Behav Ecol Sociobiol (1992) 30:83-94 Behavioral Ecology and Sociobiology © Springer-Verlag1992 A study of the genetic relationships within and among wolf packs using DNA fingerprinting and mitochondrial DNA Niles Lehman 1,,, Peter Clarkson 2, L. David Mech 3, Thomas J. Meier 4, and Robert K. Wayne 1 1 Department of Biology, University of California, Los Angeles, CA 90024, USA 2 Department of Renewable Resources, Govt. of the Northwest Territories, Inuvik, N.W.T. Canada X0E OTO 3 U.S. Fish and Wildlife Service, Patuxent Wildlife Research Center, Laurel, MD 20702, USA (mailing address: North Central Forest Experiment Station, 1992 Folwell Ave., St. Paul MN 55108, USA) 4 U.S. National Park Service, P.O. Box 9, Denali National Park, AK 99755, USA Received May 10, 1991 / Accepted October 17, 1991 Summary. DNA fingerprinting and mitochondrial DNA analyses have not been used in combination to study relatedness in natural populations. We present an ap- proach that involves defining the mean fingerprint simi- larities among individuals thought to be unrelated be- cause they have different mtDNA genotypes. Two class- es of related individuals are identified by their distance in standard errors above this mean value. The number of standard errors is determined by analysis of the asso- ciation between fingerprint similarity and relatedness in a population with a known genealogy. We apply this approach to gray wolf packs from Minnesota, Alaska, and the Northwest Territories. Our results show that: (1) wolf packs consist primarily of individuals that are closely related genetically, but some packs contain unre- lated, non-reproducing individuals; (2) dispersal among packs within the same area is common; and (3) short- range dispersal appears more common for female than male wolves. The first two of these genetically-based observations are consistent with behavioral data on pack structure and dispersal in wolves, while the apparent sex bias in dispersal was not expected. Introduction The use of hypervariable minisatellite probes, which de- tect variability at variable number of tandem repeat (VNTR) loci and produce "genetic fingerprints", has allowed the precise testing of paternity in animal popula- tions and has led to a flourish of recent studies that have questioned some conclusions based on behavioral observations (reviewed in Bruke 1989). However, uncer- tainty exists about the utility of genetic fingerprinting analyses for deducing patterns of relatedness among in- dividuals living in the same area or in a single social * Present address: Dept. of Molecular Biology, Scripps Research Institute, 10666 N. Torrey Pines Road, La Jolla, CA 92037, USA Offprint requests to : N. Lehman at the present address group (Lynch 1988). Even an assessment of paternity requires a detailed knowledge of the study population and DNA samples from parents and their possible off- spring (c.f. Wetton et al. 1987). Nevertheless, data correlating genetic relatedness of individuals with behavior in populations are essential to test hypotheses about inclusive fitness (Hamilton 1964), and genetic fingerprinting provides researchers with the ability to assess the amount of sharing of a large number of highly variable alleles within a study population. For example, Packer et al. (1991) described patterns of similarity based on genetic fingerprint pro- files in lion prides and showed that genetic fingerprinting could be used to define groups of individuals that were related at the level of siblings or parent-offspring. Yet Packer et al. (1991) needed accurate pedigrees of the studied populations as well as the history of each sam- pled lion pride to assess the degree of bandsharing rela- tedness among siblings and parent-offspring. Rarely are these parameters known in populations under study, so it would be desirable to infer relationships based primar- ily on molecular-genetic data. In this report, we use the combination of fingerprint and mitochondrial DNA (mtDNA) data to classify indi- viduals as unrelated, moderately related, or closely relat- ed at the level of siblings or parent-offspring. Our meth- od does not require detailed knowledge about the study populations but relies instead on an extrinsic calibration of relatedness. We apply this approach to three pack clusters of the gray wolf (Canis lupus) and test the follow- ing hypotheses based on observations of social behavior: (1) wolf packs are usually composed on an unrelated breeding male and female and their offspring; (2) dis- persal among nearby packs is common and; (3) dispersal is not sex-biased. The approach We first document the relationship between fingerprint similarity (bandsharing) and the coefficient of related-