The impact of ovule number on seed set calculations for self-incompatible (SI) species was investigated. Diploid Chrysanthemum was chosen for this study because accurate counts of the potential number of ovules could be made. Individuals in populations of C. carinatum, C. coron- arium, C. c. subsp, spatiosum, and C. segetum were crossed in complete diallels. All species exhibited similar results. Therefore, only the diallel data from C. coronarium subsp, spatiosum were presented. The seed set data with and without ovule counts were processed by SIGMAS, a computer program designed to analyze SI data. Incorporation of the actual number of ovules into seed set diallels provided the most realistic representation of values for self-incompatibility studies. Data derived from equations excluding ovule counts might lead to inaccurate genetic interpretations. Ovule counts were significant between and within genotypes for self (disc and ray florets), but not cross (ray florets only) pollinations. The disc florets in self-pollinations were found to be responsible for increasing the variability in ovule number. The statistics indicate that the disc and ray florets composed two distinct populations. At the diploid level with a single daisy flower type, the disc floret numbers were variable, whereas ray florets were relatively static. This was not the case with polyploid chrysanthemums, where both ovule populations were dynamic and interactive. The conservative nature of percent pseudo-self-compatibility (% PSC) deems it necessary to obtain an accurate measure of female fertility. Values for this could be obtained using a bulk pollination or a tester with unmatched S alleles.