We use the proportion of individuals with the recessive phenotype to calculate the frequency of the recessive allele (q). Why can't we calculate the frequency of the dominant allele in the same way, that is by using the proportion of the individuals with the dominant phenotype?
