The transmission/disequilibrium test (TDT) and the mean test are two of the most important methods in linkage analysis. The TDT has the disadvantage that even very close linkage will not be detected by the method if association is not present. On the other hand, the mean test has poor power performance if association is present. Both tests are sensitive to association due to linkage disequilibrium. Although a large number of studies have been made on power comparisons between the TDT and the mean test, little is concerned with the statistical relationship between the two statistics. In this paper, the conditional probabilities of transmission-nontransmission genotypes of a heterozygous parent in a case-parent tetrad family (two parents and two affected children) are derived. Based on these probabilities we theoretically show that the TDT and the mean test are valid for test of linkage. We then show that the two test statistics are uncorrelated and asymptotically independent, though they are dependent.