Breeding Systems and Pollen-Ovule Ratios inEricaSpecies (Ericaceae) of the Cape Floristic Region

Abstract

Premise of research. Breeding systems dictate the potential for self-fertilization in plants and may therefore have a strong influence on evolutionary and ecological processes. Previous investigations of breeding systems in Ericaceae indicate variation in self-sterility, but studies from the megadiverse genus Erica are almost entirely lacking. Here we quantify the degree of self-sterility and autofertility in Cape Erica species, quantify pollen-ovule ratios (P/Os), and test whether these are related to each other and vary across species with bird- versus insect-pollination syndromes.Methodology. We used controlled hand-pollinations, analyses of pollen tube development, and P/Os to investigate the breeding systems of 15 taxa (comprising 13 species) of Erica in the Cape Floristic Region of South Africa, spanning most main clades across the phylogeny.Pivotal results. In all but one species, seed set from autonomous selfing was almost absent and significantly lower than that from cross-pollination, and only one species had an autofertility index that was higher than 0.2. In 12 taxa, seed set following self-pollination was significantly reduced compared with that following cross-pollination, and the index of self-incompatibility (ISI) was higher than 0.8 for 10 taxa. Following self-pollination, pollen tubes reached the ovary in all species, and for two species with very low seed set in selfed flowers, fluorescence images revealed that self-pollen tubes entered into the ovules. P/Os ranged from 68 to 320 and are therefore low for obligate outcrossing species. These ratios did not correlate with the ISI and did not vary according to whether species conformed to insect- or bird-pollination syndromes.Conclusions. This study indicates that self-sterility is frequent in Cape Erica species and either involves a late-acting mechanism of self-incompatibility or is due to early inbreeding depression. P/Os in these species are consistently low and do not reflect variation in breeding systems and pollination syndromes.