Studies on a base population of mice were used to establish an index of components of litter size and a physiological model for measuring uterine capacity to be used subsequently in a selection experiment evaluating alternative methods for practicing selection to increase litter size. Heritability estimates of litter size, ovulation rate and ova success (fraction of ova resulting in fully formed pups) were .18, .33 and .15, respectively. No significant genetic or phenotypic correlation was found between overall ovulation rate and ova success. Phenotypic means and genetic variances were higher for characteristics measured on the right than on the left side of the reproductive tract. Linear and quadratic selection indexes, derived for a quadratic definition of breeding value, were compared. The linear index was predicted to be .99 as efficient as the quadratic one. Due to simplicity, the linear index (I = 1.21 x ovulation rate + 9.05 x ova success), scaled to have variance the same as litter size, was chosen for use. Ovulation rate in unilaterally ovariectomized females was .95 of that in females with both ovaries. No hypertrophy of the ipsilateral uterine horn in unilaterally ovariectomized females was found before implantation of embryos. Thus, unilateral ovariectomy appears to provide a physiological state to measure uterine capacity (as litter size) in the mouse.