In this thesis, I studied molecular evolution of the vertebrate genome, focusing on sex chromosomes, protein coding genes, and genome size variation. The evolutionary history of avian sex chromosomes was analyzed by comparing substitution rate among 12 gametologous gene pairs on the Z and W chromosomes. Divergence estimates were distributed into three discrete clusters, evolutionary strata, implying stepwise cessations of recombination. Stratum 3 and stratum 2 are located the intervals 1-11Mb and 16-53Mb on the chicken Z chromosome, respectively. Stratum 1 was located in the middle of stratum 2, suggesting a chromosomal inversion. Using a molecular clock, the estimated times for cessation of recombination between Z and W are 132–150 (stratum 1), 71–99 (stratum 2), and 47–57 (stratum 3) million years ago.Higher divergence rate in the Z chromosome than in autosomes (faster-Z) can be explained by positive selections on recessive alleles in hemizygous females, or by stronger genetic drift due to the smaller effective population size of the Z chromosomes. I found there was no difference in the intensity of the faster-Z effect among male-biased, female-biased, and unbiased genes, as might have been expected under a selection model. This result therefore supports the hypothesis that faster-Z is predominantly due to genetic drift.Next, I analyzed molecular evolution of protein-coding genes in birds. In the comparison of zebra finch, chicken and non-avian outgroups, I found that neutral substitution rate was highest in zebra finch, intermediate in chicken, and lowest in ancestral birds. This difference seems attributable to differences in generation time, ancestral birds being most long-lived. Several functional categories were overrepresented among positively selected genes in avian lineages, such as transporter activity and calcium ion binding. I also found that many genes involved with cognitive processes including vocal learning were positively selected in zebra finch. I also found evidence for Hill-Robertson interference acting against the removal of slightly deleterious mutations at linked loci.Finally, I studied the impact of recombination on genome size variation. I found that highly recombining regions have a more condensed genome structure, including shorter lengths of intron, intergenic spacer, transposable elements and higher gene density. In chicken and zebra finch I found that recombination rate was positively correlated with deletion bias, estimated by sequence comparisons between individual transposable elements (LINEs) and the corresponding master sequences. These observations indicate that the more compact genome structure in highly recombining region is due to a higher rate of sequence loss. Higher deletion bias in autosomes than in sex chromosomes supports this idea. I also found that sequence loss due to the deletion bias can explain nearly 20% of genome size reduction after the split of birds from other reptiles. In human, the recombination rate was positively correlated with the deletion bias estimated from polymorphic indels. These results support the hypothesis that the recombination drives genome contraction via the mutation process.