator behaviour in artiodactyls . Anim Behav . 67 : 205 – 228 . Google Scholar CrossRef Search ADS Caro TM , Lombardo L , Goldizen AW , Kelly M . 1995 . Tail-flagging and other antipredator signals in white-tailed deer: new data and synthesis . Behav Ecol . 6 : 442 – 450 . Google Scholar CrossRef Search ADS Caro TM , Stankowich T . 2010 . The function of contrasting pelage markings in artiodactyls . Behav Ecol . 21 : 78 – 84 . Google Scholar CrossRef Search ADS Cott HB . 1940 . Adaptive coloration in animals . London (UK) : Methuen . Del Brooke M . 1998 . Ecological factors influencing the occurrence of ‘flash marks’ in wading birds . Funct Ecol . 12 : 339 – 346 . Google Scholar CrossRef Search ADS Dimitrova M , Merilaita S . 2009 . Prey concealment: visual background complexity and prey contrast distribution . Behav Ecol . 1 : 176 – 181 . Edmunds M . 1974 . Defence in animals: a survey of anti-predator defences . New York (USA) : Longman Publishing Group . Edmunds M . 2008 . Flash colors . In: Capinera JL , editor. Encyclopedia of entomology . Dordrecht : Springer Netherlands . p. 1466 . Google Scholar CrossRef Search ADS Gendron RP . 1986 . Searching for cryptic prey: evidence for optimal search rates and the formation of search images in quail . Anim Behav . 34 : 898 – 912 . Google Scholar CrossRef Search ADS Kamler JANF , Ballard WB . 2006 . Ear flashing behavior of black-tailed jackrabbits (Lepus californicus) . Am Midl Nat . 155 : 402 – 403 . Google Scholar CrossRef Search ADS Kang C , Moon H , Sherratt TN , Lee S-I , Jablonski PG . 2016 . Multiple lines of anti-predator defence in the spotted lanternfly, Lycorma delicatula (Hemiptera: Fulgoridae) . Biol J Linn Soc . 120 : 115 – 124 . Kang C , Zahiri R , Sherratt TN . 2017 . Body size affects the evolution of hidden colour signals in moths . Proc R Soc Lond B . 284 : 20171287 . Google Scholar CrossRef Search ADS Langley CM . 1996 . Search images: selective attention to specific visual features of prey . J Exp Psychol Anim Behav Process . 22 : 152 – 163 . Google Scholar CrossRef Search ADS PubMed Lawrence ES , Allen JA . 1983 . On the term “search image” . Oikos . 40 : 313 – 314 . Google Scholar CrossRef Search ADS Lloyd JE . 1971 . Bioluminescent communication in insects . Annu Rev Entomol . 16 : 97 – 122 . Google Scholar CrossRef Search ADS Loeffler-Henry K , Kang C , Yip Y , Caro T , Sherratt TN . 2018 . Data from: flash behavior increases prey survival . Dryad Digital Repository. http://dx.doi.org/10.5061/dryad.g80d3j3 Merilaita S , Stevens M . 2011 . Crypsis through background matching . In: Merilaita S , Stevens M . editors. Animal camouflage: mechanisms and functions . Cambridge (UK) : Cambridge University Press . p. 17 – 33 . Google Scholar CrossRef Search ADS Olofsson M , Eriksson S , Jakobsson S , Wiklund C . 2012 . Deimatic display in the European swallowtail butterfly as a secondary defence against attacks from great tits . PLoS One . 7 : e47092 . Google Scholar CrossRef Search ADS PubMed R Core Team . 2017 . R: a language and environment for statistical computing . Vienna (Austria) : R foundation for Statistical Computing . http://www.R-project.org/. Schultz TD , Fincke OM . 2009 . Structural colours create a flashing cue for sexual recognition and male quality in a neotropical giant damselfly . Funct Ecol . 23 : 724 – 732 . Google Scholar CrossRef Search ADS Smith JNM . 1974 . The food searching behaviour of two European thrushes. II. The adaptiveness of the search patterns . Behaviour . 49 : 1 – 61 . Google Scholar CrossRef Search ADS Stoner CJ , Bininda-Emonds ORP , Caro T . 2003 . The adaptive significance of coloration in lagomorphs . Biol J Linn Soc.W . 79 : 309 – 328 . Google Scholar CrossRef Search ADS Tinbergen L . 1960 . The natural control of insects in pine woods. I. Factors influencing the intensity of predation by songbirds . Arch Nerl Zool . 13 : 265 – 343 . Google Scholar CrossRef Search ADS Umbers KDL , De Bona S , White TE , Lehtonen J , Mappes J , Endler JA . 2017 . Deimatism: a neglected component of antipredator defence . Biol Lett . 13 : 20160936 . Google Scholar CrossRef Search ADS PubMed Umbers KDL , Mappes J . 2015 . Postattack deimatic display in the mountain katydid, Acripeza reticulata . Anim Behav . 100 : 68 – 73 . Google Scholar CrossRef Search ADS Williams CR , Brodie Jr. ED , Tyler MJ , Walker SJ . 2000 . Antipredator mechanisms of Australian frogs . J Herpetol . 34 : 431 – 443 . Google Scholar CrossRef Search ADS Wyszecki G , Stiles WS . 1982 . Color science . New Jersey : Wiley . © The Author(s) 2018. Published by Oxford University Press on behalf of the International Society for Behavioral Ecology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Behavioral EcologyOxford University Press

Published: Apr 5, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off