The transition from carnivory to omnivory is poorly understood. The ability to feed at more than one trophic level theoretically increases an animal’s fitness in a novel environment. Because of the absence of light and photosynthesis, most subterranean ecosystems are characterized by very few trophic levels, such that food scarcity is a challenge in many subterranean habitats. One strategy against starvation is to expand diet breadth. Grotto Salamanders (Eurycea spelaea (Stejneger, 1892)) are known to ingest bat guano deliberately, challenging the general understanding that salamanders are strictly carnivorous. Here we tested the hypothesis that grotto salamanders have broadened their diet related to cave adaptation and found that, although coprophagous behavior is present, salamanders are unable to acquire sufficient nutrition from bat guano alone. Our results suggest that the coprophagic behavior has emerged prior to physiological or gut biome adaptations.

Guano feeding, salamander, evolution, physiology, behavior

Coprophagy is a feeding strategy commonly found in invertebrates (Weiss 2006), but much less so in vertebrates. Coprophagy sometimes exists in mammals such as rodents and lagomorphs, and to a lesser degree in pigs, horses, dogs and nonhuman primates (Thacker and Brandt 1955; Soave and Brand 1991; Marinier and Alexander 1995; Aitken 2003; Krief and Hladik 2004). In amphibians, coprophagy is rare but when present may influence larval development of some species with herbivorous larvae. For example, herbivorous tadpoles regularly feed on feces of conspecifics in captivity (Gromko et al. 1973; Steinwascher 1978; Pryor and Bjorndal 2005), even when other food sources are available ad libitum (Pryor and Bjorndal 2005). Herbivorous tadpoles have digestive morphologies and physiologies similar to other herbivorous vertebrates that rely on hindgut fermentative digestion (Pryor and Bjorndal 2005) and ingest feces to inoculate their digestive tracts with beneficial microbes (Steinwascher 1978; Beebee 1991; Beebee and Wong 1992). Growth rates are slower when feces are removed from the diet (Steinwascher 1978) suggesting that herbivorous tadpoles benefit nutritionally from coprophagy even though feces are lower in energy (Gromko et al. 1973).

The literature is scarce when it comes to coprophagy in predatory amphibians (Fenolio et al. 2006). However, faces consist of a readily available food resource for animals living in energy-limited environments, such as caves. Food and nutritional resources in caves are derived from surface inputs and can be limited both temporarily and spatially within these systems (Culver and Pipan 2014). Likewise, foraging in aphotic habitats of caves presents significant challenges for animals that potentially may go weeks to months between feeding bouts. Guano produced by seasonally roosting bats represents an important food source for both terrestrial and aquatic invertebrates (Howarth 1983; Poulson and Lavoie 2000), which in turn are prey for fishes and salamanders (Poulson and Lavoie 2000; Graening 2005; Niemiller and Poulson 2010; Fenolio et al. 2006, 2014). Salamanders have been known to be strictly carnivorous but Fenolio et al. (2006) showed that obligate cave-dwelling Grotto Salamander larvae (Eurycea spelaea (Stejneger, 1892), Fig. 1) ingests bat guano. This behavior is not incidental to the capture of aquatic invertebrate prey. Stable isotope signatures suggest nutrients from bat guano could be incorporated into salamander tissues, and nutritional analyses revealed that bat guano is comparable to potential prey items in nutritional and energy content, suggesting that bat guano could be a viable alternative food source in some energy-poor cave systems. Since the relative importance of guano in the diet of subterranean salamanders is unknown, the aim of this study was to determine whether subterranean salamander larvae could persist on an exclusive guano diet compared to the typical carnivorous diet of salamanders.

All experiments were conducted under the approval of animal protocol #15022 by the Rutgers Newark Institutional Animal Care and Use Committee that handles NJIT research. We collected 46 specimens of Eurycea spelaea from January-Stansbury Cave located in the Ozark Plateau National Wildlife Refuge in Delaware County, Oklahoma. Salamanders were housed individually in mesocosms submerged in the cave stream. Each mesocosm consisted of a 500 ml plastic bottle with small holes so that the salamanders had continuous access to fresh cave water. All lids were connected to a central line via a short string. Salamanders were collected June 6th but the study did not until June 22^nd, salamanders were fed amphipods until the start of the study. During the study, salamanders were fed every four days either a strict diet of live amphipods, bat guano, or nothing. We collected amphipods and bat guano fresh on the day of feeding from the cave. The cave is inhabited by a maternity colony (ca. 15,000 individuals) of federally endangered Gray bats (Myotis grisescens A.H. Howell, 1909) from late April through early November (Fenolio et al. 2006, 2014). Salamanders were randomly assigned to a negative control group, or one of two prey types and one of four feeding treatments based on percentage of initial body mass: 0% (control) 2.5%, 5%, and 10%. Salamanders were massed before feeding to track body mass loss or gain and fed the corresponding percentage of initial body mass of amphipods or guano. We used a milligram- accuracy scale (Ohus, Parsippany, NJ, USA). Care was taken to remove any food remnants before the next feeding, although in most cases salamander ate all the food. Salamanders that lost ≥30% of initial body mass were removed the study. We used ANCOVA to compare body mass of the different treatments in MatLab with an alpha level of 0.05. Normality of ranked data was verified via the Kolmogorov-Smirnov test. Salamanders were released back into the cave after the study per permitting regulations.

Figure 1. 

Eurycea spelaea showing troglobitic characters, lack of pigmentation and microphthalmy. Scale bar: 0.5 cm.

All data associated with this study are available from the figshare digital repository: https://doi.org/10.6084/m9.figshare.4805656

Loss of body mass in treatment groups: All treatment groups lost some body mass during the study (34 days; Fig. 2). Animals in the control group (n = 10) were removed from the study earlier (27 days) than the other groups (34 days) due to body mass loss. Salamanders in the control group experienced the steepest loss of body mass. In general, salamanders fed guano lost more body mass than salamanders fed amphipods, and at 34 days, most guano-fed salamanders had reached the 30% loss limit. Body mass was more variable in amphipod-fed groups with both gains and losses. Salamanders fed 2.5% of initial body mass (IBM) lost an average of 31.3% (±14.3%) body mass eating guano compared to an average of 9.6% (±36.0%) eating amphipods. Salamanders fed 5% IBM lost 30.2% (±4.8%) body mass when eating guano compared to 7.9% (±28.1%) eating amphipods. Body mass loss was least for salamanders fed 10% IBM, and guano-fed salamanders lost 28.3% (±4.4%) body mass compared to 7.6% (±11.5%) for the amphipod group.

Figure 2. 

Regression lines based on body mass loss of different diet types and amounts. Salamanders were fed nothing (green), live amphipods (red) or guano (blue). Groups were fed every four days based on their initial body weight, with 2.5% (A), 5% (B) or 10% (C). The calculated regression lines were as follows: Control -1.16x+96.01 R² = 0.54, n = 10; 2.5%_amphipod -0.26x+98.49, R² = 0.39, n = 6; 2.5%_guano -0.70x+93.58, R² = 0.02, n = 6; 5%_amphipod -0.28x+102.22, R² = 0.03, n = 6; 5%_guano -1.12x+98.89, R² = 0.77, n = 6; 10%_amphipod -0.35x+103.36, R² = 0.21, n = 6; 10%_guano -0.70x+96.01, R² = 0.53, n = 6.

Comparisons of weight loss: For salamanders fed 2.5% IBM, body mass loss rates for guano-fed and amphipod-fed groups were slower than the control group (Guano-fed: F = 6.82, P = 0.01; Amphipod-fed: F = 12.14, P = 0.0007) but not different from each other (F = 2.86, P = 0.09). For salamanders fed 5% IBM, guano-fed animals lost body mass at a slower rate than amphipod-fed animals (F = 11.05, P = 0.0012) and control animals (F = 14.75, P = 0.0002), while amphipod-fed animals lost body mass similarly to control animals (F = 0.07, P = 0.795). For salamanders fed 10% IBM, the amphipod-fed group lost body mass slower than the guano-fed group (F = 6.4, P = 0.131) and control group (F = 26.26, P = 1.09e-6), while the guano-fed group was similar to the control group (F = 8.02, P = 0.005).

All amphipod groups had individuals that lost and gained weight within the period of the study. The variability in body mass was smaller in the 10% amphipod IBM compared to the 5% and 2.5%, but a few individuals lost weight making the overall average mass at the end of the study slightly less than the IBM. The large variability in the 5% and 2.5% amphipod groups suggests that individuals may be behaving differently or some individuals may be stressed in the mesocosm. Nonetheless, since guano groups resemble more closely control groups we conclude that amphipods are a better food source for the salamanders.

Shifts in habitat are often linked with dietary shifts, as environmental changes frequently cause organisms to alter foraging behaviors (Rosalino et al. 2005; McMeans et al 2015). The transition from surface to subterranean habitats involves dramatic morphological, physiological, and behavioral changes associated with life in complete darkness and often limited energy resources, including a predicted increase in dietary breadth (Culver 1982, 1994; Holyoak and Sachdev 1998; Fenolio et al. 2006). In subterranean salamanders, the evolution of coprophagy may be an unusual foraging strategy to exploit a nutritious and seasonally abundant resource (i.e., bat guano) in an otherwise food-limited environment. While it has been demonstrated that Grotto salamander larvae will regularly employ coprophagy of calorically-rich bat guano (Fenolio et al. 2006), our study suggests that Grotto salamander larvae are unable to thrive on a guano-exclusive diet for a prolonged period. So in this case the coprophagous behavior has emerged in evolution prior to the necessary physiological changes to gain nutrition from it.

The apparent disagreement between coprophagous behavior in Grotto salamanders and the lack of apparent absorption may have several possible explanations. First, Grotto salamander larvae, and salamanders in general, do not possess the morphological and physiological digestive traits necessary to exploit guano as a food resource. Salamanders in general are strict carnivores with short digestive tracts and have buccal enzymes with low amylolytic activity (Stevens and Hume 2004). In contrast, coprophagy is most often associated with herbivory, which predominately utilize post-gastric (hindgut) fermentation and the consumption of feces increases the absorption of nutrients and inoculate the hind gut with microbes (Clauss et al. 2007). The selective consumption of predigested material is a form of omnivory. We know relatively little about the adaptive advantages of and the selective drivers that favor omnivory, and by proxy coprophagy, in vertebrates (but see Diehl 2003). Coprophagy requires the evolution of not only a coprophagous behavior but also the evolution of morphological and physiological digestive traits to process feces. It is unknown whether these traits are linked, but theoretically behavioral evolution can precede physiological and morphological evolution. Second, since Grotto salamanders are ingesting feces with high protein content (54%; Fenolio et al. 2006) of insectivores (bats) rather than feces from herbivores, a vastly different gut microbiome is needed to efficiently digest feces. So in addition to lacking the morphological and physiological traits, Grotto salamanders may not possess the necessary gut flora to digest and fully process the contents of bat guano. Ley et al. (2009) found that diet can impact gut microbiome diversity in mammals, which increases with evolution from carnivory to omnivory. Digestive evolution in amphibians, as well as their gut biomes and the gut’s propensity for evolution, is yet to be examined in detail. Finally, coprophagy may reflect mistaken identity due to an innate feeding response for moving prey. In subterranean habitats, aquatic salamanders and cavefishes rely heavily on mechanosensation to detect and capture moving prey. Guano falling into a pool and settling on the substrate may elicit a similar feeding response as crustaceans and other aquatic invertebrates. Guano may not be immediately rejected but ingested instead because of the high protein and fat content of the insectivorous guano. Alternatively, guano may possess a micronutrient, vitamin or mineral otherwise scarce in the subterranean habitat (see Fenolio et al. 2006). While guano may not prevent a loss in mass, it may still offer some nutritional benefit.

We thank Ozarks Plateau National Wildlife Refuge for lodging and cave access, Gal Haspel for help with the analysis, Sheilah Roenfeldt for field assistance, and Oklahoma Department of Wildlife Conservation for collection and research permits. This work was in part a product of Project E-22 entitled “Management and Protection for the Ozark Big-eared Bat, Gray Bat, and Stygobitic Fauna in Oklahoma,” and was funded by the Oklahoma Department of Wildlife Conservation.