Previous research has shown that SAR "hotspots" are induced within the laboratory rat and that the resulting thermal hotspots are not entirely dissipated by bloodflow. Two experiments were conducted to determine if hotspot formation in the body and tail of the rat, which is radiation frequency specific, would have behavioral consequences. In the first experiment rats were placed in a plexiglas cage one side of which, when occupied by the rat, commenced microwave radiation exposure; occupancy of the other side terminated exposure. Groups of rats were tested during a baseline period to determine the naturally preferred side of the cage. Subsequent exposure to 360-MHz, 700-MHz or 2450-MHz microwave radiation was made contingent on preferred-side occupancy. A significant reduction in occupancy of the preferred side of the cage, and hence, microwaves subsequently occurred. Reduced exposure to 360-MHz and 2450-MHz microwaves at 1, 2, 6 and 10 W/kg were significantly different from 700-MHz microwaves. In the second experiment semichronic exposures revealed the threshold for reduced exposure of 2450-MHz microwaves to be located between whole-body SAR's of 2.1 and 2.8 W/kg.