It has been suggested that time selection and precedence in arrival order are more important during spring than autumn migration. Migrating birds are expected to fly at faster airspeeds if they minimize duration rather than energy costs of migration, and they are furthermore expected to complete their journeys by final sprint flights if it is particularly important to arrive at the destination before competitors. We tested these hypotheses by tracking-radar studies of nocturnal passerine migrants during several spring and autumn seasons at Lund (56 degrees N) and Abisko (68 degrees N) at the southern and northern ends of the Scandinavian Peninsula, respectively. The samples from these two sites represent migrants that are mostly rather far from (Lund) or close to (Abisko) their breeding destinations. We found that the birds were flying at clearly faster airspeeds in spring than in autumn at both study sites, with spring speeds exceeding autumn speeds by, on average, 16%, after taking effects of wind conditions and vertical flight speeds into account. This difference in speeds could not be explained by seasonal differences in body mass or wing morphology and thus supports the hypothesis of time-selected spring migration. There was also a significantly larger seasonal difference in airspeed at Abisko than at Lund, suggesting that the birds may have shown an inclination to sprint on their final spring flights to the breeding destinations, although this possible extra sprint effort was modest. (C) 2011 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.