BIOLOGICAL Cambridge REVIEWS Philosophical Society

Biol. Rev. (2009), 84, pp. 203–223. 203 doi:10.1111/j.1469-185X.2008.00070.x Land as key drivers in tropical coastal forest recruitment

Erin Stewart Lindquist1*, Ken W. Krauss2, Peter T. Green3, Dennis J. O’Dowd4, Peter M. Sherman5, and Thomas J. Smith, III6 1 Meredith College, Department of Biological Sciences, 3800 Hillsborough Street, Raleigh, North Carolina 27607, USA 2 U.S. Geological Survey, National Wetlands Research Center, 700 Cajundome Boulevard, Lafayette, Louisiana, 70506, USA 3 Department of Botany, La Trobe University, Bundoora, Victoria 3086, Australia 4 Australian Centre for Biodiversity, School of Biological Sciences, Monash University, Victoria 3800, Australia 5 University of Redlands, Department of Environmental Studies, 1200 East Colton Avenue, P.O. Box 3080, Redlands, California 92373, USA 6 U.S. Geological Survey, Florida Integrated Science Center, 600 Fourth Street, South, St. Petersburg, Florida, 33701, USA

(Received 17 October 2008; revised 11 November 2008; accepted 19 November 2008)

ABSTRACT

Plant populations are regulated by a diverse assortment of abiotic and biotic factors that influence seed dispersal and viability, and seedling establishment and growth at the microsite. Rarely does one guild exert as significant an influence on different plant assemblages as land crabs. We review three tropical coastal ecosystems– , island maritime forests, and mainland coastal terrestrial forests–where land crabs directly influence forest composition by limiting tree establishment and recruitment. Land crabs differentially prey on seeds, propagules and seedlings along nutrient, chemical and physical environmental gradients. In all of these ecosystems, but especially mangroves, abiotic gradients are well studied, strong and influence plant distributions. However, we suggest that has primacy over many of these environmental factors by acting as the first limiting factor of tropical tree recruitment to drive the potential structural and compositional organisation of coastal forests. We show that the influence of crabs varies relative to tidal gradient, shoreline distance, canopy position, time, season, tree species and fruiting periodicity. Crabs also facilitate forest growth and development through such activities as excavation of , creation of soil mounds, aeration of soils, removal of leaf litter into burrows and creation of carbon-rich soil microhabitats. For all three systems, land crabs influence the distribution, density and size-class structure of tree populations. Indeed, crabs are among the major drivers of tree recruitment in tropical coastal forest ecosystems, and their conservation should be included in management plans of these forests.

Key words: biotic control, ecological filter, environmental gradient, environmental engineer, , island maritime forest, predation, seed, seedling, terrestrial mainland forest, tree.

CONTENTS

I. Introduction ...... 204 II. Mangroves ...... 208 (1) Propagule predation ...... 208 (2) Seedling predation ...... 209 (3) Spatial variation ...... 209 (4) Temporal variation ...... 211 (5) Crab filter interactions ...... 211 III. Island maritime forests ...... 211

* Address for correspondence: Tel: 01 919 760 8754; Fax: 01 919 760-8761; E-mail: [email protected]

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works 204 Erin S. Lindquist and others

(1) Fruit and seed predation ...... 212 (2) Seedling predation ...... 212 (3) Spatial variation ...... 212 (4) Temporal variation ...... 213 (5) Crab filter interactions ...... 214 IV. Mainland coastal terrestrial forests ...... 214 (1) Seed predation ...... 214 (2) Seedling predation ...... 216 (3) Spatial variation ...... 216 (4) Temporal variation ...... 216 (5) Crab filter interactions ...... 217 V. Implications ...... 217 (1) Implications across ecosystems ...... 217 (2) Conservation implications ...... 218 VI. Conclusions ...... 219 VII. Acknowledgements ...... 219 VIII. References ...... 219

I. INTRODUCTION physiological, morphological, ecological and behavioural to terrestrial environments. For example, land Limitations on plant recruitment have been a popular crabs minimise water loss by inhabiting damp, deep avenue for ecological research (see Mu¨nzbergova´ & Herben, burrows, being active at night or in high humidity and 2005 and Hermy & Verheyen, 2007 for recent reviews), through the evolutionary development of lungs, in addition particularly in the tropical literature. Limited dispersal to gills, for gas exchange (McMahon & Burggren, 1988). (Hubbell et al., 1999), seed availability (e.g. Norden et al., Land crabs vary in their level of terrestrial adaptation, even 2007) and low microsite availability (e.g. de Steven & Wright, within one family. Discoplax () species require 2002; Doust, Erskine & Lamb, 2006) are all known to regular immersion in water whereas and Gecarcoi- minimize the number of successfully recruited individuals in dea species can obtain water from food, dew, or soil sub- tropical tree populations. Focusing on microsite limitation strata through absorption (; Hartnoll, 1988). alone, abiotic factors such as light (e.g. Kyereh, Swaine & Land crabs include one family of hermit crabs, Coenobi- Thompson, 1999; Uriarte et al., 2005), desiccation (e.g. tidae (Coenobita and Birgus), which are terrestrial as adults Veenendall et al., 1996), fire (Janzen, 1985), salinity and salt but like the majority of land crabs are planktonic as lar- spray (Ceron et al., 2002) and soil characteristics (Swaine, vae. Mangrove land crabs in the families Grapsidae 1996) all have been found to limit tropical tree recruitment. (Neosarmatium and Goniopsis) and Ocypodidae (Ucides) In an examination of the literature, a large percentage (14 at or above the high tide line and are generally active out of out of 27) of studies found that abiotic conditions at the water during low tide or by climbing vegetation (Hartnoll, microsite lower growth rates and the probability of survival 1988). Land crab species which burrow and forage above (Lindquist, 2003). Of the same studies, ten also found that high tide or on the beach dunes are restricted to tropical predation or herbivory of seeds and seedlings had a signifi- and subtropical ecosystems because cold temperatures limit cant impact on tree recruitment. However, studies have their ability to survive while inactive in burrows and their mostly focused on insect and mammal seed predators (see foraging activity above ground (Wolcott, 1988). In this Janzen, 1971 for insects; Silman, Terborgh & Kiltie, 2003 for review we focus on land crab species which inhabit tropical mammals) and herbivores (see Coley & Barone, 1996 for ecosystems because the majority of research on the impact a review). of crabs on forest recruitment has focused on these habitats. Here we focus on defining the mechanisms by which Land crabs are found in tropical coastal ecosystems a poorly exposed group of seed predators and herbivores, around the world. Information on the role of abiotic factors the land crabs, affect recruitment of tree species in tropical in controlling plant growth in these systems is abundant. We coastal forests. Prior to the early 1990s, the impact of land argue that in addition to environmental gradients, coastal crabs on plant communities was not well known (Wolcott, ecosystems share a biotic factor, land crabs, which can have 1988). Since the publication of Biology of the Land Crabs as much, or more, of an effect on plant recruitment. Studies (Burggren & McMahon, 1988), investigators have advanced that focus on the environmental gradients of coastal systems our understanding of the ecological role of land crabs in minimise, by often ignoring, the important role of land tropical coastal ecosystems through a combination of crabs. Through excavation of burrows (Gutierrez et al., descriptive and manipulative experiments. Our review is 2006), collection of leaf litter in burrows (O’Dowd & Lake, the first to synthesise the findings made by these land crab 1989; Sherman, 2003) and predation of seeds and seedlings studies over the last 20 years. (e.g. Smith, 1987c; Cannicci et al., 2008), land crabs We define land crabs following Hartnoll (1988). Land determine the quantity, and sometimes the quality (species), crabs are crabs which are able to remain active outside of tropical coastal forest recruitment. The relative effects of of water for an extended period of time because of these local abiotic and biotic factors on forest recruitment in

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works Land crabs and tropical forest recruitment 205 relation to larger scale factors can be represented by a group Table 1. Ecosystem comparison of coastal gradients among of nested filters (Fig. 1). Ecological filters, such as crab mangrove, island maritime and mainland coastal terrestrial predation on seeds or seedlings (a local biotic condition) and forests. ‘]’ signifies that the factor varies (direction nonspecific) soil salinity (a local abiotic condition), cause mortality of relative to the coastal gradient (distance from shore or tide) in recruits as individuals pass from one life stage to another. By a particular system, ‘‘o’’ signifies that the factor does not vary causing differential species mortality by selecting particular with the gradient and ‘‘na’’ signifies not yet investigated in species’ traits, ecological filters can impact the species system or data are inconclusive. composition of communities (see Statzner, Doledec & Hugueny, 2004 for a review). Factor for comparison, relative To contrast the role of land crabs against abiotic factors to distance from shore or tide Mangrove Island Mainland in forest ecosystems, we selected three tropical coastal Physical environment ecosystems: mangroves, island maritime forests and main- Salinity ] oo land coastal terrestrial forests. These systems vary greatly in Flooding ] oo the strength of their environmental gradients (Table 1). For Soil texture, type, nutrient ]]] example, mainland coastal terrestrial forests experience availability ]]] seasonal droughts and floods but do not have significant Light, canopy cover Forest structure and composition Tree density (juveniles oo] A Pool of species and adults) Tree size-class distribution ] o ] Tree species distribution ]]] Crab assemblage Climate Density ] o ] Species distribution ] o ] Temporal activity na o ] Landscape Seed or propagule predation ],o o ],o Seedling predation ] o ],o ] ] Local abiotic conditions Litter removal o

Local biotic conditions

salinity gradients (Lindquist & Carroll, 2004). Conversely, salinity and water depth can dominate tree establishment B Pool of species in mangroves, but drought and rain-induced flooding by themselves are seldom important to plant recruitment except for their linkages to salinity and propagule dispersal Climate (Sousa et al., 2007). These ecosystems also differ in the prevalence and diversity of land crabs. Land crabs represent a significant proportion of the faunal biomass in all three Landscape ecosystems (Table 2), but their densities are highest in island and mainland maritime forests and they are most speciose Local biotic conditions in mangroves. Tree species richness varies among these ecosystems with mangroves having the fewest species, island maritime forests intermediate and mainland coastal terres- Local abiotic conditions trial forests the most. While considering these variations in the tree and crab assemblages (Table 2) and environmental gradients Fig. 1. Schematic diagram of the effects of ecological filters on (Table 1), we review here how the positive and negative recruitment of species (after Poff, 1997). The presence and impacts of crabs on tree recruitment are shared across three abundance of species (represented by black lines) is determined types of tropical coastal ecosystems. In order to highlight by their ability to pass through multiple filters (represented by these similarities, we present our findings in a systematic ovals). Filters are placed in a hierarchy based on the ability to limit recruitment; the upper filters (i.e. climate and landscape) order: (1) crab predation of seeds, propagules or fruits; (2) limit recruitment more than the lower filters (i.e. local abiotic crab predation of seedlings; (3) spatial variation in crab and biotic conditions). Any species that lacks traits which would predation; (4) temporal variation in crab predation, and (5) allow it to pass through an upper filter will not be available to crab predation, or filter, interactions with other factors pass through the lower filters. Previous tropical coastal forest influencing forest structure and composition. By highlight- literature has argued that the upper diagram (A) is most ing one shared, limiting factor in these ecosystems, we hope representative of tree recruitment limitation. We argue that to encourage other investigators to find commonality in where crabs are present, the lower scheme (B) is more their studies investigating limitations in plant recruitment representative; crabs, as biotic filters, may limit tree recruit- and their effects on the structure and composition of forest ment as much or more than local abiotic conditions. ecosystems.

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works ilgclReviews Biological 206

Table 2. Comparison of three tropical coastal ecosystems (mangroves, island maritime forests and mainland coastal terrestrial forests) for: (I) trophic assemblages including trees, crabs and natural enemies of crabs; (II) ecological effects on tree recruitment including seeds or propagule predation, seedling predation, litter removal and soil manipulation; (III) spatial variation of crab predation relative to the tidal/coastal location and canopy cover; and (IV) temporal variation of crab predation relative to time of day, season and year. 84 20)203–223 (2009) Mangroves Islands Mainland I. Trophic assemblages: Tree species richness d 1-7 species in Neotropics d 3-5 tree species on Enewetak Atoll, d 600-700 tree species in Corcovado, Costa (Chapman, 1976) Marshall Islands (Louda & Zedler, 1985) Rica (Quesada et al., 1997)

Ó d 18-21 species in Indo- d Approx. 50 canopy tree species d 60-150 tree and 250

09Junlcompilation Journal 2009 South-Pacific (Duke, 2006) on , Indian Ocean plant species in Cabo Blanco, Costa Rica (Green et al., 1997) (Lindquist, 2003; Camacho-Cespedes & Lindquist, 2007)

Crab densities 1 d Grapsid crabs along d natalis on Christmas Island: d Gecarcinus lateralis at La Mancha, : Murray River, NE Australia: 0.2-2.5 crabs m-2 (Green et al., 1997) 0.5-1.0 crabs m-2 (Kellman & Delfosse, 1993) -2 0-2 m (0.77 crabs per trap d Gecarcinus planatus on Clipperton Atoll: 6.0 d Gecarcinus quadratus in Corcovado, Costa night) (Frusher et al., 1994) crabs m-2 (Ehrhardt & Niaussat, 1970), Rica: 0 – 2.5 crabs m-2 (Sherman, 2002) d Xanthid crabs in Rookery -2 and up to 0.8 m on Socorro Island, d Gecarcinus quadratus in Cabo Blanco, Costa

Ó -2 Bay, FL USA: 0.3 – 4 m Mexcio (Jimenez et al., 1994; Perez-Chi, Rica: 0.8 – 3.3 crabs m-2 (Lindquist & 09CmrdePioohclScey ocamt rgnlU oenetworks government US original to claim No Society. Philosophical Cambridge 2009 (4.17 crabs per trap night) 2005). Carroll, 2004) (McIvor & Smith, 1995) d Gecarcinus ruricola on San Andres and d Gecarcinus lateralis in Veracruz, Mexico: 0.5 d in SE Providencia Islands: 0.08-0.21 crabs crabs m-2 (Capistra´n-Barradas et al., 2003) Florida USA: 1.85 per m-2 m-2 (Hartnoll et al., 2006) (Herreid & Gifford, 1963) d Cardisoma guanhumi in South Florida: d Cardisoma guanhumi on Andros Island -2 d Uca annulipes in East : -2 1.8 crabs m (Gifford, 1962) -2 (Bahamas): 1.0 crabs m (Lutz & 10-175 per m (Skov et al., Austin, 1983) 2002) d on Aldabra (Indian d Neosarmatium meinerti in East -2 -2 Ocean): 0.4 crabs m (Alexander, 1979) Africa: ;5 per m and Ifaluk (Caroline Is.): 1.2 crabs m-2 (Skov et al., 2002) (Bates & Abbott, 1958) Natural enemies d Herons, ibis, raccoons, humans d Intraguild predation (Birgus latro on d Raccoons, coatis, cats, foxes, of prey, of crabs (Diele et al., 2005) Gecarcoidea natalis) on Christmas Island herons in Costa Rica (Sherman 2002;

(Hicks et al., 1990) Lindquist & Carroll, 2004) others and Lindquist S. Erin d Introduced pigs and rats on Gecarcinus planatus on Clipperton Atoll (Pitman et al., 2005) d Flightless rails on land hermit crabs on Aldabra (Alexander, 1979). d Endemic birds including yellow crowned night heron Nyctanassa violacea gravirostris on Socorro (Perez-Chi, 2005) II. Ecological effects on tree recruitment: % seed or propagule d 0-100% propagule predation d 0-100% destroyed by hermit crabs (Coenobita d 40% removed in Cabo Blanco, Costa predation (removal) dependent on species and location perlatus) on Enewetak Atoll, Marshall Islands Rica (Lindquist & Carroll, 2004) (Smith, 1987c; Smith et al., 1989) (Louda & Zedler, 1985) adcasadtoia oetrecruitment forest tropical and crabs Land ilgclReviews Biological

d 54.3% propagule predation d >80% handled and 47% destroyed by (Allen et al., 2003) Gecarcoidea natalis on Christmas Island (0-94% by seed species; Green et al., 1997)

d d d

84 Seedling predation No damage to seedlings Gecarcoidea natalis grazed on 25 seedling Seedling density increased 144% in

20)203–223 (2009) (Siddiqi, 1995) species; seedling density was 20-fold higher, crab exclosures (Sherman, 2002) d Cotyledons removed from and seedling richness was five-fold higher in d 74-100% seedling removal outside of Avicennia seedlings crab exclosures than in unfenced exclosures (Lindquist & Carroll, 2004) (Smith, 1987b) control plots (O’Dowd & Lake, 1990; Green et al., 1997)

Ó Litter removal d Litter removal and burial by crabs d Gecarcoidea natalis processes 39-87% of annual d ;Five-fold higher litter mass in crab

09Junlcompilation Journal 2009 is important in nutrient cycling leaf litterfall on Christmas Island exclosures than in open plots (Robertson, 1986) (Green et al., 1999b) (Kellman & Delfosse, 1993) d Influence of litter removal on d 5.6 times higher leaf litter accumulation in propagule establishment has crab exclosures than in controls not been studied. (Sherman, 2003)

Soil manipulation d Elevate soils 5-15 cm d Gecarcoidea natalis relocates leaf litter on and d Crab burrows nutrient- and carbon- rich; (Minchinton, 2001) into burrows; mean turnover time plant root densities higher around d Soil sulphide & ammonium of burrow > 4.4 years (O’Dowd & Lake, burrow microsites (Sherman, 2006)

Ó levels are lowered by crab 1989; Green, 2004) 09CmrdePioohclScey ocamt rgnlU oenetworks government US original to claim No Society. Philosophical Cambridge 2009 burrowing (Smith et al., 1991) III. Spatial variation of crab predation relative to: Location relative to d Dependent on species of d No variation relative to coastal lowland and d Higher in coastal zone (0-100 m tide/coastline propagule (Osborne & inland plateau forest (Green et al., 2008) from coast) than in inland zone (>100 m Smith, 1990) d Density of Gecarcinus planatus high at low and from coast) (Lindquist & Carroll, 2004) d Higher in lower intertidal high altitude, and low at mid-altitudes, (Sousa & Mitchell, 1999) on Socorro Island (Perez-Chi, 2005) d No differences relative to tidal position (Clarke & Myerscough, 1993; Allen et al., 2003)

Canopy cover d Higher in understorey; d No difference in crab densities and seedling d No difference in seed and seedling removal lower in large gaps predation in shaded understorey and in by crabs between gaps and understorey (Osborne & Smith, 1990; light gaps (Green et al., 1997) sites (Lindquist & Carroll, 2004) Clarke & Kerrigan, 2002; d Lower crab densities in sparse canopy Clarke, 2004) cover (effect on seed/seedling predation d No change relative to gap unknown; Jimenez et al., 1994) size or understorey (Sousa & Mitchell, 1999; Allen et al., 2003; Krauss & Allen, 2003) IV. Temporal variation of crab predation relative to: Time of day d Crab activity determined more d Gecarcoidea natalis is diurnal on Christmas d Gecarcinus quadratus primarily nocturnal by stage of tide than time of day Island; activity is positively correlated with (Sherman, 2002; Lindquist & Carroll, 2004) (Frusher et al., 1994) relative humidity (Green, 1997) 207 208 Erin S. Lindquist and others

II. MANGROVES

Crabs influence many aspects of mangrove community n- ´ dynamics by facilitating the conversion of organic nitrogen to ammonia (Alongi, Boto & Robertson, 1992), promoting decomposition of organic matter (Robertson & Daniel, 1989; n-Barradas ´ Micheli, Gherardi & Vannini, 1991; Lee, 1998), grazing on leaf material (Onuf, Teal & Valiela, 1977; Beever, Simberloff & King, 1979), aerating anoxic soils through burrowing (Smith

primarily diurnal et al., 1991) and altering soil microtopography by creating

, 2003) mounds (Warren & Underwood, 1986; Minchinton, 2001). In addition, predation of propagules and seeds by crabs can be et al. extremely important in controlling recruitment (Smith, 1987c; Allen, Krauss & Hauff, 2003). , 2003; Sherman, 2003; Lindquist & Of the 70 distinct mangrove species or hybrids in the world (Kellman & Delfosse, 1993; Capistra rainfall is high andavailability seed/seedling is high (Capistra et al. Carroll, 2004) allow pulses of treecrab recruitment years in (Sherman, low 2002;Carroll, Lindquist 2004) & Barradas Gecarcinus lateralis Increased in wet season when humidity/ Annual periodicity in crab densities may (Duke, Ball & Ellison, 1998), an individual mangrove forest d d d commonly has one to seven species in the Neotropics (Chapman, 1976) but as many as 18 to 21 species in the tropics of Malaysia and Australia (Watson, 1928; Ball, 1998; Duke, 2006). Also, mangroves often occur as visibly distinct bands in many locales and there is strong evidence that predator guilds exert pressure on mangrove recruitment (Smith et al., 1989). Accordingly, environmental stress has , 1994; on Jarak direct effects on the distribution of both mangroves (Krauss et al., 2008) and crabs (Frusher, Giddins & Smith, 1994); at et al.

Gecarcinus planatus times both must cope with low oxygen levels, high tem-

, 1997, 2008) perature and desiccation (Hogarth, 1999). As the distribution G. lalandii of crabs changes along these environmental gradients, crabs et al. can exert a differential influence on mangrove seed ger- mination or propagule growth initiation through direct

, 1997) consumption, damage or burial (Smith, 1987a,b). The dominant members of the crab fauna in mangroves et al. belong to the families Gecarcinidae, Grapsidae and Ocypo- didae. The grapsid crabs are the primary consumers of erez-Chi, 2005) and  on Socorro Island (Jimenez P Island, Malaysia (Audy, 1950) high and seed/seedling availability(Green is high allow pulses of treecrab recruitment years in (Green low Plasticity in diel activity in Increased in wet season when humidity is Annual periodicity in crab densities may propagules in the Indo-West-Pacific region. In the eastern d d d Pacific, Atlantic and the gecarcinids (e.g. Cardisoma spp.) and Ocypodids (e.g. Ucides spp.) are more important than the grapsids (Twilley et al., 1997; Diele, Koch & Saint- Paul, 2005).

(1) Propagule predation Among studies conducted globally, crabs destroyed 54.3 % (^ 31 % S.D.) of seeds or propagules falling to the man- grove forest floor (Allen et al., 2003). Much research at- tention has been given to this influence, with general trends suggesting that predation is related variously to the chemi- propagules (Robertson, 1986) annually (Conde & Diaz, 1989), but the impactpropagule on predation and seedling establishment is unknown Increased with availability of Crab abundances can vary cal composition of propagules (McKee, 1995; Smith, 1987c; Mangroves Islands Mainland d d Allen et al., 2003), related negatively to propagule density (Sousa & Mitchell, 1999; Krauss & Allen, 2003; but see Smith, 1988), and unrelated to whether predation occurs in small canopy gaps or under a forest canopy (Sousa &

) Mitchell, 1999; Krauss & Allen, 2003; Allen et al., 2003; Clarke, 2004; but see Osborne & Smith, 1990). Central to cont. ( many predation studies are tests of what has become known as the dominance-predation hypothesis (Smith, 1987c). This hypothesis suggests that selective predation by crabs pre- vents preferred tree species from establishing and becoming See Green (1997) for an additional review of gecarcinid land crab densities in island and mainland coastal forests. Table 2. 1 Season Supra-annual dominant at particular points along the tidal gradient.

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works Land crabs and tropical forest recruitment 209

Several papers fail to support this hypothesis conclusively prevent establishment; it did, however, have negative effects (c.f., McKee, 1995; McGuinness, 1997; Sousa & Mitchell, on establishment with increasing dispersal distance and thus 1999; Clarke & Kerrigan, 2002), however many do not could limit long-distance supply of propagules. In some follow the fate of propagules through to establishment. For cases, the annual cycle of insects into and out of seedlings example, predation of Xylocarpus granatum seeds did not commonly results in mortality, as reported for the beetle differ under forests with different overstorey compositions, Coccotrypes rhizophorae on Rhizophora mangle in (Sousa but establishment, measured after 161 days, was greater et al., 2003b) and in Florida, USA (Devlin, 2004). What is in forests with X. granatum in the overstorey (Allen et al., more certain, however, is that cotyledon consumption by 2003). This result leaves the possibility of either differen- any vector reduces energy reserves needed for optimal tial consumption of seeds among habitats, or differential initial growth of mangrove seedlings (Minchinton & Dalby- environmental factors as possible explanations for the Ball, 2001; Sousa et al., 2003a). spatial distributions of tree species. In all, studies have The mechanisms of attack by insects versus crabs, how- indicated a large range of strategies among crabs, from ever, are very different. Many insects bore small holes into almost completely non-selective feeding by Neosarmatium the propagule (i.e. viviparous seedlings that germinate on meinerti (Dahdouh-Guebas et al., 1997) to strongly selective the parent tree prior to dispersal) to remove internal por- associations whereby crabs only consume propagules from tions of the seedling not associated with water-conducting mangrove species foreign to a particular site (Clarke & xylem tissue. Crabs, on the other hand, typically remove Kerrigan, 2002). large outer portions of the propagule to get to the inner, Some mangroves with small propagules and high more palatable portions. Yet, the effect of both insect and nutritive quality (e.g. Avicennia spp.) are heavily consumed crab herbivores can be similar for seed-producing man- by crabs. For these species, propagules are often excluded grove species. Both herbivores are responsible for killing completely from some intertidal zones (Smith et al., 1989). Xylocarpus granatum seeds in experimental studies (Robertson Other studies have also reported strong susceptibility of et al., 1990; Allen et al., 2003), presumably to benefit from Avicennia spp., and other small-seeded species, to crab the 50% increase in crude protein, 105% increase in crude predation (McKee, 1995; Sousa & Mitchell, 1999; Clarke & fat, 23% increase in total carbohydrates, and 22% decrease Kerrigan, 2002). It is likely that the dominance-predation in crude fibre content of cotyledon material versus entire X. model may apply strongest to mangrove species that are granatum seed (Allen et al., 2003). On the other hand, highly palatable and easily killed, and, hence, in species mangroves that produce propagules may have much less where crab diet selectivity is most intense. difficulty than seed producers in ameliorating mortality from crab damage once seedlings are established. For example, crabs did not cause damage to mangrove seedlings planted in Bangladesh, in which only six out of 336 (2) Seedling predation propagule seedlings died outside of crab exclosures while 11 Few studies have assessed the effects that partial crab pre- died within exclosures (Siddiqi, 1995). dation, or herbivory in the case of seedlings, has on the survival of mangrove seedlings. It is common to assume that 50% consumption of the vegetative portion of propagules or (3) Spatial variation the loss of a plumule equates to mortality (see Smith, 1987c for propagules; but see Allen et al., 2003 for seeds). Spatial variation in predation on propagules and seeds has However, insect herbivory studies offer some insight into focused on two principal topics, intertidal location (e.g. the resiliency of seedlings to attack on propagules and loss of upper, middle, and lower intertidal; Fig. 2A) and canopy biomass. Insect herbivory is widespread on mangrove position (i.e. gap versus understorey). Predation can vary propagules both before and after dispersal (Robertson, considerably with spatial shifts in intertidal position as the Giddens & Smith, 1990; Farnsworth & Ellison, 1997; degree of tidal inundation affects exposure time of seeds and Minchinton, 2006), and can have varying effects on the propagules on the soil surface (Osborne & Smith, 1990) and growth of seedlings. For example, insects reduced growth in influences the abundance of crab predators. Because species Avicennia germinans slightly, but either had no effect on composition often changes at different intertidal locations, individual seedlings of Laguncularia racemosa or completely many tests of the dominance-predation hypothesis de- stalled the growth of individual plants (Sousa, Kennedy & scribed above are also tests for predation differences among Mitchell, 2003a; Sousa, Quek & Mitchell, 2003b). Height or intertidal location. Panamanian mangroves, for example, biomass increments for seedlings of , have greater propagule predation in lower intertidal Bruguiera exaristata, Xylocarpus australasicus and Xylocarpus locations than in upper or middle intertidal locations granatum were reduced by insect damage, but survival was (Sousa & Mitchell, 1999). In that same study, larger crabs, affected only in the latter two species (Robertson et al., such as Ucides cordatus and Goniopsis cruentata, were better 1990). Generally, survival and establishment are affected suited to propagule grazing and were common in lower less often than growth for seedlings with herbivore-damaged intertidal locations; detritivore species predominated in propagules, though determining limits of sublethal tissue mid-to-upper intertidal locations. Crab predators removed damage is difficult and can vary widely by species (Sousa more Bruguiera gymnorrhiza propagules in low than in middle et al., 2003a). Minchinton (2006) reported that pre-dispersal intertidal locations on a Pacific island (Krauss & Allen, herbivory by insect larvae on A. marina propagules did not 2003; Fig. 2B), more Aegiceras corniculatum propagules in high

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works 210 Erin S. Lindquist and others

Fig. 2. (A) Cross section of the intertidal distribution for dominant species studied in Micronesia (after Tomlinson, 1986), and (B) predation of Bruguiera gymnorrhiza and Xylocarpus granatum across this gradient in one Micronesian mangrove forest (data after Krauss & Allen, 2003; Allen et al., 2003). Predation for B. gymnorrhiza propagules was reduced in the middle intertidal where the species maintains dominance in the overstorey. Trends were not significant for Xylocarpus granatum, even though X. granatum is a dominant component of the upper intertidal overstorey. No differences were apparent for comparisons of understorey (shaded bars) versus gap (open bars) locations within intertidal zone. Predation assessments followed by the same letter among intertidal zone for each species are not significantly different at P ¼ 0.05. Values are means ^ 1 S.E. (N ¼ 18, % out of 12 propagules in each plot).

than in low intertidal locations in Australia (Osborne & indicate no differences in rates of predation by canopy po- Smith, 1990) and more Avicennia germinans propagules in sition (Sousa & Mitchell, 1999; Allen et al., 2003; Krauss & a salt marsh zone than mangrove zone in Louisiana, USA Allen, 2003; Fig. 2B). The reasons for this result are unclear (Patterson, McKee & Mendelssohn, 1997). Several studies at present and warrant further research. indicate variable patterns of loss by species with intertidal Crabs can also alter the spatial distribution and growth location (Smith, 1987c, 1988; Dahdouh-Guebas et al., 1998; potential of mangroves. By serving as ecosystem engineers Clarke & Kerrigan, 2002) and some show no differences in (see Jones, Lawton & Shachak, 1994), crabs can have predation by intertidal location (Clarke & Myerscough, a profound effect on soil microtopography. Elevation 1993; Allen et al., 2003; Fig. 2B). change in most mangrove forests is minimal and best By canopy position, predation was higher in understorey measured in centimeters or tenths of meters along the entire locations than in canopy gaps for Avicennia marina and less intertidal range (Macnae, 1969b; Thom, 1982; Woodroffe, in large canopy gaps (> 600 m2) than in small canopy gaps 1992). Some species of crabs create large mounds while (< 300 m2) (Osborne & Smith, 1990). Interestingly, these excavating burrows. These mounds can range in area from patterns mimic herbivory by the beetle Coccotrypes rhizophorae 0.5 to 1.0 m2 (Minchinton, 2001). Atop these mounds, the in Panama (Sousa et al., 2003b); large gaps appear to serve soil elevation can be as much as 5-15 cm or more higher as refugia for propagules from predation by crabs and than surrounding soil and facilitate the establishment of herbivory by insects. Both Clarke & Kerrigan (2002) and mangrove species better adapted to lower frequencies of Clarke (2004) found additional support for the gap refugia tidal inundation, reduced flood durations and higher soil hypothesis relative to crab predation in that more oxygen status. Similarly, the mud ( anomala) propagules were killed by crabs under forest canopies than can alter soil elevations by as much as 1-2 m (Macnae, in large light gaps. Mechanistically, larger gaps have greater 1969a). Mangrove species richness is observably greater on irradiance which, in turn, increases the temperature of the top of such mounds in Palau and Papua New Guinea (K.W. soil (Smith, 1987c). Some experiments, on the other hand, Krauss and T.J. Smith III, personal observations). In Papua

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works Land crabs and tropical forest recruitment 211

New Guinea, as many as 250 T. anomala mounds can be enhanced by the more favourable soil conditions in the found per hectare, each having a mean size of 15 m2 and forest and by re-working of soils by crabs, crabs can also comprising greater than 20% of total mangrove area (T.J. serve as a detriment to the establishment of mangrove Smith III, unpublished observations). In addition, crab-free species through predation. Once an insect component is mangrove zones have higher soil sulphide and ammonium added, the relationship becomes even more complex. levels, which have also been associated with reduced Insect-related mortality, which is also related to soil mangrove growth (Smith et al., 1991). In fact, Minchinton conditions and light, can be as high as 72% under a forest (2001) argues that on sites where crabs are removed by canopy versus 10% in a gap (Sousa et al., 2003b). In situations a disturbance (e.g. oil spill), the site will have higher mangrove where propagule densities are really high, insects may be establishment and growth rates once those crabs recolonise. more effective predators than crabs (Robertson et al., 1990).

(4) Temporal variation III. ISLAND MARITIME FORESTS Timing of propagule production was important on a coral cay in Belize, where most propagule consumption by crabs Land crabs are vital inhabitants of tropical oceanic islands was evident at the beginning of major periods of propagule (Ehrhardt & Niaussat, 1970; Alexander, 1979; Hicks, fall and dispersal (McKee, 1995). Though mangroves are Rumpff & Yorkston, 1990; Jimenez et al., 1994; Ashmole & tropical and often produce seeds or propagules in some Ashmole, 2000). Their abundances commonly exceed capacity throughout the year (Williams, Bunt & Duke, 1981; one crab per m2 and 1000 kg ha-1 (Green, 1997). Their Twilley et al., 1997), there are often periods of time with biomass on some islands exceeds the total biomass of major inputs. This was the case for propagule predation reported in tropical rain forests in Costa Rica (115 studies conducted in Panama (Sousa & Mitchell, 1999), kg ha-1; Odum et al., 1970) and the central Amazon (210 kg Micronesia (Krauss & Allen, 2003) and Africa (Dahdouh- ha-1; Fittkau & Klinge, 1973). Reasons proffered for this Guebas et al., 1998), in which quantifying absolute rates of dominance include ecological release in the absence of propagule loss to crabs may have been affected by the natural enemies and competitors, high reproductive poten- availability of large quantities of propagules on the ground. tial and the ability to survive on a low-quality diet (Green, Studies are often designed to correspond to these peaks, 1997). Omnivory, combined with a low metabolic rate, may however, because we also expect crabs to have life-history be the key to the success of land crabs in island maritime strategies that take advantage of seasonal peaks in forests (Green, O’Dowd & Lake, 1997). Land crabs straddle propagule production. Robertson (1986) observed that three trophic levels and are free of many of the crabs decreased their consumption of fallen leaves as morphological and anatomical constraints that limit diet propagules became available; the crabs were eating the breadth of most consumers (Yodzis, 1984). propagules during this time of year and not the leaves. Anecdotes that mention that land crabs eat seeds, fruits Smith (1988), however, found that propagule density did not and seedlings on islands are rife (Bourne, 1886; Guppy, affect predation in at least one Australian system. Single 1890; Borradaile, 1901; English, 1913; Howard, 1950; propagules, and propagules in piles of 10 or 100, of both Degener & Gillaspy, 1955; Niering, 1956; Alexander, 1979). Avicennia marina and Bruguiera sexangula, were equally likely to The influence of land crabs in insular plant dynamics has be consumed, regardless of predator density (Smith, 1988). been quantified only recently and in just three island ecosystems (Fanning Island – Lee, 1985, 1988; Enewetak Atoll – Louda & Zedler, 1985; Christmas Island, Indian (5) Crab filter interactions Ocean – O’Dowd & Lake, 1989, 1990, 1991; Green et al., Interactive effects among canopy position, intertidal lo- 1997; Green, O’Dowd & Lake, 2008). Land crabs influence cation and predator guilds are common in mangroves. For insular plant community membership in two fundamental example, the relative importance of predation on prop- ways. They shape the probability of colonisation from the agules under the forest canopy versus in a gap was significant pool of immigrants arriving at an island (Ridley, 1930; for Aegiceras corniculatum in the lower intertidal zone only, and Louda & Zedler, 1985), or they mould community not in the upper intertidal zone (Osborne & Smith, 1990). membership throughout succession (Alexander, 1979). Propagule decay and predation by both crabs and snails Guppy (1906) attributed the early failure of Caesalpinia had a greater effect on recruitment of Avicennia germinans bonduc (a common drift seed) to establish on Anak Krakatau onto Spartina alterniflora marshes than recruitment into to the activities of land crabs. These crabs may assist Avicennia zones (Patterson et al., 1997). transport of sea-dispersed propagules from the shore to the An even more complex interaction involves alterations interior of islands (Ridley, 1930; Howard, 1950), but there is of soil conditions by vegetation, which in turn can affect little information on how land crabs influence the course of densities and foraging activities of crabs. Canopy cover and primary succession on islands, even in a well-studied system large root densities facilitate mangrove species recruitment like Krakatau (Thornton, 1995). Given the superabundance by altering the physico-chemical status of the soil (Gleason, of land crabs on many islands, it is not surprising that Ewel & Hue, 2003). This alteration, in turn, is more they can consume a large fraction of seeds and seedlings conducive to crab colonisation (Minchinton, 2001). Al- (Louda & Zedler, 1985; Lee, 1988; O’Dowd & Lake, 1990, though mangrove species establishment and growth are 1991), affecting seedling recruitment (Green et al., 1997).

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works 212 Erin S. Lindquist and others

(1) Fruit and seed predation These studies show that land crabs have a direct, differential effect on the survival of both the seed and young In all cases where fruit and seed selection have been seedling stages (Table 2). They may also act indirectly examined on islands, land crabs, including hermit and through their manipulation of leaf litter that is demonstrably gecarcinid crabs, displayed a catholic but choosy diet. Land important to seedling recruitment (Molofsky & Augspurger, crabs selectively remove fruits and seeds in controlled 1992; Benitez-Malvido & Kossmann-Ferraz, 1999; Table 2). experiments (Louda & Zedler, 1985; Lee, 1988; O’Dowd & Red crabs consume a large fraction of the annual litter and Lake, 1991; Green et al., 1997) but this does not necessarily seed input on Christmas Island, affecting litter cover and indicate differential fate of seeds. Seeds may be removed but biomass (O’Dowd & Lake, 1989; Green, O’Dowd & Lake, not eaten, and thus have the potential to germinate after 1999b). Their combined direct and indirect effect is best crab removal. However, seed fate experiments with the red demonstrated through their exclusion from 25 m2 plots land crab Gecarcoidea natalis in artificial burrows showed followed by long-term (three or six years) monitoring of differential predation on seeds of 10 species (Green et al., seedling recruitment. In the absence of G. natalis, not only did 1997). Seed predation varied from 0 to 94% over several seedling density and species richness increase approximately days. Physical protection and chemical deterrence were 20-fold and fivefold (Fig. 3A,B), respectively, but also there both likely to be important determinants of seed fate. Tough was a strong shift in seedling species composition. A sparse endocarps afforded complete protection to the seeds of two assemblage of only 1-3 species, largely unpalatable to G. tree species. For the other eight species, variation in seed natalis, populated the unfenced control plots in both coastal fate was probably related to both physical protection and lowland and upland plateau forest. These tree species also chemical deterrence (O’Dowd & Lake, 1991). Seeds of occurred on the exclusion plots, but an additional 10 to 12 some species are too small for land crabs to handle and their highly palatable species were found only in these exclusion survival probabilities on the forest floor remain unaffected. plots. These initial differences have persisted for seven years On Christmas Island, these species form a persistent and some of the earliest seedling recruits have grown into seedbank whose abundance and relative species composi- saplings, and a few to reproductive age (Green et al., 2008). tion is unaffected by even high densities of land crabs The compositional shifts were qualitatively consistent with (Green et al., 1999a). These crabs did not inhibit the predictions from the simple cafeteria and seedling transplant formation of a seedbank of 17,000 seeds m-2 of Muntingia experiments. In this island maritime system, the red land calabura, a tiny-seeded roadside weed. For larger seeds that crab regulates the dynamics of seedling recruitment in- land crabs can handle, the vast majority are completely dependently of gradients including forest type, understorey destroyed and consumed (Green et al., 1997). However, land light environment, elevation and distance from the sea. crabs may be effective dispersal agents for seeds of a few species with tough endocarps or chemical deterrents. For example, on Christmas Island, the seeds and seedlings of (3) Spatial variation Inocarpus fagifer are positively associated with burrow entrances (O’Dowd & Lake, 1991), consistent with Unlike mangrove and mainland coastal terrestrial forests, observations by Ridley (1930) who commonly saw a dozen few quantitative data are available to evaluate spatial or more seedlings growing around burrow entrances. On variation in the strength of the effect of land crabs on Fanning Island, Lee (1985) found that seedlings of Pandanus seedling recruitment on islands. Nevertheless, crab densities tectorius were, on average, further away from adults than vary within and between islands (Jimenez et al., 1994; Green undispersed seeds. She attributed this dispersion pattern to et al., 1997; Perez-Chi, 2005); therefore, the strength of the seed dispersal by Cardisoma carnifex, but other factors such as effect may vary accordingly. Local patchiness in the density seed predation by insects or limiting environmental of Gecarcinus planatus on Socorro Island is driven by habitat conditions near adults could be important, but were not patchiness; crab densities have been found to be lower in addressed in her study. areas with sparse canopy cover (Jimenez et al., 1994), or rocky substrata (Perez-Chi, 2005). On Christmas Island, (2) Seedling predation however, variation in land crab abundance is not related to edaphic factors or distance from the sea (P.T. Green, Strong evidence indicates that land crabs selectively con- unpublished observations) and may be primarily a conse- sume seedlings in transplant experiments using caged quence of variation in migration routes of adults and controls (O’Dowd & Lake, 1990; Green et al., 1997). emergence sites of juvenile crabs. Although the underlying Seedlings with higher nitrogen levels, fibre content and reasons for density variation remain obscure, patchiness in concentrations of total phenols experienced higher survivor- crab densities may generate patchiness in seedling recruit- ship, suggesting that selection of seedlings by G. natalis is ment. However, patchiness may have little functional determined by both structural and chemical attributes of significance for seedling recruitment if the crab density is seedlings (O’Dowd & Lake, 1990). Green et al. (1997) universally sufficient to limit seedling recruitment. The showed significant differences in seedling survival for 12 of probability of recruitment for many species is effectively nil the 16 species examined. Seedlings are vulnerable to land over the observed range of crab density on the island (0.2 – crab herbivory from germination to such a size where stems 2.5 burrows m-2) and it is only when the density of land are too woody or leaves are out of reach. This is certain to crabs has been artificially lowered, either through experi- vary among species but has never been quantified. mental exclusion (Green et al., 1997) or local extirpation by

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works Land crabs and tropical forest recruitment 213

A Coastal lowland Upland plateau 7 30

6 25 -2 5 -2 20 4 15 3 10 2 Seedlings m Seedlings m 1 5 0 0 B 0.5 0.8

0.4 0.6 -2 -2 0.3 0.4 0.2 Species m Species m 0.2 0.1

0 0 Control Exclusion Control Exclusion

C

Fig. 3. The effect of the red land crab Gecarcoidea natalis on seedling (A) abundance, (B) species richness and (C) species composition in coastal lowland and upland plateau rainforest types on Christmas Island, Indian Ocean. Red crabs were excluded from 4 or 5 25 m2 plots at coastal and inland sites for either three or six years, respectively (Green et al., 1997, 2008). Values in A and B are mean ^ 1 S.E. Open symbols are control plots and solid symbols are exclusion plots. Average crab density on paired control plots was 2.0 and 0.8 crabs per m2 at lowland and upland sites, respectively. For A, seedling abundance was significantly greater in exclusion plots in both lowland (t ¼ 7.67, df ¼ 3, P ¼ 0.005) and upland (t ¼ 8.67, df ¼ 4, P < 0.001) sites. For B, seedling species richness was significantly greater in exclusion plots in both lowland (t ¼ 8.72, df ¼ 3, P ¼ 0.003) and upland (t ¼ 8.57, df ¼ 4, P ¼ 0.001) sites. For the Non-Metric Multidimensional Scaling (nMDS) ordination of species composition in C, Global R and ANOSIM P were 0.73 and 0.067 at the lowland site, and 0.55 and 0.008 at the upland site, respectively. For the lowland site, only two of the four control plots had seedlings. The red crab, irrespective of forest type, distance from the ocean, and elevation, has an over-riding influence on the dynamics of seedling recruitment. an alien invader (O’Dowd, Green & Lake, 2003), that these could produce an ‘‘escape’’ window for some seedling plant species can recruit (Fig. 3A,B). species. Because land crabs are sensitive to changes in relative humidity any recruitment window could only occur during extended dry periods. However, these are the same (4) Temporal variation conditions that produce seedling stress and high mortality. On islands, land crab densities can vary diurnally (Audy, On Christmas Island, there is no evidence that seedling 1950; Page & Willason, 1982; Jimenez et al., 1994; Green, recruitment probabilities increase during the dry season 1997; Perez-Chi, 2005) and seasonally (Hicks, 1985; (Green et al., 1997). O’Dowd & Lake, 1990; Green, 1997; Perez-Chi, 2005). A decline in land crab densities over years or decades, The only time during which land crabs can affect seedling especially where they have been sustained at high densities, recruitment dynamics directly is the period between seed may dictate patterns of seedling recruitment and may be the dispersal and the attainment of a ‘‘crab-proof’’ seedling only way in which seedlings of some species recruit. Few size. This period is variable among seedling species but data are available to address this hypothesis. On Christmas typically will vary from weeks to months. Diurnal variation Island, red land crabs represent a widespread biotic filter to in crab density is unlikely to affect the probability of seed- seedling recruitment (Fig. 3). Their experimental exclusion ling recruitment, but seasonal variation in crab densities resulted in the recruitment of carpets of seedlings of a suite

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works 214 Erin S. Lindquist and others of canopy tree species, many of which never recruited on parable in terms of tree density (approximately 400 stems/ unfenced control plots (Green et al., 1997, 2008). These tree ha), basal area (approximately 50 m2/ha) and height species are especially palatable to red crabs and the (approximately 45 m) (Green et al., 1997; Lindquist & experiment predicted that they should be a minor compo- Carroll, 2004). Even with a larger pool of tree species, nent of the forest canopy, which should otherwise be mainland coastal terrestrial forests show distinct zonation in dominated by those few species whose seeds and seedlings the distribution, density and size structure of tree species are resistant to red crabs. In fact, palatable species dominate relative to distance from shore similar to that found in the forest canopy on Christmas Island, completely at odds mangrove ecosystems (Sherman, 2002; Lindquist & Carroll, with the prediction of the exclusion experiments and the 2004; Table 1; Fig. 4A,B). dominance-predation hypothesis applied to mangroves. Land crab populations in mainland coastal terrestrial However, the mismatch between the composition of the forests are less speciose than those of mangroves, yet canopy and seedlings in the presence of red crabs could be maintain similar densities to those on islands and in the result of a decline in red crab density or activity. Long- mangroves (Table 2). In two Costa Rican mainland coastal term recruitment failure of juveniles, epizootic events or terrestrial forests (Corcovado National Park, Osa Peninsula, catastrophic disturbance could cause abrupt or gradual Sherman, 2002; Cabo Blanco Absolute Nature Reserve, changes in crab densities but no historical evidence exists Nicoya Peninsula, Lindquist & Carroll, 2004), Gecarcinus for their occurrence. quadratus (Harlequin land crab, Gecarcinidae) densities decrease with increased distance from the coastline. In Corcovado, the crab zone extends from the shore to (5) Crab filter interactions approximately 600 m inland at which point the substrate shifts from sandy loam to clay and burrowing activity ceases Land crabs are susceptible to a wide range of verte- completely (Sherman, 2002). Within the crab zone, adult brate predators (Wolcott, 1988), but on islands their population densities are patchy ranging from 0 to 2.5 m-2 superabundance suggests that they have few natural with the highest burrow densities adjacent to buttressing enemies or competitors that can regulate population roots of large trees. The crab zone of G. quadratus in Cabo abundance (Table 2). Native ants compete for food with Blanco is similar but extends 7800 m inland with average land hermit crabs in (Morrison, 2002), but densities ranging from 3.3 crabs m-2 in the coastal crab there is no evidence that these interactions restrict crab zone (>100 m from coastline) to 0.76 crabs m-2 in the inland distribution and abundance, and thus the spatial gradient or crab zone (150-350 m from coastline) (Lindquist & Carroll, patchiness of their effect on plant recruitment. However, 2004; Fig. 4C). In Caribbean coastal forests of Veracruz, novel interactions between crabs and invasive alien Mexico, Gecarcinus lateralis (sometimes considered synony- predators and competitors make these ecosystems vulner- mous with G. quadratus) maintains mean densities of 0.5 able to rapid state shifts. If crab densities are severely crabs m-2 (Capistra´n-Barradas, Defeo & Moreno-Casasola, reduced by an introduced species, these island maritime 2003). In addition to G. quadratus (and its Caribbean forests have no other similar dominant seed and seedling counterpart G. lateralis; Hartnoll, 1988), two other species of predator. Plant recruitment, therefore, could be released crab inhabit Costa Rica’s Pacific coastline: a terrestrial from the natural crab filter (Denslow, 2003). On Christmas , Coenobita compressus (Coenobitidae), found Island, an invasive alien crazy ant (Anoplolepis gracilipes) has within 100 m of the shoreline, and a second Gecarcinid caused a rapid state change in the rainforest seedling land crab, Cardisoma crassum, found in frequently flooded community driven by its extirpation of the native red land clays bordering rivers within a few hundred meters of the crab across the island (O’Dowd et al., 2003). This coastline. introduced ant is also a predator of land crabs in the All of these terrestrial crabs found in neotropical Seychelles (Feare, 1999; Gerlach, 2004) and South Pacific mainland coastal forests forage principally on plant islands (Lester & Tavite, 2004). material, including leaf litter (Kellman & Delfosse, 1993; Sherman, 2003), seeds (Garcia-Franco, Rico-Gray & Zayas, 1991; Lindquist & Carroll, 2004; Capistra´n-Barradas & Moreno-Casasola, 2006) and seedlings (Delfosse, 1990; IV. MAINLAND COASTAL TERRESTRIAL Garcia-Franco et al., 1991; Thacker, 1996, 1998; Sherman, FORESTS 2002; Lindquist & Carroll, 2004). In the two mainland coastal terrestrial forests studied by the authors (Sherman, Mainland coastal terrestrial forests possess higher diversity 2002, 2003; Lindquist & Carroll, 2004), predation by land of trees and other woody plants than mangroves and most crabs affects seed and seedling survival differential to the island communities (Table 2). On the Osa Peninsula of coastal gradient and correlates with shifts in the forest Costa Rica for example, over 2100 species of plants from structure and species composition in the crab zone. over 185 families have been documented with an estimated 600 to 700 tree species (Quesada et al., 1997). In marked (1) Seed predation comparison to these mainland coastal terrestrial forests, the island maritime forest community of Christmas Island In mainland coastal terrestrial forests, land crabs are known contains approximately 50 canopy tree species (Green et al., to be significant seed predators providing a mechanism by 1997; Table 2). Forest stand structures, however, are com- which crabs may limit the establishment of plant species

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works Land crabs and tropical forest recruitment 215

(Table 2). By collecting seeds and carrying them to their the 0.3 m tall exclosure. The crab zone in Cabo Blanco is underground burrows (;1 m in depth) for consumption, deficient in terrestrial mice and rats, most likely due to land crabs prevent seeds from establishing through active competitive exclusion by the land crabs (R. Timm & predation or removal to depths inhospitable for seedling D. McClearn, unpublished observations). Seed removal in growth (Lindquist & Carroll, 2004). In one seed transplant the transplantation experiments did not differ among a experiment in Cabo Blanco, over 40% of non-protected wind-dispersed species, Terminalia oblonga, and two animal- seeds were removed by land crabs within five days of their dispersed species, Anacardium excelsum (Anacardiaceae) transplantation; however, removal of seeds transplanted into and Enterolobium cyclocarpum (Mimosoideae). By contrast, the crab exclosures was only 15% (Lindquist & Carroll, Capistra´n-Barradas & Moreno-Casasola (2006) documen- 2004). Seed removal in crab exclosures was most likely ted species preference by foraging G. lateralis among 10 caused by a combination of seed predators including species of fruits and seeds that varied in size and nutritional mammals (agoutis, coatis, raccoons, skunks, squirrels, deer, content. Crabs were found to remove species with large fruits and opposums) and birds that walked, climbed, or flew over or small seeds at lower rates than other seed types. T. Lee &

A B 3000 80

2500 -1 60 2000

1500 40

1000 20 Tree density ha 500 Tree speecies richness 0 0 Coastal Inland Coastal Inland

C 5

-2 4

3

2

Crab dennsity m 1

0 Coastal Inland

D E 100 100

80 80

60 60

40 40

20 20 Seeed survivorshipp (%) 0 Seedlling survivorshhip (%) 0 Coastal Inland Coastal Inland Fig. 4. (A) Tree density was higher (t ¼ -7.79, P < 0.0001) in the inland zone (1700 ^ 860 trees ha-1, N ¼ 36) than in the coastal zone (510 ^ 350 trees ha-1,N¼ 36) in a mainland terrestrial coastal forest in Cabo Blanco, Costa Rica (after Lindquist & Carroll, 2004). (B) Tree species richness (total in 0.28 ha sampled) was also higher in the inland zones (S ¼ 59) than coastal zones (S ¼ 18). (C) Densities of the land crab, Gecarcinus quadratus, were on average fourfold higher (t ¼ 6.3, P < 0.0001) in the coastal zone (0-100 m from coast, 3.0 ^ 1.4 crabs m-2,N¼ 22) than in the inland zone (150-350 m from coast, 0.76 ^ 0.78 crabs m-2, N ¼ 24). In coastal habitats where crab densities are higher, tree densities and species richness were lower. In the inland zone, tree density and species richness increased by almost the same magnitude (3.4 times and 3.5 times greater respectively) that crab density decreases (4.0 times less). (D) In experimental transplantations of seeds (N ¼ 36, 46 days) and (E) seedlings (N ¼ 25-47, 31 days), seed (X2¼ 28.3, P < 0.001) and seedling (X2¼ 7.85, P < 0.01) survivorships were higher in the inland zone where crab density is lower. Also note that seed and seedling survivorships were higher where crabs were excluded (dark bars) than where crabs were present (‘‘control’’, open bars).

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works 216 Erin S. Lindquist and others

P.M. Sherman (unpublished observations) also report G. crabs in mainland coastal terrestrial forests, similar to what quadratus preferences for certain fleshy fruits such as Simaba has been found in mangroves (Sousa & Mitchell, 1999). cedron (Simaroubaceae). Neither of these studies investigated the connection of crab predation specificity with tree distributions relative to the coastal gradient. (3) Spatial variation The magnitude of seed predation changes along a coastal (2) Seedling predation gradient in terrestrial mainland forests (Table 1). In Cabo In both mainland forests on the Pacific coast in Costa Rica Blanco, land crab densities and their removal rates of seeds (Corcovado, Sherman, 2002; Cabo Blanco, Lindquist & and seedlings are lower inland than they are closer to the Carroll, 2004), strong experimental evidence shows that coast (Lindquist & Carroll, 2004; Fig.4C,D,E). In the same crab predation of seedlings decreases tree seedling survivor- seed and seedling transplantation experiments mentioned ship. In Corcovado, after two years of experimental crab above (Lindquist & Carroll, 2004), crab predation was higher G. quadratus exclusion, average seedling densities within five protective for seeds and seedlings in the coastal zone (higher C. compressus exclosures increased 144% over baseline values while densities, present) than in the inland zone (lower G. quadratus C. compressus average control densities decreased for seedlings 3 to 25 densities, absent). This gradient of cm tall. Preference tests conducted in the crab zone predation pressure likely represents an important mechanism revealed a fivefold ratio of predatory removal rates for explaining the documented paucity of seedlings, saplings and seedling species only found in the crabless zone versus those adult small-stemmed individuals and reduced tree diversity in also found in the crab zone. These findings suggest that this forest’s most seaward zone. In addition, the three most those seedling species surviving in the crab zone are those dominant trees of the seaward zone are wind-dispersed Calycophyllum candidissimum, Lonchocarpus felipei, that are, for some reason, undesirable to crabs (i.e. similar to species ( P. quinata the assumptions of the dominance-predation hypothesis for ). Those species that disperse large numbers of small mangrove communities, see above). As a second control to seeds, such as wind-dispersed species, in any given fruiting this study, seedling species collected from the crabless and episode may have more seeds escape predation by crabs than sensu crab zones were transplanted into the crab zone and those species with lower fecundity ( Janzen, 1969). protected from crab predation. Both groups of seedlings In contrast to one mangrove crab study (Osborne & survived similarly well over six months with 83% and 70% Smith, 1990) and one mainland coastal forest study in G. lateralis ´ survival of the crabless zone and crab zone seedlings, Mexico of (Capistran-Barradas & Moreno- respectively. Such findings indicate that crabs are selective Casasola, 2006) which found land crabs to avoid large light > 2 and < 15% vegetation cover, respectively), in their seedling consumption and cause differential gaps ( 300 m but similar to several other mangrove studies (Sousa & survivorship of tree species propagules, and as a result Mitchell, 1999; Krauss & Allen, 2003; Allen et al., 2003; affect adult tree diversity and distributions in Corcovado’s coastal forest (Sherman, 2002). Clarke, 2004), our mainland findings show no relation be- tween crab foraging activity and canopy cover (Lindquist & In two transplantation studies in Cabo Blanco for 120 Carroll, 2004, P.M. Sherman, personal observations). Land days in 2001 and 40 days in 2002, seedling mortality was crabs removed seeds at similar rates in both gap (67-80% higher in unprotected treatments (74% or 100%, respect- canopy cover) and non-gap (86-91% canopy cover) plots in ively) than in exclosures (35% or 52%, respectively; the dry season when crab activity would be most influenced Lindquist & Carroll, 2004). As in the seed removal by microclimatic shifts in evaporation rate (Lindquist & experiments at the same site (see previous subsection), land Carroll, 2004). The more homogenous predation pressure crabs did not demonstrate predation selectivity between two reported may be due to the more nocturnal foraging species of seedlings, E. cyclocarpum and Pachira quinata (58% activity of G. quadratus compared to other species (Table 2), or 49%, respectively). Experimentation with additional seedling species should be conducted to substantiate this or by less stressful environmental conditions for crab foraging in smaller mainland forest light gaps (Sousa & initial finding. However, the results suggest that shifts in Mitchell, 1999). forest species composition along the coastal gradient are due to variation in crab predation pressure in space and In addition to the coastal gradient and canopy cover, increased crab densities in mainland coastal terrestrial time, not only species preferences. forests have been associated with: (1) increase in tree Despite the reported differences in seedling selectivity in buttress and root density (P.M. Sherman, personal obser- Sherman (2002) and Lindquist & Carroll (2004), both vations), (2) age of forest (Capistra´n-Barradas et al., 2003) studies found seedling predation by crabs to vary relative to and (3) increase in leaf litter (Capistra´n-Barradas et al., seedling size and stage of establishment. Sherman (2002) 2003). Further studies should assess how this spatial reports that seedlings at the cotyledon stage (0 – 3 cm), and variation in crab densities affects recruitment and sub- those taller (26 - 50 cm), were unaffected by crab exclusion. sequent coastal forest structure. Similarly, seedling predation by crabs did not limit survival of older seedlings (8 months after germination; Lindquist & Carroll, 2004). From these findings, we propose that crab (4) Temporal variation predation is higher for younger seedlings because they provide a higher nutritive reward. Size refugia, hence, may Land crabs have a distinct seasonality to their above-ground protect seedlings from removal of the recruitment pool by activity (Table 2). Because land crabs receive their oxygen

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works Land crabs and tropical forest recruitment 217 through both air- (branchial chambers) and water- (gills) control plots, and developed a thick humus and fungal breathing systems, their foraging is limited by higher hyphae layers. By collecting, hoarding and breaking down evaporation rates in the dry season of mainland coastal leaf litter in the burrow chambers (collections of up to 11.75 terrestrial forests (McMahon & Burggren, 1988). In Costa g dry mass have been retrieved), much of the litter falling in Rica populations, crabs emerge from their burrows to mate the crab zone decomposes below ground rather than at the during the first rains of the wet season, and remain active surface as it does in the crabless zone. This influence on throughout the next eight to nine months until the organic carbon stores and soil nutrient profiles can be beginning of the dry season in early January. At this point multiplied by the number of crabs burrowing in the forest. they retreat to their burrow chambers that provide Plant growth follows these nutrient-rich microsites; increased a homeostatic moist environment year-round. During the densities of fine to very fine roots have been found in and dry season, during which crabs are relatively inactive above around burrow chambers relative to same-depth samples ground (Capistra´n-Barradas et al., 2003; Sherman, 2003), from the crab zone that are unassociated with burrows the majority of tree species in seasonal tropical forests (Sherman, 2006). Furthermore in crab exclosures, crab zone disperse their seeds. We have observed, however, nocturnal seedlings survived better in bare soil plots, and crabless zone seed removal by crabs during the dry months (Lindquist & seedlings survived better in litter plots. The presence Carroll, 2004). Although seedling germination and estab- or absence of leaf-litter may act as an effect-modifier lishment typically occurs at the onset of the rains, those (Hosmer & Lemeshow, 1989) to the crab predation effect by seedlings that can establish early in the dry season or late in enhancing the survival of seedlings (Sherman, 1997). By the wet season when crab activities decline may benefit modifying the physical properties of the litter and soil layers, from a temporal refuge from this predation pressure. land crabs also act as ecosystem engineers facilitating plant Land crab above-ground activity also varies over a 24 h growth (Machicote, Branch & Villarreal, 2004). period in mainland coastal terrestrial forests (Table 2). In both Cabo Blanco and Corcovado, Gecarcinus quadratus are primarily nocturnal. Their nocturnal foraging activities may reflect sensitivity to the lower humidity and higher V. IMPLICATIONS daytime temperatures (Wolcott, 1988), and/or higher rates of daytime predation. Although avian and mammalian (1) Implications across ecosystems predators are active during both day (water, shore and riparian birds, birds of prey, grey foxes and coatis) and night By comparing the findings across three tropical coastal (owls, opossums, raccoons, kinkajous, olingos and ocelots), ecosystems, we have found that land crabs play similar nocturnal hours may provide crabs a greater opportunity to ecological roles and thus have similar impacts on tree escape predation. By contrast, Gecarcinus lateralis in Mexico recruitment in the different habitats. In all three ecosystems, is more diurnal, foraging between 07:00 and 11:00h crabs remove seeds or propagules at rates significant to (Capistra´n-Barradas et al., 2003). impact plant establishment (Table 2). Although crab Similarly to island maritime forests, year-to-year period- damage to seedlings is minor in mangroves, crabs further icity (cf. Wolcott, 1988) in crab densities may provide limit plant establishment through their preferential pre- a recruitment window for tree species where seedling dation of seedlings in island maritime and mainland coastal survival will be higher, particularly for species excluded terrestrial forests (Table 2). from the crab zone. The presence of adult trees of species Crab predation of seeds, propagules and seedlings can that are preferred by crabs as seeds or seedlings suggests vary in space and time, but not necessarily in the same that through some combination of spatial or temporal direction among the three ecosystems (Table 2). Findings in avoidance and predator satiation, plant individuals may mangroves suggest that there is not always a change in crab survive crab predation and recruit successfully. A species-by- predation pressure along the intertidal gradient. There is species analysis of trees, in regards to their palatability as also no shift in crab predation pressure in island maritime propagules and seedlings and their phenology of seed set, is forests; island land crabs have a large, negative impact on a recommended avenue for further research. seedling establishment throughout their range. In mainland coastal terrestrial forests, however, crab predation pressure has been found to be higher closer to the coastline where Gecarcinus spp. densities are highest and the terrestrial (5) Crab filter interactions hermit crab, Coenobita compressus, is present. Crab predation Land crabs can have indirect positive effects on seedling of seeds, propagules and seedlings has not been found to microsite conditions through leaf litter subterranean vary with canopy cover in mangroves and mainland coastal accumulation. Despite the propensity for land crabs to terrestrial forests (Table 2). Yet some mangrove studies have remove fruits, seeds and seedlings, they also forage on the observed the opposite; propagule removal was higher under abscised leaves that accumulate on the ground. In full canopy than in large gaps. This discrepancy in the Corcovado during the wet season, the removal of leaves findings may be an artifact of variable gap size used in the can result in broad expanses of forest floor (tens or hundreds different investigations and seasonality in data collections. of m2) completely or nearly devoid of accumulated litter Crabs are known to be less active outside their burrows (Sherman, 2003). Experimental exclosures accumulated during periods of low humidity and high sun intensity in more leaf litter over the two-year period (5.6 times) than did order to conserve water stores.

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works 218 Erin S. Lindquist and others

Because low water availability limits land crab activity harvesting of mangrove tree species reduces populations of above ground, island and mainland coastal terrestrial foliovorous and other sesarmid crabs (Emmerson & Ndenze, studies have found that crabs are active primarily in the 2007). Transformation of coastal forests in southern North rainy season when humidity is high (Table 2). To minimize America has obliterated populations of Gecarcinus lateralis and further water loss, some mainland studies have found land Cardisoma guanhumi (Wolcott, 1988). Habitat in many, if not crabs to be nocturnal (Sherman, 2002; Lindquist & Carroll, most insular ecosystems dominated by land crabs (Borradaile, 2004). In mainland coastal terrestrial forests on the 1901), has been drastically reduced through land trans- Caribbean coast of Mexico, however, Capistra´n-Barradas formation including urbanization, conversion to agriculture, et al. (2003) observed Gecarcinus lateralis to be diurnal. mining and war (e.g., Baine et al.,2006). Observations of land crabs in island communities also have Recent and rapid changes in land crab densities on found crabs to be either diel or diurnal with variation in islands are also occurring against a historical background of their daily activity patterns due to variable humidity levels harvesting of land crabs by humans (Moul, 1954; Bates & (Table 2). In all three ecosystems, crab activity is highest Abbott, 1958; Niering, 1963; Lutz & Austin, 1983; Ashmole & when the majority of seeds and propagules are dispersed, Ashmole, 2000; Pain et al., 2000). Most land crabs are tasty typically at the beginning of the rainy season in seasonal and in many circumstances have formed an important forests. By coinciding with the preferable period for seedling dietary component for local people (e.g. Foale, 1999), establishment, the recruitment limitation effect of crabs is frequently to the point of overexploitation. On the island of amplified. Although the negative impact of land crabs on Saba in the lesser , the mountain land crab Gecarcinus plant recruitment is universally significant among coastal ruricola formed a primary component in Amerindian diets ecosystems, even the most palatable tree species have been over 3000 years ago, but is now endangered through able to recruit to maintain their populations. Differential modern day hunting (Hofman & Hoogland, 2003). predation pressure relative to the tidal location or tree Legislation has been enacted to protect G. ruricola (Baine species would allow variable recruitment in space. Less et al., 2007). Local peoples have developed specialized traps palatable species may dominate the areas with more to catch gecarcinid crabs in both the Caribbean and the predation pressure (dominance-predation hypothesis), and Philippines (Maitland, 2002). The coconut or robber crab more palatable species may be more abundant where crab (Birgus latro) is so prized that numbers have been drastically densities are low. However, differential predation pressure reduced across its Indo-Pacific range (Lavery, Moritz & does not explain the observed presence of palatable species Fielder, 1996). In other situations, land crabs have been in areas with high crab densities. We suggest that annual hunted as a nuisance. On Ascension Island, bounties placed variability in crab densities may provide windows for on Gecarcinus (Johngarthia) lagostoma in 1879 resulted in recruitment of palatable species when crab densities are low the destruction of 80,000 individuals from croplands (Table 2). (Ashmole & Ashmole, 2000). In addition to predation of seeds and seedlings, crabs also Where land crab abundance has been drastically impact forest recruitment through the removal of leaf litter reduced, it is almost certain that the contemporary structure in all three ecosystems (Table 2). The removal of leaf litter and functioning of these forest communities has been can be detrimental to seedling establishment by minimising altered from its original state. In mangrove systems two soil nutrient replenishment and erosion of topsoil. However, genera, Cardisoma and Ucides, which have profound impacts by collecting the litter in their subterranean burrows, and on nutrient cycling within the forest, are harvested for maintaining their burrows through regular soil movement, human food. By processing leaf litter these land crabs help crabs may facilitate seedling establishment through an retain nutrients in the forest ecosystem. Ecological models increase in soil nutrient and carbon levels (Table 2). In have shown that nutrient retention increases forest pro- mangroves, crab burrows may be high enough to serve as ductivity, which in turn provides a positive feedback to crab refugia from tidal inundation. biomass (Wolff, Koch & Issac, 2000; Koch & Wolff, 2002). There is evidence that subsistence fisheries may be overharvesting these crabs in many locations. This could (2) Conservation implications lead to decreases in forest production which in turn could result in further reductions in crab populations, alterations In this review, we have argued that land crabs are a key in sediment geochemistry, followed by shifts in mangrove driver of seedling recruitment in mangrove, insular and tree species composition and dominance. The portunid, coastal forests. Although this illustrates their current , is an important species in the Indo-West-Pacific importance, it is certain that they played a much larger region although it does not consume mangrove propagules role in the past. At least two human-induced drivers of or seedlings. It does however prey on other crabs in the global change – land transformation and overexploitation – forests. Overharvest of Scylla serrata for human consumption have caused major declines in the distribution and is likely to affect the influence of crabs on mangrove abundance of land crabs by reducing the areal extent and ecosystems indirectly (Demopoulos et al., 2008). All three quality of coastal ecosystems. Mangroves have declined in tropical ecosystems highlighted in this review are in decline area by 35% globally, which is more than tropical due to rapid coastal development for tourism and foreign rainforests and coral reefs (Valiela, Bowen & York, 2001). residences around the world (Valiela et al., 2001; Rivera- Loss of mangroves has caused declines in benthic bio- Monroy et al., 2004). Given that land crabs serve as key diversity, including land crabs (Ellison, 2008). Even selective players in these coastal ecosystems, we recommend that

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works Land crabs and tropical forest recruitment 219 their ecological role discussed in this review be included in VII. ACKNOWLEDGEMENTS forest conservation and management plans. We would like to thank those contributors to a special session on this topic at the American Society of Limnology VI. CONCLUSIONS and Oceanography (ASLO) Summer 2006 meeting in Victoria, B.C., Canada, including Wayne P. Sousa (UC- Berkeley, USA), Todd E. Minchinton (University of (1) Current dogma suggests that biotic factors may be less Wollongong, Australia), Jose L. Gutierrez (Universidad important in insular ecosystems than in mainland ecosys- Nacional de Mar del Plata, Argentina), Kauoa M. Fraiola tems (e.g. Elton, 1958; Carlquist, 1965; MacArthur, 1972; (University of Georgia, USA) and Alan P. Covich Janzen, 1973). This view might also be extended to coastal (University of Georgia, USA). James B. Grace, Wayne ecosystems in general, like those reviewed here, where P. Sousa and two anonymous referees provided excellent salinity and flood gradients are assumed to be the prime reviews of earlier versions of this manuscript. regulators of plant establishment. However, the ubiquity of land crabs and their effect on plant recruitment in tropical, coastal forests is not consistent with this view. (2) Where their role has been quantified, land crabs are consistently found to be key determinants of seedling VIII. REFERENCES recruitment at large spatial and temporal scales on islands and continental landscapes. In fact, the magnitude of their ALEXANDER, H. G. L. (1979). A preliminary assessment of the role impact on seedling recruitment can be severalfold greater of the terrestrial decapod in the Aldabran than for coastal rainforests with a diverse assemblage of ecosystem. Philosophical Transactions of the Royal Society, London B seed and seedling predators (Molofsky & Fisher, 1993; 286, 241–246. Terborgh & Wright, 1994; Terborgh et al., 2001). Crabs ALLEN, J. A., KRAUSS,K.W.&HAUFF, R. D. (2003). Factors limiting are influential in this manner for at least five reasons. First, the intertidal distribution of the mangrove species Xylocarpus land crabs exert strong and consistent biotic influences granatum. Oecologia 135, 110–121. across two plant life-history stages: seed/fruit/propagule ALONGI, D. M., BOTO,K.G.&ROBERTSON, A. I. (1992). Nitrogen and phosphorus cycles. In Tropical Mangrove Ecosystems (eds. A. I. and seedling. Second, the effect of abiotic environmental Robertson and D. M. Alongi), pp. 251–292. Coastal and gradients (e.g., elevation, rainfall, salinity) is not as direct as Estuarine Studies 41, American Geophysical Union, Washington, that of crabs because abiotic factors may only decrease D.C. growth rates, and do not always extirpate certain species’ ASHMOLE,P.&ASHMOLE, M. (2000). St Helena and Ascension Island: recruitment pools (Fig. 1). Third, where land crab popula- a natural history. Anthony Nelson, Oswestry. tions are dense, they often exclude, or at least minimize, AUDY, J. R. (1950). A visit to Jarak Island in the Malacca Straits. other seed and seedling predators as well as herbivores. Malayan Nature Journal 5, 38–46. Fourth, land crabs indirectly influence seedling recruitment BAINE, M., HOWARD, M., TAYLOR, E., JAMES, J., VELASCO, A., dynamics through the removal of the leaf litter layer and GRANDAS,Y.&HARTNOLL, R. (2007). The development of subsequent manipulation of the soil nutrient profiles. management options for the black land crab (Gecarcinus ruricola) Finally, land crabs prey on the earliest stages of plant catchery in the San Andres Archipelago, . Ocean & recruitment, and thus initially restrict the long-term impact Coastal Management 50, 564–589. that environmental conditions can have on survival and BALL, M. C. (1998). Mangrove species richness in relation to growth of the seedling. salinity and waterlogging: a case study along the Adelaide River (3) Studies reviewed here provide evidence that land crab floodplain, northern Australia. Global Ecology and Biogeography assemblages shape tree species recruitment in a dominant Letters 7, 73–82. manner in several tropical locales. Crabs directly limit the BATES,M.&ABBOTT, D. P. (1958). Coral Island: portrait of an atoll. distribution and density of trees along coastal gradients and Charles Scribner and Sons, New York. facilitate plant establishment through several direct and BEEVER, J. W., III, SIMBERLOFF,D.&KING, L. L. (1979). Herbivory indirect mechanisms. Given that all three coastal ecosys- and predation by the mangrove tree crab . Oecologia tems vary in the strength of their environmental gradients, 43, 317–328. BENITEZ-MALVIDO,J.B.&KOSSMANN-FERRAZ, I. D. (1999). Litter we suggest that crabs do have primacy in many forests over cover variability affects seedling performance and herbivory. other, better studied abiotic effects, such as desiccation, Biotropica 31, 598–606. flooding and salinity (Fig. 1). Tree recruitment models for BORRADAILE, L. A. (1901). Land crustaceans. In The Fauna and these systems should incorporate both abiotic and biotic Geography of the Maldive and Laccadive Archipelagoes: being an account of factors affecting seed and seedling survivorship, but the work carried on and of the collections made by an expedition during the determine which factor(s) are the primary filters, and which years 1899 and 1900 (ed. Gardiner J. S.), pp. 64–100. Vol. 1, Pt 1 cause secondary mortality. Using our example of how one University Press, Cambridge. biotic factor can transcend abiotic variation across different BOURNE, G. C. (1886). General observations on the fauna of Diego tropical coastal ecosystems, we hope to encourage others to Garcia, Chagos Group. Proceedings of the Zoological Society of London isolate any common factor(s) driving plant recruitment that 1886, 331–334. in turn will influence forest structure and composition in BURGGREN,W.W.&MCMAHON, B. R. (1988). Biology of the Land other ecosystems not discussed here. Crabs. Cambridge University Press, Cambridge.

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works 220 Erin S. Lindquist and others

CAMACHO-CE´ SPEDES,F.&LINDQUIST, E. S. (2007). Common Trees of DEVLIN, D. J. (2004). Effects of the parasitic Scolytid beetle, Coccotrypes Cabo Blanco Absolute Nature Reserve. Instituto Nacional de rhizophorae, on population and community structure of forests dominated Biodiversidad (INBio), Costa Rica. by Rhizophora mangle in southwest Florida. Ph.D. Thesis, CANNICCI, S., BURROWS, D., FRATINI, S., LEE, S. Y., SMITH, T. J., III, University of Louisiana at Lafayette, Lafayette. OFFENBER,J.&DAHDOUH-GUEBAS, F. (2008). Faunal impacts on DIELE, K., KOCH,V.&SAINT-PAUL, U. (2005). Population structure, vegetation structure and ecosystem function in mangrove catch composition and CPUE of the artisanally harvested forests: a review. Aquatic Botany 89, 186–200. mangrove crab Ucides cordatus (Ocypodiae) in the Caete estuary, CARLQUIST,S.(1965).Island Life. Natural History Press, Garden City. North Brazil: Indications of overfishing. Aquatic Living Resources CAPISTRA´ N-BARRADAS, A., DEFEO,O.&MORENO-CASASOLA,P. 18, 169–178. (2003). Density and population structure of the red land crab DOUST,S.J.,ERSKINE,P.D.&LAMB, D. (2006). Direct seeding to Gecarcinus lateralis in a tropical semi-deciduous forest in Veracruz, restore rainforest species: Microsite effects on the early Mexico. Interciencia 28, 323–327. establishment and growth of rainforest tree seedlings on CAPISTRA´ N-BARRADAS,A.&MORENO-CASASOLA, P. (2006). Post- degraded land in the wet tropics of Australia. Forest Ecology and dispersal fruit and seed removal by the crab Gecarcinus lateralis in Management 234, 333–343. a coastal forest in Veracruz, Mexico. Biotropica 38, 1–7. DUKE, N. C. (2006). Australia’s Mangroves: the authoritative guide to CERON, R. M. B., PADILLA, H. G., BELMONT, R. D., TORRES,M.C. Australia’s mangrove plants. University of Queensland Press. Brisbane. B., GARCIA,R.M.&BAEZ, A. P. (2002). Rainwater chemical DUKE, N. C., BALL,M.C.&ELLISON, J. C. (1998). Factors composition at the end of the mid-summer drought in the influencing biodiversity and distributional gradients in man- Caribbean shore of the Yucatan Peninsula. Atmospheric Environ- groves. Global Ecology and Biogeography Letters 7, 27–47. ment 36, 2367–2374. EHRHARDT,J.-P.&NIAUSSAT, P. (1970). Ecologie et physiologie du CHAPMAN, V. J. (1976). Mangrove Vegetation. J. Cramer, Valduz. brachyoure terrestre Gecarcinus planatus Stimpson (d’apre`s les CLARKE, P. J. (2004). Effects of experimental canopy gaps on individus de l’atoll de Clipperton). Bulletin de la Societe Zoologique mangrove recruitment: lack of habitat partitioning may explain de France 95, 41–54. stand dominance. Journal of Ecology 92, 203–213. ELLISON, A. M. (2008). Managing mangroves with benthic CLARKE,P.J.&KERRIGAN, R. A. (2002). The effects of seed biodiversity in mind: moving beyond roving banditry. Journal predators on the recruitment of mangroves. Journal of Ecology of Sea Research 59, 2–15. 90, 728–736. ELTON, C. (1958). The Ecology of Invasions by Animals and Plants. CLARKE,P.J.&MYERSCOUGH, P. J. (1993). The intertidal University of Chicago Press, Chicago. distribution of the grey mangrove (Avicennia marina) in south- EMMERSON,W.D.&NDENZE, T. T. (2007). Mangrove tree eastern Australia: the effects of physical conditions, interspecific specificity and conservation implications of the arboreal crab competition, and predation on propagule establishment and Parasesarma leptosoma at Mngazana, a mangrove estuary in the survival. Australian Journal of Ecology 18, 307–315. Eastern Cape, South Africa. Wetlands Ecology and Management COLEY,P.D.&BARONE, J. A. (1996). Herbivory and plant defenses 15, 13–25. in tropical forests. Annual Review of Ecology and Systematics 27, 305– ENGLISH, T. M. S. (1913). Some notes from a West Indian coral 335. island. Bulletin of Miscellaneous Information 10, 367–373. CONDE,J.E.&DIAZ, H. (1989). The mangrove tree crab Aratus FARNSWORTH,E.J.&ELLISON, A. M. (1997). Global patterns of pre- pisonii in a tropical estuarine coastal lagoon. Estuarine, Coastal and dispersal propagule predation in mangrove forests. Biotropica Shelf Science 28, 639–650. 29, 318–330. DAHDOUH-GUEBAS, F., VERNEIRT, M., TACK,J.F.&KOEDAM,N. FEARE, C. J. (1999). Ants take over from rats on Island, (1997). Food preferences of Neosarmatium meinerti de Man Seychelles. Bird Conservation International 9, 95–95. (: Sesarminae) and its possible effects on the regen- FITTKAU,E.J.&KLINGE, H. (1973). On biomass and trophic eration of mangroves. Hydrobiologia 347, 83–89. structure of the central Amazonian rain forest ecosystem. DAHDOUH-GUEBAS, F., VEREIRT, M., TACK,J.F.,VAN SPEYBROECK, Biotropica 5, 2–14. N. & KOEDAM, N. (1998). Propagule predators in Kenyan FOALE, S. (1999). Local ecological knowledge and biology of the mangroves and their possible effect on regeneration. Marine land crab Cardisoma hirtipes (Decapoda: Gecarcinidae) at West Freshwater Research 49, 345–350. Nggela, Solomon Islands. Pacific Science 53, 37–49. DEGENER,O.&GILLASPY, E. (1955). Canton Island, South Pacific. FRUSHER,S.D.,GIDDINS,R.L.&SMITH, T. J., III. (1994). Atoll Research Bulletin 41, 1–51. Distribution and abundance of grapsid crabs (Grapsidae) in DELFOSSE, B. (1990). The effect of the red land crab, Gecarcinus a mangrove estuary: effects of salinity tolerances, osmoregula- lateralis, on the litter layer, nutrient availability and seedling recruitment in tory ability, and sediment characteristics. Estuaries 17, 647–654. a semi-deciduous seasonal dry tropical forest. M.S. Thesis. York GARCIA-FRANCO, J. G., RICO-GRAY,V.&ZAYAS, O. (1991). Seed and University, Toronto. seedling predation of Bromelia penquin L. by the red land crab DEMOPOULOS,A.W.J.,CORMIER, N., EWEL,K.C.&FRY, B. (2008). Gecarcinus lateralis Frem. in Veracruz, Mexico. Biotropica 23, 96–97. Use of multiple chemical tracers to define habitat use of Indo- GERLACH, J. (2004). Impact of the invasive crazy ant Anoplolepis Pacific mangrove crab, Scylla serrata (Decapoda: Portunidae). gracilipes on Bird Island, Seychelles. Journal of Insect Conservation 8, Estuaries and Coasts 31, 371–381. 15–25. DENSLOW, J. S. (2003). Weeds in paradise: thoughts on the GIFFORD, C. A. (1962). Some observations on the general biology invasibility of tropical islands. Annals of Missouri Botanical Gardens of the land crab, Cardisoma guanhumi (Latreille), in south Florida. 90, 119–127. Biological Bulletin 123, 207–223. DE STEVEN,D.&WRIGHT, S. J. (2002). Consequences of variable GLEASON, S. M., EWEL,K.C.&HUE, N. (2003). Soil redox reproduction for seedling recruitment in three neotropical tree conditions and plant-soil relationships in a Micronesian man- species. Ecology 83, 2315–2327. grove forest. Estuarine, Coastal and Shelf Science 56, 1065–1074.

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works Land crabs and tropical forest recruitment 221

GREEN, P. T. (1997). Red crabs in rain forest on Christmas Island, HOWARD, R. A. (1950). Vegetation of the Bimini Island group. Indian Ocean – activity patterns, density and biomass. Journal of Ecological Monographs 20, 314–349. Tropical Ecology 13, 17–38. HUBBELL, S. P., FOSTER, R. B., O’BRIEN, S. T., HARMS, K. E., GREEN, P. T. (2004). Burrow dynamics of the red land crab CONDIT, R., WECHSLER, B., WRIGHT,S.J.&LOO DE LAO,S. Gecarcoidea natalis (Brachyura, Gecarcinidae) in rain forest on (1999). Light-gap disturbances, recruitment limitation, and tree Christmas Island (Indian Ocean). Journal of Biology diversity in a neotropical forest. Science 283, 554–557. 24, 340–349. JANZEN, D. H. (1969). Seed-eaters versus seed size, number, toxicity GREEN,P.T.,HART, R., JANTUN, J., METCALFE, D. J., O’DOWD,D.J.& and dispersal. Evolution 23, 1–27. LAKE, P. S. (1999a). Red crabs in rain forest on Christmas Island, JANZEN, D. H. (1971). Seed predation by animals. Annual Review of Indian Ocean: no effect on the soil seed bank. Australian Journal of Ecology and Systematics 2, 465–492. Ecology 24, 90–94. JANZEN, D. H. (1973). Dissolution of mutualism between Cecropia GREEN, P. T., O’DOWD,DJ.&LAKE, P. S. (1997). Control of and its Azteca ants. Biotropica 5, 15–28. seedling recruitment by land crabs in rain forest on a remote JANZEN, D. H. (1985). Spondias mombin is culturally deprived in oceanic island. Ecology 78, 2474–2486. megafauna-free forest. Journal of Tropical Ecology 1, 131–155. GREEN, P. T., O’DOWD,D.J.&LAKE, P. S. (1999b). Monopolization JIME´ NEZ, C., ORTEGA-RUBIO, A., ALVAREZ-CA´ RDENA,S.&ARNAUD, of litter processing by a dominant land crab on a tropical G. (1994). Ecological aspects of the land crab Gecarcinus planatus oceanic island. Oecologia 119, 235–244. (Decapoda: Gecarcinidae) in Socorro Island, Mexico. Biological GREEN, P. T., O’DOWD,D.J.&LAKE, P. S. (2008). Recruitment Conservation 69, 9–13. dynamics in a rainforest seedling community: context-independent JONES, C. G., LAWTON,J.H.&SHACHAK, M. (1994). Organisms as impact of a keystone consumer. Oecologia 156, 373–385. ecological engineers. Oikos 69, 373–386. GUPPY, H. B. (1890). The dispersal of plants as illustrated by the KELLMAN,M.&DELFOSSE, B. (1993). Effect of the red land crab flora of the Keeling or Cocos Islands. Journal of Transactions of the (Gecarcinus lateralis) on the leaf-litter in a tropical dry forest in Victoria Institute 24, 267–306. Veracruz, Mexico. Journal of Tropical Ecology 9, 55–65. GUPPY, H. B. (1906). Observations of a naturalist in the Pacific KOCH,P.T.&WOLFF, M. (2002). Energy budget and ecological between 1856 and 1899. Plant Dispersal, Vol. 2. MacMillen Co., role of mangrove epibenthos in the Caete estuary, North Brazil. London. Marine Ecology Progress Series 228, 119–130. GUTIE´ RREZ, J. L., JONES, C. L., GROFFMAN, P. M., FINDLAY,S.G., KRAUSS,K.W.&ALLEN, J. A. (2003). Factors influencing the IRIBANE,O.O.,RIBEIRO,P.D.&BRUSCHETTI, C. M. (2006). The regeneration of the mangrove Bruguiera gymnorrhiza (L.) Lamk. contribution of crab burrow excavation to carbon availability in on a tropical Pacific island. Forest Ecology and Management 176, surficial salt-marsh sediments. Ecosystems 9, 647–658. 49–60. HARTNOLL, R. G. (1988). Evolution, systematics, and geographical KRAUSS, K. W., LOVELOCK, C. E., MCKEE, K. L., LO´ PEZ-HOFFMAN, distribution. In Biology of the Land Crabs (eds. W. W. Burggren and L., EWE,S.M.L.&SOUSA, W. P. (2008). Environmental drivers B. R. McMahon), pp. 6–54. Cambridge University Press, in mangrove establishment and early development: a review. Cambridge. Aquatic Botany 89, 105–127. HARTNOLL, R. G., BAINE, M. S. P., GRANDAS, Y., JAMES,J.&ATKIN, KYEREH, B., SWAINE,M.D.&THOMPSON, J. (1999). Effect of light H. (2006). Population biology of the black crab, Gecarcinus on the germination of forest trees in Ghana. Journal of Ecology ruricola, in the San Andres archipelago, western Caribbean. 87, 772–783. Journal of Crustacean Biology 26, 316–325. LAVERY, S., MORITZ,C.&FIELDER, D. R. (1996). Indo-Pacific HERMY,M.&VERHEYEN, K. (2007). Legacies of the past in the population structure and evolutionary history of the coconut present-day forest biodiversity: a review of past land-use effects crab Birgus latro. Molecular Ecology 5, 557–570. on forest plant species composition and diversity. Ecological LEE, M. A. (1985). The dispersal of Pandanus tectorius by the land Research 22, 361–371. crab Cardisoma carnifex. Oikos 45, 169–173. HERREID,C.F.,II&GIFFORD, C. A. (1963). The burrow habitat LEE, M. A. (1988). Food preferences and feeding behaviour of the of the land crab, Cardisoma guanhumi (Latreille). Ecology land crab Cardisoma carnifex. Micronesica 21, 274–279. 44, 773–775. LEE, S. Y. (1998). Ecological role of grapsid crabs in mangrove HICKS, J. (1985). The breeding behaviour and migrations of the ecosystems: a review. Marine and Freshwater Research 49, Gecarcoidea natalis (Decapoda: Brachyura). 335–343. Australian Journal of Zoology 33, 127–142. Also see accessory LESTER,P.J.&TAVITE, A. (2004). Long-legged ants, Anoplolepis publication: The population, biomass and behaviour of gracilipes (Hymenoptera: Formicidae), have invaded Tokelau, the terrestrial crab Gecarcoidea natalis (Pocock)(Decapoda: changing composition and dynamics of ant and invertebrate Brachyura). communities. Pacific Science 58, 391–401. HICKS, J., RUMPFF,H.&YORKSTON, H. (1990). Christmas Crabs.2nd LINDQUIST, E. S. (2003). Patterns of coastal forest composition, structure, edn. Christmas Island Natural History Association. Christmas and recruitment, Costa Rica: functions of an environmental gradient, seed Island. rain distribution, and crab predation pressure. Ph.D. Thesis. University HOFMAN,C.L.&HOOGLAND, M. L. P. (2003). Plum piece: of Georgia, Athens. evidence for archaic seasonal occupation on Saba, northern LINDQUIST,E.S.&CARROLL, R. C. (2004). Differential seed and lesser Antilles around 3300 BP. Journal of Caribbean Archeology seedling predation by crabs: impacts on tropical coastal forest 4, 12–27. composition. Oecologia 141, 661–667. HOGARTH, P. J. (1999). The Biology of Mangroves. Oxford University LOUDA,S.M.&ZEDLER, P. H. (1985). Predation in insular plant Press, Oxford. dynamics: an experimental assessment of postdispersal fruit and HOSMER,D.W.&LEMESHOW, S. (1989). Applied Logistic Regression. seed survival, Enewetak Atoll, Marshall Islands. American Journal Wiley, New York. of Botany 72, 438–445.

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works 222 Erin S. Lindquist and others

LUTZ,P.L.&AUSTIN, C. B. (1983). Land crabs: a new resource seedling communities determined by environment or by seed potential. Gulf and Caribbean Fisheries Institute Proceedings 35, 6–16. arrival? A test in a neotropical forest. Journal of Ecology 95, 507–516. MACARTHUR, R. H. (1972). Geographical Ecology: patterns in the O’DOWD, D. J., GREEN,P.T.&LAKE, P. S. (2003). Invasional distribution of species. Harper & Row, New York. ‘meltdown’ on an oceanic island. Ecology Letters 6, 812–817. MACHICOTE, M., BRANCH,L.C.&VILLARREAL, D. (2004). O’DOWD,D.J.&LAKE, P. S. (1989). Red crabs in rain forest, Burrowing owls and burrowing mammals: are ecosystem Christmas Island: removal and relocation of leaf-fall. Journal of engineers interchangeable as facilitators? Oikos 106, 527–535. Tropical Ecology 5, 337–348. MACNAE, W. (1969a). A general account of the fauna and flora of O’DOWD,D.J.&LAKE, P. S. (1990). Red crabs in rain forest, mangrove swamps and forests in the Indo-West-Pacific region. Christmas Island: differential herbivory of seedlings. Oikos 58, Advances in Marine Biology 6, 73–270. 289–292. MACNAE, W. (1969b). Zonation within mangroves associated with O’DOWD,D.J.&LAKE, P. S. (1991). Red crabs in rain forest, estuaries in north Queensland. In Estuaries (ed. G. H. Lauff), Christmas Island: removal and fate of fruits and seeds. Journal of pp 432–441. American Association for the Advancement of Tropical Ecology 7, 113–122. Science, Washington, D.C. ODUM, H. T., ABBOTT, W., SELANDER, R. K., GOLLEY,F.B.& MAITLAND, D. P. (2002). Convergent design of Caribbean and WILSON, R. F. (1970). Estimates of chlorophyll and biomass of Philippine bamboo land-crab traps. Journal of Crustacean Biology the Tobonuco Forest of Puerto Rico. In A Tropical Rain Forest: 22, 497–501. a study of irradiation and ecology at El Verde, Puerto Rico (eds. H. T. MCGUINNESS, K. A. (1997). Seed predation in a tropical mangrove Odum and R. F. Pigeon), pp. 3–18. U.S. Atomic Energy forest: a test of the dominance-predation model in northern Commission, Washington, D.C. Australia. Journal of Tropical Ecology 13, 293–302. ONUF, C. P., TEAL,T.M.&VALIELA, I. (1977). Interactions of MCIVOR,C.C.&SMITH, T. J., III. (1995). Differences in the crab nutrients, plant growth and herbivory in a mangrove ecosystem. fauna of mangrove areas at a southwest Florida and northeast Ecology 58, 514–526. Australia location: implications for leaf litter processing. Estuaries OSBORNE,K.&SMITH, T. J., III. (1990). Differential predation on 18, 591–597. mangrove propagules in open and closed canopy forest habitats. MCKEE, K. L. (1995). Mangrove species distribution and Vegetatio 89, 1–6. propagule predation: an exception to the dominance-predation PAGE,H.M.&WILLASON, S. W. (1982). Distribution patterns of hypothesis. Biotropica 27, 334–345. terrestrial hermit crabs at Enewetak Atoll, Marshall Islands. MCMAHON,B.R.&BURGGREN, W. W. (1988). Respiration. In Pacific Science 36, 107–117. Biology of the Land Crabs (eds.W.W.BurggrenandB.R.McMahon), PAIN,D.J.,BROOKE, M. D., FINNIE,J.K.&JACKSON, A. (2000). pp. 249–297. Cambridge University Press, Cambridge. Effects of brodifacoum on the land crabs of Ascension Island. MICHELI, F., GHERARDI,F.&VANNINI, M. (1991). Feeding and Journal of Wildlife Management 64, 380–387. burrowing ecology of two East African mangrove crabs. Marine PATTERSON, S., MCKEE,K.L.&MENDELSSOHN, I. A. (1997). Effects Biology 111, 247–254. of tidal inundation and predation on Avicennia germinans seedling MINCHINTON, T. E. (2001). Canopy and substratum heterogeneity establishment and survival in a sub-tropical mangal/salt marsh influence recruitment of the mangrove Avicennia marina. Journal of community. Mangroves and Salt Marshes 1, 103–111. Ecology 89, 888–902. PE´ REZ-CHI, A. (2005). Densities, diel activity, burrow shape, and MINCHINTON, T. E. (2006) Consequences of pre-dispersal damage habitat characteristics of Gecarcinus (Johngarthia) planatus Stim- by insects for the dispersal and recruitment of mangroves. pson, 1860 (Decapada, Brachyura, Gecarcinidae) at Socorro Oecologia 148, 70–80. Island, Revillagigedo, Mexico. Crustaceana 78, 255–272. MINCHINTON,T.E.&DALBY-BALL, M. (2001). Frugivory by insects PITMAN, R. L., BALLANCE,L.T.&BOST, C. (2005). Clipperton Island: on mangrove propagules: effects on the early life history of pig sty, rat hole and booby prize. Marine Ornithology 33, 193–194. Avicennia marina. Oecologia 129, 243–252. POFF, N. L. (1997). Landscape filters and species traits: towards MOLOFSKY,J.&AUGSPURGER, C. K. (1992). The effect of leaf litter mechanistic understanding and prediction in stream ecology. on early seedling establishment in a tropical forest. Ecology 73, Journal of the North American Benthological Society 16, 391–409. 68–77. QUESADA,F.J.,JIME´ NEZ, Q., ZAMORA, N., AGUILAR,R.&GONZA´ LEZ, MOLOFSKY,J.&FISHER, B. L. (1993). Habitat and predation effects J (1997). Arboles de la Peninsula de Osa. Instituto Nacional de on seedling survival and growth in shade-tolerant tropical trees. Biodiversidad, Costa Rica. Ecology 74, 261–265. RIDLEY, H. N. (1930). The Dispersal of Plants Throughout the World. MORRISON, L. W. (2002). Interspecific competition and coexistence Reeve and Co., Ashford, Kent. between ants and land hermit crabs on small Bahamian islands. RIVERA-MONROY, V.H., TWILLEY, R. R., BONE, D., CHILDERS, D. L., Acta Oecologica 23, 223–229. CORONADO-MOLINA, C., FELLER, I. C., HERRERA-SILVEIRA,J., MOUL, E. T. (1954). Preliminary report on land animals at Onotoa JAFFE, R., MANCERA, E., REJMANKOVA, E., SALISBURY,J.E.& Atoll, Gilbert Islands. Atoll Research Bulletin 28, 1–28. WEIL, E. (2004). A conceptual framework to develop long-term MU¨ NZBERGOVA´ ,Z.&HERBEN, T. (2005). Seed, dispersal, microsite, ecological research and management objectives in the wider habitat and recruitment limitation: identification of terms and Caribbean region. BioScience 54, 843–856. concepts in studies of limitations. Oecologia 145, 1–8. ROBERTSON, A. I. (1986). Leaf burying crabs: their influence on NIERING, W. A. (1956). Bioecology of Kapingamarangi, Caroline energy flow and export from mixed mangrove forests (Rhizophora Islands: terrestrial aspects. Atoll Research Bulletin 49, 1–32. spp.) in northeastern Australia. Journal of Experimental Marine NIERING, W. A. (1963). Terrestrial ecology of Kapingamarangi Biology and Ecology 102, 237–248. Atoll, Caroline Islands. Ecological Monographs 33, 131–160. ROBERTSON,A.I.&DANIEL, P. A. (1989). The influence of crabs on NORDEN, N., CHAVE, J., CAUBE` RE, A., CHAˆTELET, P., FERRONI,N., litter processing in high intertidal mangrove forests in tropical FORGET, P.-M., THE´ BAUD, C. (2007). Is temporal variation of Australia. Oecologia 78, 191–198.

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works Land crabs and tropical forest recruitment 223

ROBERTSON, A. I., GIDDENS,R.&SMITH, T. J., III. (1990). Seed SWAINE, M. D. (1996). Rainfall and soil fertility as factors limiting predation by insects in tropical mangrove forests: extent and forest species distributions in Ghana. Journal of Ecology 84, 419– effects on seed viability and the growth of seedlings. Oecologia 83, 428. 213–219. TERBORGH,J.&WRIGHT, S. J. (1994). Effects of mammalian SHERMAN, P. M. (1997). Direct and indirect effects of the land crab herbivores on plant recruitment in two neotropical forests. Gecarcinus quadratus (Gecarcinidae) on seedling density, organic carbon Ecology 75, 1829–1833. distributions and rooting profiles in Corcovado National Park, Costa Rica. TERBORGH, J., LOPEZ, L., NUN˜ EZ, P., RAO, N., SHAHABUDDIN,G., Ph.D. Thesis. University of Michigan, Ann Arbor. ORIHUELA, G., RIVEROS, M., ASCANIO, R., ADLER, G. H., SHERMAN, P. M. (2002). Effects of land crabs on seedling densities LAMBERT,T.D.&BALBAS, L. (2001). Ecological meltdown in and distributions in a mainland neotropical rain forest. Journal of predator-free forest fragments. Science 294, 1923–1926. Tropical Ecology 18, 67–89. THACKER, R. W. (1996). Food choices of land hermit crabs SHERMAN, P. M. (2003). Effects of land crabs on leaf litter (Coenobita compressus H. Milne Edwards) depend on past distributions and accumulations in a mainland tropical rain- experience. Journal of Expimental Marine Biology and Ecology 199, forest. Biotropica 35, 365–374. 179–191. SHERMAN, P. M. (2006). Influence of land crabs Gecarcinus quadratus THACKER, R. W. (1998). Avoidance of recently eaten foods by land (Gecarcinidae) on distributions of organic carbon and roots in hermit crabs, Coenobita compressus. Animal Behavior 55, 485–496. a Costa Rican rain forest. Revista de Biologia Tropical 54, 149–161. THOM, B. G. (1982). Mangrove ecology: a geomorphological SIDDIQI, N. A. (1995). Role of crabs in the natural regeneration of perspective. In Mangrove Ecosystems in Australia: structure, function mangroves in the Sundarbans forest of Bangladesh. Australian and management (ed. B. F. Clough), pp. 3–17. Australian National Journal of Ecology 20, 340–343. University Press, Canberra. SILMAN, M. R., TERBORGH,J.W.&KILTIE, R. A. (2003). Population THORNTON, I. (1995). Krakatau: the destruction and reassembly of an island regulation of a dominant rain forest tree by a major seed ecosystem. Harvard University Press, Cambridge. predator. Ecology 84, 431–438. TOMLINSON, P. B. (1986). The Botany of Mangroves. Cambridge SKOV, M. W., VANNINI, M., SHUNULA,J.P.&HARTNOLL,R.G. University Press, Cambridge. (2002). Quantifying the density of mangrove crabs: Ocypodidae TWILLEY, R. R., POZO, M., GARCIA, V. H., RIVERA-MONROY, V. H., and Grapsidae. Marine Biology 141, 725–732. ZAMBRANO,R.&BODERO, A. (1997). Litter dynamics in riverine SMITH, T. J., III. (1987a). Effects of light and intertidal position on mangrove forests in the Guayas River estuary, Ecuador. Oecologia seedling survival and growth in tropical, tidal forests. Journal of 111, 109–122. Experimental Marine Biology and Ecology 110, 133–146. URIARTE, M., CANHAM, C. D., THOMPSON, J., ZIMMERMAN,J.K.& SMITH, T. J., III. (1987b). Effects of seed predators and light level BROKAW, N. (2005). Seedling recruitment in a hurricane-driven on the distribution of Avicennia marina (Forsk.) Vierh. in tropical, tropical forest: light limitation, density-dependence and the tidal forests. Estuarine, Coastal and Shelf Science 25, 43–51. spatial distribution of parents trees. Journal of Ecology 93, 291– SMITH, T. J., III. (1987c). Seed predation in relation to dominance 304. and distribution in mangrove forests. Ecology 68, 266–273. VALIELA, I., BOWEN,J.L.&YORK, J. K. (2001). Mangrove forests: SMITH, T. J., III. (1988). The influence of seed predators on one of the world’s threatened major tropical environments. structure and succession in tropical tidal forests. Proceedings of the Bio Science 51, 807–815. Ecological Society of Australia 15, 203–211. VEENENDALL, E. M., SWAINE, M. D., AGYEMAN, V. M., BLAY,D., SMITH, T. J., III, CHAN, H. T., MCIVOR,C.C.&ROBBLEE,M.B. ABEBRESE,I.K.&MULLINS, C. E. (1996). Differences in plant (1989). Comparisons of seed predation in tropical, tidal forests and soil water relations in and around a forest gap in West from three continents. Ecology 70, 146–151. Africa during the dry season may influence seedling establish- SMITH, T. J., III, BOTO, K. G., FRUSHER,S.D.&GIDDINS,R.L. ment and survival. Journal of Ecology 84, 83–90. (1991). Keystone species and mangrove forest dynamics: the WARREN,J.H.&UNDERWOOD, A. (1986). Effects of burrowing influence of burrowing by crabs on soil nutrient status and forest crabs on the topography of mangrove swamps in New South productivity. Estuarine, Coastal and Shelf Science 33, 419–432. Wales. Journal of Experimental Marine Biology and Ecology 102, 223– SOUSA,W.P.,KENNEDY,P.G.&MITCHELL, B. J. (2003a). Propagule 235. size and predispersal damage by insects affect establishment and WATSON, J. G. (1928). Mangrove forests of the Malay Peninsula. early growth of mangrove seedlings. Oecologia 135, 564–575. Malayan Forest Research 6, 1–275. SOUSA,W.P.&MITCHELL, B. J. (1999). The effects of seed WILLIAMS,W.T.,BUNT,J.S.&DUKE, N. C. (1981). Mangrove litter predators on plant distributions: is there a general pattern in fall in north-eastern Australia. II. Periodicity. Australian Journal of mangroves? Oikos 86, 55–66. Botany 29, 555–563. SOUSA, W. P., QUEK,S.P.&MITCHELL, B. J. (2003b). Regeneration WOLCOTT, T. G. (1988). Ecology. In Biology of the Land Crabs (eds. W. of Rhizophora mangle in a Caribbean mangrove forest: interacting W. Burggren and B. R McMahon), pp. 55–96. Cambridge effects of canopy disturbance and a stem-boring beetle. Oecologia University Press, Cambridge. 137, 436–445. WOLFF, M., KOCH,P.T.&ISSAC, V.(2000). A trophic flow model of SOUSA,W.P.,KENNEDY, P. G., MITCHELL,B.J.&ORDO´ N˜ EZ L., B. the Caete mangrove estuary (North Brazil) with considerations M. (2007). Supply-side ecology in mangroves: do propagule for the sustainable use of its resources. Estuarine, Coastal and Shelf dispersal and seedling establishment explain forest structure? Science 50, 789–803. Ecological Monographs 77, 53–76. WOODROFFE, C. D. (1992). Mangrove sediments and geomorphol- STATZNER, B., DOLE´ DEC,S.&HUGUENY, B. (2004). Biological trait ogy. In Tropical Mangrove Ecosystems (eds. A. I. Robertson and D. composition of European stream invertebrate communities: M. Alongi), pp. 7–41. Coastal and Estuarine Studies 41, assessing the effects of various trait filter types. Ecography 27, American Geophysical Union, Washington, D.C. 470–488. YODZIS, P. (1984). How rare is omnivory? Ecology 65, 321–323.

Biological Reviews 84 (2009) 203–223 Ó 2009 Journal compilation Ó 2009 Cambridge Philosophical Society. No claim to original US government works