The effects of social class on birthweight, and its interactions with other maternal factors, were examined in groups of women bearing small-for-dates (SFD), average-for-dates (AFD) and large-for-dates (LFD) babies. The relative risk of a lower social class woman having an SFD baby steadily decreased from 1.75 to 1.20 as adjustment was cumulatively made for smoking, hypertension, maternal age and height. The subsequent addition of weight and weight-for-height made little change. The unadjusted risk of a lower social class woman having a LFD baby was very close to unity (0.99). Adjustment for other maternal factors, in the same order as at the other extreme, showed a steady rise to a significant level (1.45) when height was included; but there was a sharp reversal to a non-significant risk of 1.12 when weight was added. The large contribution of obesity in the lower classes seems to counterbalance those for height, age and smoking in the upper classes in the LFD group. The interactional effects of other maternal factors and social class are not operating to an equal and opposite degree at the two extremes of the birthweight range.