Predation on Posidonia australis seeds in seagrass habitats of Rottnest Island, Western Australia: patterns and predators

Seed predation is an important process governing the dynamics of many plant populations. We assessed seed mortality due to predation in the seagrass Posidonia australis by identifying predators and quantifying predation events using underwater video cameras, laboratory and field observations, and field tethering experiments in shallow-water habitats off Rottnest Island, Western Australia in 2003 (2 sites) and 2004 (5 sites). We assessed 4 dominant habitats: within seagrass meadows of (1) P. australis, (2) Amphibolis antarctica and (3) Halophila ovalis, and on (4) unvegetated sand. Video analysis and field and laboratory observations showed that at least 6 different crustacean taxa were seed predators — the portunid crabs Nectocarcinus integrifrons and an unidentified portunid similar to Thalamita crenata, the majid Naxia aurita, the isopod Cymodoce sp., and the hermit crabs Paguristes purpureantennatus and Calcinus dapsiles — with N. integrifrons and Cymodoce sp. being the principal seed predators. Various species of fishes were observed approaching (but never eating) seeds in the video analysis, and none appeared to be seed predators. In 2003, daily seed predation rates were dramatically higher in A. antarctica beds than in other habitats at Parker Point (mean 63% d–1 compared to means of <20% d–1 for the other habitats), but not at Nancy Cove, the only other site studied in 2003. In 2004, daily seed predation rates were higher in Posidonia australis (40 to 73% d–1) and A. antarctica (36 to 85% d–1) beds than in H. ovalis (9% d–1) beds and unvegetated sand (3 to 13% d–1) at all 5 sites studied. Logistic regression showed an approximately 20-fold increased chance of a predation event occurring within a 24 h period in P. australis and A. antarctica beds compared to unvegetated sand, but no increase in predation in the H. ovalis beds compared to sand. Although predation rates varied among sites and years, patterns among habitats were generally consistent. P. australis and A. antarctica, the 2 seagrass species with structurally complex canopies, had much higher rates of seed predation than the habitats with less structure, i.e. H. ovalis and bare sand, which may provide insufficient refuge for seed predators. These results may have important consequences for plant demographics and genetic structuring of local populations. Viewing seagrass seed dispersal strategies in the context of faunal-dependent predation processes, rather than simply in terms of plant–plant competitive interactions, may provide a better understanding of seagrass spatial distributions and successional dynamics.