Sex chromosomes are subject to sex-specific selective evolutionary forces 1,2 . One model predicts that genes with sexbiased expression should be enriched on the X chromosome 2-5 . In agreement with Rice's hypothesis 3 , spermatogonial genes are over-represented on the X chromosome of mice 6 and sex-and reproduction-related genes are over-represented on the human X chromosome 7,8 . Male-biased genes are under-represented on the X chromosome in worms and flies 9-11 , however. Here we show that mouse spermatogenesis genes are relatively underrepresented on the X chromosome and female-biased genes are enriched on it. We used Spo11 -/-mice blocked in spermatogenesis early in meiosis 12 to evaluate the temporal pattern of gene expression in sperm development. Genes expressed before the Spo11 block are enriched on the X chromosome, whereas those expressed later in spermatogenesis are depleted. Inactivation of the X chromosome in male meiosis may be a universal driving force for X-chromosome demasculinization.To investigate the genomic distribution of mouse genes with patterns of sex-biased expression, we analyzed the chromosomal distribution of genes expressed in sexually dimorphic tissues. We consider genes that are preferentially expressed in such tissues as testis, ovary and placenta to be sex-biased genes. We used two approaches to define the tissue specificity of genes expressed in mice: analysis of large-scale microarray-based gene expression data and analysis of the distribution of expressed-sequence tags (ESTs) in different cDNA libraries. Instead of focusing on transcripts restricted to a single tissue, our analysis targeted the identification of tissue-enriched transcripts using a 'preferential expression measure' (PEM) 13 . As an uniform source of expression data, we used publicly available data sets of mouse RNA samples representing 49 tissues hybridized to Affymetrix mouse U74A microarrays (GNF data set) 14 and 20 tissues analyzed with RIKEN cDNA microarrays (RIKEN data set) 15 .In the GNF data set 14 , expression of 624 genes (∼7% of all analyzed) in testis was three times greater than the median expression in other tissues, and expression of 361 genes in testis was five times greater (PEM values >1.58 and >2.32, respectively). These genes are considered testis-enriched genes. The chromosomal distribution of testisenriched genes was significantly different from uniform (χ 2 = 30.8, P = 0.05; Fig. 1a and Supplementary Table 1 online). The most significant deviation from the average gene density was on the X chromosome, where testis-enriched genes were 2.5 times less dense (the ratio of the density of testis-enriched genes on the X chromosome to the average density on all chromosomes, ρ X , was 0.40; χ 2 = 7.7, P = 0.006). The depletion of testis-enriched genes on the X chromosome was independent of the cut-off (three or five times greater expression) used in the selection process (Table 1). In addition, genes that were expressed most highly in testis and all the genes detected in testis were underrepresen...