A multifaceted comparison between the spider crabs Epialtus bituberculatus and Epialtus brasiliensis (Brachyura: Majoidea: Epialtidae) in the neotropical areas of the western Atlantic: morphology, morphometry and DNA markers

The morphological characters used to distinguish the two species of the crab genus Epialtus that occur in the neotropical areas of the Brazilian coast, E. bituberculatus and E. brasiliensis, are variable, casting doubt on their taxonomic status. Three approaches were chosen in order to resolve this uncertainty: morphological, molecular analysis and combined phylogenetic analysis with genetic data and morphometry. A molecular phylogenetic analysis was performed to test whether these two species can be separated at the molecular level (using the two mitochondrial genes 16S and COI – DNA barcode). The genetic distances and phylogenetic trees showed that E. brasiliensis fell within the clade of E. bituberculatus. We suggest that there is one species in Brazil and the presence or absence of the spine on the ventral surface of the propodus of the ambulatory pereiopods (a key diagnosing character) appears to be uninformative. However, all the analysis showed that specimens of Epialtus can be separated in three clades corresponding to three geographic regions: Caribbean Sea, Venezuela and Brazil. We conjectured the existence of three cryptic species that evolved recently. http://zoobank.org/urn:lsid:zoobank.org:pub:611E18BF-4CD6-40C8-B57E-01F2FA62E841


Introduction
The family Epialtidae MacLeay, 1838, includes 76 genera of marine crabs living from the intertidal zone to the deep sea. Among them, Epialtus H. Milne Edwards, 1834, contains 11 valid species. [1] Members of this genus can be found on both sides of the American continent: on the Pacific side, they can be found from southern California to Chile, while on the Atlantic coast, they eastern Pacific was unsuccessful, and like Garth, [6] we now have serious doubts that the original type locality is correct. We follow Rathbun [2], Garth [3] and Powers [7] in restricting the range of this species to the western Atlantic coast. Recently, this species was removed from the Chilean decapod fauna. [8] The two species of Epialtus reported for the Brazilian coast with overlapping distributions are E. bituberculatus, which can be found from Florida (USA) to Santa Catarina (Brazil), [2,7,[9][10][11][12][13][14][15] and E. brasiliensis Dana, 1852 which occurs on the Atlantic coast of Colombia, Venezuela and Brazil (Ceará, Espírito Santo to Rio Grande do Sul). [9,11,12,[15][16][17] However, there is a gap in the distribution of E. brasiliensis from Rio Grande do Norte to Sergipe, an area where this species has never been found. [9,17,18] Furthermore, the only report of this species occurrence in Salvador (Bahia, Brazil) [19] has been challenged based on its apparently disjointed geographical distribution and with no other records from Bahia. [20] Thus, the only confirmed record of its occurrence in northern Brazil is Ceará. [12,16] In addition, the geographic distribution of E. brasiliensis is discontinuous because it occurs separately in Colombia and Venezuela [9,11], in Ceará (Brazil) [16] and in the south and southeast regions of Brazil. [15] These patterns of occurrence raise questions about whether these gaps are due a lack of faunal surveys, [14] a misidentification of specimens that are morphologically similar to the sympatric species E. bituberculatus (present study) and/or because E. brasiliensis is a morphologically variable species but part of a species complex. Sharing similar habitats can result in morphological similarities due to convergence, which makes it more difficult to infer evolutionary relationships.
Epialtus bituberculatus and E. brasiliensis are sympatric in the southern hemisphere. They inhabit the intertidal zone to 10 m depth and can be frequently found associated with sea grasses and algae on hard substrates or sandy bottoms. [9,17,21] The larval morphologies of the two species are very similar, [22] and adults of both species have two tubercles on the gastric region and a simple rostrum. [2,9] They can be differentiated by the carapace subpentagonal and rostrum with rounded extremity in E. bituberculatus and carapace subhexagonal and rostrum equilaterally triangular in E. brasiliensis; ambulatory pereiopods of E. brasiliensis are stouter than those of E. bituberculatus and the last three pairs also very short with a spine on the ventral surface of the propodus. [2,9,23] In this study, a multifaceted approach composed of morphological characters, molecular phylogenetic analysis and combined phylogenetic analysis was used in order to clarify the taxonomic positions of E. bituberculatus and E. brasiliensis. For the combined analysis, we used numerical methods (traditional morphometric) with genetic data as a complement to the morphological and molecular data-sets.

Material and methods
Morphological data collection For the taxonomic revision and to compare Epialtus bituberculatus with E. brasiliensis, we selected 17 morphological characters that have been used in species descriptions [2,3,5,9,23]; and we added another nine characters for both sexes and across a size range ( Table 1).
The examined material was deposited in the Crustacean Collection of the Department of Biology (CCDB), Faculty of Philosophy, Sciences and Letters at Ribeirão Preto (FFCLRP), University of São Paulo Costa Rica). We measured the carapace length (CL = from the posterior to the anterior margin, including the rostrum) with a vernier caliper (0.01 mm) of each specimen. This standard measure is used to indicate the size of the specimens and facilitate comparisons with previous published results, between genus and among age groups.

Genetic data collection
Molecular analysis was used as a complement to the morphological data-set. Most of the specimens used for the molecular analysis were deposited in the CCDB/ FFCLRP/USP. Complementary specimens were obtained on loan from crustacean collections ( Table 2). Partial fragments of the large ribosomal subunit (16S mtDNA) and the Cytochrome oxidase I (COIthe barcode region) were used as genetic markers. Our methodology followed the protocols used by Mantelatto et al. [24], adjusted as follows.
Successful PCR products were purified using Sure Clean ® following vendor protocols and sequenced with the ABI Big Dye ® Terminator Mix (Applied Biosystems,  Sequences were aligned using Clustal W with an interface in BioEdit using default parameters. [28] Genetic distances calculated with the program Mega5 [29] were used to compare E. bituberculatus with E. brasiliensis and congeners. Tree building was performed using maximum likelihood (ML) [30] and executed in the program randomized axelerated maximum likelihood (RAxML) [31] as implemented on the online platform Cipres (Cyberinfrastructure for Phylogenetic Research). [32] The model of DNA evolution assumed GTR+G+I as default parameters for RAxML. Only bootstrap (1000 replicates) confidence values >50% were reported.
For the 16S and COI analysis, sequences from Gen-Bank of Taliepus nuttallii (Randall, 1840) and Acanthonyx petiverii H. Milne Edwards, 1834 were used as the out-groups because they are phylogenetically close to the genus Epialtus. [25] Then, we grouped specimens according to geographic distribution (Caribbean, Venezuela and Brazil) and calculated the genetic distances again with the aid of the program Mega5 in order to compare with the ML tree.
Traditional morphometrics data collection Eight body measurements, four from carapace and four from third pereopod (Table 3), were taken from the 12 specimens whose DNA was sequenced for both genes. All dimensions were measured to the nearest 0.1 mm using a caliper rule or under a light stereomicroscope. Individuals with injured or missing appendages were excluded. These measurements were first used for traditional morphometrics, and then they were added to molecular data on a combined analysis.
Carapace length is a standard measure which can be used to indicate the size of the specimens and facilitate comparisons with previous published results. However, we chose an alternative approach with principal components analysis using the first component (PC1) as a general descriptor of organisms' size [33] and the residual method to adjust the size through the residual of the relationship between PC1 (in this study, carapace length with rostrum) and each of the measures ( Table 3). Most of the variation attributed to the size is removed in the analysis, allowing a better approach in changes in the general shape of the body of organisms. [34,35] The software PAST 3.08 [36] was used for the regression between PC1 and each measure; residual was obtained and used in phylogenetic analysis. [34,35] The residual includes values below zero that were changed with the addition of 10 units, and transformed in logarithm (=LOG10) to standardize the data. [34] These changes did not affect the comparison since the data retained the same structure.

Combined phylogenetic analysis
The residual table along with the molecular alignment data (16S and COI separated) was used for building a combined matrix in the notepad, according to a specific script (data not show), and it was analyzed in the software TNT 1.1 (Tree Analysis Using New Technology) [37]. This software permits the use of continuous characters with values ranging from 0 to 65 with up to three decimal places, and analyzing them without making them discrete characters. [38] Heuristic search using new technology (New Technology Search) aims to analyze the largest number of possible trees and to find the most parsimonious tree, TBR algorithm (Tree bisection and reconnection; implemented in TNT) randomly performs rearrangements of branches from the initial trees, searching by chance a combination of branches that produces an even shorter tree than the other found. [39] In a more refined stage, sector search and tree drifting were used to isolate and analyze each part of the most parsimonious trees independently, which means an increase in effectiveness to find shorter trees even in analyses of matrices that could theoretically be analyzed only for heuristic search. [37,39] Branches of possible groupings not supported by synapomorphies were collapsed to prevent formation of monophyletic false groups; gaps ('-') in the molecular matrix were considered missing entries. Resampling method (1000 replicates, probability 33) was used as branch support because it is the most recommended for continuous data using TNT. [37,38] Three different values of implied weighing (k = 3, k = 7 and k = 10) of characters were chosen; we  [33] and the retention index to indicate the proportion of autapomorphies and homoplasy related to the total steps. [40] Results

Taxonomic assignments
We obtained a total of 97 specimens of Epialtus bituberculatus (44 males, 36 females, 17 ovigerous females) and 33 samples of E. brasiliensis (9 males, 8 females, 16 ovigerous females). Due to the morphological similarity between the two species and the lack of a distinguishing character, we could not separate them into two entities and our first hypothesis that there were two valid species could not be corroborated. Thus, based only in morphological data, we present a re-description of E. bituberculatus because of its older description, [5]  Diagnosis: Simple rostrum; carapace subpentagonal in both genders; cardiac region slightly elevated; absence (Figure 1(C)) or presence (Figure 1(D)) of proximal spine on the propodus ventral surface in the last three ambulatory pereopods.
Redescription: Carapace smooth, subpentagonal; 2 lateral lobes, 1 hepatic, 1 branchial separated by a concave border. Rostrum triangular or longer than broad; extremity rounded. Orbits absent; orbital region with lateral angles obtuse. Eyes small, mobile and partially hidden by carapace margin. Preorbital teeth absent or minute, not pointed, setae with variable length and thickness, sometimes absent. Postorbital teeth absent or minute, not pointed. Hepatic region with 2 lateral lobes slightly pronounced, curved upwards, smooth or with setae with length and thickness variable along margins. Gastric region smooth, with 2 small protogastric tubercles, smooth or with tufts of short/long setae. Cardiac region smooth, slightly elevated. Branchial region with 2 small teeth, not pointed, smooth or with short/long setae. Intestinal region smooth. Basal article of antenna subtriangular; following 2 articles cylindrical, hidden underneath rostrum in males, almost hidden in females; second and third articles formed by tufts of short/long setae on inner margin. Chelipeds of adult male strong, larger than first pair of ambulatory pereopods. In females, smaller or same size as first pair of ambulatory pereopods. Merus enlarged in 3 lobes in distal end, 1 central, 2 lateral lobes; inner margin of merus carinated in males larger than 8.0 mm. Inner margin of carpus lightly carinated, 2 crests on dorsal surface in males; inner margin not carinated in females, 2 crests on dorsal surface, 1 lobe on each lateral margin. Propodus enlarged in the lateral margins and high in the upper margin in males; less enlarged and less high in females. Fixed finger and dactylus almost entirely gaping in male, almost closed in female; fingers with ventral surface dentate and tufts of setae near distal end; dactylus (movable finger) with a large tooth in the middle of ventral surface. Ambulatory pereopods slender, subcylindrical; posteriorly decreasing in size. Merus with 3 lobes in distal end of first and second pereopods. Carpus shorter than merus, subtriangular in last 2 ambulatory pereopods. Propodus with 1 or 2 tufts of setae on ventral surface, 1 central, 1 near distal end; lightly dilated at outer distal end. Proximal spine on ventral margin of propodus in last 3 ambulatory pereopods may be present. Dactylus with 2 rows of minute spines intercalated with fine setae on ventral surface. Fourth and fifth abdominal somites united in male; fourth, fifth and sixth in female. Male first pleopod terminally truncate; coarse and less salient terminally, with a subtriangular lobe.
Remarks: Epialtus bituberculatus was described by Milne Edwards [5] for the coast of Chile and made the type species of the genus, but the Chilean location has been shown to be an error. [6, p.31] All subsequent records for the species have been from the Atlantic, [2,3,7,41,42] and this species has not been registered in Chilean decapod fauna since its original description. [8] It seems that E. bituberculatus was first recorded in western Atlantic as E. affinis Stimpson, 1859 in Indian River (Florida, USA), described through material from the Smithsonian Institution [43] and synonymized by Rathbun. [2] The type of E. affinis was probably lost and pointed out as not extant. [2] In Brazil, E. bituberculatus was recorded for the first time in Pernambuco, but it was mentioned with doubts and as 'brasiliensis form'. [44, p.67] However, the type locality of E. brasiliensis is Guanabara Bay, Rio de Janeiro, Brazil [23,45] and the holotype of this species has most likely been lost. [2, p.149] These facts make difficult the comparison of material analyzed herein with the typical representative of the specimen of the species, but finally we analyzed only a figure of Milne Edwards. [5] Thus, the designation of clades was problematic because in fact we do not know which entity we are dealing with. Finally, we suggested names and explain why we chose them (see Discussion).
Material of E. bituberculatus examined herein matches with the original description, [5] but that description is short and not detailed. In general, the specimens analyzed by us were similar to the original illustration [5, plate 15, figure 11], but even the smaller individuals and females have longer legs, mainly the first pereopod. Furthermore, merus and propodus in all pereopods are longer (Figure 1(A) and (B)) than the specimen in original description [5, plate 15, figure 11]. In the present study, the description of the adult male was identical to the material examined by Rathbun [2], but she did not mention females and juveniles. Moreover, we observed variation in the size of the rostrum (Figure 2) and in the presence or absence of teeth in the chelipeds fingers, which differs from the Rathbun's material [2, plate 45, figure 3].
In the original description, Dana [23] observed a small, closely appressed tooth behind the eyes and noted that the rostrum was entire and nearly equilaterally triangular in E. brasiliensis. In the material analyzed, we observed that the shape of rostrum was variable and ranged from triangular to longer than broad (Figure 2). In the carapace of E. brasiliensis, the anterior margin of the hepatic lobe is transverse, [2] but we observed that this angle can be oblique in females and may be shorter in males.
The shape of carapace and rostrum were two of the main traits used to distinguish species of Epialtus, according to Rathbun [2, p.147, figure 53] as emphasized in the figure of her study. However, the specimens of E. bituberculatus and E. brasiliensis analyzed in the present study showed a wide variation in the length/shape of the rostrum ( Figure 2) and in the carapace shape ( Figure 3). Therefore, these characters are not sufficient to distinguish E. bituberculatus from E. brasiliensis. Nonetheless, these characters allow the distinction of these two species from other species of the genus; E. dilatatus, E. hiltoni, E. minimus and E. peruvianus present deeply bilobed carapace and a bifid rostrum. [2][3][4] The ambulatory legs of E. brasiliensis are relatively stouter than those of E. bituberculatus and with the last three pairs very short. [2] However, in this study, both species have slender ambulatory legs that were subcylindrical and posteriorly decreasing in size. Epialtus brasiliensis is morphologically very similar to E. bituberculatus and differs mainly in the presence of a spine on the propodus in last three ambulatory pereopods, which we now know is a variable character (Figure 4). Finally, there are no reliable morphological characters that can be used to separate the two species or patterns that justified the genetic separation ( Figures 5 and 6), the low genetic distances between E. bituberculatus and E. brasiliensis (Tables 4and 5) or the combined phylogenetic analysis (Figure 7).
The male first pleopod has been described for Epialtus minimus and E. sulcirostris, [3] so we added it here as a comparative character within the genus; it has also been extensively used in other studies on American majoids. [3,46] The male first pleopod of E. bituberculatus and E. brasiliensis was the same and similar to that of E. sulcirostris but different from E. minimus. [

Genetic comparison
Thirteen specimens were used for molecular analysis, consisting of 9 samples of Epialtus bituberculatus, 3 individuals of E. brasiliensis and 1 specimen of E. dilatatus ( Table 2). Sequences of 16S and COI contained 410-429 and 560-640 basepairs, respectively.
The divergence between the external group T. nuttallii and genus Epialtus ranged from 20.7 to 23.6% for          16S (Table 4) and from 17.6 to 23.9% for COI (Table 5). Within Epialtus, the genetic distances ranged from 0 to 17.4% and 0 to 21.3% for 16S and COI, respectively. Interspecific distances between E. bituberculatus and E. brasiliensis ranged from 0 to 5.3% for 16S and 0 to 9.3% for COI. For 16S, the intraspecific distances ranged from 0 to 4.6% (Table 4). For COI, intraspecific values for E. bituberculatus and E. brasiliensis ranged from 0.2 to 9.3% and 0 to 0.2%, respectively ( Table 5). The phylogram constructed from the ML for 16S and COI showed a clear separation between E. dilatatus and the two species that are the focus of this study, but could not distinguish E. bituberculatus and E. brasiliensis from each other (Figures 5 and 6) as some specimens of E. brasiliensis were positioned within the group of E. bituberculatus. Overall, three distinct groups were observed: the first (upper part of phylogram) was formed by specimens from Caribbean, hereafter referred to as group A -Epialtus bituberculatus; the second clade consisted of specimens from Venezuela, hereafter group B -Epialtus n. sp.; and the third one, group C -Epialtus brasiliensis, contained specimens from Brazil ( Figures 5 and 6). All of the species with names in inverted commas refer to the original identification. All branches were well supported with bootstrap values above 50% (see Figures 5  and 6).
Divergences among clades observed in ML tree were 4.7% between group A (Caribbean) and group B (Venezuela) for both 16S and COI; 4 and 4.1% for 16S and COI between groups A and C, respectively; and 3.2% for 16S and 3.4% for COI between groups B and C.

Combined phylogenetic analysis
Our first hypothesis was that there were two Brazilian species of Epialtus, E. bituberculatus and E. brasiliensis, but morphological and molecular data did not show a clear separation between them. Thus, we add continuous and genetic data in a phylogenetic analysis that showed a correspondence with genetic structure that allowed the division of the three proposed groups (Figure 7).
For the different values of implicit weighting (k), the same one parsimonious tree was retained (Figure 7), with equal values of CI = 0.927 and CR = 0.848. We obtained different best scores, being 4.80 (k = 3), 2.37 (k = 7) and 1.72 (k = 10). Three distinct groups were observed; the first contained the specimen from Panama, and the second one consisted of specimens from Venezuela that was the sister group of the third clade with specimens from Brazil. In the Brazil clade, some specimens of E. brasiliensis were positioned within the group of E. bituberculatus; in this clade, southeast (SE) and northeast (NE) were separated. The values of branch support were different only in the clade with specimens from states of São Paulo and Rio de Janeiro (Figure 7(B) and (C)).

Discussion
The topology and small interspecific genetic distances (16S: 0 and 5.3%; COI: 0 and 9.3%) showed that E. brasiliensis falls within the group of E. bituberculatus. According to the genetic similarities among specimens in the ML tree, we infer that our initial hypothesis of two valid species was not supported and the main character (a spine in the propodus of ambulatory legs) that supposedly has been used to separate the two species and the others characters we had examined were not informative enough to support the two taxa, as they are quite variable. This genetic divergence is lower than the 5-24% interspecific genetic distance threshold in DNA barcoding for crustaceans. [47] Furthermore, it can be use to clarify cases of deep genetic divergence among individuals grouped as a single species that may indicate overlooked species, along with conventional taxonomic approaches. [47] Both E. bituberculatus and E. brasiliensis have similar habitats, coexisting on the same algae. [9,21,48, pers. obs.] The spine of ambulatory leg was conspicuous in all specimens that we called E. brasiliensis, but we observed that they also varied in size or can be reduced. This spine may be the result of environmental selection as the ambulatory legs are used to cling onto algae or even rocks [9,17,49,50, pers. obs.] and this habit perhaps required a means of attachment afforded by subchelate legs. [2] However, if the absence of the spine could be occasionally caused by accidental loss, evidences of damage or a tubercle or a bump should still remain, but none was observed. This morphological difference could be the expression of intraspecific variability [51] or we suppose it is a plastic character and the presence or absent can be at random.
Genetic distance and color can be used for species recognition. Recently, they have unmasked a cryptic hermit crab species Clibanarius symmetricus (Randall, 1840) that differs from a related species C. vittatus (Bosc, 1802) in color pattern of pereiopods with a genetic divergence ranging from 5.18 to 7.29%. [52] However, from E. bituberculatus and E. brasiliensis, we cannot infer about color because it varies according to the substrate or algae they inhabit [9,11] from dark green, brown and yellow [16] to sometimes with white spots in queliped (pers. obs.).
Analogous morphological and molecular analyses on decapods can be used to support our assertion. Pérez-Barros et al. [53] observed little genetic distance between sympatric squat lobsters Munida gregaria and M. subrugosa (0 and 0.2% for 16S and COI, respectively), which led the authors to synonymize them and suggest that they were morpho-species of a single taxon, and may be one of the plausible explanation for the Epialtus case reported here. In addition, some morpho-species that lack genetic differentiation could either be due to recent speciation or extensive hybridization. [54] For the tree-climbing mangrove crab Aratus pisonii from the same ocean (Jamaica, Hispaniola, Brazil, Costa Rica and Ecuador), no consistent morphological differences and limited genetic differentiation were found among the studied populations. [55] However, based on the morphologies of male pleopod and on genetic data (16S: 3.0-3.5% and 28S: 0.2-0.3%), populations from different oceans are distinct and required the description of a new species, A. pacificus, as the sister species of A. pisonii. [55] Unfortunately, we did not find morphological characters to justify genetic clusters observed here. If they are different species that arise recently, old traits could not be used to recognize and describe them.
The classification suggested here also found support in the larval morphology. The superfamily Majoidea presents two zoeal stages and one megalopa, and the characters are largely used to understand taxonomic and evolutionary relationships within this group. [22,[56][57][58] Similarities in the larval traits of the two species of Epialtus provide additional support to the genetic data from the branch of Brazilian specimens. The zoeae I of E. bituberculatus closely resemble the zoeae I of E. brasiliensis in several aspects: quantity and pattern of spines and setae on carapace, presence and position of tubercles in abdomen, quantity of setae maxillule and identical patterns of setae on scaphognathite, first and second maxilliped. [22,57] They share one feature (an unilobed coxal endite of the maxilla) with E. dilatatus and are different in the other characters. [22,56] The larval duration aspect is not frequently discussed by authors and some papers do not provide this information. [57] Majids are characterized by advanced development, in which the young hatch as zoeae, but in a state more developed than in other families. This is characterized by fewer stages as well as shorter duration of stages. [59] The duration of the larval stages of E. brasiliensis was 26 days, approximately as long as that of E. dilatatus and E. bituberculatus reared under similar conditions. [22,56,57] This time can be considered short; if in natural conditions larvae remain less time in the plankton stage, this potentially limits the dispersion of the group to other localities, which partially explains why the Caribbean is separated from Brazil.
These conditions and similarities in larval traits are consistent with the genetic similarity observed for our samples in group C (Figures 5 and 6) and therefore provide additional evidence of a single species of Epialtus in Brazil. Hence, we suggest that there is only one species in the Brazilian coast that is genetically distinct from the one in the Caribbean and Venezuela and we proposed to nominate it as E. brasiliensis because of its type locality is Rio de Janeiro. [23] The Caribbean and Brazilian groups were strongly supported by both the genetic distances (16S: 4.6 and 5.3%; COI: 7.9 and 8.6%) and bootstrap values (16S: 93 and 88, respectively; COI: 98 from Brazil branch). This is in agreement with the 2-3% genetic distance which is often postulated as the threshold for recognizing separate species in DNA barcoding, [60,61] including decapod crustaceans. [47] The distinction is likely because of geographical distances, intraspecific variability or perhaps the Amazon barrier. The Amazon is a potential impediment to dispersal in the coastal waters because of the freshwater and silt it deposits at the mouth, where the sea water is characterized by greatly reduced salinity and increased turbidity. [62] The coastal and algae-substrate marine caridean shrimp Hippolyte obliquimanus Dana, 1852 can be separated into Caribbean and Brazilian groups, but there are no morphological differences between them, and authors believed that the distinction could be explained by the Amazon barrier. [63] However, we did not observe morphological differences of species of Epialtus that support this separation that was justified by the combined analysis clustered with the molecular clades. Thus, identification should be based possibly by their geographical distribution. Here, we suggest the Caribbean group as E. bituberculatus because it is the older species, although its type locality has been erroneously referred to Chile [6]; all remaining occurrences were for the western Atlantic, [2,3,7,10,11,41] including the Caribbean. [12][13][14][15] Specimens from Venezuela fell along a different branch, which could be a distinct species, but at this phase, we did not find morphological diagnosing characters due the low number of specimens available for analysis. On the other hand, there are also species with only little or no morphological differences, which can only be distinguished by genetic differences; and this is of particular interest if there are cryptic species. [54] This difference may be due to the barrier formed by the rivers in Venezuela and Brazil, [63] geographical distance, variability within the population or the distinctness of the Venezuelan stock which suggests that three cryptic species evolved recently. Specimens of the foliate kelp crab Mimulus foliatus Stimpson (1860) have variable lengths of the rostral horns and the supra-orbital angles varied with age and sex, but they observed that the structure of the first pleopod of male fell within the range of species variation of the genus Pugettia Dana, 1851 and justified the change of M. foliatus to this genus with a new name Pugettia foliata (Stimpson, 1860). [46] Negri et al. [64] found that the specimens of Clibanarius vittatus (Bosc, 1802) were separated into two subgroups, one from Brazil and another from the Gulf of Mexico with high molecular divergences (COI and 16S genes) between these two subgroups, suggesting the existence of a cryptic species confounded under the name C. vittatus.
Although Epialtus affinis is synonymous of E. bituberculatus, [2] we did not resurrect this name to designate specimens from Venezuela because E. affinis was recorded only in Florida (USA) [43], the type was probably lost [2] and unfortunately, we have no specimens available from this region for molecular comparison. Furthermore, descriptions of E. longirostris and E. sulcirostris species show they are similar to E. bituberculatus and E. affinis in general characters of carapace, but differ in the shape and length of rostrum and chelipeds. [65] Both were synonyms of E. bituberculatus [42], but they were validated by the same author in 1925, and E. sulcirostris occurs only in some localities of the Gulf of California. [2,3] Thus, we propose that specimens from Venezuela are a new species and are not named so far due to the lack of morphological material and genetic analyses.
The five sympatric species of Thenus Leach, 1815 are remarkably homogenous in appearance, but with significant genetic divergence; a combination of live color patterns and morphometric ratios was used for species discrimination. [66] These authors emphasized T. australiensis and T. orientalis as the most difficult to distinguish within the genus using morphological characters, even though they are clearly distinct genetically (divergence of 2%). This low interspecific divergence clearly indicates a recent evolutionary split between these two sister taxa [66] as well as in the different genetic groups (Caribbean, Venezuela and Brazil) of Epialtus. Thus, we suggest that these groups have not had enough time to evolve morphological traits to distinguish them from each other; or the environmental selection favors the same phenotype even in different geographical areas. [67] If the phenotype is appropriate or near the optimum, there will be pressure to its maintenance and not for its change. [67,68] First, there were two species of Epialtus that occur in Brazil, E. bituberculatus and E. brasiliensis, with a main character (the spine on the propodus's ventral surface in the ambulatory pereopods) previously used to separate them from each other. However, molecular findings and combined phylogenetic analysis showed that E. brasiliensis fell within the branch of E. bituberculatus. Furthermore, based on the genetic differences and combined analysis, there were three groups: Caribbean sea, Venezuela and Brazil. Since we observed no phenotypic differentiation to validate such separation at this time, we propose a genetic and geographical new classification for the three groups: the Caribbean species is E. bituberculatus; specimens from Venezuela are a different and probably a new species; and there is a single species in Brazil, which is E. brasiliensis. Possibly, we have three recent/ cryptic species for which we have not found robust morphological characters to support. Finally, we suggest including other DNA sequences to increase the number of E. bituberculatus and E. brasiliensis specimens and populations, including species from Pacific, and new analyses with other genes to justify this separation and provide evidence for a separate species in Venezuela.

Author contribution
The authors contributed equally to the general idea, data collection, analysis, interpretation, writing and critically revising the paper for intellectual content.