Epithelial cells obtained from autopsies of full-term fetuses or infants less than 1 year old were isolated, amplified in primary cultures and inoculated in deepithelialized rat tracheas. These tracheas were then sealed and transplanted subcutaneously into irradiated athymic nude mice. Four weeks after transplantation the tracheal lumen was completely covered by epithelium, most of which was of mucociliary respiratory type. At this stage, tracheal transplants containing tracheobronchial epithelium from 20 different donors were exposed to silastic devices containing 0, 0.5, 1 and 2 mg paraformaldehyde. The tracheal transplants were examined histologically at 2, 4, 8, and 16 weeks after transplantation. Before sacrifice, all animals were injected with a single pulse of tritiated thymidine. Important epithelial alterations could be seen in the formaldehyde treated transplants with a maximum effect visible at 2 weeks after exposure. The highest dose of 2 mg produced, in most cases, numerous areas of epithelial erosion and inflammation whereas this effect was not as evident with the lower doses. All doses produced areas of hyperplastic epithelium alternating with areas of pleomorphic-atrophic epithelium. Although the differences in predominance of different types of epithelium was not clearly dose-dependent, the labeling index (LI) showed dose dependence between 2 and 4 weeks after initiation of exposure. The maximum mean LI was three to four times higher than normal, although in some focal hyperplastic-metaplastic lesions the LI was increased up to 20 times. These studies show that formaldehyde, although toxic at higher doses, is able to elicit at lower doses a proliferative response of the human respiratory epithelium that is not preceded by a massive toxic effect. This response is similar, although less intense than that of the rat respiratory epithelium in which formaldehyde proved to be a carcinogen.