The effect on birth outcome of work requiring different degrees of physical exertion was examined among 15,786 pregnant women who were followed through the Guatemalan Social Security Institute's hospital. Work inside and outside the home was ascertained through a questionnaire administered to each women before delivery. Odds ratios were adjusted for household income, maternal height and age, and birthweight of previous infant. Women with three or more children and no household help were at increased risk for small-for-gestational-age (SGA) births compared with women with family (odds ratio (OR) 1.79; 95% confidence interval (CI) 1.31, 2.47) or hired help (OR 2.0; 95% CI 1.16 to 3.33). Compared with office work, manual work increased the risk for an SGA (OR 1.32; 95% CI 1.12 to 1.56) and SGA/preterm birth (OR 2.56; 95% CI 1.10 to 5.96). Work in a standing compared with sitting position significantly increased the risk for a preterm birth (OR 1.56; 95% CI 1.04 to 2.60). There was a significant positive trend in frequency of SGA and SGA/preterm birth with an increase in the physical demands at work, as measured by an activity score. These data suggest that interventions to reduce physical exertion among pregnant women could improve birth outcome.