The frequency of low birth weight (less than or equal to 2500 g) and of preterm birth (less than 37 weeks) was studied in 22,761 single live births in relation to maternal employment, taking account of 11 nonoccupational confounding factors. There was some increase of low statistical significance in both types of prematurity in service and manufacturing sectors of industry. A substantial excess of preterm births was seen in women employed in food and beverage service (O/E = 1.29, p = 0.03) and psychiatric nursing (O/E = 2.47, p less than 0.01) and of low birth weight in food and beverage service (O/E = 1.30, p = 0.02), in chambermaids and cleaners (O/E = 1.42, p = 0.03), and in those employed in the manufacture of metal and electrical and certain other goods (O/E = 1.57, p less than 0.01). Heavy lifting and long hours of work were consistently related to both outcomes, changing shift work less consistently. Noise was associated with low birth weight in the health and manufacturing sectors. The findings of this study are unlikely to have resulted from subject or observer bias but the role of unidentified factors related to selection for work are difficult to assess.