Antiviral and symptomatic therapies for the common cold have been evaluated in studies that use either natural colds in the community or colds induced by experimental infection with rhinovirus. The purpose of this study was to compare the severity and natural history of experimentally induced and naturally occurring colds of undetermined etiology. Forty-one subjects with natural colds were compared to 24 subjects with experimental rhinovirus colds. The highest mean daily symptom score, rhinorrhea score, and nasal obstruction score in the natural cold subjects were 14.2, 2.1, and 2.5 respectively. In contrast, the highest mean daily symptom score, rhinorrhea score, and nasal obstruction score in the experimental rhinovirus colds were 9.3, 1.8, and 1.8, respectively (P P = 0.01). There was no significant difference in the lowest mean nasal airflow in the natural and experimental models, .552 l/sec and .508 l/sec, respectively. These results suggest that subjects with natural colds have more severe symptoms than subjects with experimental colds, however, these differences may be due to selection bias inherent in the natural cold study design. Both study designs have potential advantages and selection of the most appropriate design is dependent on the specific aims of the study.