PIP: The relationship between social class and fertility in Madras, India, is described by comparing the prevalence of sterility and duration of marriage among women in different classes. Differences between social classes in generational replacement is also explored. 463 families were selected from 3 city districts, yielding a satisfactorily representative sample. Complete data were available for 402 families in unbroken marriages. The number of children ever born ranged from 0-19, with a mean of 3.5 births. For the lower class (determined by the respondent's self assessment) the mean is 4.04 births, for the middle class, 3.19, and 3.29 for the upper class. Fertility of the lower class is significantly higher than that for the other 2 classes. When fertility is standardized by age, the total fertility of lower class women is 46% higher than that of upper class women and 26% higher than that of the middle class. The effect of sterility on fertility seems to be minor. After age adjustments, 8% of lower class women, 15% of middle class women, and 15% of upper class women are sterile. However, fertility of the middle class is 16% higher than that of the upper class, and in older age groupings where sterility differences are imperceptible, fertility of the lower class exceeds that of the middle class by 10% and the upper class by 50%. Marriage duration differs by social class: the median for the lower class is 13.8 years, 12.5 years for the middle class, and 18 years for the upper class. Controlling for differences in duration, fertility is inversely related to social class, with the exception of durations of 5 years or less. Total child mortality is inversely associated with social class: the standardized mean is 1.15 for the lower class, .82 for the middle class and .32 for the upper class. Equal mortality rates are observed for boys and girls. Forwomen of all ages, the standardized average number of surviving children is 3.21 for the lower class, 2.73 for the middle class, and 2.55 for the upper class. The number of surviving girls is inversely associated with social class. The general inverse correlations between social class and both fertility and child mortality seem to cooperate to keep the reproduction of the social classes in balance with their number in the contemporary generation.