DNA-based analysis of hair to identify species and individuals for population research and monitoring

Abstract

Reliable data on the distribution and abundance of mammal species can be problematice particularly for carnivores that are secretive, elusive, or highly dispersed. For these species, estimation of population abundance and distnbution has traditionally been accomplished using tracks and other sign that produce relative undices of abundance More rigorous designs require marked individuals for absolute population estimates based on mark-recapture experiments7 including applicabons of radiotelemetry, rdioisotopcs? and body tagging. Because these methods reqvure captvaxing and handling the animal at least once, they are considered invasive, and because many of them are costly and time consuming? their practical application is limited to intensive resewch in small study areas. Yet, large-scale and long-term monitoring of wlderanging carnivore species has become an important research and management focus of wildlife scientists and marlagers? as well as the Iarger conservation community. Increasing attention has focused on forest carnvores, such as Canada lyns (Lynx ccBnsdens) wolverine {Gulo galo) fisher (Martes pennant{)) and American marten (Martes americana; Buskirk et al. 1994 Ruggiero et al. 1994, Zielinski and Kucera 1995). Zielinski and Kucera (1995) described several standardized, noninvasive methods for detecting foF est carnivore species using remote photographic bait stations, baited track plates, and snowtracking. The authors carefully differentiated species detection from population monitoring, but suggested tht their detection methods might be used for programs to estimate population size and monitor population changes in carnivore species. The American marten is a small predator and management indicator species of coniferous forests in North America. As pan of a larger carnivore research eXort focusing on manen winter ecology, we are devising, testing, evaluating, and establishing forest carnivore monitoring protocols that are noninvasive, practical, and cost efiective at a variety of scales from forest stand to region. Conformmg to protocols in Zielinski md Kucer (1995) we developed sampling procedures using DNA analysis of hair snagged from carnivores with a baited, trap-like device (hairsnare). One of oux goals is to use this technique as a form of individual identification (self-marker) for estimating populatic)n abundance in a mark-recapture expenment (mark-resight binomial model; Minta md Mangel 1989, Gardner and Mangel 1996). We also hope to provide information about the movement of indieriduals between hairsnares. Hairsnaring has long been used to sample the presence of rare animals in remote areas (review in Raphael 1994). Typically, hair is identified to species or genera by macroor microscopic examlnation. This technique has great potential as a research tool because capture of the animal is not required, cost of materials is low, and snares do not restrain or harm the animal. Previously, the success mte of hairsnanng has proven uratic possibly due to lrariable methods and effort (hphael 1994). Most researchers have used mechatiical devices (e.g., barbed wire, saw blades) that are inconsistent in the amount of hair and follicuhr tissue (hair bulb) they remove from an axiimal. To circumsent these difElculties we began using a glue-like substance to snag hairs (aYailable commer-