Twin births are known to vary across seasons in several countries. It has been hypothesized that this variation may be due to seasonal changes in luminosity leading to pineal gland-mediated multiovulation among susceptible mothers. To describe seasonal variation of twin births in Washington State, all mothers residing in Washington State who gave birth to both a pair of twins and a singleton baby between 1984-1990 (n = 1168) were identified through linkage of computerized State birth certificates. Using a “matched-on-mother” case-control design, the estimated month of conception of twin gestations (the “case” events) were compared to that for their singleton siblings (the “control” events) to determine their relative occurrence during periods of high vs low sunlight in accordance with local climatological data. For the study population as a whole, there was only a slight tendency for twins to have been conceived during the period of high sunlight compared to their singleton siblings (OR = 1.3, 95% C.I. = 1.0-1.7). When stratified by concordant-sex vs discordant-sex, however, more discordant-sex twin pairs were conceived during the light period than corresponding singletons (OR = 1.7, 95% C.I. = 1.0-2.8), whereas no association was found for concordant-sex twins (OR =1.1, 95% C.I. =0.8-1.6). The presence of an association only among discordant-sex twins, all of whom are dizygotic, is consistent with the hypothesis that exposure to sunlight may stimulate multiple ovulation, and thus increase the incidence of twin gestations among twin-prone mothers.