Low birth weight infants are delivered with increased frequency in women who have elevated maternal serum alphafetoprotein values in the second trimester. Maternal serum alpha-fetoprotein elevations, however, are found more often in lighter-weight women, a group known to have lowerweight infants regardless of maternal serum alphafetoprotein levels. To clarify the association between elevated maternal serum alpha-fetoprotein levels and low birth weight independent of maternal weight, we applied a weight correction formula to maternal serum alphafetoprotein values from 9507 singleton viable pregnancies without major fetal malformations. Before adjusting for weight, 486 of the women (5.1%) had maternal serum alpha-fetoprotein values of 2.0 or more multiples of the median. The weight adjustment process removed 100 lighter-weight women from this category, added 58 heavier women, and led to an equivalent proportion of women in the various weight categories who were classified as having maternal serum alpha-fetoprotein values of 2.0 or more multiples of the median. Of the 388 low birth weight pregnancies (2500 g or less), 50 initially had maternal serum alpha-fetoprotein values of 2.0 or more multiples of the median; after weight adjustment, seven lighter-weight women were removed, four heavier women were added, the median birth weight fell from 2217 to 1956 g, and a threefold increase in risk was found for low birth weight outcome regardless of weight class. Maternal serum alpha-fetoprotein elevations predict increased risk for low birth weight outcome independent of maternal weight.