Environmental Biology of Fishes 69: 333–344, 2004. © 2004 Kluwer Academic Publishers. Printed in the Netherlands. The mating system of steelhead, Oncorhynchus mykiss, inferred by molecular analysis of parents and progeny Todd R. Seamons a , Paul Bentzen a,b & Thomas P. Quinn a a School of Aquatic and Fishery Sciences, University of Washington, Box 355020, Seattle, WA 98195, U.S.A. (e-mail: seamonst@u.washington.edu) b Department of Biology, Dalhousie University, Halifax, Nova Scotia, Canada B3H 4J1 Received 17 April 2003 Accepted 27 April 2003 Key words: protandry, precocious parr, reproductive timing, size assortative mating Synopsis The development of molecular markers has allowed behavioral ecologists to link parents to specific offspring, providing insights into breeding systems that were not apparent from direct observations of the social system. Studies of this type in fishes have focused on species with male parental care such as centrarchids, and on salmonids, a family with little parental care. In order to gain further insight into the mating system of steelhead trout, Oncorhynchus mykiss, a winter-spawning species whose reproductive system is poorly known, adults returning to spawn were captured in four consecutive years in a small, unfished, wild population. Juvenile offspring were sampled by electrofishing and parentage was determined by exclusion based on a 12 locus microsatellite genotype. Both males and females mated with multiple individuals, though single pair matings were also inferred. Females and males tended to have the same number of mates (median = 1), but males were more likely to have no apparent partner (43% vs. 23% for females) and the maximum number of mates were obtained by males (range 0–10 vs. 0–5 for females). There was no difference in median arrival date by sex, but 80% of the females mated with males that had already arrived rather than males arriving with or after the females (median = 7.5, range = 1–63 days difference). Contrary to expectations, there was no evidence of size-assortative mating; larger males and larger females did not tend to mate with each other more often than would have occurred by chance. Of the juveniles with only one identified parent, most had a known mother and an unknown father rather than the reverse (88% vs. 11%). We interpret this as indirect evidence that non-anadromous males achieved a significant number of fertilizations. Thus the steelhead mating system was complex, being more strongly structured by arrival date than fish size, and including a significant genetic contribution by mature male parr. Introduction The relatively recent development of molecular mark- ers suitable for discriminating genetic relationships among individuals has brought a wealth of comparisons between the observed reproductive behavior (social mating system) and genetic mating system of many organisms (Hughes 1998). Early work focused on the mating systems of birds, comparing observed patterns of paternity and offspring care with genetic patterns, often documenting extra-pair fertilizations in puta- tively monogamous species (e.g., Westneat 1987). The inferences about mating systems drawn from obser- vational and genetic studies often differed dramati- cally (e.g., Birkhead et al. 1990, Sillero-Zubiri et al. 1996), leading to many new theories of sexual selec- tion, sperm competition and mate choice (Birkhead & Møller 1998). There are now many examples of this type of com- parison in fishes, mostly investigating patterns of male reproductive success in species with male parental care (e.g., DeWoody & Avise 2001, Jones & Avise 2001). Many of the other studies of genetic mating systems in fishes involve patterns of paternity, number