In five years (1992, 1994 ± 97) we measured the frequency of extra-pair paternity (EPP) in a Bavarian population of polygynous Great Reed Warblers (Fra Ènkische Weiher region: 49°40'N, 10°51'E); these data were compared with corresponding findings in Sweden. Progeny from copulations with a male other than the partner (extra-pair young, EPY) were identified by multi-locus DNA fingerprinting with oligonucleotide probes. In 48 broods with 194 nestlings we found 19 EPY in 5 broods. The extra-pair fertilisation (EPF) rate as a percentage of total juveniles (9.8 %) is higher, but that with reference to nests (10.4 %) is only insignificantly higher than in Sweden. In contrast to Sweden, neighbour males in our population were never involved in EPF. For two EPF nests the genetic fathers of the EPY were identified; both were unpaired males several years old, with very distant territories (2.4 km away). Both had larger song repertoires than the pair males, and one sang a Great Reed Warbler x Reed Warbler (Acrocephalus scirpaceus) mixed song. In the other cases we were unable to detect the genetic father, either among the neighbouring males or the others tested. Presumably these EPY spring from floater males or rapid mate switching. As in Sweden, EPY were observed in nests of monogamous (2x) as well as polygynous (3x primary females) males. However, in Germany the breeding density was lower and distribution more patchy and the EPF broods were produced earlier (before the median egg-laying date) than those in Sweden. Our comparison supports the hypothesis that the EPF rate is higher in populations with greater genetic variability than in those with less variability. The relatively low EPP rate in this species seems to be due to a close correlation between male and territory characteristics.