DISCUSSION
Our results supported our hypothesis that predation on Texas horned lizards in town would be lower than in the natural ranch environment, and are consistent with other studies that have found lower predation in urban areas (Fischer et al. 2012). Although we did not explicitly measure predator densities, there is anecdotal evidence that the predator community in Kenedy and Karnes City differs in both the abundance and diversity of predators when compared to the ranch. For instance, birds of prey and snakes are rarely seen in or near our study plots in the towns but were seen daily at the ranch. Feral and pet cats and dogs are also common in town but do not occur on the ranch. Altered predator communities are a consistent result of urbanization (Prange and Gehrt 2004, Fischer et al. 2012) and relaxed predation pressure in town by certain types of predators like birds may be a contributing factor to the high density of Texas horned lizards in Kenedy and Karnes City (Ackel 2015) and their dietary shift to consuming small prey items (Alenius 2018)
Although the use of models has proven effective at measuring predation in other studies (Brodie 1993, McMilan and Irshick 2010, Vignieri et al. 2010, Farallo and Forstner 2012), a common concern is the confidence with which predation marks left behind on models are accurately categorized. Marks left behind on models by birds were easy to discern by the presence of conspicuous “V” or “U” shaped peck marks. Models that had been decapitated were also categorized as avian predation because birds tend to attack towards the head of their prey (Smith 1973). Avian predators accounted for 30% of predation events. Bite marks accounted for 21.7% and were also relatively easy to discern in models. Bites included obvious chisel marks left behind by rodent incisors and half-moon bite impressions on the dorsum and venter of models that were possibly indicative of predatory lizards (Crotophytus reticulatus. ). All other damage to models not falling into those categories were scored as unknown, which represented 48.3% of recorded predation events. These unknown predation events included models that had limbs removed but no definitive bite or peck marks. Northern grasshopper mice (Onychomys leucogaster ) are known predators of horned lizards (Munger 1986, Endriss et al. 2007) and have been shown to chew the limbs off Texas horned lizards in areas where both are common (Nathan Rains, Texas Parks and Wildlife, pers. comm.). We also found several live adult lizards on the ranch with missing limbs, although it is unknown what caused the loss of limbs. The models with missing limbs may also be the result of a typical greater roadrunner “centrifugal-slam” attack in which the bird grabs a lizard by any limb or tail and smashes it on the ground (Sherbrooke 1990). These results suggest avian predators are a significant threat to Texas horned lizards, which is similar to findings in other studies that show that avian predators like shrikes consume large numbers of horned lizards and may even be responsible for the evolution of increased cranial horn lengths in flat-tailed horned lizards (Phrynosoma mcallii ) (Munger 1986, Young et al. 2004).
Birds are highly visual (Fox et al. 1976) and avian predators of horned lizards may use their visual acuity to find their cryptically colored prey. Our results showed that when models were painted to background color-match their surroundings (i.e. they were more cryptic), avian predation events were significantly less likely than when models were not painted to color-match. This supports a long-held hypothesis that crypsis and background color-matching are the primary defensive adaptations for horned lizards against visually oriented predators (Norris and Lowe 1964, Pianka and Parker 1975).
There are inherent limitations in using models for predation studies which include the lack of movement, smell, and lack of appropriate behavioral responses and therefore models sample a subset of predators (Bateman et al. 2016). For instance, we did not expect models to be attacked by snakes because snakes rely on thermal, motion, and olfactory cues to sense prey (de Cock Buning 1983); all of which are not exhibited by model lizards. Similarly, we did not expect models to be attacked by cats since they are often attracted by motion (Ellis and Wells 2008). We set up motion detection video cameras in the urban areas for some models and filmed several instances of cats walking by and ignoring the models. Nevertheless, local residents have told us that sometimes their pet cats will bring dead horned lizards back to the house. Cats are known to be very efficient predators of small reptiles, birds, and mammals in urban areas (Loss and Marra 2017) and stationary model studies may not be well suited to sample these predation events. Another limitation of models may be that predators are attracted to the particular material used to construct the models (Bateman et al. 2016). Our models were attacked significantly more than the controls however, suggesting predators were not simply attracted to the urethane foam or paint used to construct our models. In the future, we recommend that controls be made into shapes (i.e. pyramids) that would offer no visual cues to encourage predation events such as may have occurred by Texas tortoises mistaking them for cactus pads.
Our hatchling models had fewer predation attempts than the juvenile or adult models. This result may indicate that stationary hatchling lizards are less vulnerable to certain types of predation. Conversely the hatchling models were morphologically the least realistic, due to their small size and difficulty in constructing them and predators may therefore simply not have recognized them as potential prey.
Texas horned lizards have declined throughout their historic range and urbanization is often suggested as one of the main drivers of those declines (Donaldson et al. 1994, Endriss et al. 2007, Wolf et al. 2013). This study presents the first data comparing predation of Texas horned lizards in urban and more natural environments, and may serve as a foundation for future studies. Understanding how some horned lizard populations, like those in Kenedy and Karnes City, are able to persist in urban environments may help inform conservation efforts for other populations. If these human modified environments have suitable vegetation and food resources they may in some cases provide a refuge for some prey species from predators (Keehn and Feldman 2018, Law et al. 2020). Our results further suggest that Texas horned lizards in natural environments experience high levels of predation pressure, and should be an important conservation consideration when targeting areas for potential reintroduction. Models may also be used as a conservation tool in the future by placing them in potential reintroduction areas and in areas where they are established prior to the release of Texas horned lizards to gauge the relative predation pressure the reintroduction site might experience. Knowledge of predation and the predator community at potential reintroduction sites may help increase the probability of survival for reintroduced lizards.