In humans, deviations from a 1:1 male:female ratio have been identified in both chromosomally normal and trisomic live births: among normal newborns there is a slight excess of males, among trisomy 18 live borns a large excess of females, and among trisomy 21 live borns an excess of males. These differences could arise from differential production of or fertilization by Y- or X-bearing sperm or from selection against male or female conceptions. To examine the proportion of Y- and X-bearing sperm in normal sperm and in sperm disomic for chromosomes 18 or 21, we used three-color FISH (to the X and Y and either chromosome 18 or chromosome 21) to analyze >300,000 sperm from 24 men. In apparently normal sperm, the sex ratio was nearly 1:1 (148,074 Y-bearing to 148,657 X-bearing sperm), and the value was not affected by the age of the donor. Certain of the donors, however, had significant excesses of Y- or X-bearing sperm. In disomy 18 sperm, there were virtually identical numbers of Y- and X-bearing sperm; thus, the excess of females in trisomy 18 presumably is due to selection against male trisomic conceptions. In contrast, we observed 69 Y-bearing and 44 X-bearing sperm disomic for chromosome 21. This is consistent with previous molecular studies, which have identified an excess of males among paternally derived cases of trisomy 21, and suggests that some of the excess of males among Down syndrome individuals is attributable to a nondisjunctional mechanism in which the extra chromosome 21 preferentially segregates with the Y chromosome.