be a way of establishing the property rights of the conjugal couples that make up larger house holds, in anticipation of eventual fission. In addition, it allows for status negotiation without either family being put in the others economic or social debt (cf Schlegel and Eloul 1988 There are variations within these major types, and there are additional features(such as the European dower) that are secondary and limited in distribution. In complex societies, the form of transaction may vary according to region or class. In prerevolutionary China, for example the landed or mercantile elite gave dowry while the landless peasantry gave indirect dowry whereas in modern China, marriage transactions have disappeared from urban areas while bridewealth has replaced indirect dowry among peasants(Fang 1990). When the forms differ by status, the preferred form, practiced by the elite, is the one considered here. 2 marriage transactions and the value on virginity Information on attitudes toward premarital sex for females, or the value placed on virginity, comes from two sources. the primary one is the code"Attitude Toward Premarital Sex(Fe male)"in Broude and Greene(1980). Using the Standard Sample of 186 preindustrial societies Broude and greene found information on this subject for 141 societies. Their code is divided into six levels of value: (1) premarital sex expected ( 2)premarital sex tolerated; (3)premarital sex mildly disapproved of but not punished; (4) premarital sex moderately disapproved of and ightly punished; (5) premarital sex disallowed except with bridegroom; and (6)premarital sex strongly disapproved of. For the present study, the first three categories were collapsed into virginity not valued"and the second three into " virginity valued "i have made four alterations to the code based on my own reading of the ethnographic literature: I have changed the coding for the Burmese from 3 to 5(Spiro 1977), for the Tikopia from 3 to 4( Firth 1936), for the Koreans from 3 to 6(Osgood 1951), and for the yurok from 2 to 4( Kroeber 1925). The societies are The second source is a body of data collected by herbert Barry and me on adolescent so cialization in Standard Sample societies not coded by Broude and Greene. The data were col- lected on adolescent behavior, not cultural attitude; coders were asked to assess whether pre- marital sex was or was not tolerated because the code is less detailed than broude and Greene's and because it measures behavior rather than attitude, I offer information only on societies in which premarital sex is not tolerated and thus, by definition, virginity is valued Presumably, there can be cases where premarital sex is moderately disapproved of and pun- ished when discovered, but most adolescent girls take the risk and indulge anyway.)These societies are indicated in Table 1 in parentheses but have not been included in the tests, for to do so would introduce a bias toward the set of societies valuing virgil It is clear that the value on ity is not randomly distributed among societies with all types of marriage transaction. Table 2 shows the distribution, which is statistically significant: p< 0001. Even when those societies without marriage transactions are eliminated, the distribution is still statistically significant: p <0O1 Others have also found significant associations between premarital sexual permissiveness and structural or cultural features. In studies by Murdock (1964), Goethals(1971),Eckhardt (1971), Paige(1983), Barry and Schlegel (1986), and others(see Broude's [19811 summary) sexual permissiveness is shown to be associated with the simpler subsistence technologies absence of stratification smaller communities, matrilineal descent matrilocal residence ab- sence of belief in high gods, absence of bridewealth (but bear in mind that in earlier studies bridewealth has been conflated with indirect dowry), high female economic contribution, little or no property exchange at marriage and ascribed rather than achieved status. These features are all highly intercorrelated, and some correlate with types of marriage transactions(Schlegel and Eloul 1988be a way of establishing the property rights of the conjugal couples that make up larger house￾holds, in anticipation of eventual fission. In addition, it allows for status negotiation without either family being put in the other's economic or social debt (cf. Schlegel and Eloul 1988). There are variations within these major types, and there are additional features (such as the European dower) that are secondary and limited in distribution. In complex societies, the form of transaction may vary according to region or class. In prerevolutionary China, for example, the landed or mercantile elite gave dowry while the landless peasantry gave indirect dowry, whereas in modern China, marriage transactions have disappeared from urban areas while bridewealth has replaced indirect dowry among peasants (Fang 1990). When the forms differ by status, the preferred form, practiced by the elite, is the one considered here.2 marriage transactions and the value on virginity Information on attitudes toward premarital sex for females, or the value placed on virginity, comes from two sources. The primary one is the code "Attitude Toward Premarital Sex (Fe￾male)" in Broude and Greene (1980). Using the Standard Sample of 186 preindustrial societies, Broude and Greene found information on this subject for 141 societies. Their code is divided into six levels of value: (1) premarital sex expected; (2) premarital sex tolerated; (3) premarital sex mildly disapproved of but not punished; (4) premarital sex moderately disapproved of and slightly punished; (5) premarital sex disallowed except with bridegroom; and (6) premarital sex strongly disapproved of. For the present study, the first three categories were collapsed into "virginity not valued" and the second three into "virginity valued." I have made four alterations to the code based on my own reading of the ethnographic literature: I have changed the coding for the Burmese from 3 to 5 (Spiro 1977), for the Tikopia from 3 to 4 (Firth 1936), for the Koreans from 3 to 6 (Osgood 1951), and for the Yurok from 2 to 4 (Kroeber 1925). The societies are listed in Table 1. The second source is a body of data collected by Herbert Barry and me on adolescent so￾cialization in Standard Sample societies not coded by Broude and Greene. The data were col￾lected on adolescent behavior, not cultural attitude; coders were asked to assess whether pre￾marital sex was or was not tolerated. Because the code is less detailed than Broude and Greene's and because it measures behavior rather than attitude, I offer information only on societies in which premarital sex is not tolerated and thus, by definition, virginity is valued. (Presumably, there can be cases where premarital sex is moderately disapproved of and pun￾ished when discovered, but most adolescent girls take the risk and indulge anyway.) These societies are indicated in Table 1 in parentheses but have not been included in the tests, for to do so would introduce a bias toward the set of societies valuing virginity. It is clear that the value on virginity is not randomly distributed among societies with all types of marriage transaction. Table 2 shows the distribution, which is statistically significant: p < .0001. Even when those societies without marriage transactions are eliminated, the distribution is still statistically significant: p < .001. Others have also found significant associations between premarital sexual permissiveness and structural or cultural features. In studies by Murdock (1964), Goethals (1971), Eckhardt (1971), Paige (1983), Barry and Schlegel (1986), and others (see Broude's [1981] summary), sexual permissiveness is shown to be associated with the simpler subsistence technologies, absence of stratification, smaller communities, matrilineal descent, matrilocal residence, ab￾sence of belief in high gods, absence of bridewealth (but bear in mind that in earlier studies, bridewealth has been conflated with indirect dowry), high female economic contribution, little or no property exchange at marriage, and ascribed rather than achieved status. These features are all highly intercorrelated, and some correlate with types of marriage transactions (Schlegel and Eloul 1988). the value on virginity 721