1 Background

Hermit crabs are one of the most conspicuous and ecologically important groups of animals inhabiting intertidal and subtidal habitats (Schembri [1982]). These animals are unique for their dependency on gastropod shells as a ‘mobile home’ to protect them from predators (Elwood et al. [1995]) and reduce the risk of desiccation during emersion at low tide (Bertness and Cunnigham [1981]). Despite the many world-wide studies on hermit crabs, those pertaining to hermit crab-shell interactions are more common than studies investigating habitat partitioning which are scarce especially in the Indo-Pacific region. The available literature on both macro- and micro-habitat preferences of hermit crabs including shell use pattern may indicate adaptations to reduce interspecific competition (Leite et al. [1998]). Habitat partitioning has been demonstrated since closely related species show variable use of gastropod shells depending on the shell size, shape and availability (Teoh and Chong [2014]), while a more heterogenous habitat provides more niches and ways to exploit the available resources (Bazzaz [1975]).

At the intertidal zone, landmark studies on biological zonation are well established (Knox [2001]; Harley [2003]; Veloso et al. [2003]; Rodil et al. [2006]; Sacrosati and Heaven [2007]), attributable to the accessibility of sites and the diversity of species of sessile and slow-moving animals that are readily enumerated (e.g. Connell [1972]). There are several important factors influencing intertidal biological zonation which include wave exposure (Stephenson [1961]; Knox [2001]; Harley [2003]), temperature (Wethey [1983]), salinity (Druehl [1967]) and substrate composition (Rai-mondi [1988]). Among these factors, sediment texture may invoke a relatively greater influence on distribution and maintenance of anomuran populations (Fransozo et al. [2008]) as sediment is utilized by these animals as shelter and food source (Abele [1974]). The adaptation of intertidal animals towards different environmental settings resulted in the formation of distinct ecological niches along the intertidal zone. This is exemplified by the unique features of rocky shores that exhibit prominent horizontal bands formed by different types of animals and plants (Nybakken [1982]).

Shells influence the growth and reproduction of hermit crabs (Fotheringham [1976]; Bertness [1981a]; Elwood et al. [1995]), and thus, the selection of a shell of optimum size and shape is essential for their survival. Hermit crab populations are limited by the availability, size and quality of their shells (Vance [1972]). Since they rely on empty shells and rarely predate on gastropods or remove the flesh from dead gastropods, hermit crabs compete intra- or interspecifically for the shell resource (Bach et al. [1976]). Although the availability of empty shells may be subjected to advective forces such as tides and waves as well as the hermit crabs themselves, the co-occurring assemblage of living gastropods typically reflects the availability of shells in intertidal mud habitats (Teoh and Chong [2014]).

In the western coast of Sabah, Malaysia, the heterogeneous environment of Sepangar intertidal flat (Figure 1) is characterized by distinct substrate zonation ranging from a mixture of sand-mud, gravels, rocks, coral rubbles to large rocky boulders. On the other hand, Sulaman mangrove fringe, which is located about 30 km northeast from Sepangar, is characterized by homogenous environment dominated by sand-mud substrate. Hermit crabs are one of the most conspicuous benthic macrofauna at both sites. However, the contrasting habitat settings appear to evoke different adaptations among the hermit crab species, possibly influencing their diversity, abundance and shell use pattern. In this study, spatial distribution and shell use of the hermit crab community in Sepangar and Sulaman shores were quantified to answer the following research questions: 1) Does a structurally heterogeneous habitat host greater diversity and abundance of hermit crabs than homogeneous habitat? 2) Are shells used by hermit crabs more diverse in a heterogenous habitat than homogeneous habitat? 3) How are hermit crab species ecologically partitioned (distribution and shell use pattern) in relation to substrate zonation in the heterogenous habitat?

Figure 1
figure 1

Schematic diagrams of sampling design at (a) Sepangar and (b) Sulaman shores.

Full size image

2 Methods

2.1 Field work

Field samplings were carried out in an intertidal rocky flat (N6°02′18.57″; E116°06′40.07″) at Sepangar Bay and fringing mangrove forest (N6°15′33.00″; E116°18′49.80″) at Sulaman, both sites located in the State of Sabah, Malaysia. A vast part of the intertidal rocky shore at Sepangar is uncovered during spring low tide to approximately 200 m from the highest tide mark. The Sulaman's mangrove area however has a lower, bare and muddy zone of about 50 m from the mangrove fringe to the sea which is completely exposed during spring low tide. Samplings were performed at both sites, each on three separate occasions in December 2007 and January 2008 during spring low tide.

At Sepangar shore, three 100-m transect lines were laid perpendicular to the shoreline 50 m apart between transects. Along each transect line, five stations were designated at intervals of 25 m. These stations were marked as 0, 25, 50, 75 and 100 m from the high-tide mark (Figure 1a). The 0 m marked the highest tide mark. At each station, 1 × 1 m quadrate sampling in triplicates were performed within a 5-m radius from each station point, taking care that unsampled quadrates were not disturbed while sampling. All hermit crabs that were found within each quadrate were collected starting from the sides of the quadrate to minimize escapement of hermit crabs. A similar sampling method was employed for Sulaman's bare mangrove zone, but given its narrow shore width (maximum 50 m), 30-m transect lines were laid with three designated stations set 15 m apart (Figure 1b). The 0-m mark however marked the lowest fringe of the mangrove forest. Substrates at the intertidal flat at Sepangar Bay were visually characterized based on the distinctive mixture of substrates present (Table 1) whereas the sediment at Sulaman's bare mangrove zone was characterized as homogenously dominated by sand-mud.

Table 1 Substrate characteristics of each station along a transect line at Sepangar shore
Full size table

2.2 Laboratory procedures and data analyses

Hermit crabs were removed from their shell by gently pulling and twisting the crab against the direction of shell spiral. In the event when crabs were unable to be retrieved from their shell, a light hammer was used to gently crack the shell before pulling out the crab. Both hermit crabs and their shells were identified to the lowest taxa. The cephalothoracic length (SL) of each hermit crab, measured from the rotral tip to the posterior margin of the hardened portion of the carapace by using a pair of Vernier calipers (accuracy to 0.05 mm), was recorded. Sex of hermit crabs was determined based on position of gonopores; males have gonopores on the coxae of the fifth pereopods while females have gonopores on the coxae of the third pereopods. Density of hermit crabs was estimated as follows:

Density ind / m 2 = Number of individuals for species in a quadrate Area of each quadrate 1 m 2
(1)

Diversity and species richness of hermit crabs and their shells were quantified using Shannon-Wiener diversity index and Pielou's index of evenness expressed as follows:

Shannon-Wiener index, H′ = −Σ(p i ln p i ),

where p i is the number of individuals for species i/total number of individuals

Pielou's index, J′ = H′/ln s

where s is the total number of species

One-way analysis of variance (ANOVA) at 5% significance level was performed to determine the significant difference in density among sampling stations for each hermit crab species. All data were log transformed [log (x + 1)] if requirements for normality and homogeneity in variance as checked by Kolmogorov-Smirnov's and Levene's tests, respectively, were not fulfilled (Zar [2010]). This analysis was performed using SPSS software version 12.0. Chi-square test at 5% significance level was performed to determine whether the distribution of the various hermit crab species is associated/dependent on sampling stations at both Sepangar and Sulaman shores.

Canonical correspondence analysis (CCA) was used to analyse and visualize the relationships between species of hermit crabs and their shells according to stations (Ter Braak [1986]) at Sepangar rocky shore but not at Sulaman mangrove because of the low number of hermit crabs and shell species. The species data set comprised of hermit crab species with sex. The site or sample of data set comprised of sampling stations along the transect lines. The proportion of shell species used by hermit crabs were treated as environmental variables (co-variables) which comprised of the five most common shell types used by hermit crabs (Clypeomorus batillariaeformis, Clypeomorus bifasciata, Cerithium zonatum, Rhinoclavis sinensis and Tenguella musiva). The CCA was performed using CANOCO 4.5 software (Ter Braak and Smilauer [2002]).

3 Results

3.1 Diversity and abundance of the hermit crab community

3.1.1 Sepangar shore

A total of 850 specimens were collected comprising of five species belonging to two genera from the family Diogenidae: Clibanarius merguiensis, C. striolatus, Diogenes klaasi, D. pallescens and D. tumidus. The largest species was C. striolatus (SL 5.64 ± 2.55 mm), followed by C. merguiensis (5.45 ± 2.48 mm), D. pallescens (4.37 ± 2.39 mm), D. klaasi (4.28 ± 2.37 mm) and D. tumidus (4.14 ± 1.86 mm). The greatest diversity was recorded at 100 m (H′ = 1.09) while the lowest at 0 m (H′ = 0.00) where only one species (D. klaasi) was present. Hermit crab species at 100 m were the most evenly distributed (J′ = 0.79) while they were the least evenly distributed at 0 m (J′ = 0.00) (Table 2).

Table 2 Number of species, H′ and J′ values of hermit crabs at Sepangar shore
Full size table

The overall highest density was reported for D. tumidus (8.98 ± 6.52 ind/m2) which peaked at 50 m (16.95 ± 12.35 ind/m2) but absent at 0 m (Figure 2a). This species was dominant at both 50 and 75 m stations (Figure 3a). Density of D. tumidus decreased gradually towards 100 m. On the other hand, D. klaasi was present in the first three stations (0, 25 and 50 m) and absent at 75 and 100 m. It was the sole species at 0 m with density of 6.25 ± 7.39 ind/m2. Peak density of this species was observed at 25 m (10.64 ± 15.44 ind/m2) (Figure 2b). Densities of both Clibanarius species generally increased towards the sea. At the highest station (0 m), C. striolatus was absent but gradually increased in density from 25 m (0.22 ± 0.54 ind/m2) to 100 m (5.33 ± 8.92 ind/m2) (Figure 2c). Similarly, C. merguiensis was absent at 0 m but had density peak at 100 m (6.11 ± 7.20 ind/m2). Lowest density of this species was recorded at 75 m (1.17 ± 2.07 ind/m2) (Figure 2d). The results from ANOVA showed no significant difference (p > 0.05) in density among stations for all species. The least abundant species in the study was D. pallescens which was present at 50 m (0.11 ± 0.17 ind/m2) and 100 m (0.11 ± 0.27 ind/m2). Chi-square test revealed strong dependency/association (χ 2 = 57.18, df = 16, p < 0.0001) of hermit crab species on stations/zones in Sepangar shore.

Figure 2
figure 2

Mean density (+SD) of hermit crabs by species at Sepangar (a-d) and Sulaman (e-f) shores. D. pallescens from Sepangar was not included due to very low density and was absent in most stations.

Full size image
Figure 3
figure 3

Species compositions of hermit crabs at (a) Sepangar and (b) Sulaman intertidal shores.

Full size image

3.1.2 Sulaman shore

There were only two species at Sulaman shore; the dominant D. foresti (SL 2.83 ± 1.03 mm) and another Diogenes sp. (Figure 3b) which was too small (SL 1.53 ± 0.42 mm) to be fully identified. Diversity and evenness indices were not computed for hermit crabs in Sulaman due to a low number of species. Density of D. foresti generally decreased from 3.33 ± 2.65 ind/m2 at 0 m to 0.16 ± 0.28 ind/m2 at 30 m (Figure 2e). Diogenes sp. was absent at 30 m and occurred in small number with an overall mean density of 0.21 ± 0.32 ind/m2. Chi-square test revealed no association (χ 2 = 0.15, df = 2, p > 0.05) between hermit crab species and station in Sulaman shore.

3.2 Shell use

At Sepangar shore, a total of 30 shell species were used by hermit crabs with most occupied shells being Clypeomorus batillariaeformis (75.59%) followed by C. bifasciata (8.61%) and Rhinoclavis sinensis (4.13%) while utilization of other type of shells were markedly low (Table 3). Shells used by hermit crabs at station 5 appeared to be the most diverse (H′ = 1.31, J′ = 0.51) among all stations whereas station 2 was the least diverse (H′ = 0.66, J′ = 0.30) (Table 4). The composition of hermit crabs based on the four most occupied shells is shown in Figure 4. The shell category ‘others’ represents the collective proportion of hermit crabs by species that used the remaining 26 shell types. The empty shells of C. batillariaeformis comprised the majority of shells used by all hermit crab species whereas shells of C. bifasciata represented substantial proportion of shells used by C. striolatus (17%), D. tumidus (11%) and D. klaasi (6%) (Figure 4). At Sulaman shore, only two shells were used by hermit crabs: Nassarius livescens and Cerithidea cingulata with the latter occupied by 98% of the hermit crabs (Table 3).

Table 3 Overall compositions (%) of shell types used by hermit crabs at Sepangar and Sulaman shores
Full size table
Table 4 Number of species and H′ and J′ values of occupied shells at Sepangar shore
Full size table
Figure 4
figure 4

Composition of shells commonly occupied by different hermit crab species (a-d) at Sepangar shore.

Full size image

3.3 Distribution and shell use of hermit crabs: influence of spatiality

The CCA triplots ordination is shown in Figure 5. The first two axes of the species-environment relation explained 95.5% of the variation. The dominance of D. klaasi at station 1 and station 2 resulted in the most apparent spatial separation of this species with other hermit crab species. Despite C. batillariaeformis being the most commonly used shells by all hermit crab species (Figure 4), D. klaasi appeared relatively more exclusive in the use of this shell. The proportional use of C. bifasciata shells by D. tumidus was higher at stations 3 and 4 where the hermit crab was most dominant. The species D. tumidus exhibited higher degree of habitat overlap with Clibanarius species than with D. klaasi. At station 5, Clibanarius species were dominant where the proportional use of T. musiva, C. zonatum and R. sinensis shells was the highest among all stations. Unlike the Diogenes, both Clibanarius species exhibited a higher degree of habitat overlap. There was no observable spatial separation between sexes for all hermit crab species.

Figure 5
figure 5

CCA triplots of the relationships between hermit crab species and their shells at Sepangar shore. Direction of arrow indicates the proportional increase in shell species used by hermit crabs. Solid circles indicate sampling stations/zones along transect. Filled and hollow triangles indicate hermit crabs D. tumidus (Dt) and D. klaasi (Dk), respectively. Filled and hollow squares indicate hermit crabs C. striolatus (Cs) and C. merguiensis (Cm), respectively. Cross indicates hermit crab D. pallescens (Dp). ‘M’, ‘F’ and ‘oF’ indicate male, non-ovigerous female and ovigerous female, respectively. Ellipsoid envelops the dominant hermit crab species at the indicated sampling stations.

Full size image

4 Discussion

This study reveals the apparent influence of habitat heterogeneity towards the diversity and abundance of hermit crabs. The more complex intertidal shore with distinct substrate zonation at Sepangar hosts higher diversity and abundance of hermit crabs as compared to the homogeneous (sand-mud) mangrove shore at Sulaman. Such disparity in hermit crab assemblages between habitats has also been observed by De Grave and Barnes ([2001]) at the coastal shores of Mozambique islands where Clibanarius virescens dominated the small island shores that were less heterogeneous compared to large islands where shores were more heterogeneous and had more diverse hermit crab assemblage. The species diversity in the structurally complex habitat is greater due to the presence of more niches and ways of exploiting the resources (Bazzaz [1975]). The present study recorded five species of hermit crabs in the heterogeneous habitat of Sepangar's rocky shore whereas only two species were found at the homogeneous habitat of Sulaman's bare mangrove zone. Such a difference in the hermit crab assemblage attributed to habitat complexity has also been reported in southern Thailand where corals and rocky shores had 16 and 9 species, respectively, compared to two and four species on the bare shore of mangroves and mudflats, respectively (McLaughlin [2002]). Similarly, more species were also associated with structurally complex coral reefs (24 species) and rocky shore (19 species), as compared to the structurally simple mudflats (seven species) in the South China Sea region (Rahayu [2000]).

The difference in hermit crab assemblages among habitats is likely to result from the variety (type), availability and suitability of gastropod shells. This is because occupancy of optimum (size and shape) shells is fundamental for the survival of hermit crabs (Elwood et al. [1995]). Complex habitats with variety of substrate types are known to support a greater variety of gastropod populations (Mantelatto and Garcia [2002]). Such habitats are likely to offer greater assortment of shells to hermit crabs. In this study, the influence of habitat heterogeneity on the shell use of hermit crabs is apparent such that the shells that were used by the hermit crabs were more diverse at Sepangar shore (30 shell species) than Sulaman shore (two shell species). Therefore, the more diverse shells available at Sepangar shore support greater diversity of hermit crabs whereas the low diversity of available shells at Sulaman shore limits the diversity of hermit crabs.

The spatial segregation related to habitat heterogeneity is demonstrated by hermit crab populations at the intertidal zone in Sepangar shore. The presence of clear substrate zonation either by a type or mixture of substrates at the site may have resulted in the spatial confinement of hermit crab species at different zone along the intertidal area (see Turra and Denadai [2002]; Fransozo et al. [2008]). Habitat zonation among sympatric hermit crab species was also observed by Bertness ([1981b]) on three species of hermit crabs: Calcinus obscures, C. albidigitus and Pagurus sp. on a rocky shore whereby C. obscures was distributed from a middle to low intertidal zone and C. albidigitus was distributed from a middle to high intertidal zone whereas Pagurus sp. was confined at a lower intertidal zone.

The spatial separation of D. klaasi with other species particularly at a high-tide mark (0 m) indicates that this species is more adapted to sand-mud conditions which characterized station 1 (0 m) and station 2 (25 m) whereas its sympatric congener, D. tumidus, is dominant at station 3 (50 m) and station 4 (75 m) where the substrate comprised of loose rubbles like gravel, rocks and coral remnants. Both Clibanarius species are more adapted to zones that are structurally more complex based on their increasing abundance from 50 to 100 m and their low presence at the high-tide mark where the substrate is more homogenous. The distribution and coexistence of hermit crab species as modulated by substrate type have been shown by Turra and Denadai ([2002]). They demonstrated experimentally the substrate preference of hermit crabs, under allopatric (single species) and sympatric (three species) (C. antillensis, C. sclopetarius and C. vittatus) conditions, for four substrate types: rocky, pebble, sand and mud. Both C. antillensis and C. sclopetarius showed more similarity in the pattern of substrate selection under sympatric than allopatric condition, suggesting the mutual influence of coexisting species on substrate selection, whereas substrate selection by C. vittatus differed subtly between allopatric and sympatric conditions.

Except at the highest station (0 m), there was considerable overlap in habitat use particularly between D. tumidus and the Clibanarius species. Among all species, D. tumidus appears to be more adaptable to a mixed substrate type based on its higher abundance from 25 to 100 m. Hermit crabs are able to employ different feeding modes depending on the available food sources (Schembri [1982]). The mobility and versatility of feeding habits of hermit crabs increase their capability of foraging large areas for food and, thus, bringing them into contact with a variety of substratum. Such ability to cope with the multiplicity of substratum and utilize the variety of food demonstrates the adaptive value of hermit crabs (Schembri [1982]). Hermit crab adaptability was also observed in three common intertidal hermit crabs: Pagurus geminus, Pagurus lanuginosus and Clibanarius virescens on a rocky shore in Japan which exhibited apparent but not distinct habitat partitions due to spatial overlaps (Imazu and Asakura [1994]).

The high standard deviation of the density and high coefficient of variation between 84% and 207% suggest patchy distribution due to the clustering of the hermit crabs. Clusters of particularly Diogenes species in this study were observed on substrates ranging from sand-mud to rocks and gravels. Forming clusters may be advantageous in minimizing risk of desiccation during emersion at low tide (Gherardi and Vannini [1991]). This is indicated by less mobility of hermit crabs in a cluster compared to individuals found in tide pools. It is also commonly suggested that clustering serves as a platform for shell exchange among hermit crabs to acquire optimum shell from their conspecifics through elaborate communication mechanisms (Gherardi et al [1994]).

Shells of 30 species were occupied by hermit crabs at Sepangar shore; however, C. batillariaeformis being the most common snail comprised the majority (75.6%) of the occupied shells. The availability of gastropod shells is an important factor in determining the shell selection pattern of hermit crabs (Bertness [1982]). In a Brazilian inlet, hermit crabs Clibanarius antillensis and Calcinus tibicen were observed to frequently use shells of Tegula viridula which was the most abundant gastropod in the area (Floeter et al [2000]). In this study, the number of shell species occupied by hermit crabs was high, considering the small sampling area covered (about 1 ha in total). In comparison, Nakin and Somers ([2007]) recorded 21 species of gastropod shells used by C. virescens at three separate sites of a South African coast, while Benvenuto and Gherardi ([2001]) recorded 20 species of gastropod shells occupied by C. erythropus in a rocky Mediterranean shore.

Besides its greater availability, C. batillariaeformis shell is characterized by deep spiralisation which allows greater water retention, advantageous to hermit crabs to overcome thermal stress during emersion (Bertness [1982]). Although other shells may offer a similar advantage, their use by hermit crabs may be limited by the rare occurrence of the shells and/or incompatible size. Occupying shells smaller than the optimal size, such as shells of Pyrene sp., may expose the hermit crab to predators while occupying heavier shells such as shells of Canarium sp. and Angaria sp. would incur high energetic cost on the hermit crab, slowing its growth and reproduction ability (Osorno et al. [2005]).

Despite C. batillariaeformis being the most common shells used by all hermit crab species, D. klaasi appears to be more exclusive in its use of this shell. Being the sole species at the high shore (0 m) and dominant at midshore (25 m), D. klaasi has less competition from other hermit crab species to acquire C. batillariaeformis shells, and hence, there is less need to occupy other shells. The high abundance of the Clibanarius species from 50 to 100 m put them in a direct competition with D. tumidus. This may have caused the slight shift in shell use in favour of C. bifasciata by D. tumidus and C. striolatus. Shells used by hermit crabs were most diverse at the low shore (100 m) where Clibanarius species were dominant. Clibanarius is able to use more variety of shells due to their larger size compared to the Diogenes species. In addition, some large Clibanarius may not be able to occupy the small C. batillariaeformis shells and, thus, are more dependent on other shell types.

5 Conclusions

This study shows that structurally more complex habitats will host a greater diversity and abundance of hermit crabs. This is similarly shown in their shell use where hermit crabs occurring in the more heterogeneous habitat used a greater variety of shell species. Substrate heterogeneity along the intertidal shore results in hermit crab zonation but with some overlap in habitat use that appears to influence the shell use pattern.