A large number of historical documents, proverbs, folklore and taboos show the existence of a strong preference for sons in Korea (Lee, 1973; Kim, 1969, pp. 218-374). As a result of the social and economic advantages attached to males, son preference is embedded in the traditional social organization of Korea, particularly through the kinship system and related normative culture (Kwon and Lee, 1976; Cho et al., 1982). Gender preference was so marked in traditional Korea that a wife bearing no son could be abandoned by her husband or rejected by his family. Despite rapid socioeconomic changes, son preference persists in Korea as an institutionalized value.The desire to have sons plays an important role in determining family size. Some studies have pointed out that strong son preference might impede efforts to attain low fertility (Park, 1983; Arnold, 1985; Arnold and Liu, 1986). However, this concern has proved to be groundless. In spite of a strong son preference, Korea has achieved a fertility rate way below the replacement level (1.3 in 2001), as have a number of eastern Asian countries such as China, Taiwan, and Hong Kong (UN, 2002).In the second half of the 1980s and the early 1990s, however, to accommodate both a strong son preference and low fertility, a new demographic phenomenon emerged in Korea: the rise of the male/female sex ratio at birth (i.e. excess of male births relative to female). This phenomenon of missing girls was also observed in China, Taiwan, and Hong Kong (Roy, 1994; Park and Cho, 1994).The main objective of this study is to understand the origins and the mechanisms underlying the recent increase in the sex ratio at birth in Korea. We will also examine the pattern of changing trends and regional differentials in sex ratios from 1980 to 2003. Causal factors and implications of son-selective reproductive behaviour are discussed. Two demographic simulations were conducted to show how prenatal sex screening and sex-selective abortion raise the sex ratio at birth, and, at the same time, play a role in lowering the level of fertility.I. Recent trends in sex ratioThe mid-1980s can be recorded as an important turning point in the history of Korean demography. Korea has completed the full pattern of the demographic transition, and has been approaching a stationary population with negative growth potential in the offing (Kim, 1992, 2004). Despite the cultural setting of strong son preference, rapid industrialization and urbanization have led to a widespread preference for small families (Kwon and Kim, 2002). The total fertility rate (TFR) decreased sharply from about 6.0 in the early 1960s to the replacement level (2.1 children per women) in the mid-1980s and has now dropped way below the replacement level (1.19 in 2003). The crude birth rate was estimated to be as low as 10.2 in 2003. The annual growth rate of the population was estimated at 0.57% in 2003, as low as those of developed countries (KNSO, 2004). From 2020, the population of Korea is projected to decline (under the current age structure) (KNSO, 2004).With the decline in fertility to below the replacement level in the mid1980s, distortions began to emerge in the sex composition of Korean society . This new demographic phenomenon has become a serious issue. Generally speaking, the sex structure of a population is determined by three main components: sex ratio at birth, sex differentials in mortality and sexselective migration. As regards migration, it is generally agreed that the Korean population, since the late 1950s, can be regarded as an almost closed population'^1'. Furthermore, sex differentials in child mortality are not observed in Korea (Kim, 1997b). Thus, the juvenile sex ratio has basically become a function of the sex ratio at birth (Kwon et al., 1975; Kim, 1992, 2004; Park and Cho, 1994).In most countries, sex ratios at birth vary normally between 102107 males for 100 females (UN, 2002). …