Although geographic variation in semen quality has been reported, this is the first study in the United States to compare semen quality among study centers using standardized methods and strict quality control. We evaluated semen specimens from partners of 512 pregnant women recruited through prenatal clinics in four U.S. cities during 1999-2001; 91% of men provided two specimens. Sperm concentration, semen volume, and motility were determined at the centers, and morphology was assessed at a central laboratory. Study protocols were identical across centers, and quality control was rigorously maintained. Sperm concentration was significantly lower in Columbia, Missouri, than in New York, New York; Minneapolis, Minnesota; and Los Angeles, California. Mean counts were 58.7, 102.9, 98.6, and 80.8 X 10(6)/mL (medians 53.5, 88.5, 81.8, and 64.8 X 10(6)/mL) in Missouri, New York, Minnesota, and California, respectively. The total number of motile sperm was also lower in Missouri than in other centers: 113, 196, 201, and 162 X 10(6) in Missouri, New York, Minnesota, and California, respectively. Semen volume and the percent morphologically normal sperm did not differ appreciably among centers. These between-center differences remained significant in multivariate models that controlled for abstinence time, semen analysis time, age, race, smoking, history of sexually transmitted disease, and recent fever (all p-values < 0.01). Confounding factors and differences in study methods are unlikely to account for the lower semen quality seen in this mid-Missouri population. These data suggest that sperm concentration and motility may be reduced in semirural and agricultural areas relative to more urban and less agriculturally exposed areas.