Habitat fragmentation affects entire ecological systems, causing population decline, and biodiversity loss. Habitat corridors have been used as a method to mitigate the negative effects by connecting previously fragmented habitat. An aspect of habitat fragmentation that has been poorly tested is effects on trophic dynamics. Working in a large and long-term habitat fragmentation experiment, we tested the effects of corridors on seed predation by ants, focusing on Pogonomyrmex badius, a granivorous native ant species. It provides several ecosystem services such as increasing soil fertility around mound sites and seed dispersal. It is unknown whether or not fragmentation may lead to disproportionate seed predation by P. badius and if corridors may mediate these effects. As such, the focal question of this study is: Do corridors, fragment shape and edge proximity affect the rate of P. badius seed predation? Working in a large-scale fragmentation experiment, we placed seed depots with longleaf pine (Pinus palustris) seeds in each of 3 patch types (connected, rectangular, and winged; 21 patches total) at 3 edge categories (center, single edge, double edge) chosen in a stratified random fashion across 7 study sites (n = 7) for 3 weeks and assessed the number of seeds removed.
Results/Conclusions
We found a marginally significant effect of patch type on the proportion of depots with seeds removed (Χ2 = 4.6, df = 2; P = 0.10). Rectangular patches had higher proportions of depots with seeds removed by 0.10 and 0.17 compared to connected patches and winged patches respectively. We found that edge categories did not affect the proportion of depots with seeds removed (Χ2 = 0.26, df = 2; P = 0.88). Our findings on patch type effects are supported by previous studies in our system that found greater densities of P. badius in rectangular patches. P. badius densities are typically higher away from edges so our finding that seed removal did not differ among edge categories is surprising. This may be attributed to the minimum distance from an edge used in our study (12.5m) being too far from the edge to reduce P. badius densities and activity. Future work will use P. badius data from pitfall traps as a predictor variable to further investigate seed removal patterns in our landscape. Our study provides valuable insight for both wildlife managers and researchers to best determine how to improve fragmented habitat to cater to native ant and plant communities.