It was assumed that the frequency of yeast mutants with respiratory deficiency is proportional to the gene size Tzagoloff and Dieckmann's 1990 data giving the number of obtained mutants and numerousness of complementation groups were used in this paper. In the mathematical model these mutations are treated as a distribution of balls in the cells in its nonclassical version with a differentiated probability of cell choice. It has been proved that statistical data reject the stable gene size and the probability distribution function of thre relative gene size has been estimated. The achieved distribution is characterized by a large asymmetry. The expected number of mutants unknown up to now assumed to be of the same minimal size has been also introduced.