Nutrients more than predators alter decomposition and invertebrate communities in tropical brown food web mesocosms

Background/Question/Methods

The relative roles of bottom-up and top-down forces in brown food webs remains idiosyncratic and often based on observational studies. We experimentally tested the relative impact of two concentrations of nitrogen and phosphorous and two generalist predators--lycosid spiders and Pheidole sp. ants--on decomposition and invertebrate composition using a factorial mesocosm design. We set up 80 15 x 26 cm mesocosms with 50g wet leaf litter residua, and 75g loose leaf litter from old second growth forest on Barro Colorado Island, Panama. Each mesocosm received a litter bag with cellulose filter paper and two wooden dowels. Macro predators were removed from 40 mesocosms; the other 40 received a spider and a litter ant colony. Within each predator treatment were 5 nutrient levels (n=8 each): controls, 0.26 (+N) and 2.6g (+10N) urea (CH₄N₂O), 0.24 (+P) and 2.4g (+10P) super triple super phosphate (P₂O₅) added as dry fertilizer at the outset of the experiment. We tested the hypotheses that H1: decomposition is more P- than N-limited, H2: that increased microbial activity promotes microbivore abundance, and that H3: predators hinder decomposition by decreasing the abundance and activity of microbivores.

Results/Conclusions

After 1 month, nutrient, but not predator additions had large effects on decomposition and the microbial and microbivore community. H1: Decomposition responses were qualitatively variable and dose-dependent. The low dose +P had little effect on decomposition. In contrast, the high dose +10P increased cellulose decomposition 1.3 fold but decreased woody decomposition 1.5 fold relative to controls. The low dose +N enhanced cellulose decomposition 2-fold, but had no effect on wood. The high dose +10N decreased cellulose decomposition 2-fold and increased woody decomposition 1.3 fold. Moreover, there was a 2-fold increase in fungal fruiting bodies on +10N vs. controls. H2: The +P, +10P, and +N treatments had no discernible effect on invertebrate abundance. However, in +10N mesocosms, abundance decreased 4.5 fold for oribatid mites, 7-fold for collembola, and isopods were nearly extinct relative to controls. H3: The +Predator treatments resulted in 5% increases in mesocosm wet weight and a 9% increase in litter depth litter depth. These results suggest bottom-up forces vary in complex ways at the microbial level depending on nutrient, nutrient quantity and substrate, that microbivores may suffer in response to fungal gains, perhaps through antibiosis, and that predators may decrease detrivore activity, likely through increased anti-predator behavior.