THE male-specific portion of the Y chromosome is especially useful for studies of human origins. Patterns of nucleotide variation that are neutral with respect to fitness should permit estimates of when and where ancestral Y chromosomes existed(1). However, variation on the human Y chromosome has been observed to be greatly reduced relative to the autosomes and the X chromosome(2-5). One explanation is that selection for a favourable mutation on the nonrecombining portion of the Y chromosome has resulted in the recent fixation of a single Y haplotype(5,6). A 2.6-kilobase fragment encompassing a polymorphic Alu insertion was sequenced from 16 human and four chimpanzee Y chromosomes. Patterns of nucleotide sequence diversity and divergence provide no evidence for a recent, strong selective sweep on the human Y chromosome. The time back to a common ancestral human Y chromosome is estimated to be 188,000 years, with a 95% confidence interval from 51,000 to 411,000 years. These results are consistent with autosomal and mitochondrial DNA studies that suggest a long-term human effective population size of 10,000 and a sex ratio of 1 (ref. 7). These inferences contradict predictions of the multiregional hypothesis(8) positing a widespread transformation of Homo erectus populations into Homo sapiens.