Intraspecific variation in megalopae of Clibanarius antillensis ( Anomura , Diogenidae ) among western Atlantic populations

The objective of this study was to describe the morphology of Clibanarius antillensis Stimpson, 1859 megalopae collected in the vicinity of Isla Sacrificios, Mexico, and compare it to previous descriptions originated from Brazilian and Panamanian specimens raised in laboratory conditions. We found four meristic differences between the Brazilian and the Panamanian and Mexican populations with the Brazilian population: the outer flagellum of antennule in the Panamanian and Mexican populations has more aesthetascs on the second, third, and fourth segments of the outer flagellum (0, 6, 5, 3, 0) than the Brazilian population (0, 4, 4, 2, 0); the maxilla has more setae on the scaphognathite in the Brazilian specimens (70) than in the Mexican and Panamanian specimens (49-62); the crista dentata of the third maxilliped is formed by only three denticles in the Brazilian specimens, while 4 or 6 denticles form the same structure in the Panamanian and Mexican populations; there are fewer number of setae on the endopod of the uropod in the Brazilian population. The differences may be explained by intraspecific variation.


Morphological variations on megalopa of C. antillensis
Nauplius, 26: e2018031 inTroduCTion The genus Clibanarius Dana 1852 is represented by 60 species worldwide (Lemaitre and McLaughlin, 2018); four of these species can be found in the Gulf of Mexico, Clibanarius antillensis Stimpson, 1859, C. sclopetarius Herbst 1796, C. tricolor (Gibbes, 1850), and C. vittatus (Bosc, 1802) (Felder et al., 2009).These four species have also been reported in Veracruz (Álvarez et al., 2011).Recent plankton studies performed in the Parque Nacional Sistema Arrecifal Veracruzano (PNSAV), on the southwestern Gulf of Mexico, have found megalopae of eight species of hermit crabs of the family Diogenidae: Calcinus tibicen (Herbst, 1791), Clibanarius antillensis, C. sclopetarius, C. vittatus, Dardanus insignis (de Saussure, 1858), Paguristes sericeus A. Milne Edwards, 1880, P. spinipes A. Milne-Edwards, 1880, and Petrochirus diogenes (Linnaeus, 1758) (Cházaro-Olvera et al., 2013).Bartilotti et al. (2008) described the complete larval development of two species of Clibanarius.While doing so, the authors reviewed morphological larval characters among species of the genus.In general, larval development within the genus is very homogenous, consisting of four or five zoeal stages and one megalopa.However, morphometry of the larval stages varies within the genus and species can be identified based on specific larval characters (Siddiqui et al., 1991;1993;Bartilotti et al., 2008).Although morphology of larval characters are believed to be species-specific, some morphological variations have been found among larval developmental stages described under laboratory conditions of C. antillensis from Brazil (Brossi-Garcia and Hebling, 1983) and Panama (Siddiqui et al., 1991); these morphological variations are believed to be the result of geographic differences and/or differing laboratory conditions (Siddiqui et al., 1991;Bartilotti et al., 2008).
The distribution of the hermit crab C. antillensis extends throughout the western Atlantic including Bermuda, Eastern Florida, Gulf of Mexico, Caribbean Sea, and Brazil (Felder et al., 2009).Within-species larval variations in decapod crustaceans are not uncommon in the marine environment.In particular, latitudinal variations have been explained by increased seasonality and unpredictability of primary productivity: a macroecological gradient is formed, and it modifies the number of larval instars or the morphology in the same stage of development in some marine crustaceans (Anger, 2001;Oliphant et al., 2013).The objective of this study was to compare the morphology of the megalopa of C. antillensis collected at the PNSAV to those reared in laboratory conditions from both Brazilian and Panamanian populations.

MaTerial and MeThods
The PNSAV is a coral reef system located in the northwest sector of the Bay of Campeche, in the polygon delimited by 19°00'00''-19°16'00''N and 95°45'00''-96°12'00''W.It is formed by 23 coral reefs distributed in northern and southern groups, which are separated by the mouth of the Jamapa River (Granados et al., 2007) (Fig. 1).
Megalopae were collected using light traps placed off the northwest of Isla Sacrificios for three consecutive nights, on June 2015.These collecting cycles were initiated at 21:00h and stopped at 01:00h on the next day during the new moon phase.The temperature recorded was almost invariable (28 to 29 °C), and salinity was 35 psu.All traps were constructed with plastic boxes that were 0.4 m long, 0.25 m wide and 0.30 m high, with 40 mm diameter inlet holes in the upper part of each side of the box; the box contained a white light of 38 lumens (Fig. 2).Trap contents were collected manually and placed in 500 ml plastic bottles at the end of the sampling period.All samples were preserved in 70% ethanol.

Morphological variations on megalopa of C. antillensis
Nauplius, 26: e2018031 Clark et al. (1998), while setal classification followed that proposed by Garm (2004).Images of the larvae were taken using a Leica DM750 microscope equipped with an Omax 14MP USB 3.0 digital camera.The illustrations were made with the Corel Draw V.12 program.

resulTs
We found 2,795 larvae of the genus Clibanarius: 968 of these were identified as C. antillensis, 656 as C. sclopetarius, and 1171 as C. vittatus (Tab.1).We were able to differentiate the species C. antillensis from C. sclopetarius and C. vittatus considering the total length, number of aestethascs in the external flagellum of the antennule, number of segments of the antenna, number of segments of the palp of the mandible and the number of plumose marginal setae in the endopod and exopod of uropods (Tab.2).
Remarks.We found four meristic differences among the Brazilian (Brossi-Garcia and Hebling, 1983) the Panamanian (Siddiqui et al., 1991) and Mexican populations (Tab.1).The antennule in the Panamanian and Mexican populations had more aesthetascs on the second, third and fourth segments (0. 6, 5, 3, 0) with respect to the Brazilian population (0, 4, 4, 2, 0).There were fewer plumose setae on the scaphognathite in the Panamanian and Mexican populations (49-55) with respect to the Brazilian population (70 plumose setae).In the second maxilliped, the difference was of 1 seta, whereas in maxilliped 3, there were 1-3 fewer denticles in the Brazilian population on the crista dentata.The exopod of the uropod had fewer (10) marginal plumose setae in the Brazilian population.

disCussion
We identified hundreds of larvae of C. antillensis, C. sclopetarius, and C. vittatus and the observed key characters of the megalopa of C. sclopetarius and C. vittatus corresponded to those observed in the literature (Lang and Young, 1977;Brossi-Garcia, 1987).Thus there was no doubt concerning the identity of the megalopae of C. antillensis collected; such variations in morphology of the megalopae must be explained.
We found that the Panamanian and Mexican populations were morphologically more similar in the number of aesthetascs on the antennule, setae on the scaphognathite, second maxilliped, number of denticles on the crista dentate and number of setae on the exopod of the uropod.
There are two potential explanations for such differences: either these differences are the result of intraspecific variations, probably following a latitudinal gradient or there are two cryptic taxa, currently unrecognized: one distributed mainly in Brazil and another distributed in the Caribbean, Mexico and the USA.
In the first option, McLaughlin and Gore (1988) showed that it is highly probable that considerable developmental differences can occur between geographically separated populations.The differences during larval development relate to the changes in environmental conditions, such as the mixture of water masses, food deficiency or changes in temperature Morphological variations on megalopa of C. antillensis Nauplius, 26: e2018031 and salinity.This flexibility is an evolutionary and ecological strategy which enhances survival and allows the distribution of the larvae to be extended (Welch and Epifanio, 1995;Hartnoll, 2001;Thatje and Bacardit, 2000;Negreiros-Fransozo et al., 2008).In this respect, Tirmizi and Siddiqui (1980), McLaughlin and Gore (1988) and Siddiqui et al. (1991) found differences in the setation and spination of Pagurus kulkarnii Sankolli, 1962, Pagurus hirsutiusculus (Dana, 1851) and C. antilliensis.Furthermore, the morphological variability increases in the later larval or decapodid stages.For example, in Crangon crangon (Linnaeus, 1758), in the first decapodid stage, at least two forms can be distinguished, differing primarily in the developmental state of the antennae, the maxillipeds, the pereiopods, and the pleopods (Linck, 1995).Brossi-Garcia and Hebling (1983) found more setae on the scaphognathite of the maxilla of the Brazilian population; this could be explained by the availability of food affecting the life cycle of many benthic marine invertebrates, with significant implications for their survival, growth and metamorphosis (Calado and Leal, 2015;Pechenik and Tyrell, 2015).It has been found that the morphology of the feeding structures is more developed in the megalopa and juveniles, and the mouthparts present an increasing number of setae compared to previous developmental stages, which indicates that these stages are more able to capture food and they are capable of processing solid food available in their benthonic environment (Abrunhosa et al., 2006).The ability to process small food particles increases because of the increment in the size of the foregut and the number of setae on the mouthparts (Abrunhosa and Melo, 2008).The phytoplankton is most abundant in high latitudes and in upwelling zones along the equator and near coastlines.However, phytoplankton is scarce in the subtropical regions, where nutrient levels are low and drop off in the summer (NASA, 2018).
For the second potential explanation, we were considering genetically comparing different populations of C. antillensis since it has been shown that cryptic species do occur among species of this genus.Furthermore, Negri et al. (2014) found morphological and genetic differences strong enough to resurrect C. symmetricus (Randall, 1840).This species was reported for many years under the name C. vittatus because of their close morphological similarities.The authors defined the genetic divergences between C. vittatus and C. symmetricus, identifying that the latter species may occur more broadly in the Caribbean, the Antilles, or the southern Gulf of Mexico.Similar to C. vittatus, C. antillensis has a broad range in the western Atlantic, from the southeastern coast of the United States, and from Bermuda to Brazil (Felder et al., 2009), although the distribution is not continuous, so differences might be found.While the present study was in review, we were not aware that Nishikawa (2017) already preformed a robust and representative sampling with two molecular markers (16S and COI) and found no genetic structure among different western Atlantic populations of C. antillensis, including populations from Mexico, Panama and Brazil.Thus it appears that differences observed in the megalopa stage among geographically separated populations are the result of interpopulation variation.

aCKnowledgeMenTs
We appreciate the support to the national sabbatical program 2018 of CONACYT.We also thank the authorities of SEMARNAT, CONANP and CONAPESCA-DGOPA (SAGARPA) for the permits granted (PPF/DGOPA-051/15).RR wishes to acknowledge PRODEP-SEP, Mexico, through the program "Apoyo a la Incorporación de NPTC (Ago/1/2018 -Jul/31/2019)".We are also thankful to F.L. Mantelatto and two anonymous reviewers for suggestions made to an earlier version of this paper.

Table 1 .
Abundance of Clibanarius species in the sampling area.

Table 2 .
Comparison of relevant megalopae features of Clibanarius antillensis, collected from a Mexican population to those obtained in laboratory conditions from Brazilian (Bra) . The same features are compared to C. sclopetarius and C. vittatus collected in the same area ( 1 ,Brossi-Garcia, 1987; 2 , Lang  and Young, 1977; * obtained from the figure).