Feeding Ecology of the Milksnake (Lampropeltis triangulum, Colubridae) in the Western United States

—We examined the diet of the Milksnake (Lampropeltis triangulum) in the western United States and evaluated predictions about ontogenetic shifts, sexual divergence, and geographic variation in diet. Identifiable prey items were found in 139 specimens, and 41 additional prey items were recorded from the literature, for 180 prey items in total from 175 individual snakes. Lampropeltis triangulum is a generalist predator and feeds primarily on lizards and mammals. Skinks made up a large portion of the total diet. Other lizard taxa were also important prey, whereas reptile eggs, snakes, and birds were consumed infrequently. Ontogenetic shifts in diet were documented. The upper size limit of prey increased with increasing snake size, and adult snakes continued to feed on small prey. Prey type also was related to snake size. Juveniles fed more frequently on lizards, but adults fed mainly on mammals. Although males were longer than females, there was no sexual size dimorphism in mass, and there were no differences in diet between sexes. Diet varied geographically, and the proportion of endothermic prey was greater at higher latitudes after accounting for snake size. Snakes form a clade composed entirely of predaceous species, making them valuable study models for understanding predator and prey relationships (Greene, 1983). Research on snake diets has provided insight into various aspects of evolution (Savitzky, 1983), ecology (Reynolds and Scott, 1982), and conservation planning (Holycross and Douglas, 2007). Patterns in foraging ecology of many species are well documented, but most species have not been examined. Evaluating the foraging ecology of widespread species is important to establish the general patterns of prey–predator size relationships, ontogenetic shifts, sexual divergence, and geographic variation in diet. Snakes are gape-limited predators, and snake size is a fundamental predictor of the upper size limit of the prey that can be consumed. Many snake species undergo ontogenetic shifts in diet, switching from smaller to larger prey with increasing size and age (Mushinsky, 1987), and excluding small prey from their diet in preference to larger, more energy-rich prey (Arnold, 1993). The ratio of prey-to-snake mass is an important response variable in quantifying ontogenetic shifts (Greene, 1983). Ingestion of proportionately heavier prey allows snakes to forage less frequently, a trade-off that may require a larger gape and larger body size. Two common patterns of prey to snake size relationships related to ontogenetic shifts with increases in snake body size are the ontogenetic telescope and the ontogenetic shift in lower prey size (Arnold, 1993). The ontogenetic telescope describes an increase in prey size with increasing snake size, but no exclusion of smaller prey items from the diets of adult snakes (Fig. 1A). The ontogenetic shift in lower prey size describes a pattern of snakes feeding on larger prey with increasing snake size while also excluding smaller prey (Fig. 1B). Both ontogenetic shifts are qualitatively similar and require weight ratios to distinguish the two patterns (Rodriguez-Robles, 2002). Ontogenetic shifts are often related to prey type as well as to prey size. Endothermic prey tend to be bulkier and more difficult to swallow than more slender, ectothermic prey, and snakes that prey on endotherms are larger and have a subsequently larger gape than snakes feeding primarily on ectotherms (Rodriguez-Robles et al., 1999). Ontogenetic shifts in diet for many snake species are characterized by a shift from smaller ectothermic to larger endothermic prey with increasing snake size (Mushinsky, 1987). Optimal foraging theory predicts that predators should either maximize energy gains or minimize time spent to obtain a fixed amount of energy (Kie, 1999). Dietary shifts in prey size and type may be related to differing energy and nutritional returns of different prey, balanced against costs of prey capture, ingestion, and digestion (Arnold, 1993). For example, larger endothermic prey may be preferred relative to smaller ectothermic prey because larger prey provide more energy relative to effort. Swallowing prey headfirst is an energetic adaptation to reduce the cost of prey handling. Most snakes swallow their prey headfirst to minimize the resistance of limbs, teeth, scales, fur, or feathers; reduce handling time; prevent abrasion of the esophagus; and allow ingestion of larger prey (Greene, 1976; Mori, 2006). Alternatively, factors such as differences in prey availability, habitat use, and thermal preferences between juveniles and adults also may be important to the optimization of energy acquisition, ultimately influencing dietary shifts (Shine, 1991b). Sexual size dimorphism is pronounced in many snake species (Shine, 1978) and can lead to dietary divergence between the sexes (Shine, 1991a; Glaudas et al., 2008). Shine (1991a) considered dietary divergence to be an adaptation driven by body size differences related to reproductive biology. In many snake species with limited sexual dimorphism in body size, little divergence in diet is found between males and females (Holycross et al., 2002; Gardner and Mendelson, 2003). Documentation of dietary differences or similarities is important for understanding potential niche differentiation, resource partitioning, and intraspecific competition between males and females. Some snake species are specialized predators with little geographic variability in their diets (Shine, 1984; Gardner and Mendelson, 2003), but most wide ranging species are characterized by geographic variation in diet (Rodriguez-Robles and Greene, 1999; Clark, 2002; Holycross and Mackessy, 2002; Glaudas et al., 2008). Geographic variation in diet is likely to be related to geographic variation in prey availability. Endotherm diversity increases with latitude relative to the diversity of ectotherms (Simpson, 1964; Kiester, 1971), and endotherms are increasingly available as prey at higher latitudes. Therefore, Corresponding Author. E-mail: [email protected] DOI: 10.1670/10-091 differential prey availability could potentially affect relationships between predator and prey size, ontogenetic shifts, and sexual divergence in diet. The Milksnake (Lampropeltis triangulum) is distributed widely across a latitudinal gradient of >5,600 km, representing a variety of climates, biomes, ecoregions, and habitats that support a diverse suite of potential prey (Williams, 1988). Relative to other populations throughout the distribution of L. triangulum, populations in the arid western United States, are patchily distributed, more secretive and fossorial, smaller in body size, and less abundant (Williams, 1988; Conant and Collins, 1998; Stebbins, 2003). Although male L. triangulum are generally larger than females (Fitch and Fleet, 1970; Rodriguez and Drummond, 2000), sexual size dimorphism is less pronounced in the smaller western subspecies (Williams, 1988). In several western states, L. triangulum is a species of conservation priority and is either protected from collection or its collection is regulated (Hammerson, 1999; Werner et al., 2004; Brennan and Holycross, 2006; Sievert and Sievert, 2006). In these states, observations of feeding have been generally limited to anecdotal field observations and lists of prey species accepted by captive specimens (Guidry, 1953; Wright and Wright, 1957; Kamb, 1978). We examined the diet of L. triangulum in the western United States using museum specimens, field observations, and literature records. Our primary objectives were to 1) test predictions about the feeding ecology of L. triangulum over a large geographic region and 2) provide additional dietary information on this species. Specifically, we predicted 1) a positive relationship between prey and snake mass consistent with an ontogenetic shift in diet; 2) an ontogenetic shift in prey type with larger snakes preying on endotherms; 3) dietary divergence between males and females, with males feeding on a higher proportion of endothermic prey; and 4) geographic variation in diet, with the frequency of endothermic prey increasing with latitude. MATERIALS AND METHODS Data Collection.—We examined museum and field specimens of L. triangulum from Arizona, Colorado, Kansas, Montana, Nebraska, New Mexico, Oklahoma, South Dakota, Texas, Utah, and Wyoming. These states were chosen to cover an area of similar aridity, where primary production is limited by soil moisture, and approximately coincide with the region west of the 100th meridian (Hunt, 1967). Voucher specimens requested for this study were located on the HerpNet data base portal (http:// www.herpnet.org/herpnet/portal.html, 16 October 2005; Appendix 1). Field samples were collected from feces of specimens palpated in the field in Navajo County, Arizona, in 2005 and 2006 as part of an inventory and monitoring program for reptiles and amphibians (E. Nowak, unpubl. data). For each museum specimen, we made incisions (2–5 cm) into the stomach and colon and visually searched for prey items. If prey items were found after initial examination, incisions were enlarged to allow their removal. Prey items were stored in 70% ethanol after their removal. Field specimens from northern Arizona were palpated, and fecal material was collected and stored in 70% ethanol, with subsequent release of live snakes. We noted the location of prey items in the gastrointestinal tract (GI) and considered the stomach the upper gastrointestinal tract (UGI) and the intestines and colon the lower gastrointestinal tract (LGI). Prey items that occurred across both the UGI and LGI were combined into a UGI and LGI category for analysis. Because most gape-limited predators swallow their prey headfirst (Greene, 1976), when possible, we inferred the direction of ingestion (head or tail first) from orientation in the GI. Prey items were examined under a · 40 magnification dissecting microscope. We identified prey items to the lowest level of taxonomic resolution possible by comparison with reference specimens, taxonomic keys, and range maps. Lower levels of taxonomic resolution were generally achieved for prey items from the UGI relative to those from the LGI. We grouped prey items into hierarchical categories consisting of endothermic (bird or mammal) and ectothermic prey. Ectothermic prey items were further classified into snakes, lizards, reptile eggs, and unknown reptiles. We measured snout-to-vent length (SVL) by stretching a flexible tape measure along each snake’s body from the tip of the rostral scale to the terminus of the anal scale (61 mm). After removal of prey items from the GI, snakes and prey items were blotted dry with a paper towel, and mass was determined with a digital scale (60.1 g). Sex was determined by the presence or absence of hemipenes or other reproductive structures (i.e., ova, vas deferens, or testes). However, we were unable to determine sex for many damaged and poorly preserved specimens. To expand our sample sizes beyond museum and field specimens, we searched the literature for direct observations of feeding and prey items collected from free-ranging, wild snakes. Data from the literature were included in analyses of prey FIG. 1. Hypothetical relationships between predator and prey size. (A) Ontogenetic telescope; an increase in prey size with increasing snake size but no exclusion of smaller prey items. (B) Ontogenetic shift; an increase in prey size with increasing snake size while also excluding smaller prey (Arnold, 1993). 516 B. T. HAMILTON ET AL.