Abstract Reversed sexual size dimorphism (RSD, fe- males larger than males) is commonly found in birds of prey. We used kestrels (Falco tinnunculus), breeding in western Finland in a temporally varying environment of 3-year vole cycles, to assess current hypotheses for the evolution and maintenance of RSD. Our 12-year data showed only weak correlations between parental size and breeding parameters (laying date, clutch size and the number of fledglings produced). The degree of RSD per se was unrelated to breeding success, contrary to the pre- diction of the female dominance hypothesis. Females with small males produced larger clutches in low-vole years, independently of laying date, which supports the small male (or its equivalent inter-sexual selection) hy- pothesis. Small females tended to have more fledglings, particularly in low-vole years, which is inconsistent with the hypotheses for an advantage of large female size (the starvation, intra-sexual selection, reproductive effort, and supplementary feeding hypotheses). As for males, small- er females may be more efficient hunters, the importance of which should be most pronounced under harsh breed- ing conditions. Our results suggest that the directional selection on a particular size in kestrels may be under contrasting selection pressures by the environment, and, at least in breeding females, the advantages of large size can actually be counterbalanced during harsh environ- mental conditions. Key words Body size · Breeding success · Contrasting selection · Environmental heterogeneity · Vole cycle Introduction In reptiles, mammals and birds, males are usually larger than females (Andersson 1994). Yet, conspicuous exam- ples of reversed sexual dimorphism (RSD, females larger than males) are encountered. Among avian species, RSD is common in particular taxa, such as raptors (Falconi- formes) and owls (Strigiformes). In these taxa, RSD has attracted major interest and numerous hypotheses have been developed to explain its evolutionary significance, yet there is no consensus. The main divider between the competing hypotheses for the evolution of RSD is whether they try to explain the evolution of large female or small male size. Large female size is thought to be selectively advan- tageous in several ways. The female dominance hypothe- sis (FDH) suggests a behaviorally driven explanation for large female size. Its premise is that large females play an important role in the maintenance of the pairbond (Perdeck 1960; Amadon 1975; Mueller 1986) and are able to subordinate males to ensure that males provide adequate food for the family (Cade 1960; Smith 1982). The intra-sexual selection hypothesis (Intra-SSH) states that individuals of the larger sex compete for the smaller sex and large size confers greater competitive ability (Olsen and Olsen 1984, 1987; Newton 1986; Olsen and Cockburn 1993). Other “female-size” hypotheses predict advantages at various stages of breeding, but as they are applicable also to species that are not reverse dimorphic, they are not likely to explain the evolution of RSD alone, but also need hypotheses for small male size. The repro- ductive effort hypothesis (REH) states that larger females produce more and larger eggs (Reynolds 1972; Selander 1972; Snyder and Wiley 1976; Mueller and Meyer 1985). The starvation hypothesis (SH) states that larger females can cope with longer fasting times (Lundberg 1986), which would be of adaptive significance if fe- males began to breed under harsh conditions. Hence, large females should be able to start breeding earlier than small females (Korpimäki 1986a; Meijer et al. 1988). According to the female supplementary feeding hypothe- S. Massemin ( ✉ ) · E. Korpimäki · J. Wiehn Section of Ecology, Department of Biology, University of Turku, 20014 Turku, Finland Present addresses: S. Massemin, CEPE-CNRS, UPR 9010, associé à l’Université Louis Pasteur, 23 rue Becquerel, 67087 Strasbourg Cedex 2, France, e-mail: s.massemin@wanadoo.fr, Fax: +33-3-88106906 J. Wiehn, ETH Zürich, Experimental Ecology, ETH-Zürich NW, 8092 Zürich, Switzerland Oecologia (2000) 124:26–32 © Springer-Verlag 2000 Sylvie Massemin · Erkki Korpimäki · Jürgen Wiehn Reversed sexual size dimorphism in raptors: evaluation of the hypotheses in kestrels breeding in a temporally changing environment Received: 7 May 1999 / Accepted: 20 January 2000