Paternal occupational exposures to hydrocarbons have been associated with childhood nervous system cancer, but study results have not been consistent. This population-based case-control study was designed to examine this association using a large sample size to increase the precision of risk estimates. The birth certificates of 499 children who died in Texas from intracranial and spinal cord tumors were compared with 998 control certificates randomly selected from all Texas live births. Information on parental job title and industry at the time of birth was obtained from the birth certificates. No significant associations were identified for the dichotomized variable of all hydrocarbon-related occupations combined, as variously defined in previous studies, or for most of the specific jobs affiliated with exposures to hydrocarbons. Significant, relatively stable odds ratios (OR) were found for printers and graphics arts workers (OR = 4.5; 95% confidence interval (CI) = 1.4-14.7) and chemical and petroleum workers with high exposure levels (OR = 3.0; CI = 1.1-8.5). A discussion of the biases involved in this type of study design is presented.