Phylogeographic Structure within the Fiddler Crabs Leptuca thayeri and Uca maracoani (Brachyura, Ocypodidae) along the Tropical West Atlantic

© 2022 Academia Sinica, Taiwan Open Access Special issue: Systematics and Biogeography of Fiddler Crabs Phylogeographic Structure within the Fiddler Crabs Leptuca thayeri and Uca maracoani (Brachyura, Ocypodidae) along the Tropical West Atlantic Murilo Zanetti Marochi1,*, Marcelo Marucci Pereira Tangerina2, Renata de Oliveira Rodrigues1, Claudia Laurenzano3, Wagner Vilegas1, Tânia M. Costa1, and Christoph D. Schubart3 1UNESP - São Paulo State University, Bioscience Institute, Coastal Campus, São Vicente – SP, Pça. Infante Dom Henrique Square, 11330-900, Brazil. *Correspondence: E-mail: [email protected] (Marochi). E-mail: [email protected] (de Oliveira Rodrigues); [email protected] (Vilegas); [email protected] (Costa) 2UNESP – São Paulo State University, Institute of Chemistry, Campus of Araraquara, Araraquara – Prof. Francisco Degni Street, 55, 14800-060, Brazil. E-mail: [email protected] (Tangerina) 3University of Regensburg, Zoology and Evolutionary Biology, Regensburg, Germany; Universitätsstr. 31, 93053. Fax: +49 941 9433304. E-mail: [email protected] (Laurenzano); [email protected] (Schubart) Received 30 December 2021 / Accepted 17 April 2022 / Published 16 November 2022 Communicated by Hsi-Te Shih Special issue: Systematics and Biogeography of Fiddler Crabs (articles 64–71). Editors: Hsi-Te Shih and Benny K. K. Chan Most fiddler crabs have an extended planktonic larval phase, potentially maintaining gene flow among widely separated populations, in the absence of marine barriers. Such marine barriers could be long coastal stretches without suitable habitat, freshwater plumes caused by large river mouths, or strong currents. Typically, fiddler crabs inhabit mangrove habitats, and as mangroves tend to have a patchy distribution, it is important to gather information on the connectivity between neighboring mangroves and recognize local endemisms. To detect potential genetic differentiation among mangrove-dwelling populations of Leptuca thayeri and Uca maracoani along several thousand kilometers of a tropical coastline, mtDNA sequences of different populations from Brazil and two Caribbean islands were analyzed and compared. As shown in previous studies with fiddler crabs, Brazilian populations are genetically indiscernible, and our data suggest the absence of long-standing gene flow barriers in the two studied species along the Brazilian coast. This includes both sides of the postulated biogeographic barriers corresponding to the split of the Central South Equatorial Current and to the Amazon River freshwater plume. In contrast, conspecific individuals from the Greater Antilles carried different haplotypes, suggesting a biogeographical barrier between Brazil and the Caribbean, apparently having limited gene flow between both regions for extended time periods. Key words: Cox1 mtDNA, Restricted gene flow, Population genetics, Brazil, Caribbean. BACKGROUND Genetic connectivity can be broadly defined as the degree to which populations are considered open or closed by gene flow and dispersal patterns (Hellberg et al. 2002). In most marine invertebrates, planktonic larval development is part of their reproductive strategy, resulting in high dispersal potential (Muñiz-Salazar et al. 2005; Anger 2006). The longer the time span of the planktonic larval phase, the farther the propagules can possibly be transported passively by ocean currents (up to several hundred kilometers) and the higher the Citation: Marochi MZ, Tangerina MMP, de Oliveira Rodrigues R, Laurenzano C, Vilegas W, Costa TM, Schubart CD. 2022. Phylogeographic structure within the fiddler crabs Leptuca thayeri and Uca maracoani (Brachyura, Ocypodidae) along the tropical West Atlantic. Zool Stud 61:67. doi:10.6620/ZS.2022.61-67. Zoological Studies 61:67 (2022) doi:10.6620/ZS.2022.61-67 1 © 2022 Academia Sinica, Taiwan potential gene flow can be expected among widely separated populations (Silva et al. 2010). This, in turn, can prevent evolution of genetic structuring and local adaptations (Kelly and Palumbi 2010; Bray et al. 2017). In the case of restricted gene flow, the consequent genetic differentiation is not immediately expressed and thus not recognized. Lack of phenotypic divergence may indicate that a species is in morphological stasis or differences among populations are very subtle and require careful diagnosis (Chaklader et al. 2016; Marochi et al. 2017). Barriers to dispersal in the marine environment may not result from landscape features as in terrestrial ecosystems, but they may occur due to physical and chemical features, such as salinity, temperature, currents patterns, and eddies (Cowen 2002; Banks et al. 2007; Weersing and Toonen 2009; Chapman et al. 2011). For example, the Amazon River discharge in northeastern Brazil and the Orinoco River discharge in Venezuela send great plumes of freshwater into the adjacent ocean, changing the salinity in nearby coastal areas, and thereby potentially hindering the dispersal of organisms that do not tolerate different salinity gradients (Lessios et al. 2001 2003; Hu et al. 2004). The Central South Equatorial Current (CSEC) splits into the North Brazil and South Brazil currents, carrying organisms in opposite directions (Marochi et al. 2017). Also, upwelling and downwelling processes may act as a barrier to dispersal due to temperature gradients and may result in local larval retention (Lessios et al. 2003; Taylor and Hellberg 2003). Thus, despite extensive larval dispersal potential, populations along the coast may present marked genetic differences from each other (disjunctive variation), or show gradual differences (clinal variation), which are not always evident at first, since the same genotype can produce different phenotypes following environmental traits (Scheltema 1975). Fiddler crabs are distributed in tropical and subtropical coastal areas all over the world, mostly in estuarine areas with mangrove and salt marsh vegetation (Crane 1975; Shih et al. 2016). The postlarval nonmigratory nature and low dispersal capacity of fiddler crabs after metamorphosis to benthic stages would result in potential genetic structure between widely separated populations. However, their planktonic larval stages achieve a greater level of genetic exchange among widespread littoral populations. Silva et al. (2010) found no regional genetic or morphological differentiation within a large study area (~3,300 km) and convoluted ocean current circulation patterns in Austruca occidentalis (described by Naderloo et al. 2016 as Uca occidentalis) along the East African coast. Likewise, a lack or limited genetic structuring was also observed among populations of South American brachyuran crab species along the Atlantic coast (Oliveira-Neto et al. 2007; Laurenzano et al. 2012; Hampton et al. 2014; Wieman et al. 2014; Marochi et al. 2017; Buranelli and Mantelatto 2019). On the other hand, higher genetic differentiation has been reported for Caribbean and Central American populations (Laurenzano et al. 2013; Buranelli and Mantelatto 2019; Peres et al. 2020; Thurman et al. 2021). Leptuca thayeri (Rathbun, 1900) and Uca maracoani (Latreille, 1802) are common and widely distributed neotropical fiddler crab (Crane 1975). Leptuca thayeri inhabits burrows in muddy and sandy mangrove areas (commonly among roots) in mesohaline and euryhaline habitats in the mid intertidal zone and is distributed along the Western Atlantic coast from the USA (Florida), Gulf of Mexico, Cuba, Jamaica, Puerto Rico, Guatemala, Venezuela, Trinidad and Tobago, to Brazil (Pará, Maranhão, Piauí, Ceará, Rio Grande do Norte, Paraíba, Pernambuco, Alagoas, Sergipe, Bahia, Espírito Santo, Rio de Janeiro, São Paulo, Paraná, Santa Catarina) (Crane 1975; Thurman et al. 2013; Farias et al. 2014; Mantelatto et al. 2020). The larval duration of L. thayeri from hatching to metamorphosis lasts approximately one month, and successful larval development only occurs in salinities higher than 20 ppt (Anger et al. 1989). Uca maracoani has been recorded along the Atlantic coast of South America from Trinidad and Tobago, Venezuela, and Guyanas to the south of Brazil (Melo 1996; Wieman et al. 2014), with so far dubious records from the Antilles. Uca maracoani inhabits mesohaline and euryhaline habitats (salinity varying from 14 to 32 ppt) on mud flats along bays, estuaries, mangrove areas or riverbanks close to their mouths in the low intertidal zone (Masunari 2006; Wieman et al. 2014). Even if the first zoea and megalopa stages are described morphologically (Negreiros-Fransozo et al. 2009), there is a lack of information in literature regarding the duration of the larval phase and their salinity preferences. However, a similar developmental time, from hatching to metamorphosis, as in other fiddler crab species is expected (≅ 20 to 30 days) (Wieman et al. 2014). The different habitat preferences and optimum salinities during larval development can influence the level of genetic connectivity among populations, and higher genetic structure can be expected in more oligohaline species, showing larval retention mechanisms in estuaries, favored by active larval behavior (e.g., vertical migration) (Cronin and Forward 1986; Kelly and Palumbi 2010; López-Duarte et al. 2011; Bray et al. 2017). To test the hypothesis of high genetic homogeneity among Leptuca thayeri and Uca maracoani populations, we analyzed the page 2 of 15 Zoological Studies 61:67 (2022) © 2022 Academia Sinica, Taiwan genetic differentiation, mainly based on the comparison of Brazilian populations with specimens from the Caribbean (Jamaica and Hispaniola). MATERIALS AND METHODS Sampling of Leptuca thayeri and Uca maracoani Leptuca thayeri specimens were collected from three Brazilian populations in the states of Pará (Marudá), Bahia (Acuípe), and São Paulo (Bertioga). In addition, 11 specimens from Jamaica (Priory, St. Ann) and the Dominican Republic (Luperón). Uca maracoani specimens were collected from one Brazilian population in Pará (Marudá), and one population in the Dominican Republic (Sánchez mangroves) (Table 1) (Fig. 1). All individuals sampled were deposited into the Zoologische Staatssammlung München (Table 1). DNA extraction, amplification, and sequencing A total of 45 individuals of L. thayeri and 20 individuals of U. maracoani from five and two sample sites, respectively, were used for genetic analyses (Tables 2 and 3). Genomic DNA was extracted from muscle tissue of pereiopods using the Puregene (Gentra Systems) buffer system method. For both species, DNA amplification from the mitochondrial gene cytochrome oxidase, subunit I (Cox1) was carried out by polymerase chain reaction (PCR) (40 cycles: 45 sec 94°C/1 min 48°C/75 sec 72°C denaturing/annealing/elongation temperatures). For L. thayeri and U. maracoani an 857 base pair region was amplified using the primers COL1b 5'-CCW GCT GGD GGW GGD GAY CC-3' and COH16 5'-CAT YWT TCT GCC ATT TTA GA3', and a shorter region of 640 base pairs using the Fig. 1. Sample sites of Leptuca thayeri (black dots) and Uca maracoani (grey dots) populations from throughout the West Atlantic: Luperón & Sánchez from the Dominican Republic; Priory (St. Ann) from Jamaica; Marudá (Pará), Acuípe (Bahia), Bertioga (São Paulo) from Brazil. Grey squares correspond to localities sampled by Wieman et al. (2014), with sequences of U. maracoani in GenBank. N page 3 of 15Zoological Studies 61:67 (2022) © 2022 Academia Sinica, Taiwan primers COL1b 5'-CCW GCT GGD GGW GGD GAY CC-3' and COH1b 5'-TGT ATA RGC RTC TGG RTA RTC-3' (Schubart 2009). This was necessary, because not all of the older specimens allowed to amplify the longer fragment, and in consequence two different datasets were evaluated, one maximizing the number of individuals (but with a short alignment), the other one maximizing the length of the alignment and exploring new variable positions (but with fewer representatives). Amplification results were checked by running 4 μl of PCR product on 1.5% TBE agarose gel electrophoresis. PCR products were outsourced for sequencing with the Table 1. Sampling locations, coordinates, numbers of individuals, museum collection numbers from the Zoologische Staatssammlung München (ZSM) and newly submitted GenBank accession numbers of Leptuca thayeri and Uca maracoani Species Location Coordinates Number of individuals Collection no. GenBank no. Leptuca thayeri Marudá, PA 0°37'0.23"S 47°37'94.6''W 10 ZSMA20210114 OM938770-OM938779 Leptuca thayeri Acuípe, BA 15°04'9.92"S 38°59'95.8"W 15 ZSMA20210115 OM938755-OM938769 Leptuca thayeri Bertioga, SP 23°50"13.7"S 46°09'16.5"W 10 ZSMA20210116 OM938746-OM938754 Leptuca thayeri Priory, St. Ann, JA 18°26'43"N 77°12'56"W 9 ZSMA20210117 OM938780-OM938789 Leptuca thayeri Luperón, DR 19°53'48''N 70°57'.33''W 1 ZSMA20210118 ON065543 Uca maracoani Marudá, PA 0°37'0.23"S 47°37'94.6''W 10 ZSMA20210119 OM938790-OM938799 Uca maracoani Sánchez mangroves, DR 19°13'30"N 69°37'10"W 11 ZSMA20210120 OM938800-OM938810 Table 2. Genetic diversity indices and neutrality tests for each analysed population of Leptuca thayeri and Uca maracoani based on regions of different lengths of the Cox1 gene Species Sequences length Location Population N h S Hd π Tajima’s D test Fu’s Fs test Leptuca thayeri 609 base pairs Priory, St. Ann Jamaica (JA) 9 69 0.89 0.003 -1.59* -1.96* Marudá Pará (PA) 10 7 9 0.933 0.005 0.45 -1.63 Acuípe Bahia (BA) 15 11 16 0.933 0.006 -0.7 -4.23* Bertioga São Paulo (SP) 10 7 11 0.933 0.005 -0.43 -1.62 Greater Antilles 10 69 0.844 0.003 -1.69 -1.9 Pará + Bahia + São Paulo 35 19 23 0.923 0.006 -1.13 -9.28* all sequences 45 25 40 0.947 0.011 -0.72 -7.66* Uca maracoani 606 base pairs Sánchez Dominican Republic (DR) 11 10 0 0 0 - Rio Amapá Amapá (AP) 11 7 7 0.818 0.002 -1.64* -4.05* Marudá Pará (PA) 10 7 60.911 0.002 -0.97 -3.98** São José do Ribamar Maranhão (MA) 12 10 10 0.955 0.002 -2.04** -9.06** Fortaleza Ceará (CE) 13 11 12 0.974 0.003 -1.75* -8.97** Itapissuma Pernambuco (PE) 8 7 8 0.917 0.002 -1.79* -4.2** Salvador Bahia (BA) 9 9 12 1 0.004 -1.87** -7.45** Santa Cruz Espírito Santo (ES) 10 7 7 0.911 0.002 -1.26 -3.75** Parati Rio de Janiero (RJ) 10 3 3 0.378 0.0009 -1.56* -0.45 Guaratuba Paraná (PR) 9 69 0.893 0.003 -1.72* -2.21* AP + PA + MA + CE + PE + BA + ES + RJ + PR 91 48 45 0.89 0.002 -2.54** -5.68* Dominican Republic 11 10 0 0 0 - all sequences 103 49 47 0.902 0.0046 -2.16** -4.94* Hd: haplotype diversity, h: number of haplotypes, N: number of individuals, S: number of polymorphic sites, π: nucleotide diversity. *p < 0.05, **p < 0.01. page 4 of 15Zoological Studies 61:67 (2022) © 2022 Academia Sinica, Taiwan primer COL1b to Macrogen Europe, Inc. (Amsterdam, the Netherlands). Obtained sequences were edited in Chromas Lite 3.01 (Technelysium Pty Ltd 2005) and manually aligned with BioEdit 5.0 (Hall 1999). Primer sequences and adjacent regions were omitted, resulting in an alignment of 826 or 609 base pairs for L. thayeri and 825 or 606 base pairs for U. maracoani. The software Artemis (Rutherford et al. 2000) was used to rule out the presence of stop codons, which could indicate the presence of pseudogenes. Sequences were submitted to GenBank (Table 1). Genetic data analyses The number of haplotypes, and haplotype (h) and nucleotide (π) diversities were calculated in DnaSP v5 (Librado and Rozas 2009). To assess levels of genetic differentiation among populations, pairwise ΦST values (Weir and Cockerham 1984; Loh et al. 2001) were calculated with Arlequin ver. 3.11 (Excoffier et al. 2005). A statistical parsimony network was constructed with PopArt (Polzin and Daneshmand 2003; Leigh and Bryant 2015). The variance between tested groups was assessed by Analyses of Molecular Variance (AMOVA), using Arlequin ver. 3.11. For L. thayeri two populations from Caribbean islands (Jamaica and Dominican Republic) and three from Brazil (Pará, Bahia and São Paulo) were compared with the shorter DNA alignment (609 bp). The longer alignment (826 bp) could only be used to compare the northernmost Brazilian population (Pará) with the other northeast and southeast Brazilian populations (Bahia and São Paulo). For U. maracoani, one population from the Caribbean (Hispaniola) and one new one from Brazil (Pará) were compared to a relatively large dataset from GenBank (KF666951KF666995), including Brazilian populations from Amapá, Maranhão, Ceará, Pernambuco, Bahia, Espírito Santo, Rio de Janeiro and Paraná (Wieman et al. 2014). To examine the population history and to evaluate whether the populations follow the neutrality model at the sampling sites, Tajima’s D, Fu’s Fs, and mismatch distribution analyses (Tajima 1989; Fu 1997; Schneider and Excoffier 1999) were carried out using the Arlequin ver. 3.11 software. To test for significant restriction of gene flow, we used a non-parametric permutation procedure (Excoffier et al. 1992), incorporating 10,000 permutations. RESULTS Network reconstructions and statistical analyses were based on two alignments of different lengths in order to include sequences from previous studies and GenBank records. This data treatment has the advantage of allowing the inclusion of sequences from more specimens due to the shorter alignment, and thus increases statistical power for distinguishing apparently isolated populations (e.g., Caribbean vs. Brazil), even if the alignment is shorter and may include fewer variable sites. In contrast, fewer sequences in favor of a longer alignment are preferable in the case of no or weak differentiation (e.g., within Brazil), as networks based on a longer alignment allow visualization of more mutations that may have been established between these closely related populations. For the shorter alignment (609 bp) of L. thayeri, 25 different haplotypes were recorded among 45 sequences of the five investigated West Atlantic populations, as compared to 26 different haplotypes in 34 sequences of the three Brazilian populations with the longer alignment (826 bp). The corresponding genetic distances among these haplotypes are depicted in the haplotype networks shown in figure 2. The one based on the shorter alignment clearly demonstrates that Brazilian Table 3. Genetic diversity indices and neutrality tests for each analysed population of Leptuca thayeri and Uca maracoani based on the Cox1 gene Species Sequences length Location Population N h S Hd π Tajima’s D test Fu’s Fs test Leptuca thayeri 826 base pairs Marudá Pará (PA) 10 9 12 0.978 0.005 0.23 -3.97 Acuípe Bahia (BA) 15 13 22 0.971 0.006 -1.03 -6.53* Bertioga São Paulo (SP) 9 8 12 0.972 0.004 -0.62 -3.51 Bahia + São Paulo 24 18 27 0.957 0.005 -1.38 -10.23* Pará 10 9 12 0.978 0.005 0.23 -3.97 L. thayeri all sequences 34 26 32 0.973 0.005 -1.46 -20.31* Uca maracoani 825 base pairs Sánchez Dominican Republic (DR) 11 21 0.327 0.0004 -0.1 0.35 Marudá Pará (PA) 10 9 8 0.978 0.002 -1.22 -7.35** U. maracoani all sequences 21 11 18 0.819 0.007 0.82 -0.22 page 5 of 15Zoological Studies 61:67 (2022) © 2022 Academia Sinica, Taiwan and Caribbean populations cluster separately, with ten mutational steps (1.48%) in between them (Fig. 2A). From the 45 haplotypes of this shorter alignment, 18 represent single individuals, while the others are shared haplotypes, diverging from one possibly ancestral haplotype H5. This most common haplotype is shared by eight specimens (four individuals from Bahia and two each from São Paulo and Pará). H13 is also found in individuals from all three Brazilian populations (two individuals each from São Paulo and Pará and one from Bahia). In the Caribbean, haplotype H20 is shared by four individuals (three from Jamaica and one from the Dominican Republic). The longer alignment reinforces the close clustering of the haplotypes from Brazilian populations but suggests a slight distinction of the Pará population (Fig. 2B). 22 out of 34 haplotypes of the 826 bp alignment represent single individuals, while the others are shared haplotypes, diverging from one possibly ancestral haplotype H2. This most common haplotype is shared by five specimens (three individuals from Bahia and two from São Paulo). H7 is found in all Brazilian populations (one individual per population). All these longer haplotypes found in Brazilian populations are closely related, except for a slight deviation of the Pará population not being represented in the most common haplotype (H2) but sharing a unique haplotype instead. This homogeneity stands in sharp contrast to the large and consistent genetic distances to sequences from the Caribbean Province (islands of Hispaniola and Jamaica). Haplotype diversities within the shorter alignment varied from 0.89 to 0.933 and the nucleotide diversity from 0.003 to 0.006, while the longer alignment varied from 0.957 to 0.978 and the nucleotide diversity from 0.004 to 0.006 (Tables 2 and 3). Statistical analyses based on the shorter alignment did not include the single individual from the Dominican Republic, due to insufficient sample size. Mean pairwise ΦST values among populations were relatively high (0.64) (Table 4). The majority of pairwise differences between populations were not significant, except for the comparison of Jamaican and the combined Brazilian populations of Pará, Bahia and São Paulo (Table 4). None of the pairwise differences between Brazilian populations based on the longer alignment were significant (Table 4). The AMOVA results showed no significant differences between the two tested groups of the shorter alignment (P = 0.25), although the highest amount of variation (76.07%) was in this category. The variation among populations within groups was also not significant (P = 0.44). Significant differences were found within populations (P ≦ 0.001) and were responsible for 23.87% of the total variation (Table 4). The AMOVA results based on the longer alignment also failed to show significant differences between the two tested groups (P = 0.3), accounting for 5.6% of the total variation, while -2.59% of the variation was among populations within groups (P = 0.66), and 97.08% of variation within populations (P = 0.33). The L. thayeri haplotypes from Caribbean and Brazilian populations are significantly different (Fig. 2 and Table 4). In the short alignment (606 bp) of U. maracoani, Fig. 2. Haplotype network of Leptuca thayeri constructed with Popart, with a connection limit of 95%, derived from Cox1 mtDNA with 609 basepairs (A) and 826 basepairs (B). Black dots represent missing haplotypes (one-step edges); numbered ‘H’ indicate most common haplotypes. page 6 of 15 Zoological Studies 61:67 (2022) © 2022 Academia Sinica, Taiwan 49 different haplotypes were found among 103 sequences of the ten populations. The haplotype network in figure 3 depicts genetic distances among them, showing that all Brazilian populations cluster in one single clade, five mutational steps away (0.66%) from the Caribbean haplotypes (Fig. 3A). From the 49 haplotypes, 42 are represented by single individuals, while the others are shared, diverging from the possibly ancestral haplotype H7. This most common haplotype is shared by twenty-nine specimens (five individuals from Amapá, eight from Rio de Janeiro, three from Maranhão, three from Pernambuco, three from Espírito Santo, three from Paraná, two from Ceará, one from Pará and one from Bahia). H6 is found in three individuals from Pará, two from Ceará, two from Espírito Santo, one from Maranhão, one from Amapá and one from Pernambuco. H5 is found in four individuals (one each from Amapá, Pará, Maranhão and Rio de Janeiro). H31 is found in two individuals from Pará and one from Bahia. H3 is found in one individual from Amapá and one from Maranhão, and H10 was present in one individual from Maranhão and another from Bahia. In the Caribbean, only a single haplotype (H46) was obtained, suggesting a genetic bottleneck (Fig. 3A). Overall, all haplotypes found in Brazilian populations are closely related. The corresponding haplotype diversities varied from 0.8 to 1 and the nucleotide diversities from 0.002 to 0.004, with the exception of the population of Parati (state of Rio de Janeiro), with values of 0.378 and 0.0009 (Table 2). In the 826 bp alignment, 8 out of 11 haplotypes represent single individuals from the Pará population (Fig. 3B). No haplotypes were shared between the populations of Pará and the Dominican Republic, which are separated by nine mutations (0.97%). This finding Table 4. Pairwise differences between sampled populations of Leptuca thayeri Sequence length São Paulo (10) Bahia (15) Pará (10) Jamaica (9) 609 base pairs São Paulo -------------- 0.67 0.4 < 0.001* Bahia -0.0341 -------------- 0.2 < 0.001* Pará -0.0089 0.02529 -------------- < 0.001* Jamaica 0.7964 0.76801 0.78427 -------------- Source of variation d.f. Sum of squares Variance components Variation (%) p-value Among groups 1 80.109 5.4717 76.07 0.2551 Among populations within groups 2 3.543 0.0047 0.07 0.4428 Within populations 40 68.667 1.716 23.87 < 0.001* Total 44 153.318 7.19323 Fixation indices FSC: 0.00278 (among populations within groups) FST: 0.7613 (within populations) FCT: 0.7606 (among groups) São Paulo (9) Bahia (15) Pará (10) 826 base pairs São Paulo -------------- 0.67 0.2 Bahia -0.029 -------------- 0.2 Pará 0.0289 0.03 -------------- Source of variation d.f. Sum of squares Variance components Variation (%) p-value Among groups 1 3.466 0.1319 5.6 0.3 Among populations within groups 1 1.575 -0.0633 -2.69 0.66 Within populations 31 70.9 2.2871 97.08 0.33 Total Fixation indices FSC: -0.0284 (among populations within groups) FST: 0.0291 (within populations) FCT: 0.056 (among groups) ΦST -values below dashed lines; corresponding P-values above diagonal; d.f., degrees of freedom; FSC: variance among populations within groups; FST: variance among populations; FCT: variance among groups defined a priori. Numerals in parentheses denote the sample size of each population. * P < 0.05. page 7 of 15Zoological Studies 61:67 (2022) © 2022 Academia Sinica, Taiwan Table 5. Pairwise differences between sampled populations of Uca maracoani Sequence length Dominican Republic (11) Amapá (11) Pará (10) Maranhão (12) Ceará (13) Pernambuco (8) Bahia (9) Espírito Santo (10) Rio de Janeiro (10) Paraná (9) 606 base pairs Dominican Republic ----------- < 0.001* < 0.001* < 0.001* < 0.001* < 0.001* < 0.001* < 0.001* < 0.001* < 0.001* Amapá 0.9037 ----------- 0.89 0.37 0.51 0.7 0.06 0.83 0.17 0.38 Pará 0.8841 -0.029 ----------- 0.91 0.94 0.88 0.13 0.99 0.91 0.23 Maranhão 8355 0.0041 -0.02 ---------- 0.47 0.37 0.84 0.75 0.47 0.58 Ceará 0.8958 0.0011 -0.03 -0.0037 ---------- 0.99 0.25 0.51 0.67 0.3 Pernambuco 0.8474 -0.015 -0.02 0.0034 -0.023 ---------- 0.07 0.011* 0.94 0.28 Bahia 0.882 0.0487 0.029 -0.0265 0.0126 0.0416 ---------- 0.29 0.9 0.64 Espírito Santo 0.9599 -0.022 -0.04 -0.0199 -0.013 0.0076 0.044 ---------- 0.06 0.53 Rio de Janeiro 0.8799 0.0185 -0.03 -0.0123 -0.012 -0.0362 0.044 0.0163 ---------- 0.08 Paraná 0.8942 -6E-04 0.013 0.0134 0.0134 0.0233 -0.011 0.0079 0.0488 ---------- Source of variation d.f. Sum of squares Variance components Variation (%) p-value Among groups 1 64.29 3.22 80.27 0.09 Among populations within groups 8 6.99 0.008 0.22 0.49 Within populations 93 72.94 0.78 19.51 < 0.001* Total 102 144.23 4.02 Fixation indices FSC: 0.0111 (among populations within groups) FST: 0.8049 (within populations) FCT: 0.8027 (among groups) Dominican Republic (11) Pará (10) 825 base pairs Dominican Republic ----------- < 0.001* Pará 0.8938 ----------- ΦST -values below dashed lines; corresponding p-values above diagonal; d.f., degrees of freedom; FSC: variance among populations within groups; FST: variance among populations; FCT: variance among groups define a priori. Numerals in parentheses denote the sample size of each population. * P < 0.05. Fig. 3. Haplotype network of Uca maracoani constructed with Popart, with a connection limit of 95%, derived from Cox1 mtDNA with 606 base pairs (A) and 825 base pairs (B). Black dots represent missing haplotypes (one-step edges), numbered 'H' indicate most common haplotypes. page 8 of 15Zoological Studies 61:67 (2022) © 2022 Academia Sinica, Taiwan reinforces the large and consistent genetic distances from South American and Caribbean populations. Haplotype diversities varied from 0.32 to 0.97, and the nucleotide diversities from 0.002 to 0.0004 (Table 3). All pairwise differences between populations of U. maracoani from the Dominican Republic and Brazil were significant in both alignment lengths, while only Pernambuco and Espírito Santo populations were significantly different, for the shorter alignment, when exclusively comparing Brazilian populations (Table 5). Even if the AMOVA results do not confirm these significant differences between Brazilian and the Caribbean populations (the two tested groups) (P = 0.09), this comparison revealed the highest amount of variation (80.27%). The variation among populations within groups was also not significant (P = 0.49) and responsible for 0.22% of variance. Significant differences were found within populations (P < 0.001) and were responsible for 19.51% of the total variation (Table 5). The demographic history of Brazilian and Caribbean populations of L. thayeri and U. maracoani was reconstructed using mismatch distributions and neutrality tests. Leptuca thayeri populations showed a bimodal distribution of pairwise differences (Fig. 4A), corresponding to the intra-regional differences on one hand and the inter-regional on the other, while the mismatch distribution of Brazilian populations showed a unimodal distribution (Fig. 4B). Uca maracoani populations from this study and from Wieman et al. (2014) also showed a bimodal distribution of pairwise differences (Fig. 5), corresponding to the intra-regional differences on one hand and the inter-regional on the other. The majority values of the neutrality tests, when combining all populations of Caribbean and Brazil, were negative for both species (Tables 2 and 3). Fig. 4. Mismatch distribution for two Caribbean and three Brazilian populations of Leptuca thayeri with 609 basepairs (A) and three Brazilian populations with 826 basepairs (B). Fig. 5. Mismatch distribution for one Caribbean and nine Brazilian populations of Uca maracoani with 606 basepairs (A) and one Brazilian and one Caribbean population with 825 basepairs (B). page 9 of 15Zoological Studies 61:67 (2022)