Date of Award

January 2017

Degree Type

Open Access Thesis

Degree Name

Master of Science (MS)


Biological Sciences

First Advisor

Luke E. Dodd

Second Advisor

David M. Hayes

Third Advisor

Valerie E. Peters


Lepidoptera are a core resource for many of North America's insectivorous bats. These predators consume Lepidoptera of varying sizes, and some bat species remove the wings prior to consumption. Selection of larger prey and subsequent wing removal may allow bats to optimize the energetic value afforded by Lepidoptera. In Chapter 1, I explore the relationships between caloric yield, body size, and wing presence. Laboratory-reared Trichoplusia ni moths were grouped into large and small size classes, and wings were removed from half the moths in each size class. Bomb calorimetry was used to determine the gross heat (cal/g) of moths in each treatment. To account for potential differences in energetic value among species, specimens of Malacosoma americanum, Halysidota tessellaris, and Iridopsis sp. moths were also combusted. Larvae of M. americanum were field–collected in April 2012 and reared in the laboratory. Adult H. tessellaris and Iridopsis sp. moths were wild-caught using an illuminated substrate at Mammoth Cave National Park in June – July 2015. No energetic differences were detected for size class or wing condition of T. ni. Additionally, no differences were detected in the caloric yields of the species analyzed, except between Ma. americanum and Iridopsis sp. (P = 0.03). These results suggest that Lepidoptera of various species and sizes may be of similar prey quality, and that the removal of wings by bats may be unrelated to caloric yield. Even so, I believe the lack of differences detected in this study indicate that my approach was likely too coarse of a method to capture subtle energetic differences.

Recent advances in high–throughput gene–sequencing technology have provided the opportunity for bat dietary studies to be conducted with high resolution; in Chapter 2, I describe methods for refining PCR parameters with the intent to maximize amplicon yield. Fecal pellets were collected in May and August of 2011 and 2016 from a maternity colony of Corynorhinus rafinesquii and stored in 95% ethanol at -80°C. Insect DNA was extracted on a per–pellet basis and amplified by PCR; reaction parameters and reagent quantities were experimentally manipulated to determine optimal primer concentration, annealing temperature, and number of PCR cycles. Mean amplicon yield did not differ significantly across years, indicating that samples were preserved successfully and allowing future temporal comparisons to be made. Samples amplified with 0.5 μM primers had significantly higher mean DNA yield than those amplified with 0.4 μM primers (P = 0.02). Mean amplicon yield differed significantly across annealing temperatures ranging from 50°C to 60°C in a gradient PCR (P = 0.008). Number of PCR cycles was also significant (P = 0.0002); samples amplified with 35 cycles had greater amplicon yield than those amplified with 30 cycles. These results suggest that optimal PCR conditions for bat dietary studies may include a 1:20 ratio of primer volume to total reaction volume, a 52°C annealing temperature, and 35 PCR cycles, although the optimal number of PCR cycles may be reagent-dependent.

White-nose Syndrome (WNS) has devastated the insectivorous bats of eastern North America, resulting in dramatic population declines and shifting foraging niches. In Chapter 3, I investigate the effects of WNS, as well as prescribed fire and insect availability, on bat assemblage diversity and composition. Acoustic bat surveys and concurrent insect sampling were conducted at Mammoth Cave National Park before and after the first on–site detection of WNS. Echolocation calls were classified by phonic group (low–, mid–, or Myotis–frequency). All insects were identified to order, and Lepidoptera were clustered into six classes defined by wingspan and characterized by mean dry weight and caloric values. Mean wingspan differed significantly across all size classes (P < 0.05), suggesting that my classification was effective. Model selection determined that the best–fitting model for diversity of bat phonic groups included the relative abundance of dominant insect orders as well as WNS; WNS was the only significant term (P < 0.05). A competing model included only WNS and was also significant (Δ AICc = 1.22, P < 0.05). A distance-based redundancy analysis of the bat assemblage in relation to the relative abundances of dominant insect orders, burn history, and WNS was significant (P < 0.05, 999 permutations). Model selection found that the best-fitting model for Lepidoptera size class distribution included only WNS (P < 0.05). My results implicate WNS as the primary driver of bat assemblage composition, but the magnitude of WNS masks more subtle processes. The indirect effects of WNS on Lepidoptera remain unclear, but my research suggests shifts in the composition of this assemblage following the arrival of WNS at Mammoth Cave National Park.