Long distance migrations by large mammals are increasingly imperiled by human development. We studied
autumnmigratory patterns of caribou (Rangifer tarandus) in relation to an industrial road in northwestern Alaska.
Webuilt nullmovementmodels to determine the expected time to cross the road if cariboumovementswere not
affected by the road. We then identified individuals that took longer to cross than expected (slow crossers) and
those that did not differ from that expected from the null model (normal crossers).We identified eight as slow
and 20 as normal crossers. Slow crossers took an average of 33.3±17.0 (±SD) days to cross the road compared
to 3.1± 5.5 days for normal crossers. Slow crossers had an average crossing date of 8 Nov. ± 7.7 days versus 25
Oct. ± 20.6 days for normal crossers. Movement rates of the two classes did not differ before crossing the road,
but slow crossers moved N1.5 times as fast as normal crossers after crossing the road. Movement patterns were
partially explained by environmental attributes, but were most strongly affected by how far a caribou was from
the road and whether it was classified as slow or normal crosser. While avoidance is an important aspect of the
effects of roads on populations, our results showthe importance of other factors, such as howlong individuals are
delayed in crossing when assessing the influence of development on wildlife.