Habemus seeds but they are non-viable: The importance of assessing seed viability in seed rain

Abstract

Seed rain is critical for forest recovery, especially in abandoned lands that have been converted to agriculture. However, such a critical role depends on seed viability, which has been largely overlooked in most seed rain studies. We assessed the viability of seeds in the seed rain falling into 12 plots from a Brazilian tropical dry forest: six plots exposed to slash-and-burn agriculture (‘burned plots’ = 111 seeds from 10 species) and six located in forest stands (‘control plots’ = 5267 seeds from 23 species). Using the tetrazolium test, we found that only 767 of 5378 seeds (14%) from 16 species were viable, but surprisingly, seed viability was 2.2 times higher in burned plots than in control ones. This does not imply that forest regeneration through seed rain is more effective in burned plots, as very few seeds arrived at these plots. A functional analysis revealed that the higher seed viability in burned plots may be explained by the fact that these plots were dominated by relatively large, hard-coated (orthodox), and abiotically dispersed seeds, with some type of dormancy, as seeds with these functional trait attributes showed the highest viability. To accurately assess the importance of seed rain for forest regeneration we must consider seed viability, as it can be very low in some places, especially if dominated by species with relatively small, animal-dispersed and recalcitrant seeds lacking dormancy.