In an ideal infinitely large population without mutation, migration, and selection, gene and genotype frequencies remain unchanged over generations. However in finite real population, gene frequencies fluctuate randomly over generations due to finite sampling of gametes therefore the census population cannot explain the effects of inbreeding and allele frequency drift. The concept of effective population size is considered a fundamental measure of representiveness of a sample of individuals (plants, animal, humans) in evolutionary, quantitative genetics, breeding, and genetic resources conservation. Genetic models based on the number of male and female gametes contributed by individuals of monoecious plant species have been developed based on variance effective population size applied to the context of artificial selection (Vencovsky, 1978) and to specific aspects of genetic resources conservation (i.e., collection and regeneration) (Vencovsky, 1978; Crossa and Vencovsky, 1994). Crossa and Vencovsky (1997) and Vencovsky and Crossa (1999) showed the theoretical developments and practical applications of the variance effective population size when drift occurs at two stages, (1) sampling parents for reproduction and (2) sampling gametes (offspring) from those parents for monoecious populations and for populations under mixed self and random mating. For the case of several subpopulations, variance effective population size becomes dependent on the allelic diversity among populations and the number of subpopulations. When a single population is considered and seeds sampled have a family structure, at the limit variance effective population size depends only on the number of seeds parents and the coancestry coefficient among sibs within a family. Accession regeneration is the case where reference population is finite and gametic control is a major factor allowing to recover loss up to to 20% of seeds.