Spontaneous mutation provides the substrate for evolution on one hand and for genetic susceptibility to disease on the other hand. X-linked diseases such as hemophilia B offer an opportunity to examine recent germ-line mutations in humans. By utilizing the direct sequencing method of genomic amplification with transcript sequencing, eight regions (2.46 kb) of likely functional significance in the factor IX gene have been sequenced in a total of 60 consecutive, unrelated hemophiliacs. The high frequency of patient ascertainment from three regions in the midwestern United States and Canada suggests that the sample is representative of hemophiliacs of northern European descent. Twenty-six of the delineated mutations are reported herein, and the group of 60 is analyzed as a whole. From the pattern of mutations causing disease and from a knowledge of evolutionarily conserved amino acids, it is possible to reconstruct the underlying pattern of mutation and to calculate the mutation rates per base pair per generation for transitions (27 x 10(-10)), transversions (4.1 x 10(-10), and deletions (0.9 x 10(-10)) for a total mutation rate of 32 x 10(-10). The proportion of transitions at non-CpG nucleotides is elevated sevenfold over that expected if one base substitution were as likely as another. At the dinucleotide CpG, transitions are elevated 24-fold relative to transitions at other sites. The pattern of spontaneous mutations in factor IX resembles that observed in Escherichia coli when the data are corrected for ascertainment bias. The aggregate data hint that most mutations may be due to endogenous processes. The following additional conclusions emerge from the data: (1) Although in recent decades reproductive fitness in individuals with mild and moderate hemophilia has been approximately normal, the large number of different mutations found strongly suggest that these levels of disease substantially compromised reproduction in previous centuries. (2) Mutations which putatively affect splicing account for at least 13% of independent mutations, indicating that the division of the gene into eight exons presents a significant genetic cost for the organism. In one individual a "silent" mutation at lysine 5 is likely to cause hemophilia by generating a perfect splice donor consensus sequence in exon b. (3) All the missense mutations occurred at evolutionarily conserved amino acids. As additional data are generated on the pattern of mutations caused by specific mutagens, it will be possible to utilize the pattern of spontaneous mutation to estimate the maximal contribution of that mutagen during the past century.