How does the seed fate of Crotalaria podocarpa DC, a highly competitive herbaceous legume in arid rangelands, contribute to its establishment probability?

• Seed predation by invertebrates affects the seed fate of C. podocarpa.
• Pre-dispersal predation rates are 30% and post-dispersal predation rates are 90%.
• C. podocarpas main dispersal mode is explosive dehiscence reaching an area of 50 m2.
• Secondary seed dispersal by animals account for rare long-distance dispersal events.
• 400 left seeds m−2 seem to be enough, to allow successful C. podocarpa recruitment.

Plant recruitment can be constrained by processes shaping its seed fate, such as pre- and post-dispersal seed predation, while facilitated by higher seed production and more effective dispersal. These activities are controlled by specific predator and disperser guilds that respond to different microsite conditions such as vegetation density with varying predation and removal rates. In Namibia's arid rangelands, the native, annual herbaceous legume Crotalaria podocarpa has considerably increased after several years of elevated rainfall, thereby reducing density and vitality of economically important fodder grasses. To investigate whether this proliferation entails the risk of a permanent establishment, we studied pre- and post-dispersal seed predation, as well as primary and secondary seed dispersal of the legume and quantified the impact of different predator groups on post-dispersal seed removal. We experimentally excluded (i) neither vertebrates nor invertebrates, (ii) birds, (iii) vertebrates (birds and rodents), and (iv) vertebrates and large invertebrates on plots with different C. podocarpa density (0–35% cover). Seed predation was mainly caused by invertebrates, with 30% pre-dispersal seed predation in pods and up to 90% post-dispersal seed removal for free seeds on the ground which is the most persistent seed stage. C. podocarpa density did not affect post-dispersal seed removal or secondary seed dispersal. As main dispersal mode for the study species explosive dehiscence was identified, with seeds reaching dispersal distances of up to 4 m, while wind or rolling dispersal played a minor role. Subsequent secondary dispersal by animals accounted for dispersal distances up to 19 m. Our findings highlight the combined effects of pre- and post-dispersal seed predation to determine the total seed fate. With a contribution of about 370 seeds m−2 to the species persistent soil seed bank, seed numbers seems to be more important for C. podocarpas vagility than the species dispersal ability. Supplemented by seedling fate and adult mortality, this data can contribute to reliably predict the risk of a further spread of the species.