INTRODUCTION
Pseudapocryptes elongatus (Cuvier, 1816) is a goby (Gobiidae sensu Gill & Mooi, Reference Gill and Mooi2012 or Gobionellidae sensu Thacker, Reference Thacker2009) in the subfamily Oxudercinae. Oxudercine gobies include highly amphibious genera known as mudskippers (genera Boleophthalmus, Periophthalmodon, Periophthalmus and Scartelaos), but the others also show varying degrees of amphibious behaviour. The taxonomy of the Oxudercinae was reviewed by Murdy (Reference Murdy1989) and the taxonomic status of P. elongatus was described by Ferraris (Reference Ferraris1995). According to Murdy (Reference Murdy1989), P. elongatus has the capacity to breathe air through its gills, by modified buccopharyngeal and opercular epithelia, and through its skin, but lacks the ability to survive for short periods out of water. However, we observed that this fish not only gulped air while in shallow water but also made short excursions on the surface of mudflats (authors’ personal observation). This fish is highly euryhaline, showing no mortality during 96-h exposure to salinities up to 50 (Bucholtz et al., Reference Bucholtz, Meilvang, Cedhagen, Christensen and Macintosh2009). Our field records showed that the salinity of adult habitats ranged from 10 to 29.
Pseudapocryptes elongatus is one of the most important aquaculture fish species in southern Vietnam and recruiting juveniles are intensively captured for use in aquaculture. Aquaculture of P. elongatus in the Mekong Delta of Vietnam has developed rapidly during the past decade due to relatively high demand and high market value (Bucholtz et al., Reference Bucholtz, Meilvang, Cedhagen, Christensen and Macintosh2009). However, increasing juvenile catch has raised concerns about a potential decline of wild populations. Therefore, it is crucial to understand the early life history of P. elongatus, such as larval duration and migration, and to develop rearing techniques that will enable sustainable fisheries and aquaculture management.
There is little information on the life history of P. elongatus, particularly of their spawning grounds, larval stage, movements and migrations. Previous studies of P. elongatus have addressed the morphology and systematics of adults (Das, Reference Das1933, Reference Das1934; Murdy, Reference Murdy1989; Yadav & Singh, Reference Yadav and Singh1989; Ojha & Haque, Reference Ojha and Haque1994), and basic biological traits such as body size, condition or early maturation status of sub-adults (Ambak et al., Reference Ambak, Dinh, Hassan and Phuong2007; Bucholtz et al., Reference Bucholtz, Meilvang, Cedhagen, Christensen and Macintosh2009). No morphological descriptions of juveniles have been published, probably due to technical problems of species identification among oxudercine gobies during larval and juvenile stages.
Thus, the aim of this study was to describe diagnostic morphological traits and obtain information on the larval and juvenile phases of P. elongatus, with the goal of forming a scientific basis for further ecological studies. We describe the morphology of juvenile P. elongatus based on genetically confirmed specimens, and report results on larval duration and migratory history of P. elongatus as inferred from daily rings and Sr:Ca ratios using otolith microstructural and chemical analyses.
MATERIALS AND METHODS
Study site and fish collection
Recruiting juvenile oxudercine gobies were purchased in June (wet season) and October (beginning of dry season) 2012 from fishermen catching them with a bag net in canal estuaries in Bac Liêu Province, Vietnam (09°15′N 105°45′E, Figure 1). The climate of the Mekong Delta region is governed by monsoons-steady winds, and the wet season begins in May and continues until late September (Hoanh et al., Reference Hoanh, Guttman, Droogers and Aerts2003). The specimens were immediately preserved either in 95% ethanol for genetic species identification and otolith analysis (see below) or in 10% formalin for description of chromatophores on the body surface. To determine the diagnostic characters that discriminate P. elongatus at the juvenile stage (Figure 2A, B for P. elongatus), we measured standard length (SL), total length (TL), head length (HL) and pre-anal length (PAL) to the nearest millimetre using digital calipers. We estimated caudal fin length (CFL) as the difference between TL and SL. The number of anal fin rays, one of the key characteristics to the species identification in Murdy (1989), was also counted for oxudercine juveniles (N = 10) after precisely identifying individuals by DNA nucleotide sequencing. Otoliths were extracted from a total of 26 randomly chosen juvenile P. elongatus from each sampling month, identified on the basis of the morphological traits determined in this study (see Results section).
Fig. 1. Locations of fish collection in the Mekong Delta, southern Vietnam.
Fig. 2. Photographs of juvenile Pseudapocryptes elongatus at an unpigmented stage (A), an early stage of pigmentation (B), and a well-pigmented stage (C). Juveniles (A) and (B) were collected with a bag net in Bac Liêu, and (C) was collected in a mangrove swamp.
To obtain the reference DNA nucleotide sequence for genetic identification of juveniles, adult P. elongatus were collected upstream of the same canal in October 2012 (N = 5). Adult individuals of another oxudercine goby, Parapocryptes serperaster (Richardson, 1846), were also obtained from a bag-net catch in Năm Căn, Cà Mau Province, Vietnam (08°45′N 105°00′E) in August 2014 (N = 8) for genetic identification of some juveniles, because there were no available DNA nucleotide sequences of this genus in the DNA database for the similar body shape as P. elongatus in the subfamily Oxudercinae of this region. In addition, one juvenile P. elongatus at a well-pigmented stage was collected in a mangrove swamp using a hand net (Figure 2C) and preserved in 10% formalin for morphological description.
Species identification
Genetic species identification was used to confirm the species of juvenile fish from bag net samples. Total genomic DNA was extracted from the muscle of adult P. elongatus (N = 5) and caudal tissue of oxudercine juveniles (N = 10) using a QIAGEN DNeasy blood & Tissue Kit (Qiagen) following the manufacturer's instructions. A partial fragment of the cytochrome oxidase I subunit gene (COI) (about 600 bp) was amplified by PCR using fish versatile primers L5956 (5′-CAC AAA GAC ATT GGC ACC CT-3′) and H6558 (5′-CCK CCW GCK GGG TCA AAG AA-3′). PCR was performed using Model 9700 (Applied Biosystems) or iCycler (Bio-Rad) thermal cyclers in a reaction volume of 10 µl containing 1.0 µl of KOD plus ver.2 Buffer (TOYOBO), 0.2 mM dNTP, 1.5 mM MgSO4, 0.3 µM each of forward and reverse primers, 0.3 U KOD plus v.2 polymerase (TOYOBO) and 1 µl of template DNA. The thermal cycling profile was as follows: initial denaturation at 94°C for 2 min, followed by 35 cycles of denaturation at 98°C for 10 s, annealing at 64°C for 30 s, extension at 68°C for 40 s, and an additional extension at 68°C for 7 min. Double stranded PCR products were purified using LaboPass™ PCR (Cosmo Genetech) or ExoSap-IT (Affymetrics), and were subsequently used for direct cycle sequencing with the BigDye™ Terminator v3.1 Ready Reaction Cycle Sequencing Kit (Applied Biosystems). DNA nucleotide sequences at the 3′-end were determined using the L5956 primer. Labelled fragments were analysed on an ABI PRISM model 3730 or 3130xl genetic analyser (Applied Biosystems). DNA nucleotide sequences were edited with 4Peaks (Griekspoor and Groothuis, mekentosj.com) or Geneious (Biomatters) software. All sequences were deposited in DDBJ/EMBL/GenBank with accession nos. LCO10470–LCO10479 (oxudercine-juvenile_01-10) and LCO10480–LCO10484 (P.elongatus-adult_01-05). In addition, the DNA nucleotide sequences were determined for adult Pa. serperaster (LCO10485–LCO10492; Pa.serperaster-adult_12-19), which were morphologically identified according to Murdy (Reference Murdy1989) and Tran et al. (Reference Tran, Shibukawa, Nguyen, Ha, Tran, Mai and Utsugi2013).
For phylogenetic analyses, available DNA nucleotide sequences of other oxudercine species were obtained from the database as follows: Oxuderces dentatus Eydoux & Souleyet, 1850 (JN631361.1), Scartelaos histophorus (Valenciennes, 1837) (EU595271.1–EU595281.1, FJ238025.1–FJ238032.1), Boleophthalmus pectinirostris (Linnaeus, 1758) (JX679027.1, JN631362.1), Boleophthalmus boddarti (Pallas, 1770) (KF874277.1), Periophthalmus modestus Cantor, 1842 (JX679050.1), Periophthalmus barbarus (Linnaeus, 1766) (AF391339.1), Periophthalmus magnuspinnatus Lee, Choi & Ryu, 1995 (JX679049.1). The sequences of two sicydiinae species, Sicyopterus japonicus Tanaka, 1909: JX628620.1 and Sicyopterus lagocephalus (Pallas, 1770): KF482068.1, were used as an out-group for the neighbour-joining tree analysis. The DNA nucleotide sequences were aligned, and the NJ tree was constructed based on a Tamura-Nei model with bootstrapping for 1000 replications using the Geneious software. Tamura-Nei distances were calculated using MEGA v.6.0.2.
Otolith analysis
Extracted otoliths from juvenile P. elongatus (June: N = 15, October: N = 11) were embedded in epoxy resin (Epofix, Struers), ground to near the otolith core using a grinding machine equipped with a diamond cup-wheel (Discoplan-TS, Struers), and then polished with a colloidal silica suspension (OP-S, Struers) on an automated polishing wheel equipped with a semi-automatic specimen mover (MD-Chem and Rotopol-35 equipped with PdM-Force20, Struers). The number of rings was counted under a light microscope to estimate daily age of the juveniles. Rings of the otoliths were assumed to be deposited daily according to the previous studies of other oxudercine gobies (Mazlan & Rohaya, Reference Mazlan and Rohaya2008).
Otoliths of the three specimens (TL: 16.7–17.1 mm) collected in June 2012 were further cleaned using distilled water and ethanol, and dried at 50°C in an oven prior to elemental analysis. Otoliths were Pt-Pd coated by a high-vacuum evaporator (E-1030, Hitachi). These otoliths were used for line analysis of Sr and Ca concentrations, which were measured in a line along the longest axis of each otolith from the core to the edge using a wavelength-dispersive X-ray electron microprobe analyser (JXA-8230, JEOL). CaSiO3 and SrTiO3 were used as assay standards. The accelerating voltage and beam current were 15 kV and 12 nA, respectively. The electron beam was focused on a point 2 µm in diameter, with measurements every 2 µm from the core to edge. All Sr:Ca ratios were calculated as molar ratios (mmol mol−1).
Data analyses
Statistical analyses (linear model framework such as t-test for means, F-test for variances, ANCOVA for slopes) were performed with the programs Prism v. 6.0c (GraphPad Software, Inc., La Jolla, CA, USA) and the lm library in R v. 3.0.2 (R Core Team, 2013). Welch's correction was used for the t-test to adjust for small sample sizes and slightly different variances of samples. Values are shown as means ± SD, wherever possible.
RESULTS
DNA nucleotide sequences of the juvenile oxudercine gobies divided into three genetically distinct groups in the NJ tree, indicating that they were composed of three species (Figure 3). Among the specimens analysed, four juveniles were identified as P. elongatus because their DNA nucleotide sequences were unambiguously nested within a clade of P. elongatus. Similarly, five juveniles were identified as Parapocryptes serperaster and one as S. histophorus on the basis of the NJ tree. The sequences in each clade were closely matched with the reference sequences for P. elongatus, Pa. serperaster and S. histophorus. The genetic distances within each species (Tamura-Nei distance = 0–0.0025) were much smaller than the intraspecific distances compared with reference sequences and the interspecific distances among the oxudercine species (0.123–0.222). Divergences between genera had quite low bootstrap values, except for Parapocryptes, and the monophyly of genus Periophthalmus was not supported by the NJ tree.
Fig. 3. Neighbour-joining tree of COI sequences of Pseudapocryptes elongatus, Parapocryptes serperaster and Scartelaos histophorus from the Mekong Delta, Vietnam for seven Oxudercinae species, and two Sicydiinae species (out-group). Bootstrap values and accession numbers for DDBJ/EMBL/GenBank are given. Asterisks depict newly registered sequences to DDBJ/EMBL/GenBank by this study.
Morphological observations of the genetically identified juvenile oxudercine gobies revealed the several morphological features among three species of this study. Juvenile P. elongatus were characterized by the following five morphological features (Figures 2 & 4): (1) a slender body form, (2) the anterior edges of both the second dorsal fin and the anal fin located at the midpoint of the body axis, (3) relatively low numbers of chromatophores over the body surface, (4) melanophores in the parietal region between the eyes and (5) a single row of melanophores at the base of the anal fin (Figure 4A, B, E, F). Juvenile specimens of Pa. serperaster were characterized by the following morphological features (Figure 4C, G): relatively high numbers of fringe-like chromatophores over the body surface, melanophores in the parietal region not only between the eyes, and similar features to P. elongatus: a slender body form, the anterior edges of both the second dorsal fin and the anal fin located at the midpoint of the body axis, a single row of melanophores at the base of the anal fin. Juvenile specimen of S. histophorus had a slender body form like other species but clearly different morphological features to other species of this study (Figure 4D, H): high numbers of chromatophores at the base of the dorsal fin, dense melanophores in the parietal region evenly, and absence of the single row of melanophores at the base of the anal fin.
Fig. 4. Photographs of genetically identified juveniles of Pseudapocryptes elongatus, Parapocryptes serperaster and Scartelaos histophorus collected by bag nets in Bac Liêu, Vietnam. Head of pigmented juveniles of P. elongatus (A, B) (SLs: 13.0 mm), Pa. serperaster (C) (SL: 12.3 mm) and S. histophorus (D) (SL: 12.5 mm). Pigment spots at the caudal part of P. elongatus (E, F), Pa. serperaster (G) and S. histophorus (H).
Morphological measurements of genetically identified oxudercine juvenile specimens are summarized in Table 1. The range of the morphological characteristics of the genetically identified juveniles overlapped among species, except SL and CFL between P. elongatus and Pa. serperaster. Standard lengths of the juveniles were quite similar among three species, although the SLs of P. elongatus (13.0 ± 0.1 mm) were slightly larger than that of Pa. serperaster (11.7 ± 0.5 mm) and S. histophorus (12.5 mm) (Table 1). Caudal fin length of P. elongatus (19.0%SL) was significantly smaller than that of Pa. serperaster (27.7%SL) (t-test with Welch's correction, P < 0.01, Table 1). In addition, juvenile P. elongatus can be distinguished from the Pa. serperaster by the absence of melanophores on ventral body surfaces, except along the bases of the anal and caudal fins. Although morphological characteristics of adult P. elongatus in this study were within the ranges or very close to those reported by Murdy (Reference Murdy1989) (Table 1), relative HL and CFL of juvenile P. elongatus were slightly different than those of adults (Table 1).
Table 1. Morphological characteristics of oxuderine gobies in the Mekong delta, Vietnam in comparison with Murdy (Reference Murdy1989).
Note: Values were mean ± SD (upper) and range (lower). P: Pseudapocryptes; Pa: Parapocryptes; S: Scartelaos.
Species were identified by the Cytochrome Oxydase I subunit gene. See text for more details.
Pseudapocryptes elongatus with the typical patterns of melanophores and relatively short caudal fins (CFL: 16.8 ± 4.0%, SL: 15.8 ± 0.84 mm, N = 26) had an average of 38.0 ± 4.1 otolith rings (range: 31–45). Variation in the number of rings was small and not significantly different between juveniles collected in June and October (F test, P = 0.97). In contrast, there was a significant difference in the number of otolith rings between juveniles collected in June (39.7 ± 3.6, N = 15) and October (35.5 ± 3.5, N = 11) (t-test, P < 0.01, Figure 5). There was also a slight but significant difference in SLs between June (14.1 ± 0.4, N = 15) and October (12.8 ± 0.6, N = 11) (t-test, P < 0.01). However, the number of rings and SLs were not correlated (significance of the slope, P = 0.12 for June and P = 0.58 for October) and these relationships were not significantly different between June and October (ANCOVA, P = 0.22). Otolith Sr:Ca ratios of the three specimens of P. elongatus were nearly constant along the otolith transects, ranging from 8.9, 9.0, 9.9 mmol mol−1 on average, and 5.2, 5.6, 6.4 mmol mol−1 for the lowest value in each individual (Figure 6).
Fig. 5. Estimated daily ages of recruiting juvenile Pseudapocryptes elongatus collected by bag nets in Bac Liêu, southern Vietnam inferred from otolith rings.
Fig. 6. Sr/Ca ratios from otolith core to the edge of otoliths of recruiting juvenile Pseudapocryptes elongatus collected by bag nets in Bac Liêu, southern Vietnam.
DISCUSSION
Species identification and morphology of juvenile P. elongatus
The juvenile oxudercine gobies that sympatrically recruit to estuaries of the Mekong Delta include at least three species. The genetic species identifications in this study are highly reliable because the sequences of the oxudercine juveniles nested within a single species, satisfying the strict tree-based assignment criteria described in Wilson et al. (Reference Wilson, Rougerie, Schonfeld, Janzen, Hallwachs, Hajibabaei, Kitching, Haxaire and Hebert2011). Although the number of specimens used for species identification of the bag-net catch was small (N = 10), the data indicated that the catch included at least three species of oxudercine gobies: P. elongatus, Pa. serperaster and S. histophorus. Fishermen suffocate the catch by leaving it in non-aerated water in order to harvest only P. elongatus, which can survive hypoxia exposure through aerial oxygen uptake. Currently, no data are available on the air-breathing capacities of P. elongatus or Pa. serperaster, but the latter is presumably less tolerant to hypoxia, and therefore typically dies during this procedure. Dinh et al. (Reference Dinh, Qin, Dittmann and Tran2014) reported that Pa. serperaster rarely emerged from its burrow in a mudflat. Thus it is likely that these fish include several species having different tolerance to hypoxia.
The NJ tree based on the COI gene sequences of oxudercine gobies in this study was different, especially in the location of Periophthalmus and Oxuderces, from the cladogram of Murdy (Reference Murdy1989), which was based on morphological traits. The NJ tree in this study was generated only by COI gene sequences, and this was not sufficient genetic information to determine the phylogenetic relationships among genera at the divergence between two clades – one including Pseudapocryptes, Scartelaos and Boleophthalmus and the other including Periophthalmus and Oxuderces. The bootstrap values for the two clades were quite low, but the two groups differ in body form. The former clade (Pseudapocryptes, Scartelaos and Boleophthalmus) consists of fishes with a slender body form while the latter (Periophthalmus and Oxuderces) have stouter body shapes. This may imply that the difference in body shape among oxudercine species has evolved monophyletically within the lineage. However, this study did not examine a sufficient number of DNA gene regions, and comprehensive taxon sampling would be required to determine reliable phylogenetic relationships among genera in the subfamily Oxudercinae.
Among the morphological characters observed in this study, patterns of the chromatophores in the parietal region and body surface and presence/absence of melanophores at the base of the anal fin would be important to distinguish among P. elongatus, Pa. serperaster and S. histophorus in the Mekong Delta. In addition to these characteristics, the CFL (%SL) of P. elongatus was significantly smaller than that of Pa. serperaster in juvenile specimens among various morphological measurements in this study. However, this study showed differences in CFL ratio between juvenile and adult P. elongatus, probably due to allometric changes during development. The CFL ratio would be used for species identification in combination with pigmentation patterns at least in the juvenile stage. To establish a morphological species identification of oxudercine species, the CFL value needs to be considered in conjunction with development stages or life history phases.
Larval duration and migration
This study provides one of the first clues about the larval phase of P. elongatus such as larval duration and migration. Annual ages of mudskippers were investigated previously for Boleophthalmus dussumieri Valenciennes, 1837 (described as B. dentatus) in Gujarat, India (Soni & George, Reference Soni and George1986) and growth zones in mudskipper otoliths were validated as annual increments in Periophthalmus argentilineatus Valenciennes, 1837 collected from Tanzania (Kruitwagen et al., Reference Kruitwagen, Hecht, Pratap and Wendelaar Bonga2005). The number of rings on the sagittal otolith of recruiting juvenile P. elongatus showed little variance between June and October. The larval duration of P. elongatus was estimated as 38 days on average, which was similar to the larval duration of other oxudercine species in captivity: 39 days for Boleophthalmus pectinirostris (Zhang et al., Reference Zhang, Hong, Dai, Zhang, Cai and Huang1989) and 45 days for Periophthalmus modestus (Kobayashi et al., Reference Kobayashi, Dotsu and Miura1972). Certainly, diurnality of otolith ring deposition in P. elongatus needs to be validated by further studies. Bucholtz et al. (Reference Bucholtz, Meilvang, Cedhagen, Christensen and Macintosh2009) estimated the age of 1.5 cm post-larval P. elongatus as ~5 weeks in the same region, which agrees well with our data, although they gave no account of the method used.
Otolith Sr:Ca ratios of P. elongatus indicated that all juveniles spent their entire larval period in a marine environment. The Sr:Ca values of P. elongatus otoliths ranged from 8.9 to 9.9 mmol mol−1, which matches typical values of seawater (8.5–8.6: de Villiers et al., Reference de Villiers, Shen and Nelson1994; de Villiers, Reference de Villiers1999). Taken together, our data on the Sr:Ca ratio and the number of rings in P. elongatus otoliths imply that the fish reproduce in a saline environment not far from the site of capture. In comparison of estimated age at recruitment between seasons, there was a slight but significant difference of juvenile daily age. This may be due to (1) seasonal differences at the main spawning grounds, which are at different distances from the sampling site, (2) current velocities along the Mekong Delta coast, (3) physiological and/or behavioural mechanisms that determine larval duration or (4) some combination of these factors.
Life history and spawning grounds of P. elongatus
Fully mature specimens of P. elongatus have not been reported in previous studies. Furthermore, we interviewed numerous local fishermen and officers of the Fisheries Department and none reported having seen P. elongatus crowding for reproduction, or even gravid females. This seems to indicate that P. elongatus spawn in areas not easily observable by humans. Ambak et al. (Reference Ambak, Dinh, Hassan and Phuong2007) reported that the gonado-somatic index (GSI) of P. elongatus collected in the Mekong Delta coast peaked between June and August, but the peak value for females was only ~2.5%. Bucholtz et al. (Reference Bucholtz, Meilvang, Cedhagen, Christensen and Macintosh2009) also reported a very low GSI of females up to 0.29% for P. elongatus collected in February to May. These values are much lower than those reported for some other oxudercine species (8% in Scartelaos gigas Chu & Wu, 1963; Kim et al., Reference Kim, Baek, Kim, Chang and Kim2011; 10% in Boleophthalmus pectinirostoris, Hong et al., Reference Hong, Chen, Zhang and Wang2007; 13% in Periophthalmus minutus Eggert, 1935; Takeda et al., Reference Takeda, Hayashi, Toba, Soyano and Ishimatsu2011), but was similar to some other species (3.5% in Periophthalmodon schlosseri, Mazlan & Rohaya, Reference Mazlan and Rohaya2008).
Although the air-breathing capacity of adult P. elongatus strongly indicates their likely occurrence in hypoxic, stagnant, inland brackish waters, the otolith Sr:Ca analyses revealed that larval hatching undoubtedly takes place in the marine environment. Inferring from the reproductive strategy of other oxudercine gobies (Periophthalmus modestus, Ishimatsu et al., Reference Ishimatsu, Yoshida, Itoki, Takeda, Lee and Graham2007; Periophthalmus minutus, Takeda et al., Reference Takeda, Hayashi, Toba, Soyano and Ishimatsu2011; Periophthalmodon schlosseri, Ishimatsu et al., Reference Ishimatsu, Takeda, Tsuhako, Gonzales and Khoo2009; and Boleophthalmus pectinirostris, Hong et al., Reference Hong, Chen, Zhang and Wang2007; Toba & Ishimatsu, Reference Toba and Ishimatsu2014), it may be hypothesized that spawning of P. elongatus occurs in burrows excavated in soft, fine sediment in coastal mudflats.
Knowing that the larval period of P. elongatus is ~30–45 days in the saline environment, both oceanographic conditions affecting larval transportation by currents and active swimming of larvae relative to moving water are critically important components of their dispersal to reveal the reproductive biology and recruitment scenario of P. elongatus in the Mekong Delta. At the start of the pelagic larval period, fish larvae are generally poorly developed and typically drift passively with the currents, although previous studies showed that some fish larvae have better locomotory abilities, including the ability to regulate their vertical distribution (Leis, Reference Leis2006). Thus, currents must be a major factor in dispersal of larvae of this stage. Determining the velocity and direction of coastal currents around the Mekong Delta over the course of the year could potentially help understand the transport from the spawning grounds. Several papers have reported surface circulation patterns of the South China Sea, including the east coast of Vietnam (Shaw & Chao, Reference Shaw and Chao1994; Hu et al., Reference Hu, Kawamura, Hong and Qi2000; Morimoto et al., Reference Morimoto, Yoshimoto and Yanagi2000). These studies showed that the current off the Mekong Delta flows from south-west to north-east in August, while the direction of the current is reversed, flowing from north-east to south-west in December. These studies indicate that early-stage larvae of P. elongatus are probably transported from the south-west of the main juvenile-catch locations (Bac Liêu Province and its environs) in the peak season, possibly in Cà Mau Province. At the latter end of the pelagic larval period, fish larvae are better adapted to regulating their position in the water column and in locating appropriate habitats for settlement (Leis, Reference Leis2006). Therefore, investigation is needed of the behavioural traits exhibited by the various developmental stages of P. elongatus larvae in addition to the determination of coastal hydrodynamics in the Mekong Delta to understand their life history and to ensure the conservation of this unique fish species.
ACKNOWLEDGEMENTS
We are grateful to Mr F. Mori, Dr M. Yagi and local fishermen for assistance in field sampling. We also sincerely thank officers of the Fisheries Division and local governments of southern Vietnam for coordination in field sampling. Anonymous reviewers are appreciated for their valuable comments on the manuscript.
FINANCIAL SUPPORT
K.Y. was financially supported by a Research Fellowship for Young Scientists (No. 245842) from the Japan Society for Promotion of Science. This study was supported by grant-in-aid for scientific research abroad (B) (Nos. 23405033, 15H05252) from the Japan Society for the Promotion of Science to K.Y., M.W., T.T. and A.I.
