In 10 (10.8%) of 93 laboratory-confirmed or probable cases of Rocky Mountain spotted fever seen at Duke University Medical Center from 1969 to 1991, illness without rash or fleeting or atypical skin eruptions were noted. Data from these 10 cases and 33 similar previously reported cases of Rocky Mountain "spotless" or "almost spotless" fever support the premise that human Rickettsia rickettsii infection has a broader spectrum than that indicated by its classic description. Sixty-one percent of patients with Rocky Mountain spotless or almost spotless fever have been men. Two-thirds have been black. Although in some cases the absence of rash may be due to the prompt institution of therapy with chloramphenicol or tetracycline, in others long delays in recognition of the disease and treatment occurred, resulting in the deaths of 53% of patients. On the basis of outcome in the cases reviewed, empirical administration of chloramphenicol or a tetracycline to selected patients is justifiable (including those patients with severe organ dysfunction), even if they have no rash, a transient rash, or a rash of unusual distribution. Indeed, without such therapy, patients with spotless or almost spotless fever may die. Recognizing such atypical cases is analogous to seeing a wolf in sheep's clothing--intuition, experience, knowledge, and a high index of suspicion are required.