The relationship between the pollen–ovule ratio and pollen size: another comparative test of a sex allocation hypothesis

Hypothesis: Sex allocation theory predicts that the pollen–ovule ratio should decrease linearly with increasing pollen size among seed plants (Charnov, 1982). Data examined: We retrieved data for pollen–ovule ratio, pollen size, pollen grain number, and mating system from published literature for 311 angiosperm plant species. Methods: We used model II regressions on cross-species data as well as on phylogenetically independent contrasts (PIC) to quantify the relationship between the pollen–ovule ratio and pollen size. Partial correlations were applied to test if an association between these two traits arises because of a correlation with a third variable, the number of pollen grains. Results: A linear negative correlation between the pollen–ovule ratio and pollen size does exist for these plant species, both in phylogenetically corrected and uncorrected data. However, the correlation was not consistently found at the taxon and mating system levels. For virtually all groups investigated, the correlation disappeared when we controlled for the effect of pollen grain number. Thus the correlation between the pollen–ovule ratio and pollen size is spurious. Conclusions: Considering male function, the pollen–ovule ratio depends on the number of pollen grains produced by a flower but not on the size of the pollen grains. For the ‘male part’ of Charnov’s model, its validity can be called into question.