has diminished and, starting in 1960, marital disruption has slightly increased in the region.
The findings of diminishing celibacy and no trend toward later marriage are against the expectation that demographic strains might produce 'preventive checks' in marriage patterns. The increasingly rapid demographic growth in the post-war years did not produce the multiphasic response in nuptiality postulated by Davis. These results are also contrary to the hypothesis that modernization brings about, more or less mechanically, delays in age at marriage in developing countries. Moreover, they suggest that nuptiality and fertility are influenced by quite different sets of factors.
Given that Latin America has undergone an important process of urbanization and modernization, the pattern of a more or less constant age at marriage suggests that the entry into unions depends more on cultural factors than on socio-economic change. This is somewhat puzzling given the wide socio- economic differentials in marriage patterns documented by the WFS and other studies in Latin America and elsewhere ( Casterlineet al. 1984). The combination of these differentials with a modernization process should produce an upward trend in age at marriage. The lack of such a trend in Latin America might indicate that the socio-economic differentials do not denote a causal relationship. It also suggests that within some socio-economic strata, age at marriage might have declined.
A final finding that should be pointed out is the upward trend in the prevalence of consensual unions. Since the relationship between type of union and fertility is complex and poorly understood ( Merrick 1986), the demographic effect of this rise is hard to disentangle. What does seem clear is that a surge in illegitimate births is probably under way in Latin America.
Assuming no out-of-union births, the following relation links fertility rates (f) with marital fertility rates (g) and the proportion in union (m):
f = mg (1)
which results in the following first-order difference equation for changes from time 0 to 1:
Δf = g0 Δm + m0 Δg + Δm Δg (2)
where the operator Δ represents changes over time
Δf = f1 - f2; Δm = m1 - m0; Δg = g1 - g0. (3)
The three terms at the right hand side of equation (2) decompose fertility change into that resulting from nuptiality change alone, marital fertility change alone, and the joint contribution of both factors, respectively. The third term in the equation--the joint