Two species of kestrel, the common and lesser, were caught each month at three geographically defined locations in Israel over a 12-month period, and a total of 306 blood samples were examined for West Nile virus neutralizing antibodies. The prevalences and mean antibody titers were analyzed statistically by the multiple linear regression model and were shown to be significantly affected by two of the independent variables, location and age of the bird. The season had no overall effect on prevalence and titer but a comparison of the mean monthly titers revealed that April was highest and July and August the lowest statistically for the common kestrel which is a resident species. In contrast, the migrating lesser kestrel was caught only in the spring months and principally at the Jerusalem location, where eight out of 29 birds were seropositive. By comparing the serology of the non-migrating, common kestrel with the migrating, lesser kestrel, the effect of seasonality was evaluated in relation to their ecological patterns and yielded evidence for the entry in April of a small number of previously infected common kestrels into Israel. This serological approach based on continuous sampling over an extended period could be used to forecast in the coming years the timing and dispersion of West Nile virus in both Old and New Worlds if surveys are based on a limited number of informative (flag) species.