To explore the etiology of testicular cancer, cases of testicular cancer were identified among members of a cohort of Danish boys born between 1941 and 1957 (inclusive), who had attended schools in Copenhagen and Gentofte and whose school health records were contained in an archive under the supervision of the Danish Cancer Registry. One hundred and eighty-three cases of testicular cancer diagnosed before 31 December 1984 were identified; 366 controls, matched to cases by sex and age, were selected from the same cohort. Information on potential risk factors and confounders was obtained from two sources: school health records and midwife protocols, both of which were recorded prior to the diagnosis of testicular cancer in cases. Relative risks (RR) approximated by the odds ratios were calculated and, in logistic regression analyses, adjustments were done for known or suspected confounders. A decreasing risk of testicular cancer with increasing birth order was observed (P = 0.020). Compared with being firstborn, being number four or more in birth order was associated with a significantly decreased RR for all testicular cancers (RR = 0.3, 95 percent confidence interval [CI] = 0.3-0.8) and testicular seminoma (RR = 0.1, CI = 0.02-0.9). No association was observed between high social class and the risk of testicular cancer (RR = 1.4, CI = 0.8-2.3); neither was age at which the study subjects had mumps or measles related to risk of testicular cancer. No cases of mumps orchitis were observed before or during school years. A slightly increased RR for testicular cancer among boys from small families could be explained by the association between family size and birth order.(ABSTRACT TRUNCATED AT 250 WORDS)