PROBLEMS WITH TESTING INBREEDING AVOIDANCE - THE CASE OF THE COLLAREDFLYCATCHER

Four year's data on collared flycatchers, Ficedula albicollis, breedin g in a nestbox plot on the island of Gotland, Sweden, was used to inve stigate whether individuals avoid mating with close kin (i.e., parents or sibs). Only one case of close inbreeding (0.5% of all pairs) was o bserved during the years of study. The observed frequency of close inb reeding was compared to expected frequencies based on two different nu ll models. Assuming no inbreeding avoidance behaviors (e.g., dispersal or kin recognition), but taking into account the fact that mortality, and different arrival and pairing times of individuals reduce the pro bability of mating with close kin, the expected frequency of close inb reeding is 10% and 15% for female and male recruits (i.e., born in the study plot), respectively. However, assuming mating to be random with in the study plot reduced the expected frequency of close inbreeding t o 1% or less for both males and females. Consequently, conclusions dra wn concerning inbreeding avoidance depend on the null model used. Cont rasting estimated costs of tolerating close inbreeding with those of a voiding it (by dispersal to other plots), however, suggests that the c osts of avoiding close inbreeding are substantially greater than those of tolerating it. Therefore, although inbreeding avoidance cannot be rejected as a cause of dispersal of this species, it is not the primar y cause, and particularly not for sex-biased dispersal. The general pr oblems of investigating inbreeding avoidance are discussed. It is argu ed that all previous null models based on random mating in finite popu lations produce expected frequencies of close inbreeding that in fact include inbreeding avoidance, since they implicitly assume random disp ersal within a finite population. Thus, comparisons between observed a nd expected frequencies of close inbreeding based on random mating are inadequate. The most promising method of investigating inbreeding avo idance is to experimentally study individual movements and mating pref erences in the presence and absence of close kin.