Aquaculture as a tool to support goby-fry fishery? Current knowledge on biology and ecology of the red-tailed goby Sicyopterus lagocephalus

Sicydiinae species are amphidromous gobies, adults spawn in freshwater, whereas free embryos undergo a pelagic open sea phase. Post-larvae or juveniles are caught for human consumption when entering in freshwater after their pelagic larvae life in sea-water. Such goby-fry fisheries are existing since centuries and widespread in tropical areasovertheworld.Thereareuncertaintiesrelatedtocaughtvolumesandtrendsbut, overall, go-fry fisheries are declining and their sustainability is questionable. Aquacul-ture is a potential tool in conservation and management of wild goby stocks. Among Sicydiinae species, the red-tailed goby Sicyopterus lagocephalus is the most spread and used as a model species in numerous works involving amphidromous fish. The aim of the study was to evaluate the potential in aquaculture of the red-tailed goby through the analysis of the literature available on this species. We found that this species has some assets to be a potential candidate for aquaculture such as a short production cycle and a high fecundity as well as potential high market values. Nevertheless, given the small size at hatching and the long pelagic larval life, larval rearing is likely to be a challenging rearing phase but appears to be feasible based on past experiences with other goby species. Throughout the paper, we provided recommendations for future research in red-tailed goby aquaculture.

exploited genus is Sicyopterus, especially in the Philippines and Reunion Island (Bell, 1999;Delacroix, 1987;Manacop, 1953).These amphidromous species share a very distinctive life cycle.The Sicydiinae species have amphidromous life histories with spawning in freshwater.Their larvae drift downstream to the sea, where they have a marine larval life before migrating back into the rivers to grow and reproduce (Manacop, 1953;McDowall, 1988).Harvest occurs when large schools of postlarvae ascent into the mouths of rivers using nets, diverse gears and traps (Bell, 1999;Delacroix, 1987).Still today, the impact of fisheries on Sicydiinae populations remains largely unknown.
Post-larvae of Sicydiinae play an important socio-economic role by providing food for local populations.Therefore, goby post-larvae are prized products in some regions and can reach very high prices on local markets (e.g.> $80 kg -1 in Reunion Island [Thomas, 2017]; i.e. 4% of the average local month salary per kg harvested).Nevertheless, there is only limited information regarding the volumes harvested.Several studies agree that many of the goby-fry fisheries have declined considerably (Bell, 1999;Castellanos-Galindo et al., 2011;Manacop, 1953) and recommend management and conservation measures such as banning fisheries during specific periods of the year.Nevertheless, high discrepancies in harvested volumes and trends have been highlighted.For example, in FishStatJ, the FAO's dataset on global aquaculture production, fishery data on Gobiidae reported for Reunion Island, one of the most studied goby-fry fisheries were 1-2 tons per year from 2002 to 2008 (FAO, 2021).Over the same period, Thomas (2017), based on surveys and literature review, estimated average harvested volumes from 30 to 45 tons annually.Such mismatch production statistics between different sources have been reported for other species (Garibaldi, 2012;Metian et al., 2014) which can distort decision-making for fish stocks management.
Fisheries are not the only human pressure on Sicydiinae species.
They also faced other anthropogenic pressures that often impact aquatic species from tropical islands (Artzrouni et al., 2014).As amphidromous species, free movement between freshwater and the marine environment is crucial for completing their life cycle.River mouth closures caused by inadequate river flow management can lead to long-term fish extirpation (McDowall, 1995).A lack of recruitment at the river mouth may cause a decrease in the adult population in the catchment.The presence of dams along the rivers further limits both the upstream colonization of fish and the chances of larvae reaching the sea, but quantitative assessments of such anthropogenic impacts are still limited (see Jarvis & Clos, 2019 for review).Such studies used different methods based on the intensive fishway monitoring to evaluate population dynamics (Lagarde et al., 2015), the biometric and morphometric analysis to measure the effects of dams on the individual morphological selection (Lagarde et al., 2020a) or the direct assessment of the passage of obstacles by the biota (Kreutzenberger et al., 2021).As a general rule, studies conclude that goby-fry fisheries are not sustainable under current conditions mainly because of the lack of knowledge on the biology and ecology of these species (Delacroix, 1987;Keith et al., 2015;Valade et al., 2009).The same studies recommend a reflection regarding the management of wild goby stocks.
Aquaculture techniques offer valuable tools for the production of fish for human consumption and can potentially help to manage wild fish stocks (Froehlich et al., 2017;Patterson, 2019).Thus, the development of goby aquaculture may be a valuable way to limit pressure on wild stocks: 1. Through releasing captive-bred individuals to improve wild recruitment (i.e. stock enhancement); 2. By providing the local production of captive-bred individuals for human consumption to counteract the declines of local goby-fry fisheries, as Robert (1977) suggested.
Properly implemented and well-managed aquaculture techniques can benefit fish species (Le Vay et al., 2007;Mustafa, 2003).Interestingly, to our knowledge, there is no report of amphidromous goby aquaculture neither as a conversation tool nor for human consumption.
The present work focused on the red-tailed goby S. lagocephalus.
This widespread species is one of the main harvested species with fisheries in both Indian and Pacific oceans (Bell, 1999;Manacop, 1953) and the most studied Sicydiinae as a model species for amphidromous fish for studying life-history traits, biology and physiology (Lord et al., 2019).The main objective of this review was to evaluate the potential in aquaculture of the red-tailed goby based on the current knowledge on its biology and ecology and provide guidance for further research in goby aquaculture.The establishment of goby aquaculture requires a clear definition of the objectives (i.e.species conversation and/or human consumption) and may involve a market analysis, which is still premature and out of the scope of this review.However, we discussed some critical socio-economic points.

MATERIALS AND METHODS
The search for bibliographic data was conducted in four steps.The first was collecting peer-reviewed literature from 1950 to May 2021, using Web of Science (WOS) and Google Scholar.Although WOS covers >12,000 scholarly journals and provides a fair representation of international mainstream scientific research (Moed, 2005), Google Scholar helps in the retrieval of most difficult information to find, even in non-English language journals (Falagas et al., 2008).Searches (on abstract, keywords and titles) were performed individually using the scientific name (accepted binomial taxonomic name and synonyms) as well as vernacular name in different Latin languages listed from Froese and Pauly (2021).
In the second step, the data were supplemented by searching for relevant information in the grey literature.The documentation of gobyfry fisheries began early (at least 18 th century) and often reported in non-peer-reviewed works.In the same way, information on biology and ecology has been mostly gathered from reports and Ph.D. thesis and not necessarily in peer-reviewed articles.
The third step was performed after analysing the relevance of the documents collected in the previous steps.Once the literature was sorted, the reference list of each of the papers was screened for additional scientific titles and journals until no further relevant reference could be found ('snowballing' references;Sayers, 2007).
From the search results, non-relevant records (i.e.studies that did not explicitly address topics on the targeted species) were removed.
Selected references (n = 141; Figure 1) were included in a database containing the following information: (1) the type of publication, (2) first author name, (3) year of publication, (4) journal (when appropriate), (5) country or region (based on the location of the experiment or sampling or, if any, the affiliation of the first author) and ( 6) the complete reference of the publication (see Supporting Information for the list of the selected references).
In the fourth and last step, the search for bibliographic data was extended to other Sicydiinae species when information about specific aspects (e.g.reproduction biology, larval development) was missing for the red-tailed goby.

GEOGRAPHICAL RANGE AND TAXONOMY
Among amphidromous species, the gobies of the Sicydiinae subfamily are a species-rich group of >100 species (Taillebois et al., 2014).In the Sicydiinae, the genus Sicyopterus is the most diversified with 24 species (Lord et al., 2019) distributed in the Indo-Pacific area from west to east from Comoros Islands in the Indian Ocean to French Polynesia in the Pacific Ocean covering more than 18,000 km (Keith et al., 1999).
However, only one species, the red-tailed goby S. lagocephalus (Pallas, 1770), is distributed throughout this entire range (Keith et al., 1999;Watson et al., 2000), within latitudes from -30 to 31 (Keith et al., 2011(Keith et al., , 2012)).The other Sicyopterus species generally have a more restricted distribution area, and they are endemic to a small group of islands (Keith et al., 2005).Sicyopterus lagocephalus was initially described as Gobius lagocephalus by Peter Simon Pallas (1770) based on a single specimen examined in the St. Petersburg Museum.The specimen has been described first as a Gobius by Joseph Koelreuter (1764) (Sparks and Nelson, 2004).However, the taxonomic status was questioned by Sparks and Nelson (2004), and Smith and Sparks (2007) proposed to suppress the name G. lagocephalus commonly used.Because of its wide geographical distribution range, some ichthyologists have sometimes considered several different species before analysing morphological and meristic variations by Watson et al. (2000).Thus, S. lagocephalus was supposed to be restricted to islands in the western Indian Ocean.In contrast, many similar species were considered to exist in the Pacific Ocean (e.g. S. taeniurus) in French Polynesia (Keith et al., 2005).Keith et al. (2005)  F I G U R E 2 Amphidromous life-cycle of the red tailed-goby Sicyopterus lagocephalus.The oceanic pelagic phase is in dark blue, whereas the freshwater steps are highlighted in light blue

BRIEF DESCRIPTION
The red-tailed goby is among the largest Sicyopterus species, achieving a maximum total length of 13 cm (Keith et al., 1999(Keith et al., , 2015)).In adults and juveniles, the body is somewhat elongated and sub-cylindrical.The snout is rounded, and the mouth is slightly inferior (Keith et al., 2015), typical of a 'grazing' feeding behaviour.In adults, sexual dimorphism is well pronounced.Males have prolonged second, third and fourth spines of the first dorsal fin.The membrane almost extends to the tip of these long spines.Males have a more slender and longer body than females (Teichert, 2012).Furthermore, males have a triangular urogenital papilla with a distal rounded tip, whereas females have bulbous urogenital papilla with fimbriate projection around the distal opening (Keith et al., 2015).Females are brown during the breeding season, and males are blue-green with a rounded caudal fin turning bright red (Keith et al., 1999;Teichert, 2012).

Life cycle characteristics
The red-tailed goby is an amphidromous species meaning that adults spawn in freshwater, whereas free embryos undergo a pelagic open sea phase (Lagarde et al., 2015;Teichert et al., 2012Teichert et al., , 2014b;;Thomas et al., 2018).The description of the life-cycle is presented in Figure 2.
Females laid eggs in freshwater rivers (Manacop, 1953;Delacroix, 1987;Delacroix and Champeau, 1992;Keith et al., 1999).The newly hatched embryos passively drift down to the sea within a few days after hatching (Lagarde et al., 2017;Valade et al., 2009).The duration of the drift downstream coincides with the vitelline resorption and the opening of the mouth (Delacroix and Champeau, 1992).The larvae then begin their marine planktonic life, allowing them to disperse potentially over long distances (McDowall, 2010).At recruitment, postlarvae swim in schools of million individuals in coastal areas towards the rivers' mouths, limiting predation risks (Keith, 2003;Teichert et al., 2014a).In freshwater, for a few days, the post-larvae can keep a pelagic swimming and a herd behaviour with small schools of 10-100 individuals (Hoareau, 2005;Keith et al., 2015).Metamorphosis into juveniles occurs in freshwater (Keith et al., 2012;Taillebois et al., 2011), and fish become sexually mature after 90-130 days in rivers (Lagarde et al., 2020b).The red-tailed goby life expectancy in freshwater remained unknown but was estimated at 2 years (Artzrouni et al., 2014) by comparison with another similar species, S. stimpsoni (Blob et al., 2010).
Among the life-cycle traits of red-tailed goby, Artzrouni et al. (2014) highlighted the need to prioritize the data acquisition on the life expectancy in rivers and the mortality at river mouths and during the drift of larvae to sea to describe wild population dynamics better.

Habitat
From hatching to recruitment, red-tailed goby lives in a marine environment (Table 1).To date, knowledge of the ecology and biology of this species during their marine phase remains very limited.Because investigations in coastal and open ocean habitats are methodologically complex, critical aspects such as their habitat preferences are scarce.
Nevertheless, the increasing use of analytical techniques, including microstructural and Sr:Ca or Ba:Ca ratio analysis in otoliths on recruiting post-larvae caught at river mouths (Lord et al., 2010(Lord et al., , 2011) ) and stable isotope analysis ( 13 C, 15 N and especially 34 S; Dubé and Benoy, 2005;Sorensen and Hobson, 2005), can improve knowledge on their marine phase.Environmental DNA may also be a promising tool for identifying preferential habitats and facilitate fine-scale geographic and temporal mapping of fish populations at relatively low cost (e.g.Berry et al., 2019;Buxton et al., 2018;Collet et al., 2018;Stoeckle et al., 2017).
The knowledge about habitats of the red-tailed goby during its freshwater life period is derived mainly from those associated with studies performed in Reunion Island, where amphidromous gobies have been extensively studied for decades (Figure 1).Overall, redtailed goby post-larvae colonize the streams independently of their river of birth, as shown by Berrebi et al. (2005), meaning that there is no 'homing' (McDowall, 2010) and suggesting the plasticity of this species in its environmental requirements.

Reflections on socio-economical aspects
• determination of potential bottleneck in production (e.g.larval feeding, mortalities); • cost and production time optimization (e.g.aquaculture operations, investments, cycles).
In the case of conservation and/or repopulation/restocking programs, aquaculture techniques/protocols must be adapted to socioecological needs (e.g.strategies for retention and/or control of genetic resources; assessment, management and control of environmental biosecurity).

Broodstock maintenance and sexual maturation
Non-invasive sexing, such as examination by direct observation or photography of the urogenital region, has been used successfully in redtailed goby (100% reliable; Balon and Bruton, 1994).
In captivity, broodfish can be kept into small aquariums (∼40 L) in a recirculating system dedicated to reproduction for environmental condition control and to limit the risk of diseases, as shown for other goby species (Lindstrom, 1988).Although Delacroix (1987) observed a sex ratio slightly biased in favour of males (1 male for 0.85 females), a sex ratio of 1 male to 1-2 females is likely the best compromise to avoid male competition for spawning sites.In addition, because the female seems to play a determining role in the pairing of red-tailed goby (Teichert, 2012), the presence of two females may maximize the chances of fertilization.
TA B L E 2 Main characteristics related to the biology and ecology of juveniles and adults Sicyopterus lagocephalus for future research in aquaculture
From the few studies performed on the nutrition of red-tailed goby adults, the current formulated feeds developed from omnivorous fish (such as carp or tilapias), usually with low protein content (≤30%) and formulated mainly from plant ingredients, may be appropriate for goby aquaculture.Additional feeding with a frozen diet such as Artemia sp. may be an asset (Lindstrom, 1988).Research in goby aquaculture should focus primarily on nutrition, a parameter that significantly influences the gametogenesis of fish, particularly females (Cabrita et al., 2008).
Females lay an entire clutch in a unique event (from 14,000 to 232,000 eggs), and then another batch of oocytes is recruited (Teichert, 2012;Teichert et al., 2013a).They can spawn several times throughout the year with an inter-spawning interval of 1-2 months, but the spawning frequency can vary according to environmental cues and over the spawning period (Teichert et al., 2014b).The high fecundity of redtailed goby (i.e.7000-15,000 oocytes per gram; Table 2) is an asset for aquaculture production.
Among the abiotic factors influencing spawning, the most critical water parameter that stimulates spawning is the temperature (Teichert et al., 2013a).Other factors may play a significant role in triggering spawning events, including the availability of nest supports and the decreasing conductivity of the water due to rainfall (Teichert, 2012).
More recent studies focused on larval drift (Lagarde et al., 2017(Lagarde et al., , 2018) ) also hypothesized that the increase of discharge during austral summer is the primary driver enhancing spawning activity.These potential cues need to be tested in controlled conditions.Some studies report injecting human chorionic gonadotropin (hCG) in mature red-tailed goby females (Delacroix, 1987;Delacroix and Champeau, 1992) to obtain eggs from captive broodfish.Nevertheless, the success of induction by hormone injection is limited (>30% of spawning success after injection), and mortality of injected fish is high (20% in less than 48 h post-injection; Delacroix, 1987;Delacroix and Champeau, 1992).Based on current knowledge, it is reasonable to expect spontaneous spawning in captivity if appropriate conditions are provided (see above).

Incubation and larval rearing
In the wild, the incubation of the eggs, laid in adhesive clusters, is very short (i.e.50-60 h; Table 1).This rearing step should be performed in well-oxygenated and well-circulated water.The male shows parental care behaviour (Teichert, 2012;Teichert et al., 2013a).Thus, in captivity, the incubation of eggs can be performed in the broodfish tanks and likely does not require dedicated structures.
Like other Sicydiinae species, larval rearing can be initiated with the transfer of the PVC shelter with attached eggs that are ready to hatch in larval rearing tanks (e.g.Bell and Brown, 1995).Alternatively, newly hatched larvae can be collected and then transferred (e.g.Archambeault et al., 2015).The latter option has a higher risk of mortality.Valade et al. (2009), monitoring the effect of temperature on unfed larvae survival time, reported the highest survival rate from 20 to 23 • C, with mortality increasing at higher temperatures.Ellien et al. (2011) showed no effect of stocking density (between 30 and 150 larvae per litre) on the development and the mortality of newly hatched larvae.
In the wild, goby larvae migrate passively down to the sea very quickly.Thus, several studies have demonstrated the need to transfer larvae to seawater before the age of 4 days; otherwise, development cannot continue, and the larvae die (Ellien et al., 2011;Valade et al., 2009).The transfer does not necessarily need to be progressive with a gradual increase in salinity (Valade et al., 2009) and can be carried out as early as 24 h after hatching (Bell and Brown, 1995;Delacroix, 1987).In a recent study, Teichert et al. (2021) highlighted that the survival of red-tailed goby larvae in freshwater exceeded 150 h at 21 • C, whereas it dropped below 50 h at 31 • C. In seawater, the larval development of unfed larvae was affected by temperature with gradual decrease in survival with increasing temperatures.The mouth and anus opened after 30-50 h in seawater depending on the water temperature (Delacroix, 1987;Ellien et al., 2020;Teichert et al., 2021;Valade et al., 2009) suggesting that larvae need to find suitable prey in seawater within a short time.Field studies highlighted that larvae and post-larvae of Sicydiinae are secondary consumers and omnivorous during their pelagic phase and mainly feed on zooplankton and plant/macroalgae detritus (Baptista et al., 2020;Keith et al., 2008).
Nevertheless, the exact nature of the first feeding of wild larvae is largely unknown, and given the small size of the larvae (total length: 1.4-1.7 mm; Ellien et al., 2020;Keith et al., 2015;Valade et al., 2009), this is the most critical point for the larval rearing of the goby species.
Although this parameter has not been accurately measured, based on the analysis of published photographs (Delacroix, 1987;Valade et al., 2009) through ImageJ software (Abràmoff et al., 2004), the aperture of the mouth can be estimated at 80-85 µm in red-tailed goby larvae at the presumed first-feeding stage (Figure 3).According to Allen et al. (2006), the critical prey size depends on width rather than length.
(∼50 µm of width; Kline and Laidley, 2015) are likely to be suitable prey for young life stages.Kondo et al. (2013)  Altogether, these previous findings suggested that the red-tailed goby larvae diet should be first based on copepod nauplii and/or smallsize rotifers, and then include brine shrimp nauplii, which is easier to cultivate, previously enriched with phytoplankton or commercial products.The quality of the first diet is, as in many species, a key parameter to explore through aquaculture research on goby species.
Nevertheless, the studies on the larval development of Sicydiinae species carried out larval rearing for a few days or a few weeks only.So, the growth and survival performances during this rearing phase remain largely unknown and require further investigation.
The amphidromous nature of the red-tailed goby is a technical constraint.Although larvae can be transferred from freshwater to seawater relatively easily (Delacroix, 1987;Delacroix and Champeau, 1992;Valade et al., 2009), the transfer requires rearing structures that can operate with seawater (larval phase) and with freshwater (broodfish).
Nevertheless, aquaculture of amphidromic species exists on a commercial scale.One of the closest examples is the aquaculture of ayu Plecoglossus altivelis in Japan.Ayu is an amphidromous fish in Japan, commercially important for freshwater fisheries.A decline in the abundance of wild ayu populations in rivers due to habitat destruction and blockage of migratory routes has promoted stock enhancement strategy from hatchery populations (Iguchi et al., 2003).

From post-larvae to adults
The gregarious behaviour of the juveniles red-tailed goby until their sexual maturity (Teichert, 2012) suggests a possible maintenance in a common tank.During the first days after recruitment, post-larvae do not feed (Keith et al., 2012); afterward, the juveniles must switch from a planktonic feeding mode to a benthic feeding mode (Keith et al., 2008).
Therefore, commercial pellets used mainly for demersal omnivorous species such as cyprinids are likely appropriate for gobies.Nevertheless, the suitability of commercial pellets needs to be confirmed experimentally.In the wild, metamorphosis occurs approximately 2 weeks after recruitment in freshwater (Keith et al., 2008).In aquaculture, this step can be challenging; nevertheless, Keith et al. (2008)

CONCLUDING REMARKS
The present study provides a critical overview of the scientific literature on the red-tailed goby S. lagocephalus.It is an attempt to foster further investigations in the aquaculture of these species.We do not advocate aquaculture as an obvious alternative to goby-fry fisheries because more research is needed to study the technical feasibility and sustainability of goby aquaculture.Nevertheless, this review lays the foundations for this research area, which has not been performed to date.
Below is a summary of key findings and recommendations for future research: • Aquaculture is an area of research that should provide an alternative to fisheries and a new tool for the enhancement of goby post-larvae recruitment.• Goby aquaculture can only be considered after a socio-economic analysis.
• The current knowledge on the biology and ecology of red-tailed goby suggests the potential of this species and, at this stage, holds promise for aquaculture.
• The efficiency of the transition from freshwater to seawater in captivity needs to be confirmed.
• Although it is reasonable to expect the spawning in captivity from broodfish, larval rearing may be more challenging.
• The nutrition of the early larval stages should be among the priorities of research for goby aquaculture.

F
Bibliometric analysis of the papers published on the red-tailed goby Sicyopterus lagocephalus from 1950 to May 2021.(a) Publications over time ranking as peer-reviewed articles, book and book chapters, and Ph.D. thesis.The cumulative number of publications is indicated on the right.(b) The total number of studies performed by geographical area (i.e. the location of the experiment/sampling or, if any, the affiliation of the first author) confirmed, by phytogenetic analysis, the ubiquity of S. lagocephalus in the Indo-Pacific area and identified two different clades, one clustering most of French Polynesian haplotypes and the other clustering most of Mascarene (including Comoros) haplotypes.
Teichert et al. (2014a) found that both adults and juveniles have weak habitat preferences based on the characterization of microhabitat variables (depth, velocity and predominant substrate) and the presence of conspecifics and sympatric species.Overall, in the investigated red-tailed goby habitats, depths ranged from 3 to 450 cm, and water TA B L E 1 Main characteristics related to the biology and ecology of eggs and larvae Sicyopterus lagocephalus for future research in aquaculture Goby aquaculture may require a socio-economic analysis out of the scope of the present review.Determining the potential market, social acceptance (Ruiz-Chico et al., 2020) and production costs are significant points in the development of goby aquaculture.However, given the scarce data on the red-tailed goby's reproductive biology and husbandry techniques, research and development are required beforehand.Based on the data collected and analysed during this study, the research and development phase should at least include: succeeded in the larval rearing of three Rhinogobius species into 30-L transparent polycarbonate tanks by feeding larvae with the super-small rotifers (Brachionus rotundiformis) immediately after hatching.Phytoplankton (Nannochloropsis oculata) was added three or four times per day in the tank to enhance the nutritional values of the rotifers.Then, brine shrimp (Artemia sp.) were added as additional live prey after 7 days at 19-21 • C. Lindstrom (1988) tested different first feeds for Lentipes concolor and performed gut content analysis.The study showed that larvae ingested all of F I G U R E 3 Measurement of mouth opening (white lines) in red-tailed goby Sicyopterus lagocephalus larvae 5 days after hatching from published photographs: (a) Valade et al. (2009) and (b) Delacroix (1987) them (i.e.phytoplankton, oyster trochophores, small rotifers, copepod nauplii and undetermined wild plankton).
maintained individuals for approximately 1 month (until they reached adult size) in mesocosms without indicating any episodes of mortality related to metamorphosis.