Microarray gene expression data provide a wealth of information for elucidating the mode and tempo of molecular evolution. In the present study,we analyze the spatial expression pattern of human duplicate gene pairs by using oligonucleotide microarray data,and study the relationship between coding sequence divergence and expression divergence. First,we find a strong positive correlation between the proportion of duplicate gene pairs with divergent expression (as presence or absence of expression in a tissue) and both synonymous (KS) and nonsynonymous divergence (KA). The divergence of gene expression between human duplicate genes is rapid, probably faster than that between yeast duplicates in terms of generations. Second,we compute the correlation coefficient ® between the expression levels of duplicate genes in different tissues and find a significant negative correlation between R and KS. There is also a negative correlation between R and KA,when KA ≤ 0.2. These results indicate that protein sequence divergence and divergence of spatial expression pattern are initially coupled. Finally,we compare the functions of those duplicate genes that show rapid divergence in spatial expression pattern with the functions of those duplicate genes that show no or little divergence in spatial expression.