To determine the nature of the delayed response to exercise, we had 20 atopic asthmatics perform cycle ergometry on 2 occasions while breathing either frigid or hot-humid air in a random fashion. The latter served as a sham control. Forced expiratory volume in one second (FEV1) was measured serially for 10 h after each trial. Subjects developing a second wave of obstruction after recovery from the initial asthma attack were rechallenged on a third day with methacholine and followed in an identical fashion. Cold-air exercise produced an immediate 28% fall in FEV1 for the group as a whole, after which 2 distinct patterns of recovery developed. In 13 subjects, the initial obstruction resolved over several hours. Thereafter, lung function remained constant. In the remaining 7 subjects, the early attack was followed by an initial improvement and then progressive deterioration. This pattern occurred at the same times and to the same magnitude both in the hot-humid experiment in which the initial obstruction was absent, and when the obstruction was induced with methacholine. Based on these observations, it appears that the late asthmatic reaction that follows physical exertion in some subjects is a nonspecific epiphenomenon that is neither dependent upon the existence of an early response nor is unique to exercise.