Consecutive visitors to a general medicine outpatient clinic were screened with a hypochondriasis questionnaire. Two thirds (n = 41) of those exceeding a preestablished cutoff met the criteria for DSM-III-R hypochondriasis when given a structured diagnostic interview, while the other third (n = 22) did not. The latter group showed significant decreases in their hypochondriasis questionnaire scores over a 3-week interval. They had less psychiatric disorder and more medical morbidity than the DSM-III-R hypochondriacs. They also viewed their medical care more positively, and their physicians considered them less hypochondriacal. The transiently hypochondriacal group was also compared with a random sample (n = 75) of the patients below the cutoff on the screening instrument. They had more Axis I disorder, more personality disorder, reported higher levels of somatosensory amplification, and more medical disorder. The differences in psychiatric comorbidity and amplification persist when medical morbidity is controlled for by matching and analysis of covariance. This is consistent with the hypotheses that among patients confronted with a medical illness, those with personality disorder and those who are sensitive to somatic sensation are more likely to develop transient hypochondriasis.