Using data from the Oxford Record Linkage Study (ORLS), we conducted a case-control study to estimate the sex-specific risks of inguinal hernia in siblings of children with this condition. There were 1921 male and 347 female cases born during 1970-86 who were operated on for inguinal hernia at ages 0-5 years during 1970-87, with 12,886 male and 2534 female control subjects. The relative risk of inguinal hernia was 5.8 [95% confidence interval 4.0-8.4] for brothers of male cases and 4.3 [2.1-8.7] for brothers of female cases (both relative to brothers of control subjects). The relative risk was 3.7 [1.8-7.9] for sisters of male cases and 17.8 [6.9-46.3] for sisters of female cases (both relative to sisters of control subjects). The pattern of sex-dependent risks, particularly the large risk for sisters of female cases, suggests a multifactorial threshold model for the disease. Girls have much lower rates of inguinal hernia than boys, and if these rates are low because of a low susceptibility to disease due to the absence of a sex-related risk factor, then those girls who develop disease might have a potentially large contribution to susceptibility from genetic or intrauterine risk factors unrelated to sex.