In a single-prey loader. the load-size effect may select for transportation of large prey to the nest and consumption of small prey at capture sites. Sonerud suggested that analyses of prey items delivered by avian predators to the nest may profoundly bias the shape of functional response (FR) curves. Therefore, FR may be underestimated, if alternative prey types are smaller than the primary prey type, whereas in the reverse situation FR may be overestimated. Data from Eurasian kestrels Falco tinnunculus breeding in western Finland during 1989-92 showed that the proportion of voles of the genus Microtus and Clethrionomys (the primary prey) increased to the same extent from low through intermediate to high vole densities among both prey items captured and transported from capture sites to nests. Thirty-six male kestrels feeding their mates and nestlings and 18 females feeding their nestlings did not carry small prey items less than large prey items. Therefore, we conclude that the bias due to the load-size effect in the FR curve is probably negligible and that the model apparently overestimated its importance in biasing the FR curve. The probable reason is that decisions by central-place foraging avian predators are also influenced by factors other than load-size effect: for example, prevention of extra-pair fertilizations, defence of offspring and reduction of flight costs during hunting.