In order to better understand somaclonal variant rate evolution in plant tissue culture, a statistical approach has been adopted. According to this approach, the variant percentage could be calculated by: %V = [1 - (1 - p)n] X 100, where %V is the percentage of variant, p the probability of variation and n the number of multiplication cycles. A numerical estimation was performed to characterize the variance of this function. It has been demonstrated that a wide scale of variance is associated with '%V', due to the occurrence of variations after a variable number of multiplication cycles in the different lines of culture. Two main conclusions can be drawn from this model: (1) a variant rate increase can be expected as an exponential function of the number of multiplication cycles; (2) after a given number of multiplication cycles, variable off-types percentages can be expected. Due to the complexity of biological systems, this statistical approach could obviously not be applied directly for the calculation and forecasting of variant rates in tissue culture. However, this approach results in a better understanding of two apparently confusing experimental features often reported in tissue culture: the increase of the variant rate as a function of the length of the culture period on the one hand, and, on the other hand, the observations of different variant rates among lines cultured for the same lengths of time under strictly identical culture conditions. This approach also underlined that the comparison of somaclonal variant percentage between batches of plants from different in vitro treatments could be, in some cases, insufficient for ascertaining a difference of variability generated by tissue culture.