Intraspecific variation in carapace morphology among fiddler crabs ( Genus Uca ) from the Atlantic coast of Brazil

Isolation due to geographical barriers should promote genetic and morphological divergence among populations. Marine currents flowing in opposing directions along landmasses can constitute barriers that isolate populations dependent upon aquatic dispersal. The distribution of fiddler crabs (genus Uca) is regulated primarily by the oceanic transport of their planktonic larvae and by available adult habitat. Along the Brazilian coast of eastern South America, the flow of 2 major oceanic currents separates northern from southern Uca populations, which may promote intraspecific divergence in ‘trans-Brazilian’ species. Populations of 10 Uca species were sampled at 64 locations north and south of the Ponta do Calcanhar, Rio Grande do Norte, Brazil. Carapace shape was assessed using geometric morphometrics to analyze 12 surface landmarks in 1319 female crabs. Carapace shape differs significantly in each species. In morphospace, the carapace forms of the 10 species appear to separate into traditional subgeneric clusters. Within the 8 species exhibiting trans-Brazilian distributions, northern and southern populations show distinct carapace differences. Depending on species, either the hepatic or the branchial region is larger in northern populations. Since significant genetic variability among such populations has not been confirmed, divergence in carapace shape suggests significant ecological modulation of phenotype within each species. Apparently, environmental differences between northern and southern localities exert a greater impact on carapace morphology than impeded gene flow. The drivers under pinning diversification of carapace shape remain unknown, however.

RIGHTS CHECKING: 4/6/2016 --Sherpa/Romeo green journal, doja as an open access journal, author can archive publisher's version/pdf, on author's personal website, institutional website and institutional repository, Gold' Open Access = article published by publisher with free availability to all users; available from IR as Open Access (or formerly Free Access). Publisher's PDF may be posted on the Author's personal or institutional website or deposited into the Author's institutional Open Access repository any time after publication only if the article is published with 'Gold' Open Access. In all cases, the article on a website or in the repository must have a link to its abstract page on the Inter-Research website, and cite the article doi.
This Article is brought to you for free and open access by the Faculty Work at UNI ScholarWorks. It has been accepted for inclusion in Faculty Publications by an authorized administrator of UNI ScholarWorks. For more information, please contact scholarworks@uni.edu.

INTRODUCTION
The coast of Brazil constitutes a composite of 5 contiguous tropical and subtropical biomes, stretching from above the Amazon River in the north to the border with Uruguay in the south (Thurman et al. 2013). Two of these biomes occur along the northern coast, and 3 along the southern coast, where they are separated by the Ponta do Calcanhar in Rio Grande do Norte state. At the Ponta do Calcanhar, the westward flow of the Central South Equatorial Current (CSEC) to be unimpeded by significant freshwater outflow. To illustrate, molecular studies have shown littoral populations of the mangrove crab Ucides cordatus (Linneaus, 1763) between Ponta do Calcanhar and São Paulo to be genetically homogeneous, implying a high degree of connectivity and panmixia among populations (Oliveira-Neto et al. 2007).
Fiddler crabs are fossorial, semi-terrestrial crustaceans that live primarily in the littoral zone of protected bays, estuaries and lagoons, and particularly in mangroves (Crane 1975). Twenty species of Uca are known from the shores of the western Atlantic, Gulf of Mexico and the Caribbean Sea (Beinlich & von Hagen 2006). The geographical range of the adult crabs appears to be partly dependent on the dispersal of planktonic larvae carried by ocean currents and tides (Epifanio et al. 1988, Christy 2011, López-Duarte et al. 2011, Shih 2012. Of the 10 species inhabiting the coastline of the western South Atlantic Ocean, 8 occupy both the northern and the southern coasts of Brazil (Melo 1996, Bezerra 2012, Thurman et al. 2013 and represent 4 subgenera. When adult females along the northern coast release zoeae into estuarine or coastal waters, the larvae are transported toward the Amazon River and the Caribbean or out into the mid-Atlantic Ocean. In contrast, larvae released along the southern coast are transported in the opposite direction toward Cabo Frio and Uruguay (Boltovskoy et al. 1999, Psuty & Mizobe 2005.
While the geographical distribution of Uca (sensu lato) in Brazil is influenced primarily by regional hydrology, geomorphology and climate (Thurman et al. 2013), ocean current patterns help regulate larval dispersal and thus direct gene flow and affect connectivity. Connectivity in turn can influence intraspecific variation both within and among marine populations (Kelly & Palumbi 2010, Sanford & Kelly 2011, Ituarte et al. 2012. The division of the major currents at the Ponta do Calcanhar, in particular, may significantly control gene exchange between the northern and southern populations of individual fiddler crab species. Should this be the case, we would expect to see divergence among northern and southern populations. By the early Miocene (22 million years ago), fiddler crabs were genetically distinct from other Ocypodidae , and all Pacific and North American clades of Uca were genetically distinct by the late Miocene (17 million years ago). Rosenberg (2001) analyzed the phylogeny of 88 species of Uca using 236 morphological traits, providing results similar to those previously reported in molecular studies. Nevertheless, the pre-vailing view holds that speciation in Uca from the western Atlantic has proceeded without significant morphological divergence (Salmon et al. 1979, Levinton 2001. Crane (1975) reported difficulty in distinguishing among females of 3 sympatric Brazilian species belonging to the subgenus Leptuca. She felt that the females of U. leptodactyla Rathbun, 1898, U. cumulanta Crane, 1943and U. uruguayensis Nobili, 1901 differed only in relative proportions. However, very few studies have quantitatively addressed intraspecific phenotypic variation in Uca over a wide geographical area. Silva et al. (2010) found the widespread species U. annulipes (Milne-Edwards, 1837) from southeast Africa to exhibit very little morphological or genetic structure, suggesting that populations distributed over 3300 km between Mikindani, Kenya, and Mlalazi, Republic of South Africa, are sufficiently connected by high larval transport to maintain panmixia. In contrast, along 13 500 km of coastline in the USA and Mexico, several endemic Uca species show detectable morphological variation (Hopkins & Thurman 2010) even though the widerspread species do not necessarily exhibit greater variation than species with smaller ranges.
In this study, we examine the impact of the major oceanic currents along the eastern coast of South America on phenotypic variation in several species of fiddler crabs from Brazil. Specifically, we address 2 questions: (1) Based on carapace structure, are the various Uca species morphologically distinct? (2) Does the Ponta do Calcanhar constitute a significant geographical feature coincident with phenotypic diversification within each Uca species? After collecting and preserving specimens from numerous locations, we performed geometric morphometric analyses to quantify variation in carapace shape both within and among the 10 Uca species from Brazil. We expanded Rosenberg's (2001) phylogenetic analysis to include all the Brazilian species, finding that morphological variation largely corresponds to the known phylogenetic relationships. Further, the 8 trans-Brazilian species exhibited morphological differences between their northern and southern populations, suggesting that the Ponta do Calcanhar might represent a biogeographical feature that underpins intraspecific divergence. In general, as carapace width broadens, length shortens when comparing specimens from northern to southern populations. The greatest variation occurs in the branchial and hepatic regions. However, given available genetic information for Uca species from Brazil, such differentiation does not appear to correlate with underlying genetic structure (Wieman et al. 2013).

Sampling
Over 7000 fiddler crab specimens were collected from habitats along 9600 km of the Brazilian coast between 2009 and 2010 (Thurman et al. 2013). Field collections, authorized by the Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis (IBAMA, permit nos. 2009/18559-1 and2010/23976-1) were made at 64 localities between the states of Amapá (AP) and Santa Catarina (SC) (Fig. 1). Fiftyfour of these locations were sampled between Pernambuco (PE) (Itapissuma, Ilha de Itamaracá, and northernmost point) and SC (Palhoça, Barra da Passagem, and Rio Massiambu) from June to November 2009. The remaining 10 localities were sampled from AP (Calçoene, Rio Cacoal) to Ceará (CE) (Fortaleza, Rio Cocó) between June and August 2010. These sites were not randomly chosen but constitute convenient points of access to littoral habitats by road, track or boat; the collections may be biased for particular or rare species. The habitat character (salinity, substrate), species composition and precise location of the sites are provided in Table S1 in the Supplement (available at www.int-res.com/articles/suppl/ b020p053_supp.pdf).

Species identification
The 10 species of fiddler crabs collected during this study were identified using traditional morphological characteristics. Since 5 of the species occur in North America, details of their morphologies have been described elsewhere (Barnwell & Thurman 1984). A dichotomous key (Melo 1996) was used to identify the Brazilian fiddler crabs, together with supplementary descriptions of the more cryptic species (Crane 1943, Holthuis 1967, Chase & Hobbs 1969, Coelho 1972, von Hagen 1987, Tavares & de Mendonça 2003, Bedê et al. 2007). Although there is some overlap between the morphological characters used to discriminate among species and those captured by the landmark data (see below), specimens were assigned to species prior to digitizing landmarks. Thus, the character set used to identify specimens and that used to estimate variation are not coincident. Further, species were identified largely by one investigator (C.L.T.) while morphometric analysis was performed by others (M.J.H., K.R.H.).

Phylogenetic analysis
A phylogenetic reconstruction for the 10 species of Uca from Brazil was created based on a parsimony analysis of morphological characters. A scored list of 236 characters for each of 11 species was taken from Rosenberg (2001) Thurman, 1981, U. (M.) marguerita was also included to clarify and assess the rapaxvictoriana relationship (von Hagen 1987). Two maximum parsimony trees were found employing PAUP software, version 4.0b10 (Swofford 2000) and the PaupUp, version 1.0.3.1 graphical interface (Calendini & Martin 2005), using the Branchand-Bound search algorithm and U. Afruca tangeri (Eydoux, 1835) as the out-group (see Rosenberg 2001). All characters were unordered and equally weighted (121 parsimony-informative characters, 66 constant characters). Multi-state taxa were considered polymorphic. A bootstrap tree was constructed from 500 pseudo-replicates and using a full heuristic search (Fig. 2). The final tree is a strict consensus. Excluding U. (M.) marguerita, all tree branches were equal in length from one node and no boot-strap values were greater than 50%. Consequently, this working phylogeny with 12 species is tentative, and we have chosen not to rename any taxa, although the subgenera Minuca and Leptuca may be paraphyletic.

Specimen preparation
After collection, live specimens were transported by air or car to a laboratory at the Centro de Biologia Marinha (CEBIMar/USP), São Sebastião, São Paulo, Brazil, where they were used in physiological experiments. Subsequently, the crabs were quickly killed by chilling and preserved in 80% ethanol (Rufino et al. 2004). Lots labeled by location for each species were deposited at the Museu de Zoologia of the Universidade de São Paulo. Since male fiddler crabs have one greatly enlarged cheliped, their carapace shape may be distorted as a structural response to claw asymmetry (Yerkes 1901, Duncher 1903, Huxley 1971, Miller 1973. Given statistical considerations, the asymmetrical component of variation within largely bilaterally symmetric organisms must be analyzed separately from the symmetrical component (Bookstein 1996, Klingenberg et al. 2002. While the asymmetrical component is frequently of interest in studies of variation within individuals (e.g. that due to developmental instability), the symmetrical component represents the shape variation among individuals. Since we were interested in this aspect of variation, and in changes related to biogeography rather than sexual selection, only female specimens (n = 1319) were used, and only the symmetrical component of variation was analyzed here.
Specimens were oriented for photography so that the carapace was horizontal in frontal view and its anterior-and posterior-most edges lay in the same horizontal plane in lateral view (Fig. 3). Orientation and digitization error was assessed by repeatedly mounting and digitizing a single random specimen for each species analyzed (Hopkins & Thurman 2010). A single investigator (M.J.H.) performed all photography, and all digitization was carried out by another (K.R.H.). Error was quantified by comparing disparity in the error samples with that in the species data using DisparityBox6i (Sheets 2001(Sheets −2007. The disparity in each error sample was less than an order of magnitude smaller than that for the entire species; thus, measurement error was deemed negligible.

Morphometric analysis
Morphological variation was examined using geometric morphometric techniques (Bookstein 1991, Zelditch et al. 2004. A total of 21 landmarks were chosen to capture the overall shape of the carapace using the program 'tpsdig2' (Rohlf 2010). The landmarks utilized here reflect the 3-dimensional nature in the Supplement). As paired landmarks represent redundant information regarding the symmetric component of variation, we reflected the landmarks across the midline, leaving 12 landmarks for analysis (Fig. 3). The landmark data were standardized using Standard6 (Sheets 2001(Sheets −2007 to eliminate variation due to allometric growth (Fig. S1) and were transformed using thin-plate spline decompositions; the resulting partial and uniform warp scores were then used to perform statistical analyses (Rohlf 1990, Bookstein 1991, Zelditch et al. 2004. Partial warps describe the same variation as Procrustes residuals; each can be transformed into the other by rotation of the coordinate system. Variation among and within species was evaluated using a principal components analysis (PCA) of the warp scores (Zelditch et al. 2004). The morphometric shape data were mapped onto the phylogeny using unweighted squared-change parsimony to reconstruct values at internal nodes from the shape averages of the species at the terminals (Maddison 1991, Klingenberg & Gidaszewski 2010. Phylogenetic signal in the morphological data was assessed using a permutation test. The morphological shape data are randomly swapped between terminals and the resulting tree lengths calculated. When phylogenetic signal is absent from the shape data, randomly swapping these data is equally likely to produce a greater or smaller tree length. If phylogenetic signal is present, then randomly swapping the shape data should result in a greater tree length. Thus, the null hypothesis holds that the shape data show no phylogenetic signal, and the empirical p-value for the test is the proportion of the permuted data sets in which the sum of squared changes is shorter or equal to the value obtained for the original data (Klingenberg & Gidaszewski 2010). Both analyses were performed using MorphoJ (Klingenberg 2011).
The Procrustes distance between species' means was used to evaluate relatedness of groups in morphospace, and a resampled Goodall's F-test was performed on the standardized data to test for statistically significant differences between the northern and southern populations. A canonical variates analysis (CVA) was used to describe the morphological variation between northern and southern populations. To assess the proportion of original specimens from the northern or southern populations matching the CVA discrimination, a jack-knife assessment test was performed a posteriori within each species. The better the assignment matches the original grouping, the better the CVA is able to discriminate between the 2 populations. A 2-way multivariate analysis of variance (MANOVA) was also conducted to test for differences in carapace shape within species and among localities (northern versus southern populations), and for any species−locality interaction. Finally, PC 2 (24.3%) shape trajectories between northern and 12 southern populations were compared in terms of magnitude (vector length) and 10 direction (angular difference between vectors). Differences were tested using a 8 permutation test where the null hypothe-6 sis is that the difference is greater than expected from random pairs of vectors

Interspecific distinctions
Principal components (PC) 1 and 2 account for 63.2% of variation across all the Uca (sensu lato) species examined (Fig. 4). Each species is statistically distinct from its congeners (Table 1), and species considered to be members of the same subgenus cluster closely together in morphospace (Fig. 4). The working phylogeny was projected onto the species' distributions described by the PCA and displayed in morphospace (Fig. 5). Each point represents the mean shape of the species. The morphometric data show significant phylogenetic signal (p ≤ 0.0002 after 10 000 permutations). In the reconstruction, clear divergence is seen along the first principal component axis between species in the subgenera Minuca and Leptuca. Divergence between U.
(U.) maracoani and all other species is clearly evident along the second principal component axis. Within the subgenus Minuca, there is clear divergence of species along both the first and second principal components axes (lower left portion of Fig. 5).
In general, the distribution in morphospace of the 10 species of Brazilian fiddler crabs is consistent with the older evolutionary relationships suggested by Crane (1975), and supported by Rosenberg (2001) and Beinlich & von Hagen (2006). Differences among

Intraspecific variation
Based on the CVA, each of the 8 species with a trans-Brazilian distribution exhibits morphological divergence between populations from the northern and southern coasts. Variation is due primarily to the relative size and shape of the hepatic and branchial regions of the carapace (Fig. 6). Intraspecific differ-In general, southern populations are broader in the antero-lateral ridge and shorter in carapace length than are the northern populations (Fig. 6). Variation in the hepatic and branchial regions manifests as a swelling and broadening of the carapace. However, there are differences in the patterns of divergence between northern and southern populations among the 8 species. For example, Uca Minuca mordax, U. (M.) burgersi, and U. Leptuca leptodactyla show primary swelling in the hepatic region (Fig. 6L,N (Fig. 6M). As a result, a number of species show significant differences in magnitude and direction of shape change between their northern and southern populations (Table 3). Notably, magnitude differs far more frequently than direction, suggesting that species differ mostly in the degree to which they have diverged across this geographical boundary. These differences may be due to different rates of divergence over time, to different lengths of time during which divergence has taken place, or to different degrees of plasticity among the species.
The jack-knife assignment test indicated that individuals initially 'assigned correctly' to a region were frequent but not remarkable (Table 4). In each of the 8 cases, this might be expected since the 2 populations belong to the same species. While some differences are present, the populations overlap morphologically. Again, the populations of Uca Minuca mordax appear to express the greatest degree of intraspecific divergence while those of U. Boboruca thayeri and U. (M.) vocator show the least divergence. However, the latter case may derive from the small number of specimens used (N, north = 34, N, south = 16).  In traditional classification schemes, species of the genus Uca can be separated into 2 morphological groups consisting of the 'narrow-fronted' and the 'broad-fronted' species (Rathbun 1918, Bott 1954. Crane (1975) divided the genus into 9 subgenera creating a unique subgenus, Boboruca, for U. thayeri and U. umbratila Crane, 1941. Owing to cheliped armature, lethargy and form of waving display, she considered Boboruca to be related to the Indo-Pacific subgenera Paraleptuca Bott (= Amphiuca Crane) and Tubuca Bott (= Deltuca Crane). Crane's classi-DISCUSSION

Interspecific variation
Our presumptive phylogeny for the 10 Uca (sensu lato) species from Brazil is in consonance with that proposed for fiddler crabs from Trinidad by Albrecht & von Hagen (1981). However, neither U. Leptuca uruguayensis nor U. Minuca victoriana were included in the latter phylogeny, and U. (U.) major (Herbst, 1804) was not incorporated in the present scheme since it is not ecologically relevant for Brazil. Here, all 10 Uca species are morphologically distinct to varying degrees (Figs. 4 & 5). We found U. (L). leptodactyla to be very similar morphologically to both U. (L.) cumulanta and U. (L.) uruguayensis (Fig. 5, Table 4). fication system was modified recently by Rosenberg (2001) and Beinlich & von Hagen (2006), and Boboruca was incorporated into the subgenus Minuca.
Some findings support Crane's supposition that the subgenus Boboruca is distinct. Salmon (1987) compared courtship behavior, reproductive biology and ecology of Uca (B.) thayeri to broad-fronted (U. Leptuca pugilator (Bosc, 1802)) and narrow-fronted (U. Gelasimus vocans (Linnaeus, 1758)) species. Intertidal ecology, habitat usage and female behaviors (burrow defense, incubation and mate selection) were most similar to the subgenus Gelasimus. However, female reproductive physiology (opercula decalcification, receptivity and clutch periodicity) was more similar to the subgenus Leptuca. Emphasizing distinctness, the ultrastructure of spermatozoa from U. (B.) thayeri is unique for the genus (Benetti et al. 2008). Salmon & Zucker (1987) thus offered an alternative hypothesis that U. (B.) thayeri is a broadfronted species demonstrating convergence in behavior, ecology and physiology with species in the narrow-fronted subgenera, arguing that U. (B.) thayeri is not derived from the subgenus Minuca but Table 3. Uca spp. Comparison of shape vectors between northern and southern populations among species. Differences in magnitude (vector length) shown at upper right; differences in direction (angular difference between vectors) shown at bottom left. Although the angles between vectors are large, only a few are significant. This is likely due to the degree of within-group variation relative to the orientation of angles between vectors. *p < 0.05, ** p < 0.01, *** p < 0.001 based on permutation test (1000 iterations). See Table 1  represents an independent, converging evolutionary trajectory. Additionally, employing a multivariate analysis of 12 meristic characteristics in 6 species, Diniz-Filho (1990) sorted adult male Uca from Brazil into 3 distinct clusters along PC 1 (size) and PC 2 (shape) axes. Specimens from the subgenera Uca and Leptuca were clearly distinct from those of the subgenera Results from molecular phylogenetic analyses are contradictory. Using DNA or 16S ribosomal RNA, Sturmbauer et al. (1996) and Levinton et al. (1996) found that Uca Boboruca thayeri is most closely related to several species in the subgenus Leptuca. Landstorfer & Schubart (2010) compared U. (B.) umbratila to 9 other tropical Minuca species from the Pacific shores of Costa Rica. Their parsimony network analysis of a 619 bp DNA sequence for 28S ribosomal RNA from species in the subgenus Boboruca revealed marked differences compared to those of the 9 species of the subgenus Minuca. However, analysis of the relationships among the species using Bayesian Inference of 658 bp DNA sequences for a cytochrome oxidase subunit (COX-1) suggests that U. (B.) umbratila is most closely related to U. Minuca brevifrons (Stimpson, 1860), a sibling species of U. (M.) mordax. Based on these molecular studies, the exact relation of U. (B.) thayeri to other New World 'broad-fronted' Uca remains unresolved.
In the present study, the 10 Brazilian species generally form subgeneric clusters in morphospace, supporting previously hypothesized evolutionary relationships (Crane 1975, Rosenberg 2001, Beinlich & von Hagen 2006. Uca (U.) maracoani is distinct in morphospace from the Minuca and Leptuca clades. The Uca (sensu stricto) clade is thought to be basal in the phylogeny to both the subgenera Minuca and Leptuca (Rosenberg 2001), which share a more recent relationship (Albrecht & von Hagen 1981. Our findings also support Albrecht & von Hagen's (1981) suggestion to abandon the subgenus Boboruca and incorporate its member species (U. thayeri and U. umbratila Crane, 1941) into the subgenus Minuca. The New World fiddler crabs would then form 3 clades (Beinlich & von Hagen 2006): subgenus Uca (narrow-fronted), subgenus Minuca (broad-fronted) and subgenus Leptuca (broad-fronted). The frontal (interocular) width is approximately 20% of carapace width in the subgenus Boboruca, 10% in the subgenus Uca, and between 30% and 40% in the subgenera Minuca and Leptuca. Interestingly, in terms of its osmotic physiology, U. (B.) thayeri exhibits the regulatory pattern seen in members of the subgenus Uca (i.e. U. (U.) major and U. (U.) maracoani) rather than in species from the subgenera Minuca or Leptuca (Lin et al. 2002, Thurman 2005, Faria et al. 2011. U. (B.) thayeri has an elevated hemolymph isosmotic concentration and does not osmoregulate well in low or high salinities, unlike most Leptuca and Minuca. Thus, there may be some convergence in osmotic physiology between Boboruca and Uca.

Intraspecific variation
The 8 species with trans-Brazilian distributions exhibit significant intraspecific variation in carapace morphology. These findings contrast with a study from the east coast of Africa in which Silva et al. (2010) found a lack of both morphological and genetic variation across very remote, isolated populations of Uca (Leptuca) annulipes (Milne-Edwards, 1837) in the western Indian Ocean. However, the present findings are consistent with those of Hopkins & Thurman (2010) for fiddler crabs along the eastern shore of North America, where similar patterns of morphological variation were found in geographically separated populations in the Gulf of Mexico and Atlantic Ocean. The differing results for the American and African studies may derive from the choice of landmarks, since those used by Silva et al. (2010) allowed analysis of carapace perimeter but not of the dorsal surface. However, the findings most likely differ due to the spatial scale assessed: the coastlines sampled by Hopkins & Thurman (2010), and here, are much longer (∼ 13 500 and ∼ 7500 km, respectively) than the African coastline (∼ 3300 km) sampled by Silva et al. (2010).
Divergent carapace morphology in the 8 trans-Brazilian species suggests that the Ponta do Calcanhar may be a disruptive geographical barrier. It bisects the coast of Brazil into 2 regions where conspecific crabs differ significantly in carapace shape. Intraspecific variation is localized primarily in 2 anatomical regions: branchial and hepatic. As the specific site of carapace variation is not unique to subgenus or species, there may be a relationship between carapace shape and environmental factors. For example, habitat differences in humidity may affect gene expression and morphological variation via an unidentified epigenetic mechanism. The branchial region of the carapace overlying the gill chambers assures water conservation. Thus, enlarging the branchial chambers would likely serve as a safeguard against desiccation (Jones 1941). Fiddler crab species from arid regions in the western Gulf of Mexico show similar adaptations (Thurman 1998, Hopkins & Thurman 2010. A detailed examination of the relationship between various environmental factors and carapace shape in several species is currently in progress. Beyond latitude, we expect these studies to demonstrate that various components of shape variation are related to habitat salinity and substrata grain size. Climate along the northeastern Brazilian coast between Pernambuco (PE) and São Luis (Maranhão, MA) is arid, while the southern coast is more humid (Espenshade & Morrison 1974, Boltovskoy et al. 1999, Psuty & Mizobe 2005. However, this may not account for morphological variation within species. For example, among the 8 trans-Brazilian species, only Uca Boboruca thayeri and U. Leptuca cumulanta show similar salinity and substrate preferences (Thurman et al. 2013), and they exhibit equivalent variation in branchial and hepatic carapace regions. No other pairs or group of species exhibiting ecological similarities display a common pattern of morphological variation. ) vocator inhabits fine-grained substrates. In general, variation in a specific carapace region does not correlate with differences in salinity or substrate preference among the species. Also, there appears to be no obvious difference in habitat vegetation that might influence divergence among the northern and southern crab communities (Thurman et al. 2013, Table S1 in the Supplement).
Presently, no known factor appears to drive morphological variation in populations of Brazilian Uca (sensu lato). In fiddler crabs, gene flow is promoted by larval transport on oceanic currents and tides (Epifanio et al. 1988, Neethling et al. 2008, Weersing & Toonen 2009, López-Duarte et al. 2011, and larvae of long planktonic duration are expected to disperse over greater distances (Grantham et al. 2003, Lester et al. 2007, Shanks 2009). This should promote extensive communication among populations, maintaining uniformity in morphology and genotype across the species' range. Morphological divergence among populations should be either random and unstructured or related to environmental differences between localities (Sanford & Kelly 2011). However, several studies have found genetic structure and dispersal potential to be uncorrelated (Weersing & Toonen 2009). Further, population networks of small effective population size may receive little or no influx of novel genes, and eventually diverge through inbreeding and drift (Fisher 1958, Dobzhansky 1959, Wright 1969. Thus, genotypic diversity among isolated populations may increase as variation within each population declines. At each location across the range, environmental factors may act selectively on phenotypes, altering genotype frequencies or even producing unique genotypes. Finally, certain habitat conditions can produce a variety of phenotypes from a single genotype (Miner et al. 2005, Vogt et al. 2008. Consequently, depending upon habitat−organism interac-tions, both genetic variability and phenotypic plasticity can drive diversity.
Historically, low intraspecific genetic diversity is a hallmark in fiddler crabs. Felder & Staton (1994) reported minimal divergence in several allozyme systems among 8 trans-Floridian populations of Uca Minuca minax (LeConte, 1855). Like east African U. annulipes (Silva et al. 2010), no genetic structure has yet been found across coastal populations of fiddler crabs in Brazil. Wieman et al. (2013) found little genetic variation in the DNA sequences of cytochrome oxidase-1 haplotypes in U. (U.) maracoani distributed between Amapá (AP) and Paraná (PR). Studies on other crab species also suggest that marine populations along the southern coast of Brazil are intimately connected by gene flow (Oliveira-Neto et al. 2007, Laurenzano et al. 2012. However, other brachyuran species along this coast may exhibit strong genetic differentiation in patterns that indicate isolation-by-distance over about the same length of coastline (Ituarte et al. 2012). For fiddler crabs, at least, the absence of demonstrated genetic structure across populations implies that intraspecific variation results from phenotypic plasticity attributed to either epigenetic change (Miner at al. 2005, Vogt et al. 2008, variation at other genetic loci, or currently unknown variables.