Short term patterns of germination in response to litter clearing and exclosure of large terrestrial vertebrates along an Amazon forest regrowth gradient

Efforts to restore tropical forest ecosystem services depend on understanding the barriers to germination of species of economic and cultural interest. Here we use two important non-timber forest product species (NTFP: Inga capitata. Desv. - Fabaceae and Euterpe oleracea Mart. - Arecaeae) to compare germination across a forest regrowth gradient in the northeast Brazilian Amazon. Experimental treatments were used to examine the effects of mid to large-bodied vertebrates and litter cover on seed fate (germination, removal and invertebrate infestation) in 15 lowland sites within small-holder properties. Regrowth stage was classified into three groups, with five sites each of: late second-regrowth forest, early second-regrowth forest and abandoned pasture. We conducted a paired split-plot experiment using experimental plots composed of a vertebrate exclosure versus an open treatment and subplots with and without litter. We used Generalized Linear Mixed-Effects Models (GLMMs) to compare additive and interaction effects of treatments across regrowth stages compared with 15 paired control sites. We found that the effects of regrowth stage and exclosure were species specific and these effects differed between responses (germination, removal and invertebrate infestation). Clearing litter generated a significant effect only for invertebrate infestation, which increased in the cleared plots. Our findings show that seed removal limits germination success for both species across forest regrowth stages, with invertebrate infestation also having important but secondary effects. Increased removal and unfavorable abiotic conditions make direct seeding unviable in pasture sites. We suggest that direct seeding is a viable alternative for the establishment of these widely available NTFP species in late and early regrowth forests.