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

INTRODUCTION

The importance of consumers in structuring natural communities is the subject of continuing debate (Hunter and Price 1992, Strong 1992, Paine 1995). Although a wide variety of animals can regulate prey abundance and composition in a variety of habitats (Menge et al. 1994), generalizations about the properties of communities in which consumers exert an overriding influence are not yet possible (Hunter and Price 1992, Power 1992). Species richness may be one important factor. Trophic links in speciose communities may be sufficiently interconnected so that control by just one or a few consumer species is improbable (Strong 1992). If so, consumers would be more likely to impart structure in relatively simple communities, like those seen on oceanic islands.

Current perceptions of the biota on oceanic islands are not consistent with this possibility. Phenomena such as the dissolution of protective mutualisms (Janzen 1973), density compensation in the absence of predators, herbivores, or competitors (MacArthur 1972), and the susceptibility of insular biotas to invasion (Elton 1958, Carlquist 1965, Pimm 1990) have contributed to a general viewpoint of insular communities as groups of weakly interacting species. Further, oceanic islands typically lack large mammalian herbivores (Wallace 1911, Carlquist 1974, Williamson 1981) and a retinue of specialized seed-eating insects (Janzen 1971, 1975). Both groups can be important determinants of seedling recruitment across a variety of mainland habitats (e.g., Howe et al. 1985, Brown et al. 1986, Terborgh et al. 1993). Their paucity on oceanic islands suggests that seed and seedling predators do not contribute significantly to structuring insular plant communities.

Land crabs, including the gecarcinids (Brachyura, Gecarcinidae) and hermits (Anomura, Coenobitidae) are often dominant consumers on tropical islands (Louda and Zedler 1985, Green 1997). They are omnivorous, and despite recurrent anecdotes hinting at their influence on insular plant dynamics (e.g., Guppy 1890, Niering 1956, Grubb 1971), only a few short-term studies have examined their effects (Lee 1985, Louda and Zedler 1985, O'Dowd and Lake 1990, 1991). Here we examine the role of the red land crab, Gecarcoidea natalis (Brachyura: Gecarcinidae) in seedling dynamics in rain forest on Christmas Island, Indian Ocean. We tested the hypothesis that red crabs alter the density and species composition of seedlings, because red crabs are abundant throughout the forest, and previous studies have shown them to be important consumers of fruits, seeds, seedlings, and leaf litter on the forest floor (O'Dowd and Lake 1989, 1990, 1991). Our study centers on a large, long-term experiment that excluded red crabs from plots in both light gaps and the shaded understory and assessed the impact of seed and seedling predation on the entire assemblage of seedlings. Few comparable experiments have been conducted in diverse continental tropical forests (e.g., Terborgh and Wright 1994), perhaps because of the difficulty in pinpointing sources of mortality for every species and identifying scores of tiny seedlings (Howe 1986, Terborgh 1992). We report results from the first 2 yr of crab exclusion, over which seedling recruitment was intensively monitored.

MATERIALS AND METHODS

The study system

Christmas Island (105 [degrees] 40[minutes] E, 10 [degrees] 30[minutes] S) is an elevated, oceanic limestone island, 360 km south of Java in the Indian Ocean. The island covers 134 [km.sup.2], and rises to a maximum elevation of 361 m. Average annual rainfall is [approximately]2000 mm, most of which falls between November and May (Stoddart 1971). Daily minimum and maximum temperatures vary between 20 and 30 [degrees] C, and relative humidity (RH) rarely falls below [approximately]70%.

Our study was conducted near Murray Hill (MH) in "plateau" rain forest, one of two major forest types on the island (Du Puy 1993). This forest type is widespread above [approximately]200 m elevation, and is dominated by a dozen or so canopy species whose frequencies vary locally. Green (1993) recorded 13 tree species on a 0.3-ha plot at MH, several times fewer species than on similar-sized plots in diverse continental tropical forests (e.g., Gentry 1988, Whitmore 1990); however, density (404 stems/ha) and basal area (48 [m.sup.2]/ha) were comparable to mainland forests (Gentry 1982, Swaine et al. 1987). The canopy at MH reaches 40 m and is evergreen, although it thins noticeably during the dry season.

The red land crab (Gecarcoidea natalis) is the most conspicuous ground-dwelling animal on the island. These large crabs (to 120 mm carapace width and [greater than]500 g) occur at all elevations, and are abundant wherever rain forest is still intact; Green (1997) estimated a mean density of 1.3 crabs/[m.sup.2] and a mean biomass of 1454 kg/ha at MH. Except for some activities associated with annual breeding at the coast (Hicks 1985), these crabs are completely terrestrial. They are diurnally active, and retreat into shallow burrows at night and when RH is too low for surface activity; crabs are inactive below 77% RH, and high levels of activity are only seen at [greater than or equal to]95% RH (Green 1997). Robber crabs (Birgus latro) sometimes eat red crabs, but predation does not appear to regulate the adult population.