Shell occupation by the land hermit crab Coenobita violascens ( Anomura , Coenobitidae ) from Phuket Island , ailand

Shell occupation by the land hermit crab Coenobita violascens Heller, 1862 was investigated from January 2011 to March 2012 on Phuket Island in the Andaman Coast of Th ailand. Th e samples of C. violascens were collected monthly using multiple quadrat sampling. Twenty shell species from 11 families were found occupying by C. violascens, which were mainly marine gastropods (90%). Th e three most common occupied shell species were Chicoreus brunneus (Link, 1807) (21.3% of hermits), followed by Filifusus fi lamentosus (Röding, 1798) (14.9%) and Laevistrombus canarium (Linnaeus, 1758) (14.9%). Biconical shells and those with ovate apertures were the most commonly occupied shell types. Furthermore, individual C. violascens probably occupy the shells of at least three diff erent gastropod species during their lifetime. Interestingly, C. violascens shows a tendency of occupying specifi c categories of shells in relation to shell species, shape and aperture shape. Other aspects of shell occupation by C. violascens compared to congeneric species are also discussed. keY words Shell use, shell quality, terrestrial hermit crab.


Shell occupation by Coenobita violascens
Nauplius, 26: e2018004 benefits of using shells as refuges by these crabs include protection from desiccation and predators (Greenaway, 2003).Additionally, land hermit crabs are able to replenish water into internal space of their occupied shells to maintain body moisture, allowing them to forage further inland (Wilde, 1973).
The shell utilization patterns of land hermit crabs have been studied in several areas of the world: e.g. the western Atlantic by Morrison and Spiller (2006); the eastern Pacific by Abram (1978), Guillen and Osorno (1993) and Laidre and Vermeij (2012); North Pacific by Willason and Page (1983) and Szabo (2012); the western Pacific by Boneka et al. (1995); Red Sea by Volker (1967), Sallam et al. (2008) and Sallam (2012); and the western Indian Ocean by Barnes (1999;2001;2002).Although a recent investigation by Bundhitwongrut et al. (2015) was carried out on shell use by Coenobita rugosus H. Milne Edwards, 1837 at Cape Panwa of Phuket Island in the Andaman Sea, eastern Indian Ocean, little is known about shell occupation by other land hermit crabs in this area.There are 382 gastropod mollusc species recorded from the Andaman coast of Thailand including Cape Panwa (Tantanasiriwong, 1978).However, intense shell exploitation and trade in both local and imported mollusks have been reported from Phuket Island (Aungtonya and Tantichodok, 1991;Aungtonya and Hylleberg, 1992;Aungtonya and Khokiattiwong, 1992;Bussarawit, 1995).In addition, unoccupied shells in good condition are scarce in the supralittoral zone of this site (Bundhitwongrut et al., 2015).These situations presumably influence shell resources for land hermit crabs in the area.
Seventeen valid species of land hermit crabs in the genus Coenobita Latreille, 1829 have been recognized throughout the world (De Grave et al., 2009;McLaughlin et al., 2010;Rahayu et al., 2016).Coenobita violascens Heller, 1862 is distributed in the Indo-Pacific region from the east coast of Africa to the Philippines and Japan (McLaughlin et al., 2007).McLaughlin (2002) reported the occurrence of C. violascens from coasts of the Andaman Sea and Gulf of Thailand.Coenobita violascens mainly inhabits beaches and areas associated with mangroves (Nakasone, 1988;McLaughlin et al., 2007;Doi et al., 2016).Limited information on the life history of C. violascens has been known and recorded.Only some recent studies on C. violascens have been recently reported, including functional morphology related to water uptake (Becchi et al., 2015), emigration behavior and molting during the sea-to-land transition (Hamasaki et al., 2015a), larval development under laboratory conditions (Hamasaki et al., 2015b;Kato et al., 2015), larval release and tree-climbing behavior (Doi et al., 2016) and thermal adaptation of embryos (Hamasaki et al., 2016).The objectives of the present study are to reveal shell occupation and the relationship between the characteristics of occupied shells and crab characteristics of C. violascens at Cape Panwa, Phuket Island, Andaman coast of Thailand.

Study area
The study site was Cape Panwa (7°48'26"N 98°24'35"E) in the southeast part of Phuket Island on the Andaman coast of Thailand.The climate is wet tropical and is influenced by the wet southwesterly monsoon from May to October and the dry northeasterly monsoon from November to April (Khokiattiwong et al., 1991).The study site is exposed to the semidiurnal tide with amplitude of 2.15-2.27m at spring tide to 0.85-1.15m at neap tide (Limpsaichol, 1981).The study beach was located in the supralittoral zone in the area of the Phuket Marine Biological Center (PMBC).This beach is behind the office of PMBC.The width of this beach is about 50 m and the distance between the mean sea level and the PMBC office is approximately 45 m.The beach is characterized by open sand scrub vegetation, consisting of rather coarse sand patches of shale or phylitte (Nielsen, 1976a).The inland side is covered with dense vegetation alternating with sparse vegetation in front of a steep cliff.A map of the study area and environmental data during the study period are given in Bundhitwongrut et al. (2014).

Specimen sampling and analysis
The sampling period was from January 2011 to March 2012.Coenobita violascens was collected three days per month.The hermit crabs were collected by hand by the same person (the author) at low tide from the supralittoral area in the early morning (Sallam et al., 2008).The sampling technique used in this study was multiple quadrat sampling (Barnes, 1999).Four temporary line transects were randomly placed at 15-m intervals perpendicular to the shoreline from the supralittoral zone to the inland area.Moreover, 16 temporary quadrats of area 1 m 2 were placed at 5-m intervals on transects from 5 m above the mean sea level to 45 m further inland.Five to six quadrats were sampled on each sampling day.Quadrats with 10-cm high walls were used to prevent the agile crabs from escaping.
All hermit crabs sampled were brought to the laboratory of PMBC.Each hermit was carefully pulled out of its occupied shell while holding the crab in the air and waiting until most of its body extended from the shell in order to investigate the crab and its inhabited shell characters.If the crab's uropods still held onto inside the shell, especially those inhabiting shells with a long spire, a metal wire was used to tickle the crab's abdomen to induce it to vacate the shell.To avoid negative impacts on native animals, removal of crabs from the population at the study area was not permitted according to the policy of the PMBC.Furthermore, the current sampling method prevented unnecessary hermit mortality in order to preserve this population of C. violascens.Consequently, all hermits were allowed to reinhabit their previously-occupied shells after the observations, and they were kept in several aquaria with food and water until the investigations were finished in each sampling month.
The additional marking method following Bundhitwongrut et al. (2015) was used in this study.All sampled hermits were marked after investigation in each sampling month before being released into the natural habitat from where they were collected.To prevent cheliped and pereiopod autotomy, each hermit was gently held and marked after being pulled from its shell.The markings were made with a waterproof pen and then coated with nail varnish on the outer surface of both major cheliped of each hermit and the body whorl near the outer lip of its occupied shell.In every month, if recaptured hermits were found, they were marked again.
Measurement and investigation of characters of C. violascens were carried out, including cephalothoracic shield length (CSL) and width (CSW), major chela length (MCL) and width (MCW), weight (CW), sex and reproductive stage (male, non-ovigerous female and ovigerous female).Species of occupied shells were identified according to several references (Brandt, 1974;Nielsen, 1976b;Tantanasiriwong, 1978;Middelfart, 1997;Poutiers, 1998).Additionally, those shells were compared with specimens deposited in the reference collection of PMBC to confirm their identities.Quantitative shell characters were measured and examined, including shell length (SL) and width (SW), weight (WW), internal volume (SIV), aperture length (SAL) and width (SAW).The internal volume of shells was evaluated using a graduated syringe (Floeter et al., 2000).All quantitative measurements were recorded to the nearest 0.01 mm for size using digital vernier calipers, 0.01 g for weight using digital weighing scales, and 0.1 ml volume using graduated syringes.
Qualitative shell characters were examined, including shell shape, aperture shape and shell quality.The shell and aperture shapes were categorized following Bundhitwongrut et al. (2015).The categories of shell shapes consisted of biconical, conical, elongately conical, fusiform, globose, oval, pyramidal, pyriform, turban and vermiform.The categories of shell aperture shapes comprised elongately ovate, irregular, ovate, round and semicircular.The categories of shell damage were undamaged and damaged.Damaged shells were shells with a hole, a broken apex, damaged inner lip of the body whorl, or greater damage in a large portion of the shell (Barnes, 1999).Moreover, the value of shell internal volume per weight (SIV/W ratio) was calculated for each shell inhabited by individuals of C. violascens as a predictor of shell quality (Osorno et al., 1998).

Statistical analysis
If recaptured C. violascens were found, their data were excluded to avoid possible pseudoreplicates.Individuals of C. violascens were separated into groups according to sex and reproductive stage as male, female, non-ovigerous female, or ovigerous female.Hermits were categorized into 0.5-mm CSL size classes to facilitate the comparison in shell use as a function of hermit crab size according to Nakasone (2001) and Bundhitwongrut et al. (2014).Regression analysis was utilized to determine relationships between quantitative characters of hermit crabs and occupied shells using the power function equation (Y=a.X b ) (Sallam et al., 2008;Bundhitwongrut et al., 2015).The critical significance level adopted in all statistical tests was p < 0.

Shell occupation in relation to shell shape
Shell occupation of C. violascens varied in relation to shell shape (Tab.2).All hermits occupied seven

Shell occupation in relation to shell aperture shape
Shell occupation of C. violascens varied in relation to shape of shell aperture (Tab.3).All C. violascens occupied four categories of shell aperture shape, including elongately ovate, irregular, ovate and round.Shells with ovate apertures (74.5%) were the mostoccupied by all crabs.Males and non-ovigerous females were found occupying all four aperture shapes and ovigerous females were found inhabiting one aperture shape (ovate).

Shell occupation in relation to shell damage
All C. violascens occupied undamaged shells (46.8%) slightly fewer than damaged shells (53.2%).Males and non-ovigerous females were found occupying both damaged and undamaged shells and ovigerous females were found inhabiting only undamaged shells.

Shell occupation in relation to SIV/W ratio
The range of SIV/W values of shell inhabited by C. violascens was 0.120-4.604(Tab.1).According to the ratio of SIV/W, the most-occupied shell species at the study area was not the lightest shell species.Chicoreus brunneus, the most-occupied shell species, had a SIV/W ratio (mean ± SD) equal to 0.213 ± 0.056, but had a very low ranking SIV/W ratio (19 th out of 20 shell species).However, the lightest-occupied shell species was P. canaliculata with a SIV/W ratio equal to 4.604, but ranked 8 th in terms of occupation by C. violascens.

Relationship between crab and shell characteristics
The relationship between characters of C. violascens

Shell occupation by Coenobita violascens
Nauplius, 26: e2018004 and occupied shells are shown in Tab. 4. The values of the determination coefficient (r 2 ) from the regression analysis ranged between 0.62 and 0.93.Strong correlations were observed between characters of hermits and internal volume and aperture width of occupied shells (r 2 > 0.91).Shell aperture width was the most correlated with hermit characters (r 2 > 0.92), whereas shell weight and aperture length were least correlated with characters of hermits (r 2 < 0.66), except for cephalothoracic shield width (r 2 = 0.83).

discussion
Coenobita violascens was found sympatrically with C. rugosus and Coenobita brevimanus Dana, 1852 at the study site.The results from this study are the first record on shells occupation by C. violascens.This coenobitid species shows a tendency of occupying specific categories of shells in relation to shell species, shape and aperture shape.Coenobita violascens individuals probably occupy the shells of at least three different gastropod species during their lifetime.Moreover, other aspects of shell occupation by C. violascens compared to other congeneric species are also discussed.
Coenobita violascens was found occupying fewer shell species (n = 20) than sympatric C. rugosus (n = 63) studied by Bundhitwongrut et al. (2015).This probably implies that C. violascens possess more specific requirements or less plasticity of use of shell resources.At least, the evidence is that C. violascens were not found using shells of neritids of the genus Nerita Linnaeus, 1758 although these shell groups were usually encountered as living individuals and discarded shells in the study area, which were commonly occupied by sympatric C. rugosus in high numbers reported by Bundhitwongrut et al. (2015).
Different sized individuals occupied different shell species.It is possible that C. violascens at the study site inhabits the shells of at least three different gastropod species as they grow.For instance, the most-occupied shell species, Chicoreus brunneus, were inhabited by medium-sized individuals (6.5-12.0mm).However, smaller individuals (2.5-6.5 mm) occupied other Table 4.The relationship between characters of hermit crab Coenobita violascens and the inhabited shells sampled represented by regression equations.r 2 = determination coefficient; CSL = cephalothoracic shield length; CSW =cephalothoracic shield width; CW = crab wet weight; MCL = major chela length; MCW =major chela width; SL = shell length; SW = shell width; WW = shell wet weight; SIV = shell internal volume; SAL = shell aperture length; SAW = shell aperture width; N = 47.

Shell occupation by Coenobita violascens
Nauplius, 26: e2018004 smaller-shell species (e.g.Clypeomorus batillariaeformis, Drupella rugosa), whereas larger individuals (> 12.0 mm) inhabited other larger species [e.g.Chicoreus ramosus (Linnaeus, 1758)].This inference is similar to the previous studies on Coenobita clypeatus (Fabricius, 1787) by Morrison and Spiller (2006) and C. rugosus by Bundhitwongrut et al. (2015), in that these land hermit crab are required to use the shells of a few different gastropods during their lifetime.To preserve these hermit crabs, therefore, the conservation of shell diversity is necessary because these hermits need different sizes and types of shells to complete their life cycle (Bundhitwongrut et al., 2015).
Coenobita violascens occupied particular categories of shell and aperture shapes.Frequently occupied shell shapes were higher-spired shells, which were mid-and high-spired shells (e.g.biconical and fusiform shells).This pattern of shell shape use by C. violascens is similar to those of small Coenobita perlatus H. Milne Edwards, 1837 living on exposed sandy beach, reported by Willason and Page (1983), and Coenobita cavipes Stimpson, 1858 investigated by Barnes (1999).More available space inside high-spired shells retains water much better than that of low-spired shells, thereby providing more resistance to desiccation during exposure to thermal stress (Bertness, 1982;Willason and Page, 1983;Barnes, 1999).Conversely, C. rugosus at the study site more frequently occupied low-spired shells (Bundhitwongrut et al., 2015) due to its behavioral ecology as a burrowing species (Barnes, 1999).The most-occupied shell aperture shape by C. violascens at the study site was ovate aperture as previously reported in C. rugosus at the same location by Bundhitwongrut et al. (2015).This aperture shape may allow this land hermit crab to fully seal the aperture with the major chela to avoid desiccation like other coenobitid crabs (Szabo, 2012).Interestingly, there was no C. violascens found occupying shells with semicircular aperture.This aperture shape is the diagnostic character of intact neritid shells.This occurrence is relevant to that there was no occupation of neritid shells by C. violascens.
The proportions between damaged and undamaged shells occupied by C. violascens were not significantly different at the study site as previously reported in sympatric C. rugosus by Bundhitwongrut et al. (2015).However, most shells inhabited by C. violascens were in worn condition and had lost the columella (unpublished data) as previously recorded by Kinosita and Okajima (1968), Ball (1972), Sripathi et al. (1977), Laidre (2012), Szabo (2012) and Bundhitwongrut et al. (2015).The missing columella of occupied shells may be due to mechanical or chemical abrasion (Kinosita and Okajima, 1968), resulting in light-weight and more space inside shells available for hermit crabs.These shells may be used previously by other hermits over a period of many years (Ball, 1972;Abrams, 1978;Boneka et al., 1995).Moreover, available empty shells as limited resource were scant at the study site (Bundhitwongrut et al., 2015), as also recorded earlier in other locations (Ball, 1972;Morrison and Spiller, 2006;Laidre and Vermeij, 2012).Consequently, these old and worn shells may be the main shell resources, especially for adult hermit crabs as a result of shell facilitation rather than competition (Abrams, 1978).
According to the energy saving hypothesis proposed by Osorno et al. (1998), Chicoreus brunneus as the most occupied shell species by C. violascens in this study was not the lightest species.This finding is similar to shells inhabited by sympatric C. rugosus investigated by Bundhitwongrut et al. (2015).Although hermits occupying the lightest shell species available, P. canaliculata, might obtain some advantages such as saving energy and having more internal space, hermits are probably at risk because of the thin shell wall from potential predators, such as the rough red-eyed crab, Eriphia smithii MacLeay, 1838, which was often seen at the study site (Bundhitwongrut et al., 2014).
Morphological characters of inhabited shells and C. violascens were significantly correlated.This finding is similar to previous studies on other coenobitid crabs, including C. rugosus by Boneka et al. (1995) and Bundhitwongrut et al. (2015) as well as Coenobita scaevola (Forskål, 1775) by Sallam et al. (2008).These shell characters, especially internal volume and aperture width, seemingly comprise the main determinants for C. violascens shell occupation.The strong correlations between shell internal volume and hermit characteristics possibly means that internal volume is crucial for C. violascens in providing more space that may allow hermits to grow rapidly or retain more fertilized eggs during reproduction (Osorno et al., 1998).Moreover, spacious internal volume of inhabited shells provides more capacity for storing water inside to maintain body moisture that is very important for life on land (Wilde, 1973;Greenaway, 2003).Additionally, the strong relationship between

Shell occupation by Coenobita violascens
Nauplius, 26: e2018004 shell aperture width and hermit characters possibly indicates that C. violascens can use the major chela and walking legs to effectively seal the aperture firmly when retreating into the shell, which provides more protection from desiccation and against predators (Ball, 1972;Abrams, 1978;Sanvicente-Anorve and Hermoso-Salazar, 2011).
The habitat type of this study area may have resulted in relatively small numbers of specimens of C. violascens in the samples because this coenobitid is mostly recorded from habitats related to mangrove and terrestrial areas near estuaries (Nakasone 1988;McLaughlin et al. 2007;Doi et al. 2016).The present study site is supralittoral sandy beach that is not the most preferred habitat of this species.Therefore, further study in its main habitat in estuarine areas is recommended.Furthermore, C. violascens probably has a more cryptic lifestyle than the most common species, C. rugosus, at the study site.The foraging period and other activities of C. violascens presumably occur at different times to reduce competition with other species.Consequently, sampling methods may be more effective for collecting C. violascens in other habitats such as bait pitfall traps (Morrison and Spiller, 2006;Hsu and Soong, 2017).

acknowledgeMents
The present study was financially supported by a Thai Government Science and Technology Scholarship of the National Science and Technology Development Agency (NSTDA), Ministry of Science and Technology of Thailand.I would like to thank PMBC and the staff for allowing me to use the facility for research throughout the study period.I wish to thank Miss Vararin Vongpanich (PMBC) for her help in checking recent scientific names of some gastropod shell species.My special thanks are given to Dr. Warren Y. Brockelman (BIOTEC) for constructive suggestions and grammatical review on the manuscript.I am deeply grateful to three anonymous reviewers, the associate editor and editor for their invaluable comments.

Figure 1 .
Figure 1.Shell occupation by Coenobita violascens as a function of hermit crab size.

Table 1 .
Percentage and mean SIV/W ratio of shell species inhabited by Coenobita violascens at Cape Panwa, Phuket Island, from January 2011 to March 2012.The numbers in parentheses after the percentage of shells occupied are the numbers of crab individuals.SIV/W = shell internal volume/weight ratio as mean ± SD for species that N > 1 and as value for species that N = 1.

Table 2 .
Percentage of shell shape categories occupied by Coenobita violascens at Cape Panwa, Phuket Island, from January 2011 to March 2012.The numbers in parentheses after the percentage of shells inhabited are the numbers of crab individuals.

Table 3 .
Percentage of aperture shape categories of shells occupied by Coenobita violascens at Cape Panwa, Phuket Island, from January 2011 to March 2012.The numbers in parentheses after the percentage of shells inhabited are the numbers of crab individuals.