Mixed model nested ANOVAs comparing mean temperature-adjusted pollen performance traits between the sister species exilis and unguiculata following cross- and self-pollinations. In these analyses, p-values < 0.05 are associated with statistically significant fixed effects. We used the restricted maximum likelihood (REML) method to estimate variance components for the random effects of Pollen Recipient and Pollen Donor. Random effects associated with variance components whose 95% confidence intervals (CI) do not include zero are statistically significant and indicated in bold italics.