The variance effective size (Ne) was formulated for populations of monoecious plant species that are partly asexually propagating with discrete or overlapping generations. It was shown that partly asexually reproducing populations have larger or smaller effective sizes (ratios to the census size N) than fully sexually reproducing populations, according to whether the term Vc/c¯ is smaller or larger than the term (Vk/k¯ +1−β)/2, where c¯ and Vc are the mean and variance of the number of progeny asexually produced per plant per year, respectively, k¯ and Vk are the mean and variance of the number of gametes contributed per plant per year, respectively, and β is the selfing rate of each plant. Asexual reproduction has no effect on Ne when the two terms are equal, as is true when the numbers of both sexually and asexually produced progeny per plant per year are Poisson-distributed (Vc/c¯=1 and Vk/k¯=1+β). Populations with a larger generation length (L) tend to have a smaller effective size: for a population model of age-independent survival and fecundity with an annual rate δ of asexual reproduction, Ne declines asymptotically to N(2−β)/{3−β +Vk/k¯ +(2Vc/c¯ −Vk/k¯ −1+β).δ} as L gets large, which simplifies to N(2−β)/4 under a Poisson-distributed reproductive contribution. The trade-off relation of Ne and L, however, does not always hold: for stage-structured populations, increase in the survival rate of juveniles may act to increase both Ne and L.