Genetic Structure of the Endemic Fiddler Crab Uca (Xeruca) formosensis on the West Coast of Taiwan

© 2023 Academia Sinica, Taiwan Open Access Genetic Structure of the Endemic Fiddler Crab Uca (Xeruca) formosensis on the West Coast of Taiwan Kun-Chin Hung1, Ching-Yu Liou2, Chih-Chung Wen3, and Hui-Chen Lin1,* 1Department of Life Science, Tunghai University, Taichung 407, Taiwan. *Correspondence: E-mail: [email protected] (Lin). E-mail: [email protected] (Hung) 2Endemic Species Research Institute, Council of Agriculture of the Executive Yuan, Jiji, Nantou 552005, Taiwan. E-mail: [email protected].tw (Liou) 3Department of Safety, Health and Environmental Engineering, Hungkuang University, Taichung 433, Taiwan. E-mail: [email protected] (Wen) Received 20 January 2022 / Accepted 17 February 2023 / Published 1 June 2023 Communicated by His-Te Shih Xeruca formosensis is the only endemic species of fiddler crab on the west coast of Taiwan. However, its natural habitats and populations have been compromised by excessive anthropogenic activities and improper land use over the past four decades. In light of these changes, we sought to evaluate the genetic diversity and gene flow of the species by examining the genetic variation of X. formosensis at different sampling locations. To this end, we performed molecular analyses of three endonuclease-amplified fragment length polymorphisms (TE-AFLP) and the cytochrome oxidase subunit I (COI) marker from leg muscle samples. We found that the genetic variation within sampling locations was higher than that among sampling locations, and the expected heterozygosity of genetic diversity (Hj) was 0.152 for TE-AFLP data. Meanwhile, the COI marker showed high haplotype diversity (h = 0.976 ± 0.008) and a low genetic differentiation level (FST = 0.021) in X. formosensis populations. Importantly, the genetic connectivity of X. formosensis may be influenced by larval-stage crabs drifting between coastal and marine habitats. As such, crab gene flow is promoted among populations by larval exchange via nearshore currents. Although X. formosensis has high gene flow, the species could undergo an extinction crisis if the population sizes continue to decline, as with most endangered species. In order to maintain the natural habitats and population size of X. formosensis, long-term monitoring and investigation will be necessary. Key words: Xeruca formosensis, Genetic variation, Larvae drifting, Genetic diversity. BACKGROUND In Taiwan, there are 15 known species of fiddler crabs, with Xeruca formosensis being the only endemic species on the west coast. The first record of this species was made at Lugang in Changhua County in 1918 (Shih et al. 1999), and it was announced as a new species (originally Uca formosensis) in 1921 (Shih et al. 1999). In 2016, the genus was changed to Xeruca based on morphological and molecular evidence, including mitochondrial 16S rDNA, cytochrome oxidase subunit I (COI), and nuclear 28S rDNA (Shih 2015; Shih et al. 2016), which has been further supported by the mitogenome analysis (MY Liu and Shih 2022). Previous studies indicated that X. formosensis inhabits tidal areas where there are wide flats with very clayey-muddy substrate and no mangroves on the supratidal zone (Shih et al. 1999; Liou 2012). In the past, large populations of this species existed at several locations, including Zengwen estuary, Dadu estuary, Xiangshan area, and Shengang area (Fig. 1). In particular, the Shengang area of Changhua County was once known as the “hometown of X. formosensis” (Shih 1997; Liou 2012). However, recent anthropogenic activities and improper land use Citation: Hung KC, Liou CY, Wen CC, Lin HC. 2023. Genetic structure of the endemic fiddler crab Uca (Xeruca) formosensis on the west coast of Taiwan. Zool Stud 62:24. doi:10.6620/ZS.2023.62-24. Zoological Studies 62:24 (2023) doi:10.6620/ZS.2023.62-24 1 © 2023 Academia Sinica, Taiwan have led to marked declines in the natural habitats and populations of X. formosensis. Human endeavors such as landfills, fish farms and artificial mangrove planting have nearly eliminated X. formosensis from many areas (Shih et al. 1999; Chen 2008; Liao et al. 2008; Liou 2012). For example, inappropriate dike construction along the coastline of Shengang area has displaced X. formosensis and allowed both Austruca lactea and Tubuca arcuata to become two of the most dominant species in the wetland (Liou 2010). In addition, expansion of aquaculture facilities near the Zengwen estuary in the Tainan City greatly damaged the habitats of X. formosensis such that only three stable populations remain in this location (Liou 2012). Meanwhile, construction of a water recycling center and anthropogenic planting of a mangrove forest in the Xiangshan area has led to a gradual decrease of the X. formosensis population in the past decade (Wu and Chang 2008). According to the International Union for Conservation of Nature and Natural Resources (IUCN), regulations for protecting focal species should make use of gene flow, genetic variation and genetic differentiation as key indexes (Frankel 1974; Araujo Fig. 1. Sampling sites for Xeruca formosensis in the west coast of Taiwan. Red circles represent the sampled populations, the red star represents the proposed population in Bazhang River at Jiayi County (this population was not yet sampled), and the black triangle represents the Dadu estuary. Taichung Changhua Yunlin Nantou Tainan Chiayi Miaoli Kaohsiung Gaomei Dacheng Mailiao Qigu Xiangshan Xianxi Shengang (Zengwen estuary) Dadu estuary 80 604020100 23°N 120°E 121°E 24°N Kilometers N page 2 of 15Zoological Studies 62:24 (2023) © 2023 Academia Sinica, Taiwan and Ramos 2000; Geist 2010). Moreover, excessive anthropogenic development and inappropriate land use not only impact the natural habitats of organisms but also lead to decreased genetic diversity and increased risk of population extinction (Dixo et al. 2009; Haag et al. 2010). As such, research on the genetic structure and diversity among geographically distinct populations can be highly useful for identifying management units for vulnerable species (Baeza et al. 2019; Mendiola and Ravago-Gotanco 2021). DNA fingerprinting and COI markers are robust established methods that can be used to efficiently produce large amounts of accurate genetic data for genetic diversity analyses and further conservation purposes. Many DNA fingerprinting methods have been implemented, including amplified fragment length polymorphisms (AFLP), and three endonucleaseamplified fragment length polymorphisms (TE-AFLP). These fingerprinting methods have been widely applied in different areas, including genetic structure research on crustaceans (Fratini and Vannini 2002; Herborg et al. 2007; Wieman et al. 2014; Seeb et al. 2002). Similarly, the COI marker of crustacea is useful for detecting and evaluating intraspecific diversity (Tang et al. 2003; Aoki et al. 2008). Since knowledge of genetic diversity and structure is important for species conservation, detailed studies to gather genetic structural information on the endemic fiddler crab X. formosensis are warranted. However, up to now, beside the molecular studies mentioned above (Shih 2015; Shih et al. 2016; MY Liu and Shih 2022), most studies on this species have focused on population size, behavior, habitat selection, adult and larval morphology (Takahasi 1935; Shih et al. 1999; Shih et al. 2005; Liao et al. 2008; Chen and Lee 2008; YC Zhang and Shih 2022). Only two studies have reported allozyme analyses, and these were conducted in 1984 and 1999 (Zhang 1984; Shih 1999). Thus, it is important to identify the current state of genetic variation at high-resolution using modern methods in order to better understand the genetic structure of X. formosensis among populations. The habitats of X. formosensis are mostly found within the wetlands of the eastern Taiwan Strait, which is located between the South China Sea and the East China Sea on the western coast of Taiwan. The Taiwan Strait contains several major currents (Hsiao et al. 2011), including the China Coastal Current, the extension of the South China Sea Warm Current, and the intrusion from the Kuroshio Current (Hsieh et al. 2004; Lan et al. 2004; Guan and Fang 2006; Dur et al. 2007). These ocean currents are subject to strong seasonal influence by monsoons (Wyrtki 1961; Liang et al. 2003). In the summer, a northward intrusion of the current from the South China Sea and branch of the Kuroshio arises as the southwest monsoon appears from May to November. In winter months, the intrusion from the Kuroshio is blocked by the northeasterly monsoon, which lasts from approximately December to February depending on climatological conditions (Hu et al. 2010). Based on these environmental conditions and its life history, the breeding season of X. formosensis occurs in the summer, from April to September. In light of recent habitat destruction and population declines, we sought to evaluate the species genetic diversity and structure for X. formosensis. To this end, we investigated the genetic variation across seven sampling locations that were geographically divided into three groups (north, central and south). We analyzed the expected heterozygosity of genetic diversity and population structure (including genetic clustering among locations) based on TE-AFLP DNA fingerprinting (Pritchard et al. 2000; Wieman et al. 2014). Furthermore, we determined the nucleotide diversity (π) and haplotype diversity (h) from mitochondrial DNA sequencing data. Multiple statistical methods, such as analysis of molecular variance (AMOVA), mismatch distribution, principal coordinate analysis (PCoA) and historical effective population estimation, were used to explore the endemic genetic status of X. formosensis. Overall, our findings contribute to understanding the current status of the endemic species, X. formosensis, in Taiwan. MATERIALS AND METHODS Sample collection In total, 280 samples were collected from seven different locations in mudflat intertidal zones on the west coast of Taiwan with natural distributions of X. formosensis (Table 1). Adult X. formosensis individuals were collected from the following seven different coastal locations: Xiangshan area of Hsinchu City, Gaomei area in Taichung City, Shengang area, Xianxi area, and Dacheng in Changhua County, Mailiao area in Yunlin County, and Qigu area in the Tainan City; hereafter, these areas are respectively referred to as Xiangshan, Gaomei, Shengang, Xianxi, Dacheng, Mailiao, and Qigu. The X. formosensis sampling locations were further grouped according to geographic area: north (Xiangshan); central, (Gaomei, Shengang, Xianxi, Dacheng, and Mailiao); and south (Qigu) (Table 1). Total genomic DNA was extracted from adult individuals using the MasterPureTM DNA Purification Kit (Epicentre, United States) and stored in Tris-EDTA at -20℃ for later use. page 3 of 15Zoological Studies 62:24 (2023) © 2023 Academia Sinica, Taiwan Molecular methods DNA fingerprinting TE-AFLP genotypes data Based on AFLP, the TE-AFLP method consists of three parts, as described by van der Wurff et al. (2000). First, total genomic DNA (~100 ng) was digested by XbaI, BamHI, and RsaI, and two sets of adapters (XbaI adapter sequences were 5'-ACGTTGTGGCGGCGTCGGACTAGA-3' and 3'-CCGCCGCAGCTCTGATCT-5'; BamHI adapter sequences were 5'-ACGAAGTCCCGCGCCAGCAA GATCC-3' and 3'-GGGCGCGGTCGTTCTAG-5') were selectively ligated using an enzyme-specific sequence to restriction fragments in a single reaction volume. Each reaction contained 2 μl 10X Ligase buffer, 2.0 μl 500 mM NaCl, 7.5 U ligase (NEB, USA), 1.25 U XbaI (Promega, USA), 4.0 μl BamHI adapter (1 pmol/μl and 4.0 μl XbaI adapter (1 pmol/μl). Second, a labelled XbaI primer (5'-GGCGTCGAGACTAGACC-3' *, * represents the fluorescein-labeled site) and unlabeled BamHI-CC primer (5'-GTTTCGCGCCAGCAAGAT CC-3') were used for selective amplification. Each reaction included 0.5 μl digestion/ligation DNA sample, 2.5 μl 5× PCR buffer, 0.25 μl BamHI-C primer (10pmol/ μl), 0.25 μl XbaI-CC primer (10 pmol/μl), 0.125 μl Taq polymerase (Go-Taq, Promega, USA), 0.25 μl 10 mM dNTPs and 11 μl distilled water (van der Wurff et al. 2000). Third, polymerase chain reaction (PCR) was performed on a thermal cycler (Eppendorf Master cycler gradient 5331, Eppendorf AG, Germany) using an initial denaturation step at 95°C for 3 min, followed by 10 cycles (denaturation at 95°C for 30 s, annealing at 70°C for 30 s, elongation at 72°C for 1 min), 40 cycles (denaturation at 95°C for 30 s, annealing at 60°C for 30 s, elongation at 72°C for 1 min), and final elongation at 72°C for 20 min. After PCR, the samples were stored at -20℃ for later experimentation. PCR products were used for SNP genotyping (130 loci for each sample). Gene-Mapper software was used to transfer SNP data to a 1 and 0 matrix format by detecting the presence of each band in the gel (Mission biotech Ltd, Taipei, Taiwan). Mitochondrial DNA COI sequencing A 648 base-pair region of the COI gene was amplified using the following universal primers: HCO2198: 5'-TAAACTTCAGGGTGACCAAAAAAT CA-3' and LCO1490: 5'-GGTCAACAAATCATAAA GATATTGG-3' (Folmer et al. 1994; Shih 2015). PCR amplification conditions consisted of an initial denaturation step at 95°C for 5 min, followed by 35 cycles of 30 s at 95°C, 30 s for primer annealing at 51°C, and extension at 72°C for 15 min; after the reaction, samples were held at 4°C (Shih 2015). PCR products were separated by electrophoresis on a 2% agarose gel to check quality. Molecular data analyses TE-AFLP genotypes Genetic clusters were assessed based on TEAFLP genotype data of X. formosensis by Bayesian assignment test using STRUCTURE 2.3.4 (Pritchard et al. 2000). The program operation (number of subpopulations) was set as follows: 3,000 lengths burnin period; 30,000 MCMC reps after burn-in; admixture model of ancestry info; and range of genetic demes (K) from 1 to 7, with three repetitions per genetic cluster (K) for allele frequencies correlated with the population of the frequency model. The best value of K was detected by calculating ΔK (Evanno et al. 2005). The genotype data were also used to estimate the molecular variance among/within sampling location (Excoffier et al. 1992), Table 1. Sampling information on the fiddler crab Xeruca formosensis, including sampling location, group assignment (according to geographic distance), county, geographic coordinates, number of DNA samples for mtDNA (COI) and TE-AFLP markers analyses Sampling location Group County Geographic coordinates Number of DNA samples mtDNA (COI)TE-AFLP analyses Xiangshan North group Hsinchu 24°48'02.9"N 120°54'57.5"E 29 40 Gaomei Central group Taichung 24°18'39.7"N 120°32'59.7"E 21 40 Shengang Central group North Changhua 24°09'54.3"N 120°27'54.6"E 21 40 Xianxi Central group North Changhua 24°07'31.0"N 120°25'59.2"E 11 40 Dacheng Central group South Changhua 23°51'27.7"N 120°15'43.3"E 20 40 Mailiao Central group Yunlin 23°49'44.0"N 120°14'10.2"E 22 40 Qigu South group Tainan 23°03'27.9"N 120°03'36.0"E 28 40 page 4 of 15Zoological Studies 62:24 (2023) © 2023 Academia Sinica, Taiwan and PCoA was performed using GeneAlex 6.501 (Peakall and Smouse 2006) based on an Euclidean distance matrix. Nei’s heterozygosity (Hj) (Nei 1987) and the percentage of polymorphic loci (%P) for each location were calculated using AFLP-SURV version 1.0 (Lynch and Milligan 1994; Vekemans et al. 2002). A consensus tree of X. formosensis was constructed by the Phylip 3.6 software (Felsenstein 1993) based on the unweighted pair group method with arithmetic mean (UPGMA). Mitochondrial DNA COI sequencing data The analysis of molecular variance (AMOVA) was performed on data from 152 individuals using Arlequin 3.5.2.2. Molecular diversity indices, including the number of haplotypes (Nh), haplotype diversity (h), nucleotide diversity (π), pairwise fixation index (FST) estimates, and the average number of nucleotide differences (k), were estimated using DnaSP, version 6.0. Two neutral hypothesis tests, the Tajima D test (Tajima 1983) and Fu’s Fs test (Fu 1997) were performed to detect the expansion of the whole population of X. formosensis. The mismatch distribution (Lavery et al. 1996) was determined using Arlequin with a spatial expansion model, and the sum of square deviation (SSD) of observed frequencies was greater than the expected frequencies (P > 0.05), indicating that this population conforms to an expansion model (Harpending 1994; Deli et al. 2016). In addition, the unit of mutational time (τ) and absolute time since expansion (T) at the same time were assessed according to mismatch distribution, following analyses described in previous studies (Rogers and Harpending 1992; Harpending 1994; Tokuyama et al. 2020). The effective population size was estimated based on formulae from a previous study (Watterson 1975), and the mutation rate (μ = 2.1 × 10-8) was adopted from a previous study on the terrestrial crab genus Sesarma spp. (Wares and Cunningham 2007; Wieman et al. 2014). Furthermore, a haplotype network was constructed using PopART 1.7 by the minimum spanning network (MSN) method (Leigh and Bryant 2015; Hardianto et al. 2022). The nucleotide sequences of X. formosensis from this study were deposited into GenBank (National Center for Biotechnology Information, NCBI) under accession numbers ON692540-ON692691. RESULTS TE-AFLP data Total TE-AFLP data were obtained from 280 individuals and used for AMOVA. The results showed 93% genetic variation within location and 98% genetic variation within groups (Table 2). The Bayesian assignment test revealed two genetic clusters for X. formosensis (ΔK = 2) (Fig. 2). Comparing the results of the three groups of X. formosensis in this study (north, central and south groups) indicated that both genetic diversity and polymorphic loci in the central group were lower than those of the north and south groups (Table 3); those for Shengang (Hj = 0.033; %P = 16.92%) were the lowest. The highest genetic diversity and polymorphic loci were seen in the south group, especially those in Mailiao. For the whole population, the percentage of polymorphic loci was 86.48%, and the expected heterozygosity of genetic diversity index was 0.152 (Table 3). PCoA analysis did not show obvious clustering of the seven X. formosensis sampling locations (Fig. 3). Moreover, the UPGMA tree only indicated that the distance of the north group was relatively farther than others; the analysis did not clearly distinguish the southern and central groups too (Fig. 4). Mitochondrial DNA COI sequencing data A 648 base-pair coding region of the COI gene was sequenced from 152 adult individuals collected from seven locations of X. formosensis. According to COI sequencing results, we found nucleotide frequencies of A = 27.2%, T = 36.5%, C = 20.1%, and G = 16.1%. The results of AMOVA showed 96–98% genetic variation within sampling locations (Table 2), and mean FST value obtained from the AMOVA was 0.021 (Table 4), with a significant P-value (P < 0.05). Pairwise comparisons showed negative FST values between Xiangshan and Mailiao and between Shengang and Mailiao (Table 5); negative values indicate essential genetic homogeneity. The pairwise FST values comparing between the north, central and south groups were positive (Table 6). Notably, when comparing between the north and central groups, and between the north and south groups, the pairwise FST values were significant (P < 0.05). However, comparing between the south and central groups yielded a non-significant pairwise FST value (P value = 0.23). According to our molecular diversity analysis of the COI marker, haplotype diversity (h) was high in all locations, with Xiangshan having the highest value (h = 0.993 ± 0.014). In addition, the nucleotide diversity values (π) for Xiangshan, Xianxi, Qigu and Mailiao populations were higher than those for Gaomei, Shengang and Dacheng. As shown in table 3, 118 haplotypes were detected among all samples. The overall haplotype diversity index was 0.976 ± 0.008 (range, 0.924–0.993), while the nucleotide diversity page 5 of 15Zoological Studies 62:24 (2023) © 2023 Academia Sinica, Taiwan index was 0.010 ± 0.004 (range, 0.007–0.016), and the average number of nucleotide differences (k) was 5.95 (range, 4.65–7.37). Further, the nucleotide diversity indexes for Gaomei and Shengang were lower than those for the rest of the examined locations. Haplotype network analysis revealed no obvious patterns (Fig. 5). The results of Tajima’s D test (P < 0.05) and Fu’s Fs test (P < 0.02) showed negative values for our X. formosensis samples (Table 4). Mismatch distribution plots exhibited a single peak for X. formosensis, which indicates a past expansion (Fig. 6). Furthermore, the units of mutational rate prior to the present (τ) was 4.684, and the absolute time (T years) since the expansion of the X. formosensis was 108,861 years ago (Table 4). We also determined that the population size post-expansion (θ1 = 28.44) was higher than before the expansion (θ0 = 2.02), and the mode value of theta in the MCMC parameters was 0.09823. The effective population size (Ne) was 4.68 × 106 for X. formosensis (Table 4). Furthermore, according to the mutation rate we adopted, the X. formosensis population is predicted to have expanded in the past 100,000 years. Table 2. Analysis of molecular variance for seven sampling locations, three groups (north, central and south) and two subpopulations (north and others) of Xeruca formosensis based on TE-AFLP and mitochondrial DNA cytochrome oxidase subunit I (COI) data DNA data type Source d.f. Percentage of variation Sum of squares Variance components TE-AFLP Among 7 sampling locations 6 7.00 236.22 0.73 Within sampling location 273 93.00 2771.10 10.15 Total 279 100 3007.32 10.88 Among 3 groups 2 2.86 71.17 0.31 Within group 277 97.14 2936.15 10.60 Total 279 100 3007.32 10.91 Among 2 populations 1 1.51 22.06 0.17 Within populations 278 98.49 2985.26 10.74 Total 279 100 3007.32 10.90 mtDNA COI Among 7 sampling locations 6 2.07 25.55 0.06 Within sampling location 146 97.93 426.94 2.92 Total 152 100 452.48 2.99 Among 3 groups 2 2.06 1.45 0.08 Within group 149 97.94 588.10 3.95 Total 151 100 602.79 4.03 Among 2 populations 1 3.39 10.45 0.14 Within populations 150 96.61 592.34 3.95 Total 151 100 602.79 4.09 Fig. 2. Genetic analysis as inferred by the assignment test and using a Bayesian approach in STRUCTURE 2.3.4. A) Estimation of populations using a ΔK value of 2 based on TE-AFLP markers. B) The result of Bayesian assignment test indicted two genetic clusters for seven locations of X. formosensis. page 6 of 15Zoological Studies 62:24 (2023) © 2023 Academia Sinica, Taiwan Table 3. Nuclear and mitochondrial COI DNA diversity indexes for the fiddler crab Xeruca formosensis Genetic marker TE-AFLP mtDNA (COI) Sampling location and groups Hj%P Nhh ± SD π ± SD k Xiangshan 0.197 82.31 % 25 0.993 ± 0.014 0.016 ± 0.008 7.12 Gaomei 0.085 43.08 % 11 0.924 ± 0.033 0.007 ± 0.004 4.65 Shengang 0.033 16.92 % 18 0.976 ± 0.023 0.009 ± 0.005 5.23 Xianxi 0.106 46.92 % 10 0.946 ± 0.066 0.012 ± 0.007 6.98 Dacheng 0.221 80.00 % 18 0.984 ± 0.024 0.010 ± 0.005 5.36 Mailiao 0.221 96.92 % 14 0.939 ± 0.035 0.010 ± 0.005 5.91 Qigu 0.206 88.46 % 22 0.980 ± 0.014 0.010 ± 0.006 6.22 North group 0.197 82.31 % 25 0.993 ± 0.012 0.016 ± 0.008 7.12 Central group 0.147 59.23 % 71 0.972 ± 0.011 0.012 ± 0.006 5.41 South group 0.206 88.46 % 22 0.982 ± 0.015 0.012 ± 0.006 5.75 North populations 0.197 82.31 % 25 0.993 ± 0.012 0.016 ± 0.008 7.12 Other populations 0.160 67.69 % 85 0.980 ± 0.007 0.011 ± 0.006 7.37 Whole population 0.152 86.48 % 118 0.976 ± 0.008 0.010 ± 0.004 5.95 Hj: expected heterozygosity of genetic diversity. %P: percentage of polymorphic loci. Nh: number of haplotypes. h: haplotype diversity. π: nucleotide diversity. k: average number of nucleotide differences. Fig. 3. Results of principal coordinate analysis (PCoA) of TE-AFLP markers. Coord. 2 (32.41%) Coord. 1 (27.05%) Gaomei Shengang Dacheng Mailiao Qigu Xiangshan Xianxi Fig. 4. X. formosensis consensus tree constructed by the UPGMA method based on TE-AFLP markers. page 7 of 15Zoological Studies 62:24 (2023) © 2023 Academia Sinica, Taiwan DISCUSSION Genetic variation and structure Our analysis of TE-AFLP data and COI marker from X. formosensis showed that the genetic variation within the sampling location was higher than it was among sampling locations. These results indicate that X. formosensis possesses characteristics of genetic structure that are typical of decapods (Tracey et al. 1975; Brian et al. 2006). Importantly, X. formosensis currently has greater than 90% genetic variation within location and group, and the mean FST value (0.021, with significant P-value) suggests the possibility of an X. formosensis metapopulation (Wang et al. 2019; Tokuyama et al. 2020). In addition, this result indicates that X. formosensis still has high gene flow among sampling locations, even though it faces survival pressures and habitat destruction. Notably, the pairwise FST among the three groups showed that the northern group was significantly different from the other two (P < 0.05), but there was no significant difference between the southern and central groups. In line with this result, the UPGMA tree only indicated a grouping of the northern group and others (Fig 4). Thus, we concluded that X. formosensis should be divided into two populations (north and others) rather than three (north, central and south). This conclusion led us to further conduct the AMOVA analysis and calculate diversity indexes for the two populations (north and others); the results are shown in tables 2 and 3. The genetic diversity value (Hj = 0.152) we observed for X. formosensis is higher than that of another fiddler crab species, Uca maracoani (0.062), reported in Wieman et al. 2014. However, it is lower than the reported values for other brachyurans, such as swimming crab (Portunus trituberculatus) and Cancer setosus (Gomez-Uchida et al. 2003; Liu et al. Table 4. Neutrality test and mismatch distribution values for Xeruca formosensis NeTDFus τTθ0θ1SSD FST Whole population 4,680,000 -2.109 ** -3.332 ** 4.684 108,861.37 2.02 28.44 0.0013 0.021 ** TD: Tajima D test. Fus: Fu’s Fs test. τ: time since expansion in units of mutational time. T: absolute time since expansion in years. θ0: estimated value of population size before expansion. θ1: estimated value of post-expansion population size. SSD: sum of squared deviation value. FST: fixation index. ** mean significant value (P < 0.05). Table 5. The pairwise FST value (below the diagonal) and pairwise FST P-value (above the diagonal) for the COI marker among 7 sampling location of Xeruca formosensis Xiangshan Gaomei Shengang Xianxi Dacheng Mailiao Qigu Xiangshan - 0.0371 ± 0.006** 0.2168 ± 0.010 0.0156 ± 0.004** 0.0537 ± 0.006 0.9043 ± 0.013 0.0156 ± 0.004** Gaomei 0.0362 -0.1602 ± 0.012 0.0664 ± 0.007 0.0898 ± 0.010 0.1641 ± 0.014 0.1426 ± 0.011 Shengang 0.0090 0.0165 - 0.1113 ± 0.011 0.3291 ± 0.018 0.5889 ± 0.014 0.2940 ± 0.019 Xianxi 0.0547 0.0344 0.0263 -0.2832 ± 0.015 0.0566 ± 0.008 0.1660 ± 0.012 Dacheng 0.0322 0.0257 0.0053 0.0063 -0.1953 ± 0.012 0.2539 ± 0.014 Mailiao -0.0291 0.0250 -0.0156 0.0674 0.0177 - 0.1055 ± 0.010 Qigu 0.0435 0.0133 0.0055 0.0130 0.0053 0.0341 - ** mean significant value (P < 0.05). Table 6. The pairwise FST value (below the diagonal) and pairwise FST P-value (above the diagonal) for the COI marker among 3 groups (north group, central group, and south group) of Xeruca formosensis North group Central group South group North group - 0.01367 ± 0.0037 ** 0.01758 ± 0.0039 ** Central group 0.028 - 0.22852 ± 0.0131 South group 0.043 0.0035 - ** mean significant value (P < 0.05), and for details of groups, see Table 1. page 8 of 15Zoological Studies 62:24 (2023) © 2023 Academia Sinica, Taiwan 2012). According to mitochondrial COI DNA diversity analyses, X. formosensis has higher levels of haplotype diversity and nucleotide diversity (Table 3) than fiddler crabs and other crabs from mangrove, coastal and marine habitats (Azuma et al. 2007; Aoki et al. 2008; Darling et al. 2008; Wieman et al. 2014). Nevertheless, the diversity of X. formosensis is lower than that of its sympatric species, A. lactea (Tokuyama et al. 2020). In addition, the highest value of genetic diversity (Hj) was found in the Mailiao and Dacheng sites (Table 5). The Mailiao sampling site is a large, abandoned fish farm near the Sixth Naphtha Cracking Complex, south of the Zhuoshui River. According to our personal observations, the distribution area was approximately Fig. 5. TCS network showing the relationships among the recorded haplotypes of X. formosensis based on the analysis of a 648 base pair region of the mitochondrial gene COI using PopART. page 9 of 15Zoological Studies 62:24 (2023)