Evol Ecol https://doi.org/10.1007/s10682-018-9931-x ORIGINAL PAPER
Predation and cryptic coloration in a managed landscape Richard W. Orton1 • Eric J. McElroy2 • Lance D. McBrayer3
Received: 17 August 2017 / Accepted: 27 January 2018 Ó Springer International Publishing AG, part of Springer Nature 2018
Abstract Protective forms of animal color, such as crypsis, are thought to reduce the probability of detection by visual predators. However, because crypsis is ostensibly intuitive, the working hypothesis of cryptic coloration is seldom tested. Additionally because crypsis is a background-specific adaptation, events which alter habitat structure and substrate composition are likely to affect rates of predation on cryptic animals; animal colors that are cryptic against one visual background may be conspicuous against different visual backgrounds. Populations of Sceloporus woodi, a cryptic diurnal lizard, occupy clear-cut stands of sand pine scrub and prescribe-burned longleaf pine habitat within the Ocala National Forest. Here, we used a combination of clay models resembling S. woodi, and spectral analysis, to examine the effects of spatial heterogeneity and model-substrate contrast on rates of predation. The rate of attack on clay models differed between substrate types and habitats, and was highest when clay models were conspicuous against the local visual background. The dorsal color of models greatly contrasted open sand and dead wood, but had similar reflectance values to leaf litter, suggesting that models were most cryptic on leaf litter. We conclude that crypsis is adaptive in this species, and that variation in rates of attack between sampling locations is related to changes in substrate composition due to management history. For instance, the data suggest that the rate of attack on clay models would decrease in response to succession in sand pine scrub, because aging in sand pine scrub results in increased amounts of leaf litter and decreased amounts of open sand. Overall, the results of this study support the theory of protective coloration. Keywords Differential predation Protective coloration Habitat alteration Visual background Selection pressure
& Lance D. McBrayer
[email protected] 1
Florida Fish and Wildlife Conservation Commission, 3377 US-90, Lake City, FL 32055, USA
2
Department of Biology, College of Charleston, 66 George Street, Charleston, SC 29424, USA
3
Department of Biology, Georgia Southern University, PO Box 8042-1, Statesboro, GA 30460, USA
123
Evol Ecol
Introduction Predation is a reoccurring focus of evolutionary studies because predators reduce fitness and can act directionally on specific phenotypes (as reviewed by Lima and Dill 1990). For example, animals with colors and patterns that blend with the visual background (crypsis) are often found in environments with high rates of predation (Kettlewell 1965; Endler and Basolo 1998). Crypsis is also often observed in smaller taxa that are frequently exposed to visual predators, such as amphibians (Storfer et al. 1999), small terrestrial mammals (Hoekstra and Nachman 2003), moths (Endler 1984), and reptiles (Stuart-Fox et al. 2004; Rosenblum 2006; Stuart-Fox and Moussalli 2009; Farallo and Forstner 2012). It is hypothesized, and perhaps intuitive, that crypsis increases the probability of survival for prey animals, by decreasing the likelihood of detection by visual predators (Storfer et al. 1999; Stuart-Fox et al. 2003; Troscianko et al. 2016). However, because crypsis is a substrate-specific adaptation (Ruxton et al. 2004), predation rates on cryptic animals are likely to vary between different visual environments. Events that alter substrate composition, such as habitat management, may then impact predation rates on cryptic species. Yet, the fitness advantages and function of crypsis have only been tested in a few contexts, and seldom from the perspective of visual predators, such as birds (but see Farallo and Forstner 2012). A growing body of literature suggests that geographic variation in the color of cryptic species can result from variation in substrate composition (Cott 1940; Norris and Lowe 1964; Storfer 1999; Stuart-Fox et al. 2004; Rosenblum et al. 2010). For example, there is compelling evidence for both local adaptation and convergent evolution in several species of sympatric lizards, residing across the desert ecotone of White Sands, New Mexico (see Rosenblum 2006; Rosenblum et al. 2010; Rosenblum and Harmon 2011). However, relative rates of predation are rarely measured in studies of color adaptation and the visual environments are often homogenous (e.g., desert environments). Indices of relative predation rates between environments may be useful when studying geographic variation in animal color, because differential predation, in itself, can drive phenotypic divergence (as reviewed by Endler 1995). Furthermore, in heterogenous environments, it is feasible that rates of predation from visual predators vary between different substrate types. This knowledge is important in both evolutionary and ecological contexts, because cryptic animals may forgo activity in environments where predation risk is increased, in turn, impacting home range size (Lagos et al. 1995), ability to forage and/or thermoregulate (Holmes 1991; Pitt 1999; Ahnesjo and Forsman 2006), and mating success (Sih et al. 1990). In landscapes subject to habitat alteration, cryptic animals may be abruptly exposed to increased rates of predation, and likely, increased selection pressure. The impacts of anthropogenic disturbance are rapidly reducing the extent of suitable habitat for many species (Greenberg et al. 1994; Byers 2002; Seabloom et al. 2006; Didham et al. 2007; Dinnage 2009; Schlossberg and King 2009). For example, much of sand pine scrub (SPS) habitat, which has high levels of endemism (Christman and Judd 1990), has been lost to citrus development (Myers 1990). Species that fail to quickly adapt to changes in habitat structure may suffer decreases in population size and potential extinction. The Ocala National Forest (ONF) of central Florida is a mosaic of managed stands of longleaf pine (LLP) and SPS. Naturally, these habitat types contrast in several environmental variables, including dominant substrate type (Wells 1928; Jackson 1973; Kaunert and McBrayer 2015). However, the ONF is an anthropogenically-altered landscape, where LLP and SPS are managed with different strategies (Enge et al. 1986;
123
Evol Ecol
Greenberg et al. 1994; Tiebout and Anderson 1997). Longleaf pine is burned on an annual or biannual schedule, while SPS is managed with a combination of clear-cutting and firesuppression. Thus, SPS management results in increasing vegetation density (and leaf litter) over time. In contrast, LLP is managed to remain in early stages of succession, with leaf litter and ground cover ebb increasing between burns (Tiebout and Anderson 2001). The timing and type of management practice in the ONF may exaggerate differences in substrate composition between LLP and SPS, as well as among stands of each habitat type in various stages of succession. The ONF harbors a large meta-population of Florida scrub lizards (Sceloporus woodi), a species of conservation concern, where populations are fragmented among stands of LLP and regenerating SPS (Jackson 1973; Enge et al. 1986). Additionally, color polymorphism has been recently observed in these populations. Because scrub lizards in ONF are also exposed to avian predators, such as Florida scrub jays (Aphelocoma coerulescens), scrub lizards in the ONF provide an ideal opportunity in which to study the relationship between crypsis, predation, and habitat. In this study, we used clay models to examine the influence of habitat and substrate type on predation in Florida scrub lizards. Without prior knowledge of color polymorphism in scrub lizards, we used uniformly-colored clay models and software designed to measure color from the perspective of avian predators, to quantify the contingency of predation on model-substrate contrast. We tested the effects of model-substrate contrast and habitat type (LLP vs. SPS) on rates of avian predation on clay models. We also compare patterns in light of management history. Our methods followed several published studies that have successfully used clay models to examine the influence of animal color and/or environmental variables on predation (Pfennig et al. 2001; Wuster et al. 2004; Husak et al. 2006). Clay models provide an advantage for testing hypotheses related to predation by removing individual variation such as size and behavior (Paemelaere et al. 2013). We predicted that attacks from visually-oriented predators would be related to the degree to which models contrasted substrate and that the rate of attacked models would vary between LLP and SPS due to differences in substrate composition.
Materials and methods Clay models We constructed models (Fig. 1a) by molding a thin layer (1–2 mm) of pre-tinted oil-based modeling clay (Roma Plastilina; Van Aken International, Rancho Cucamonga, CA, USA), around commercially available plastic lizard replicas. The replica lizards [snout-vent length (SVL) = 60 mm] were within the size range of the average adult male scrub lizards (mean = 55 mm SVL; range = 50–65 mm SVL). We modeled our clay replicas (n = 380) after male scrub lizards because the species is sexually dimorphic and males are thought to incur higher rates of predation than females. Males have reduced dorsal patterning and bright blue throat badges, which are thought to play a role in sexual selection but may inadvertently attract predators (Husak et al. 2006). The dorsal side and color badges of clay models were painted with acrylic paint (Liquitex, Cincinnati, OH, USA) to match the coloration of adult male scrub lizards. The color of models was selected by eye prior to the experiment, and was later verified with a spectrophotometer after the completion of the experiment. Additionally, we constructed 38 control models, shaped as spheres, from the same clay and painted with the same paint used for the dorsal surface of
123
Evol Ecol
Fig. 1 a Photograph of an intact clay model. b Photograph of a clay model that was attacked multiple times by an avian predator on the head and torso, or by multiple avian predators. Models that were attacked multiple times were only counted as a single attack due to uncertainty in the number of predators. c Close-up of a model attacked by an avian predator on the head. Data collected from camera traps suggest the model may have been attacked by a wild turkey. d A photograph taken by a camera trap showing a wild turkey near one of our models placed on dead wood (in foreground)
clay lizard models. Controls were used to assess if lizard models were attacked due to predation or other motivations.
Site selection Scrub lizards prefer early successional habitats of SPS and LLP (Tiebout and Anderson 2001). At the start of the 2015 scrub lizard active season, we extensively scouted the ONF for stands occupied by scrub lizards [totaling 60 ? man hours; average time spent scouting each sampling location (stand) = 3 h]. In addition to the timber harvesting schedule of the United States Forest Service (USFS), the transient nature of SPS only provides a narrow window in which scrub lizards occupy stands (approximately 0–9 years; Tiebout and Anderson 1997). Thus, it is impossible to find occupied replicate stands that were clear-cut at the same time point and that contain lizards (see Kaunert and McBrayer 2015). We are confident that our sampling efforts represent the best available stands to compare within both LLP and SPS habitats. Yet, we acknowledge that the management practices of the USFS in the ONF do not permit replicated management treatments across SPS and LLP (e.g. a 5 year post burn LLP stand, or a burned SPS stand of any age).
123
Evol Ecol
Field placement We placed a total of 380 clay models among five LLP and five SPS stands between June and August 2015 (Fig. 2). Temporal effects related to predation are unlikely, because fluctuations in temperature, rainfall, and vegetation density are minimal in the ONF during this period (Dirac 2012). Along with lizard clay models, we placed a total of 38 control models in the field to quantify rates of attack on models that may have been due to curiosity, or other motivations aside from predation. Models were placed in typical basking and foraging locations of scrub lizards, being thereby available to avian predators (e.g.,
Fig. 2 Sample locations (stands) and the rate of attack on models per each stand in the Ocala National Forest. Open triangles represent the total rate of attack on models per stand and are in relative proportion to one another in size. The letter ‘‘X’’ represents zero attacks for the stand indicated. Stands of longleaf pine (LLP) are labeled by their given name (e.g., Kerr Island), while sand pine scrub (SPS) stands are labeled 1–5. One of the five SPS stands had zero attacks, while three of five LLP stands had zero attacks. Longleaf pine habitat is colored in green, while SPS habitat is colored in yellow. Open water sources are colored in blue. (Color figure online)
123
Evol Ecol
merlins [Falco columbarius], Florida scrub jays, and wild turkeys [Meleagris gallopavo]). Forty models were placed at nine of ten sampling locations, and 20 were placed at the tenth (SPS 4) due to the stand’s much smaller size. The area of SPS 4 is 2.9 ha, where the mean stand size of our sampling locations is 5.3 ha the range is 2.9–5.7 ha (Bronson 2016). At each stand, models were spaced 10–20 m apart to characterize encounter rates of scrub lizards in the ONF (R. Orton, personal observation). To test for predation and crypsis of scrub lizards, we placed models on the three dominant substrates found in both LLP and SPS: open sand, dead wood, and leaf litter (Kaunert and McBrayer 2015). These substrates contrast in color and are typical substrates for scrub lizard activity (L. McBrayer, personal observation; Kaunert and McBrayer 2015). Each of the three substrate types were found in every stand and models were randomly placed in the field in proportion of each substrate type within stands. Statistical analysis comparing the placement of our models with data collected on substrate composition in the ONF, showed that the proportions of clay models placed in the field per substrate type did not significantly differ from the proportions of each substrate type found in each habitat type in the ONF (Kaunert and McBrayer 2015). For each sampling period, we simultaneously placed 40 models in one LLP stand and 40 models in one SPS stand (except SPS 4, see above). In addition, we placed one control model per ten lizard models at each sampling site (i.e., four controls placed at each sampling site except SPS 4, which received 2 clay models). Controls were placed in the field using the same methodology as lizard clay models. This process was repeated five times, for a total of five sampled stands of LLP and five sampled stands of SPS, and a total of 380 models and 32 controls placed in the field. No stand was sampled more than once. All models and controls were left in the field for 72 continuous hours of fair weather, to allow predators sufficient time to locate clay models, but to avoid predator acclimation. Clay models were not checked during this time and were retrieved between the 73rd and 74th hour. Upon retrieval, we scored all models as ‘‘attacked’’ or ‘‘not attacked’’ according to the presence or absence of tooth or beak marks retained in the malleable clay, which were easily differentiated from one another (Fig. 1b, c).
Spectrophotometric measurements We used a bifurcated fiber optics probe (Ocean Optics, Dunedin, FL, USA) connected to an Ocean Optics Flame spectrometer and a xenon light source (Ocean Optics) to measure the spectral reflectance of model dorsal color, open sand, and leaf litter. We attempted to measure dead wood, but its irregular shape made accurate measurements impossible. The fiber optic probe was held with a standard probe holder at 1 mm from the surface of all substrates and models at 45°. All measurements were recorded with Ocean View software v1.5.2 (Ocean Optics) and were taken relative to a certified 99% diffuse white reflectance standard. Dark current and white standard measurements were taken before each spectrophotometric reading. To reduce noise, the average of two scans with a boxcar width of two was used for each measurement.
Spectrophotometric analysis and visual model use Thirty five of 36 (approximately 97%) attacks on clay models were from avian predators. Thus, we used AVICOL software (Gomez 2006) to analyze spectral data because the AVICOL analysis integrates cone sensitivity curves that provide the sensitivities to wavelengths for each of the four single cone types that birds possess (Bowmaker et al. 1997). In AVICOL, Vorobyev and Osorio’s (1998) physiological model was used because
123
Evol Ecol
it accounts for chromatic and achromatic contrast between colors as perceived by an organism with tetrachromatic vision. This model is accepted to be consistent with the capabilities of avian vision (Osorio et al. 1999; Endler and Mielke 2005) and considers the chromatic and achromatic properties as independent sources of visual stimulation (Vorobyev and Osorio 1998). Acrhomatic contrast is crucial for distinguishing between small objects or objects visualized from large distances (Vorobyev and Osorio 1998; Osorio et al. 1999; Lind and Kelber 2011). For example, achromatic contrast might be expected to be important for avian predators that search for small prey, such as lizards, while midflight or while perched on tree branches. However, we initially analyzed both chromatic and achromatic contrast. We found achromatic contrast to be a more reliable measure. Thus, we focused our spectrometric analysis on achromatic contrast, which, according to previous studies on the visual systems of avian predators, is expected to be of higher relevance to this study (Osorio et al. 1999; Lind and Kelber 2011). The achromatic contrast between colors of two different samples is expressed as a just noticeable difference (JND) value, where a JND value = 1 is the critical value for discrimination between two colors or objects of varying brightness (Gomez 2006). We also analyzed 15 samples of sand and 15 samples of leaf litter, collected from each habitat type (3 from each of 5 stands). Three spectrophotometric measurements were taken on each sample and averaged. This sample average was then averaged across the other two samples of the same substrate type collected from the same stand, to obtain a measure of reflectance for each substrate type from each stand. We also took measurements of three lizard models (to ensure accuracy of measurements) and found the difference in reflectance measurements between models to be negligible (0.4–0.6 just noticeable difference, JND units). Because the difference in reflectance between the three lizard models was negligible and all models were colored with identical paint, there was no need to measure the spectral reflectance of additional models. The average from the three lizard model spectral measurements was then used to determine the contrast against each substrate type from each stand where models were placed.
Cameras Camera traps (Browning model BTC-5, Browning Arms Co, Morgan, UT, USA) were deployed to aid in predator identification and to potentially observe predation events. Four camera traps were deployed at each stand for the same 72 h as our models. The camera traps were focused on one model per substrate type (n = 3) and one camera trap was focused on a random model. The camera traps were set to take five photographs in rapid succession following the triggering of the camera sensor, set off by movement. Also, cameras were set to take one photograph every minute from 0600 to 2030 h, the period when diurnal scrub lizards are active. The photographs were downloaded and reviewed using compatible Browning software.
Statistical analysis We were interested in the effect of habitat and substrate types on attack rates. We fit a general linear model (GLM) with habitat type and substrate type as fixed factors and attack as the response (0 or 1) using a binomial distribution and a logit link function. However, habitat type is confounded with stand in that five stands were LLP and five different stands were SPS. Additionally, the geographic location of stands may result in spatial
123
Evol Ecol
autocorrelation of attack rates. To account for these issues, we fit a generalized linear mixed model (GLMM) with habitat type and substrate type as fixed factors, the location (i.e. latitude and longitude) of stand as a random effect, attack (0 or 1) as the response using a binomial distribution and a logit link function. We computed confidence intervals around parameters estimates and also computed odds ratios by taking the exponential of each parameter estimate. For GLMM, we used Bartlett-corrected likelihood ratio tests, based on 1000 bootstrap replicates, to assess the significance of the fixed effects. These tests compared a model with only the random effect (i.e. location) with (1) a model with habitat and location, (2) a model with substrate and location and (3) a full model with habitat, substrate and location. We did not analyze the interaction term between substrate and habitat because this model suffers from quasi-separation (Allison 2008). These analyses were done using the spaMM package (Rousset and Ferdy 2014) in R-Studio 1.1.383 (RStudio Team 2016) and R 3.4.2 (R Core Team 2017). Additional statistical tests were performed using the JMP Pro v12.1 statistical package (SAS Institute, Carry, NC, USA). Because the distribution of JND values and attack rate did not meet the assumptions of normality, a nonparametric correlation was used to determine if the rate of attack on models was associated with the degree to which the dorsal color of models contrasted the color of substrate. A Wilcoxon rank-sum test was then used to determine if the spectral values of the substrate types differed between LLP and SPS. Last, a Chi square test was used to determine if our method for placing models represented the relative proportions of each substrate type in LLP and SPS. Post hoc analyses were performed using the adjusted residuals (Delucchi 1993).
Results Of the 380 models placed in the ONF, five models were lost, either through experimental error, or from being carried away by a predator. These lost models were excluded from analyses due to uncertainty. No control models were attacked, thus we excluded the 38 controls from analyses. Evidence of animal disturbance was detected on 36 of the 375 the models retrieved. Of these 36, 35 models were attacked by an avian predator, and one model was determined to have been attacked by a mammal. Because we were interested in predation rates from visual predators (i.e., avian predators), we counted this model as ‘not attacked’. Thus, 9.3% (35 of 375) of the clay models were attacked. The GLM (without location as a random effect) showed that attack rate is significantly greater on dead wood (b = 2.180, 95% CI 0.537–5.078, odds ratio 8.85) and sand (b = 2.112, 95% CI 0.530–4.996, odds ratio 8.26) compared to leaf litter. This was also true of SPS stands (b = 1.162, 95% CI 1.162–2.063, odds ratio 3.20) when compared to LLP (Fig. 3). Generalized linear mixed models including habitats (Bartlett-corrected likelihood ratio test: v2 = 11.43, df = 2, P = 0.0004), substrate types (Bartlett-corrected likelihood ratio test: v2 = 19.91, df = 3, P = 0.0010) and both (Bartlett-corrected likelihood ratio test: v2 = 18.92, df = 4, P = 0.0008) were significantly better fits to the data then the model with only the random effect of location. Parameter estimates from the full model suggest that clay models found on dead wood (b = 2.065, 95% CI 0.306–5.013) had 7.8 greater odds of being attacked than clay models found on leaf litter. Clay models on sandy substrates (b = 1.950, 95% CI 0.246–4.880) had a 7.0 greater odds of being attacked than those on leaf litter. Clay models in SPS (b = 1.396, 95% CI -0.741 to 4.449) had a marginally greater probability and a 4.0 greater odds of being attacked than those in LLP.
123
Evol Ecol
Fig. 3 Comparisons among the rate of attack on models between substrate types and habitat types (± one standard error of mean). The rates of attack on models placed on dead wood and open sand were significantly higher than for models placed on leaf litter. The rate of attack on models was also higher in sand pine scrub (SPS) than in longleaf pine (LLP)
There was no significant difference in spectral contrast (between models and their substrates) for open sand (Z = -0.054, P = 0.6139) or leaf litter (Z = -1.095, P = 0.2731) between LLP and SPS (Fig. 4a) (i.e., there was no variation in sand between LLP and SPS). Furthermore, the rate of attack on models across the ONF was associated with the achromatic contrast between models and their associated substrate (q = 0.55; P = 0.013) (Fig. 4b), but not chromatic contrast (q = 0.11; P = 0.337). We also determined that the placement of models upon different substrates significantly differed between LLP and SPS (X2 = 49.062, P \ 0.001*), with the frequency of models placed on leaf litter being greater in LLP and the frequency of models placed on open sand higher in SPS. Data published by Kaunert and McBrayer (2015) confirms that there is more open sand in SPS and more leaf litter in LLP. We report the rate of attacks on models in light of the management history of each stand in Fig. 5. However, due to the impossibility of replication of management history between SPS and LLP, no analysis is included. Nonetheless, these data are suggestive that the youngest SPS stands have the highest attacks rates. Camera traps placed near models captured over 104,400 min of photographs, yet no predation events were recorded on cameras. However, camera traps did capture the presence of multiple taxa including lizards (Florida scrub lizards, six-lined race runners [Aspidescelis sexlineatus], and skinks [Plestiodon spp.]), birds (red-bellied wood peckers [Melanerpes carolinus], Bachman’s sparrows [Paucaea aestivalis], wild turkeys [Fig. 1d]), and mammals (Florida mice [Podomys floridanus], white-tailed deer [Odocoileus virginianus], and wild boar [Sus scrofa]) near our clay models. Of the species photographed by our camera traps, wild turkeys were the most common (captured in six different photographs) followed by Florida mice (captured in three different photographs). All lizards and birds were captured by camera traps before 2130 and all mammals were captured between 2130 and 0600 when scrub lizards would be inactive. Of the models retrieved with evidence of disturbance, only one was placed with an accompanied camera trap. This
123
Evol Ecol
Fig. 4 a Achromatic contrast (measured in just noticeable difference) between models and leaf litter, and models and open sand, in longleaf pine and sand pine scrub (± one standard error of mean). b Relationship between model-substrate contrast (measured in JND) and rate of attack on models (q = 0.55; P = 0.013)
model was obviously attacked by a mammalian predator (likely a racoon [Procyon lotor] or opossum [Didelphis virginiana], judging by the distance between incisors [2.0 cm]). However, the associated camera trap was disturbed before the attack and retrieved with the lens flush against the substrate.
123
Evol Ecol
Fig. 5 Rate of attack on models by management history (± one standard error of mean). Management categories: 1–2 burned = 1–2 years post prescribe burn in longleaf pine). Sand pine scrub (SPS) is clear-cut (CC). 7CC = 7 years post clear-cut SPS; 3–6CC = 3–6 years post clear-cut SPS; and 1–2CC = 1–2 years post clear-cut SPS. Zero models were attacked in stands 7 years post clear-cut. Note that management differs between habitat types. Long leaf pine is prescribe burned on an annual to bi-annual schedule, while SPS is managed via fire-suppression and clear-cutting after stand maturity (20 ? years)
Discussion The theory of adaptive coloration predicts that cryptic individuals are more likely to evade visual predators than individuals that are more conspicuous (aside from aposematic species) (Cott 1940; Endler 1984). However, the results of this study suggest that cryptic individuals may only be more likely to evade visual predators given specific environmental parameters, or potentially, under certain predator regimes. Models found in SPS were more likely to be attacked than those in LLP (Table 1). It is unlikely that differences in attack rates between LLP and SPS were driven by variation in conspicuousness between habitat types, because model-substrate contrast did not vary according to the location from which substrates were collected. Thus, color variation between populations of cryptic species may be influenced by more than geographic variation in substrate color. In addition to predator evasion, animal colors and patterns also increase fitness for sexual selection and thermoregulation, which is vital for ectothermic organisms (Kettlewell 1973; Endler 1978; Kingsolver and Wiernasz 1991; Bittner et al. 2002; Stuart-Fox et al. 2004; Stuart-Fox and Moussalli 2009; Strugariu and Zamfirescu 2011). Thus, in environments with reduced predation intensity, fitness may be more likely to be increased by colors that more strongly satisfy the adaptive response to the thermal environment or sexual selection, as opposed to predation (Gibson and Falls 1979; Andren and Nilson 1981). For example, studies have shown that defensive morphology and behavior are reduced, and conspicuous coloration is increased in different fish species when predation is relaxed (see Moodie 1972; Seghers 1974; Endler 1978; Maan et al. 2008). In ectothermic organisms, the importance of environmental characteristics has also been described in relation to tradeoffs between the thermal environment and crypsis (Ahnesjo and Forsman 2006; Stuart-Fox and Moussalli 2009; Castella et al. 2013). Recent data indicates that
123
Evol Ecol Table 1 A list of each sample locations (stand) with the habitat type, management history, rate of attack on models per substrate type, and total rate of attack on models for each stand Habitat type
Longleaf pine (LLP)
Sand pine scrub (SPS)
Total
Stand
Management
Substrate type
Total
Leaf litter
Open sand
Dead wood
Riverside Island
1–2 burned
0/20
0/12
0/6
Norwalk Island
1–2 burned
0/20
0/10
0/9
0/38 0/39
Salt Springs Island
1–2 burned
0/20
0/11
0/9
0/40
Kerr Island
1–2 burned
0/12
3/18
2/8
5/38
Hughes Island
1–2 burned
2/20
2/14
0/6
4/40
SPS 1
7CC
0/17
0/14
0/9
0/40
SPS
3–6CC
0/11
3/23
0/6
3/40
SPS 3
3–6CC
0/6
8/28
2/6
10/40
SPS 4
1–2CC
0/5
2/7
1/8
3/20
SPS 5
1–2CC
0/4
4/20
6/16
10/40
2/135
22/157
11/83
35/ 375
The categories of management combine the type of management (burned vs. clear cut and roller chopped) and time since last disturbance. Categories: 1–2 burned = 1–2 years post burn in long leaf pine; sand pine scrub (SPS) is clear-cut (CC). 7CC = 7 years post clear-cut SPS; 3–6CC = 3–6 years post clear-cut SPS; and 1–2CC = 1–2 years post clear-cut SPS
thermal quality of habitat in the ONF differs between LLP and SPS, and that the thermal physiology of scrub lizards respectively shifts to maximize performance (Neel 2016). Because rates of predation on scrub lizards from visual predators are likely reduced in LLP, these populations may be, on average, more conspicuous than lizards inhabiting SPS. However, one of the less explored areas of crypsis is its optimization and utility in heterogeneous environments (but see Merilaita et al. 1999). The optimization of crypsis in heterogenous environments has been proposed to occur through two different mechanisms, though these processes are not unrelated. In the first scenario, Endler (1978) accurately predicted that the value of crypsis is highest in microhabitats where risk of predation is greatest. In the second case, Merilaita et al. (1999) proposed a model where Endler’s prediction is the outcome of a specific set of parameters, ultimately arguing that crypsis in heterogenous environments can either be optimized for a specific microhabitat, or simultaneously for multiple microhabitats, depending on contrasts between different visual environments, allocation of available resources, and predator regime. For example, it may be difficult to simultaneously optimize crypsis for microhabitats of extreme contrast. Furthermore, there may be little benefit gained from universal crypsis if sufficient resources are available in one specific microhabitat or if animals actively forgo activity in microhabitats where predation risk is increased. Conversely, there may be substantial benefit of universal, even if moderate crypsis, in environments where predation rate from visual predators does not vary between different microhabitats. In the ONF, scrub lizard encounter rates appear to be less restricted in LLP compared to SPS, where only in SPS do scrub lizards actively select leaf litter substrate, despite it being far less abundant than open sand and dead wood (Kaunert and McBrayer 2015). Because
123
Evol Ecol
cryptic animals will often forgo activity in environments where their crypsis is compromised (Cott 1940; Werner et al. 1983; Ahnesjo and Forsman 2006), this suggests that either crypsis is optimized differently between habitat types, that resource availability in SPS is greatest in microhabitats with leaf litter, or that overall reduction in predation rate in LLP (or lack of variation in predation rate between substrate types) allows scrub lizards access to a variety of microhabitats without increased risk of predation. It is difficult to know if scrub lizard resource availability in the ONF is higher in some microhabitats as opposed to others, because this has not been measured. Additionally, it is likely that resource availability per microhabitat varies by resource. However, compared to lizards in LLP, lizards in SPS were shown to thermoregulate less effectively (Neel 2016). This may suggest that scrub lizards compromise ability to thermoregulate in SPS to avoid elevated predation risk, because selection pressures from the thermal environment and visual predators create conflict that governs microhabitat selection (Ahnesjo and Forsman 2006). Additionally, a posteriori spectrophotometric measurements showed that the average contrast between clay models and three adult scrub lizards (6.97 JND) was similar to the contrast between clay models and leaf litter (6.87 JND). If scrub lizards from LLP and SPS are equally cryptic on leaf litter, then variation in microhabitat use between habitat types is most likely related to differential predation. Future work will measure the spectral reflectance of dorsal color in a large sample size of scrub lizards collected from LLP and SPS to test for color polymorphism and local adaptation for cryptic coloration. Because the ONF is a managed landscape, it is also important to consider the potential implications of management practices in affecting predation on scrub lizards. Clear-cutting in SPS exposes bare sand and dead wood, and decreases vegetation density (Tiebout and Anderson 1997). However, in the absence of a fire regime, vegetation density (and leaf litter) increases and the amount of open sand decreases (Tiebout and Anderson 1997, 2001). In turn, habitat succession in SPS likely decreases the probability that prey animals will be detected by visual predators, especially if prey animals are cryptic on leaf litter. In contrast, prescribe burning of LLP on an annual or bi-annual schedule, maintains habitat features characteristic of early to mid-succession (Wells and Shunk 1931), where open sand and leaf litter are both common and extensively used by scrub lizards (Kaunert and McBrayer 2015). Thus, management practice in the ONF, may affect rates of predation on scrub lizards through changes to substrate availability and vegetation density (Fig. 5). However, a future study, designed to replicate management practice (e.g., proportion of available substrate types), is needed to quantitatively assess any relationships between management and predation rate. In addition to substrate composition and habitat structure, changes in avian species abundance and richness often result from anthropogenic disturbance (Herkert 1994; Gray et al. 2007). For example, habitat fragmentation was shown to decrease gene flow and increase the need for greater dispersal distance for Florida scrub jays (Coulon et al. 2010). Furthermore, scrub jays prefer habitat with open sand and young scrub oaks (Fitzpatrick and Woolfenden 1984), which is characterized by recently-disturbed SPS (Breininger et al. 1995, 2006). During the present study, the highest scrub jay abundance that was anecdotally observed was in a recently disturbed SPS stand (Fig. 2; SPS 5), which was also the location of the highest predation rate on our models. It is plausible that the relative abundance and spatial distribution of scrub jays was a determining factor of predation rate in this study. However, management practices in the ONF are also likely to affect the spatial distribution of wild turkeys, as impacts from anthropogenic disturbance are documented in other understory species (Barlow et al. 2006). Camera traps captured several instances of potential avian predators near clay models. The most common of which, were
123
Evol Ecol
wild turkeys, and most often in stands of LLP. Wild turkeys forage terrestrially, where the advantage of cryptic coloration to prey may be overridden by a turkey’s proximity. The relative abundance of terrestrially-foraging avian predators, may describe the reduced variation in predation rate between substrate types within LLP. Further study on avian density and spatial distribution within the ONF may be needed to fully understand variation in predation rate between LLP and SPS. In conclusion, the cryptic appearance of scrub lizards is likely related to avian predation, though avian predation appears to vary between habitat types in the ONF. In turn, variation in avian predation may translate into color variation between scrub lizard populations. We suspect that differences in the rate of attack on models between habitat types was related to differences in prey detection or foraging strategy by different avian predators. We show that the rate of attack on models placed on leaf litter substrates is the lowest, and the rate of attack on models placed on highly contrasting substrates (e.g., open sand) is the highest. It appears that management practice generates differing proportions of these substrates (Kaunert and McBrayer 2015), that in turn, affect rates of avian predation in the ONF. Thus, there may be unintended consequences on the survival of species that are dependent on environmental variables subject to management practices. Given the environmental impact of anthropogenic disturbance, this may become an important consideration for future studies of evolution and ecology. Acknowledgements We would like to thank Lauren K. Neel and Chase T. Kinsey for their help in the field collecting lizards. Additionally, we owe a debt of gratitude to the Bedore Sensory Ecology Lab at Georgia Southern University for their aid in measuring spectral reflectance. We thank Dr. John Steffen for comments in earlier versions of this manuscript. Research in the Ocala National Forest was conducted with permission from the USDA Forest Service (USFS permit # SEM540). All applicable international, international and/or institutional guidelines for the care and use of animals were followed under protocol with the Institutional Animal Care and Use Committee (IACUC permit # I15011). Funding for this research was provided by a Graduate Student Professional Development Grant from the College of Graduate Studies at Georgia Southern University.
References Ahnesjo J, Forsman A (2006) Differential habitat selection by pygmy grasshopper color morphs; interactive effects of temperature and predator avoidance. Evol Ecol 20:235–257 Allison PD (2008) Convergence failures in logistic regression. In: SAS Global Forum 2008, vol 360 Andren C, Nilson G (1981) Reproduction success and risk of predation in normal and melanistic color morphs of th adder, Vipera berus. Biol J Linn Soc 15:235–246 Barlow J, Peres CA, Henriques LMP, Stouffer PC, Wunderle JM (2006) The responses of understorey birds to forest fragmentation, logging and wildfires: an Amazonian synthesis. Biol Cons 128:182–192 Bittner TD, King RB, Kerfin JM (2002) Effects of body size and melanism on the thermal biology of garter snakes (Thamnophis sirtalis). Copeia 2002:477–482 Bowmaker JK, Heath LA, Wilkie SE, Hunt DM (1997) Visual pigments and oil droplets from six classes of photoreceptor in the retinas of birds. Vis Res 37:2183–2194 Breininger DR, Larson VL, Duncan BW, Smith RB, Oddy DM, Goodchild MF (1995) Landscape patterns of Florida scrub jay habitat use and demographic success. Conserv Biol 9:1442–1453 Breininger DR, Toland B, Oddy DM, Legare ML (2006) Landcover characterizations and Florida scrub-jay (Aphelocoma coerulescens) population dynamics. Biol Cons 128:169–181 Bronson K (2016) Ocala National Forest land management record. In: USDA, editor, Florida, USA Byers JE (2002) Impact of non-indigenous species on natives enhanced by anthropogenic alteration of selection regimes. Oikos 97:449–458 Castella B, Golay J, Monney JC, Golay P, Mebert K, Dubey S (2013) Melanism, body condition and elevational distribution in the asp viper. J Zool 290:273–280 Christman SP, Judd WS (1990) Notes on plants endemic to Florida scrub. Fla Sci 53:52–73
123
Evol Ecol Cott HB (1940) Adaptive coloration in animals. Methuen, London Coulon A, Fitzpatrick JW, Bowman R, Lovette IJ (2010) Effects of habitat fragmentation on effective dispersal of Florida scrub-jays. Conserv Biol 24:1080–1088 Delucchi KL (1993) On the use and misuse of chisquare. In: Keren G, Lewis C (eds) A handbook for data analysis in the behavioral sciences. Lawrence Erlbaum, Hillsdale, pp 294–319 Didham RK, Tylianakis JM, Gemmell NJ, Rand TA, Ewers RM (2007) Interactive effects of habitat modification and species invasion on native species decline. Trends Ecol Evol 22:489–496 Dinnage R (2009) Disturbance alters the phylogenetic composition and structure of plant communities in an old field system. PLoS ONE 4:e7071 Dirac P (2012) Florida climate center. Florida State University Tallahassee, Florida Endler JA (1978) A predators view of animal color patterns. Evol Biol 11:319–364 Endler JA (1984) Progressive background in moths, and a quantitative measure of crypsis. Biol J Linn Soc 22:187–231 Endler JA (1995) Multiple-trait coevolution and environmental gradients in guppies. Trends Ecol Evol 10:22–29 Endler JA, Basolo AL (1998) Sensory ecology, receiver biases and sexual selection. Trends Ecol Evol 13:415–420 Endler JA, Mielke PW (2005) Comparing entire colour patterns as birds see them. Biol J Linn Soc 86:405–431 Enge KM, Bentzien MM, Percival HF (1986) Florida scrub lizard status survey. Technical report no. 26. US Fish and Wildlife Service, Jacksonville, Florida, USA Farallo VR, Forstner MRJ (2012) Predation and the maintenance of color polymorphism in a habitat specialist squamate. PLoS ONE 7:e30316 Fitzpatrick JW, Woolfenden GE (1984) The helpful shall inherit the scrub. Nat Hist 93:55–63 Gibson AR, Falls JB (1979) Thermal biology of the common garter snake Thamnophis sirtalis. Oecologia 43:99–109 Gomez D (2006) AVICOL, a program to analyse spectrometric data. Last updated October 2011. Free program. http://sites.google.com/site/avicolprogram or from the author at
[email protected] Gray MA, Baldauf SL, Mayhew PJ, Hill JK (2007) The response of avian feeding guilds to tropical forest disturbance. Conserv Biol 21:133–141 Greenberg CH, Neary DG, Harris LD (1994) Effect of high-intensity wildfire and silvicultural treatments on herpetofaunal communities in sand pine scrub. Conserv Biol 8:1047–1057 Herkert JR (1994) The effects of habitat fragmentation on midwestern grassland bird communities. Ecol Appl 4:461–471 Hoekstra HE, Nachman MW (2003) Different genes underlie adaptive melanism in different populations of rock pocket mice. Mol Ecol 12:1185–1194 Holmes WG (1991) Predator risk affects foraging behaviour of pikas: observational and experimental evidence. Anim Behav 42:111–119 Husak JF, Macedonia JM, Fox SF, Sauceda RC (2006) Predation cost of conspicuous male coloration in collared lizards (Crotaphytus collaris): an experimental test using clay-covered model lizards. Ethology 112:572–580 Jackson JF (1973) Phenetics and ecology of a narrow hybrid zone. Evolution 27:58–68 Kaunert MD, McBrayer LD (2015) Population density of the Florida scrub lizard (Sceloporus woodi) in managed Florida scrub and longleaf pine sandhill habitats. Herpetol Conserv Biol 10:883–893 Kettlewell HB (1965) Insect survival and selection for pattern—most camouflage and survival mechanisms though highly perfected can be adapted to changing environments. Science 148:1290–1296 Kettlewell HB (1973) The evolution of melansim: a study of a recurring necessity. Oxford University Press, Oxford Kingsolver JG, Wiernasz DC (1991) Seasonal polyphenism in wing melanin pattern and thermoregulatory adaptation in pieris butterflies. Am Nat 137:816–830 Lagos V, Contreras L, Meserve P, Gutie´rrez J, Jaksic F (1995) Effects of predation risk on space use by small mammals: a field experiment with a Neotropical rodent. Oikos 74:259–264 Lima SL, Dill LM (1990) Behavioral decisions made under the risk of predation: a review and prospectus. Can J Zool 68:619–640 Lind O, Kelber A (2011) The spatial tuning of achromatic and chromatic vision in budgerigars. J Vis 11:2 Maan ME, Eshuis B, Haesler MP, Schneider MV, van Alphen JJM, Seehausen O (2008) Color polymorphism and predation in a Lake Victoria cichlid fish. Copeia 2008:621–629 Merilaita S, Tuomi J, Jormalainen V (1999) Optimization of cryptic coloration in heterogeneous habitats. Biol J Linn Soc 67:151–161
123
Evol Ecol Moodie G (1972) Predation, natural selection and adaptation in an unusual threespine stickleback. Heredity 28:155–167 Myers RL (1990) Scrub and high pine. Ecosystems of Florida, vol 765. University of Central Florida Press, Orlando, pp 150–193 Neel LK (2016) The effects of thermal opportunity and habitat on ecologically distinct populations of the Florida scrub lizard Sceloporus woodi. Georgia Southern University, Statesboro Norris KS, Lowe CH (1964) Analysis of background color-matching in amphibians and reptiles. Ecology 45:565–580 Osorio D, Vorobyev M, Jones CD (1999) Colour vision of domestic chicks. J Exp Biol 202:2951–2959 Paemelaere EAD, Guyer C, Dobson FS (2013) The role of microhabitat in predation on females with alternative dorsal patterns in a small Costa Rican anole (Squamata: Dactyloidae). Rev Biol Trop 61:887–895 Pfennig DW, Harcombe WR, Pfennig KS (2001) Frequency-dependent batesian mimicry—predators avoid look-alikes of venomous snakes only when the real thing is around. Nature 410:323 Pitt WC (1999) Effects of multiple vertebrate predators on grasshopper habitat selection: trade-offs due to predation risk, foraging, and thermoregulation. Evol Ecol 13:499–516 Rosenblum EB (2006) Convergent evolution and divergent selection: lizards at the White Sands ecotone. Am Nat 167:1–15 Rosenblum EB, Harmon LJ (2011) ‘‘Same same but different’’: replicated ecological speciation at White Sands. Evolution 65:946–960 Rosenblum EB, Ro¨mpler H, Scho¨neberg T, Hoekstra HE (2010) Molecular and functional basis of phenotypic convergence in white lizards at White Sands. Proc Natl Acad Sci 107:2113–2117 Rousset F, Ferdy J-B (2014) Testing environmental and genetic effects in the presence of spatial autocorrelation. Ecography 37:781–790 R Core Team (2017) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna. https://www.R-project.org/ RStudio Team (2016) RStudio: integrated development for R. RStudio, Inc., Boston. www.rstudio.com Ruxton GD, Sherratt TN, Speed MP (2004) Avoiding attack: the evolutionary ecology of crypsis, warning signals, and mimicry. Oxford University Press, New York Schlossberg S, King DI (2009) Postlogging succession and habitat usage of shrubland birds. J Wildl Manag 73:226–231 Seabloom EW, Williams JW, Slayback D, Stoms DM, Viers JH, Dobson AP (2006) Human impacts, plant invasion, and imperiled plant species in California. Ecol Appl 16:1338–1350 Seghers BH (1974) Schooling behavior in the guppy (Poecilia reticulata): an evolutionary response to predation. Evolution 28:486–489 Sih A, Krupa J, Travers S (1990) An experimental study on the effects of predation risk and feeding regime on the mating behavior of the water strider. Am Nat 135:284–290 Storfer A (1999) Gene flow and local adaptation in a sunfish-salamander system. Behav Ecol Sociobiol 46:273–279 Storfer A, Cross J, Rush V, Caruso J (1999) Adaptive coloration and gene flow as a constraint to local adaptation in the streamside salamander, Ambystoma barbouri. Evolution 53:889–898 Strugariu A, Zamfirescu SR (2011) Population characteristics of the adder (Vipera berus berus) in the Northern Romanian Carpathians with emphasis on colour polymorphism: is melanism always adaptive in vipers? Anim Biol 61:457–468 Stuart-Fox D, Moussalli A (2009) Camouflage, communication and thermoregulation: lessons from colour changing organisms. Philos Trans R Soc Lond B Biol Sci 364:463–470 Stuart-Fox DM, Moussalli A, Marshall NJ, Owens IPF (2003) Conspicuous males suffer higher predation risk: visual modelling and experimental evidence from lizards. Anim Behav 66:541–550 Stuart-Fox DM, Moussalli A, Johnston GR, Owens IPF (2004) Evolution of color variation in dragon lizards: quantitative tests of the role of crypsis and local adaptation. Evolution 58:1549 Tiebout HM, Anderson RA (1997) A comparison of corridors and intrinsic connectivity to promote dispersal in transient successional landscapes. Conserv Biol 11:620–627 Tiebout HM, Anderson RA (2001) Mesocosm experiments on habitat choice by an endemic lizard: implications for timber management. J Herpetol 35:173–185 Troscianko J, Wilson-Aggarwal J, Stevens M, Spottiswoode CN (2016) Camouflage predicts survival in ground-nesting birds. Sci Rep 6:19966 Vorobyev M, Osorio D (1998) Receptor noise as a determinant of colour thresholds. Proc R Soc B Biol Sci 265:351–358 Wells BW (1928) Plant communities of the coastal plain of North Carolina and their successional relations. Ecology 9:230–242
123
Evol Ecol Wells BW, Shunk IV (1931) The vegetation and habitat factors of coarser sands of the North Carolina coastal plain: an ecological study. Ecol Monogr 1:465–520 Werner EE, Gilliam JF, Hall DJ, Mittelbach GG (1983) An experimental test of the effects of predation risk on habitat use in fish. Ecology 64:1540–1548 Wuster W, Allum CS, Bjargardottir IB, Bailey KL, Dawson KJ, Guenioui J, Lewis J, McGurk J, Moore AG, Niskanen M, Pollard CP (2004) Do aposematism and Batesian mimicry require bright colours? A test, using European viper markings. Proc Biol Sci 271:2495–2499
123