Formaldehyde vapor induces cancer of the nasal passages in laboratory animals. In this case-control epidemiologic study, occupational information was obtained for 198 persons with sinonasal cancer (SNC), for 173 with nasopharyngeal cancer (NPC) identified as incident cases by the Connecticut Tumor Registry over 41 years among Connecticut males dying of any cause, and for 605 controls sampled from Connecticut death certificates. City directories and death certificates provided information on job, industry, employer, and year of employment for exposure classification. Without knowledge as to case-control status, an industrial hygienist particularly experienced in epidemiologic studies of formaldehyde classified each study subject with respect to probability and degree of formaldehyde exposure. For those with probable exposure to the high level 20+ years prior to death the odds ratio for NPC was 2.3 [95% confidence limits (CL): 0.9, 6.0], and for those with this same risk factor among men dying at age 68+ (the median study age at death or older) the odds ratio was 4.0 [95% CL: 1.3, 12.0--with two-sided P = .015, unadjusted for multiple significance tests, and with two-sided P = .129 in testing for interaction between this risk factor (never any exposure vs. probable exposure to high level 20+ years prior to death) and age (age less than 68 yr vs. age 68+ yr)]. Odds ratios were close to unity for 9 of 13 industries. Formaldehyde-related occupations in printing appeared to be associated with any type of nasal cancer (either SNC or NPC).