abstract: An understanding of diet habits is crucial in implementing proper management strategies for wildlife. Diet analysis, however, remains a challenge for ruminant species. Microhistological analysis, the method most often employed in herbivore diet studies, is tedious and time consuming. In addition, it requires considerable training and an extensive reference plant collection. The development of DNA barcoding (species identification using a standardized DNA sequence) and the availability of recent DNA sequencing techniques offer new possibilities in diet analysis for ungulates. Using fecal material collected from controlled feeding trials on pygmy goats, (Capra hicus), novel DNA barcoding technology using the P6-loop of the chloroplast trnL (UAA) intron was compared with the traditional microhistological technique. At its current stage of technological development, this study demonstrated that DNA barcoding did not enhance the ability to detect plant species in herbivore diets. A higher mean species composition was reported with microhistological analysis (79%) as compared to DNA barcoding (50%). Microhistological analysis consistently reported a higher species presence by forage class. For affect positive species identification, microhistology estimated an average of 89% correct detection in control diets, while DNA barcoding estimated 50% correct detection of species. It was hypothesized that a number of factors, including variation in chloroplast content in feed species and the effect of rumen bacteria on degradation of DNA, influenced the ability to detect plant species in herbivore diets and concluded that while DNA barcoding opens up new possibilities in the study of plant-herbivore interactions, further studies are needed to standardize techniques and for DNA bar-coding in this context. / Dissertation/Thesis / M.S. Applied Biological Sciences 2012

Knowledge of the diet of wildlife can aid wildlife biologists to better understand how a species functions within a given ecosystem. Numerous studies have identified various avenues to examine diet for species throughout the world. Wildlife biologists have used diet composition variables as a means to better understand habitat use and aid in the management and conservation of mule deer, Odocoileus hemionus. The complexity of deer diet is still unknown, and local wildlife management agencies could improve conservation strategies with more information regarding the breadth of plant selection in deer diet. Researchers have used non-invasive methods, such as microhistology via fecal analyses, to assess diet composition for mule deer. However, microhistology has several drawbacks that include accuracy in identification and differentiation of plant species, and even genus, as well as determination of accurate proportions of taxa ingested. Genetic techniques, such as next-generation sequencing (NGS), present new avenues for analyzing herbivore diets, especially through the amplification and analyses of specific regions of chloroplast DNA (cpDNA). Additionally, few studies have directly compared microhistological and NGS diet analyses results for any wildlife species. My objectives were to compare diet composition results of both microhistological and NGS diet analyses through estimating diet richness, taxonomic resolution, percent diet, and frequency of occurrence of plant taxa across samples. Mule deer fecal samples were collected on the Navajo Nation from summer and winter ranges of two distinct mule deer populations, Chuska and Carrizo. I found far greater richness and resolution from NGS of plant taxa through the identification of a greater number of species and genera among all populations, within seasons. Upon testing both methods for both populations, no significant agreement was identified for percent of families identified in the diet with both methods, across all samples. I found trends of positive correlation in the occurrence of families between both methods for Carrizo summer diet, as well as among genera and families in Carrizo winter diet. Upon further statistical analyses, I found no significant positive correlation in the occurrence of genera and families identified with both methods among all samples. Genetic techniques may present innovative methods for determining mule deer diet in various ecosystems, and may also be applied to a broad range of herbivore diet studies.