Control of seedling recruitment by land crabs in rain forest on a remote oceanic island

Ecology, Dec, 1997 by Peter T. Green, Dennis J. O'Dowd, P.S. Lake

Escape in space is also doubtful. Distance and density-dependent variation in seed and seedling predation (Janzen 1970, Connell 1971) is unlikely because red crabs are ubiquitous throughout rain forest on Christmas Island, and the impact of crabs did not depend on seedling density, which varied over a 10-fold range on the exclusion plots. Our data also suggest that habitat-dependent predation (e.g., Augspurger 1984, Osborne and Smith 1990, Osunkoya et al. 1992) is an unlikely avenue of escape, because seedling recruitment was equally low in both light gaps and the shaded understory. Nevertheless, seedling recruitment could depend on gap size. Relative humidity is typically lower in larger gaps than smaller ones (Whitmore 1990, Brown 1993), and low RH ([less than]85%) suppresses the surface activity of red crabs (Green 1997). Even though crab densities may not differ, foraging activity of crabs could be reduced in large gaps to the extent that some vulnerable seedlings may recruit.

Similarly, most hypotheses for escape in time seem implausible. Seedlings might recruit when crab activity is relatively low, either during the dry season, or annual moulting and breeding periods (Green 1997). However, neither period is sufficient for seedlings to grow to a crab-resistant size, and seedlings did not recruit during two years on the control plots at MH. A widespread decline in the red crab population lasting several years could allow broadscale seedling recruitment to a crab-resistant size. Such a decline could result from repeated recruitment failure in juvenile crabs, or mass mortality of adults induced by climatic extremes or disease. However, there is no evidence that such a phenomenon has ever occurred on Christmas Island.

One plausible hypothesis incorporates variation in both space and time. Crab densities could shift at a local scale, producing a mosaic of varying predation pressure, resulting in patchy seedling recruitment. Long-term data on crab densities do not exist, but we know that densities vary over at least a fourfold range, from 0.4 to 1.8 burrows/[m.sup.2] (Green 1997). This lower value may approach a threshold below which seedlings can recruit, and localized seedling establishment could occur during fluctuations in crab densities.

Do red crabs structure rain forest?

Ours is one of a growing number of studies demonstrating that animals can have large effects on advance regeneration (Molofsky and Fischer 1993, Terborgh et al. 1993, Terborgh and Wright 1994), and that processes acting prior to gap formation could have primacy in determining the spatial structure, species composition, and diversity of trees in tropical forests (Brokaw and Scheiner 1989, Connell 1989). Still, it may prove impractical in forest ecosystems to assess the relative influence of these early impacts against the mechanisms that operate after seedling establishment. Experimental manipulations have linked consumer effects on seedling recruitment to the mature phase (sensu Watt 1947, Whitmore 1978) where plant generation times are relatively short (e.g., Brown et al. 1986, Brown and Heske 1990), but this is clearly difficult in forested ecosystems where plant generation times are considerably longer (Andersen 1989, Terborgh et al. 1993).