This paper examines the roles of primary subsistence base and women's contributions to food production in the etiology of polygyny. Cross-cultural data from the Ethonographic Atlas are employed. Polygyny is differentiated into categories of absent (monogamy only), occasional (less than 20% of all marriages), and frequent (20% or more of all marriages). The results show that polygyny is very likely to be allowed, but infrequently practiced, in exploitative and incipient agriculture economies. It is most likely to be widespread in herding and extensive agriculture economies, but decreases in" frequency as agricultural technology develops further. These patterns are explained by variation in the roles of men and women in subsistence production.in combination with variation in the labor-intensiveness of subsistence activities.