Comparison of Kullback's information on nuclear pedigrees with different numbers of offsprings was performed in a case when a marker gene was codominant and the recessive or dominant control with complete penetrance was supposed for the localized gene. This comparison resulted in some conclusions about optimization of the sample collection. These conclusions are found to be invariant with respect to the recombination fraction value determined by the simple alternative hypothesis, to the population frequencies of the genes considered and, within certain limits, to the disequilibrium parameter values.