Reproduction, Larviculture and Early Development of the Bluebanded

Aquaculture Research, 2015, 1–18 doi:10.1111/are.12648

Reproduction, larviculture and early development of the Bluebanded goby, Lythrypnus dalli, an emerging model organism for studies in evolutionary developmental biology and sexual plasticity

Sophie Archambeault1, Eric Ng1, Lyle Rapp1, David Cerino2, Bradford Bourque2, Tessa Solomon-Lane3, Matthew S. Grober4, Andrew Rhyne2 & Karen Crow1 1Department of Biology, San Francisco State University, San Francisco, CA, USA 2Department of Biology and Marine Biology, Roger Williams University, Bristol, RI, USA 3Neuroscience Institute, Georgia State University, Atlanta, GA, USA 4Department of Biology, Georgia State University, Atlanta, GA, USA

Correspondence: S Archambeault, University Of Washington, Molecular and Cellular Biology Program, Box 357275, Seattle, WA 98195-7275, USA. E-mail: [email protected]

tool that could be utilized for a variety of similarly Abstract diminutive species and to address a greater breadth The Bluebanded goby, Lythrypnus dalli, is a popular and depth of biological questions. ornamental aquarium species and a key organism for the study of several fundamental biological Keywords: larviculture, co-culture, Lythrypnus questions, most notably reversible sex change in dalli, acanthomorph development, fused pelvic adults. To maximize the tractability of this species fins as an emerging model system, it is essential to have an optimized propagation system and a detailed Introduction developmental staging scheme. One limitation to the larviculture of L. dalli is the relatively small size The rearing and development of marine fishes is of of the larvae, which makes the transition from yolk interest to hobbyists, commercial aquarists and to feeding challenging. We developed a protocol biologists for a variety of reasons including the and successfully reared three generations of L. dalli following: to increase our understanding of larval in the laboratory. The protocol contains several key development and comparative embryology, innovations for the rearing of diminutive fish lar- to increase ichthyoplankton identification, to vae, including tank design and co-culturing of strengthen inferences of homology of ontogenetic microalgae (Isochrysis galbana) with copepods (Par- characters and to increase the number of species vocalanus sp.) in the larval rearing tanks. In that can be bred in captivity for ornamental prop- addition, we describe the embryonic and larval agation and research (Kendall, Ahlstrom & Moser development of L. dalli under controlled conditions 1984; Collazo & Bolker 1994). Rearing fish in and in comparison with the model organism Danio captivity provides access to embryos, which is a rerio. We found that at 21°C L. dalli larvae hatch in limiting factor in evolutionary developmental biol- 4 days, reach flexion in 18–25 days and are sexu- ogy research. The ability to rear larvae of various ally mature by 3 months. Overall, the embryonic species in captivity is essential for comparative and development of L. dalli is remarkably similar to manipulative investigations of the genetic and D. rerio with several striking differences, including developmental basis of diversity and novelty. In the position and shape of the blastomere, size of the addition, observations of development may help neuromasts and corresponding cupula, and relative aquaculturists identify and address significant timing of pigmentation and brain subdivision. The hurdles in rearing these and similar species. ability to rear this species in captivity is a valuable Finally, the number of ornamental marine species

© 2015 John Wiley & Sons Ltd 1 Larviculture and development of the Bluebanded goby S Archambeault et al. Aquaculture Research, 2015, 1–18 commercially produced in captivity is currently L. dalli are attached to the substrate, cared for by limited, especially compared to freshwater species the guarding male (Rodgers, Earley & Grober (Andrews 1990; Meirelles, Tsuzuki, Ribeiro, Medei- 2006), and develop into transparent embryos that ros & Silva 2009; Olivotto, Planas, Simoes, Holt, are hardy and easily monitored. They are abundant Avella & Calado 2011; Sweet 2014), and this rear- throughout most of their range, in coastal marine ing protocol may aid in the development of captive habitats in the Eastern Pacific from the Gulf of Cali- breeding programmes that could potentially allevi- fornia to Morro Bay, CA (Miller 1972), and are not ate collection pressures on a variety of wild marine listed as a species of concern in any category (IUCN fishes targeted by aquarium trade collections. Red List, van Van Tassell, Lea & Bearez 2010). The developmental processes of teleost fishes, Finally, this species displays several derived traits with the exception of a few taxa, have not been that make them ideal for studies of evolution and well characterized, representing an important gap development, such as high contrast pigmentation in our understanding of animal biology. This defi- patterns, pelvic fins that are fused into a sucking cit is particularly striking as teleosts represent the disc (a common feature of this family), and bidirec- majority of vertebrate diversity, with approxi- tional sex change. Bidirectional sex change repre- mately 27 000 species (Nelson 2006). The most sents the extreme in sexual plasticity, and L. dalli commonly studied teleost is the zebrafish, Danio re- are commonly used in studies of sex determination rio, a freshwater species widely used in laboratory and sex change (St. Mary 1993, 1994, 1996, research. Accordingly, D. rerio development has 1998, 2000; Black, Reavis & Grober 2004; Black, been well characterized (Kimmel, Ballard, Kimmel, Balthazart, Baillien & Grober 2005, 2011; Rodgers, Ullmann & Schilling 1995; Metscher & Ahlberg Drane & Grober 2005; Lorenzi, Earley & Grober 1999), which provides a valuable basis for com- 2006, 2012; Rodgers et al. 2006; Rodgers, Earley & parative developmental studies. However, D. rerio Grober 2007; Lorenzi, Earley, Rodgers, Pepper & is not a member of the most speciose and derived Grober 2008; Lorenzi, Carpenter, Summers, Earley group of teleost fishes, the Acanthomorpha sensu & Grober 2009; Lorenzi & Grober 2012; Maxfield, Near, Eytan, Dornburg., Kuhn, Moore, Davis, Van Tassell, St Mary, Joyeux & Crow 2012). Wainwright, Friedman and Smith (2012), or Despite their popularity and tractability as a spiny-rayed fishes. This group includes many eco- model for biological enquiry, the early develop- nomically important species such as tuna, halibut, ment of gobies is relatively poorly studied (Nak- rockfish, tilapia and mullet (Stiassny, Wiley, John- atsuji, Kitano, Akiyama & Nakatsuji 1997; son & Carvalho 2004; Nelson 2006; Eschmeyer & Patzner, Van Tassell, Kovacic & Kapoor 2011), Fricke 2012). Therefore, building model systems and L. dalli have previously been considered diffi- for developmental studies among acanthomorph cult to rear in captivity (Moe 2008). This difficulty species is of particular interest to both evolution- is largely due to the small size of L. dalli at yolk ary biologists and aquaculturists. depletion and transition to exogenous feeding (L. The Gobiidae are the most diverse family of dalli larvae measure 2.1 mm at hatching, Patzner Acanthomorpha and the second-most speciose fam- et al. 2011), and the difficulty of providing appro- ily of vertebrates (Moyle & Cech 2004; Nelson priately small live food at first feeding. We devel- 2006). Many are of interest to the aquarium trade, oped a protocol to address this hurdle that can be including the Bluebanded goby, Lythrypnus dalli, adapted to rear a variety of ornamental or because they are small, brightly coloured and exhi- research species with small larvae. bit interesting social and reproductive behaviours The objectives of this study were to develop a (Sunobe & Nakazono 1993; Scaggiante, Grober, larval rearing system and protocol for L. dalli, and Lorenzi & Rasotto 2004; Drilling & Grober 2005; to describe their development in an evolutionary Meirelles et al. 2009). Adult L. dalli are readily col- context. We have successfully reared three genera- lected within their range on shallow reefs using tions of L. dalli in captivity and characterized their SCUBA. In addition, adults are readily maintained embryonic and larval development in comparison in captivity, breed readily and females lay hundreds with D. rerio. This comparison highlights the of eggs per clutch. Their size at maturity is relatively developmental variation among teleosts and the small (adult TL less than 6 cm, Miller 1972), mak- need for greater representation of the Acanth- ing them tractable with respect to space limitations omorphs for studies that address the bases of novel in aquaculture and laboratory settings. The eggs of traits in a phylogenetic context.

inspected, staged and returned to their designated Materials and methods tubes and tanks for paternal care until hatching, between 4–6 days later at ~20°C. Animals (breeding colonies) Eye pigmentation, yolk colour and age of All animal procedures were conducted in accordance clutches were monitored to predict hatching. Just with protocols approved by the Institutional Animal before hatching, clutches were moved to individual Care and Use Committees of San Francisco State Uni- rearing tanks and placed in submerged hatching versity (SFSU), Georgia State University, and Roger chambers (Fig. 1c). Hatching was generally Williams University. Adult gobies were collected induced within a few hours after covering larval using SCUBA from Catalina Island, California and tanks to eliminate light penetration. If a partial or transferred to SFSU or the Georgia State fish facility. no hatch occurred by the next morning, transpar- Embryos were then shipped from Georgia State Uni- encies were removed from hatching chambers, left versity to Roger Williams University. Embryonic and in the air for up to 5 min, shaken mildly, and early larval development descriptions were based on replaced. fish housed at SFSU, while the rearing protocol was developed at Roger Williams University. Larval rearing system Breeding colonies were established with five adult females housed in 38-litre (10-gallon) rect- Hatched larvae were reared in 100-L circular angular tanks with darkened sides. The largest tanks connected to a 380-L recirculating system female in each group usually changed sex within (Fig. 1d). The system contained a biofilter and pro- 2 weeks, resulting in operational sex ratios of 1:4 tein skimmer, and was maintained under the same within each colony. Sex change was confirmed by temperature and water quality conditions as the the presence of fertilized eggs. adult broodstock. Rearing tank water levels were Adult tanks were maintained at ambient tem- controlled by external T-barbs to maintain the vol- perature (~20°C, range = 14.4°C–22.2°C), with ume at 75 L. A 40-watt aquarium strip light illu- + salinity between 30–35, NH3/NH4 and NO2 minated approximately ⅓ of the tank surfaces,

<1 ppm, and NO3 <20 ppm. Each tank had iso- 24 h a day. An 8-cm-diameter cylindrical stand- lated, recirculating water filtered with an Emperor pipe screen of 40-lm mesh was attached to the 280 BIO-Wheel filter (Marineland Aquarium Prod- centre drain of each larval tank (Fig. 1e). A bubble ucts, Moorpark, CA, USA) and received 25% ring around the bottom of the screen provided weekly water changes with filtered, sterilized natu- constant gentle aeration and kept larvae and food ral seawater (Mt. Hope Bay, Rhode Island). Tem- away from the standpipe drain. Water inlets were perature, salinity, flow, fish health and behaviour placed against the sides of the rearing tanks to were monitored daily and water quality was moni- minimize turbulence and were fed with constant tored weekly. A photoperiod of 14:10 hours light: flow of 0.1–0.15 L minÀ1. Daily maintenance dark was provided by ambient fluorescent lighting, included system water changes of 10–25%, clean- mimicking natural conditions during the breeding ing of microalgae drip lines and standpipe screens, season. Tank bottoms were syphoned every day, and syphoning of debris and deceased larvae from or as needed. Adults were fed a variety of food rearing tanks. including Otohime Fish Diet (Marubeni Nisshin Feed Co., Ltd., Tokyo, Japan), a gelatin based diet Plankton Cultures with added seafood products, frozen Mysis shrimp and frozen bloodworms 3 to 4 times daily until Larval L. dalli require extremely small foods at the satiated. onset of exogenous feeding. We found that the copepod nauplii of the genus Parvocalanus sp. were well suited for the relatively small larvae of L. dalli, Egg production, incubation and hatching similarly to what has been reported for the lemon- Each tank contained two 2-inch long, ¾-inch peel angelfish (Olivotto, Holt, Carnevali & Holt diameter PVC tubes lined with a plastic transpar- 2006). Parvocalanus crassirostris (Reed Mariculture ency for egg deposition (Fig. 1a). PVC tubes were Inc., Campbell, CA, USA) were cultured in 150- visually checked daily for new egg clutches litre tanks at densities of ~10–30 copepods per (Fig. 1b). When eggs were found, the clutches were mL. To maintain these cultures, the microalgae

(a) (b) (c)

(d)

(e)

Figure 1 Recirculating larval rearing system. (a) Two-inch long segments of ¾ inch PVC pipe were fitted with transparencies for egg deposition. (b) Bright orange yolks made newly deposited eggs easy to spot within the PVC tubes. (c) Hatching chambers were 3.8-cm-diameter glass tubes (flower vases) fed with a rigid water line 1 cm from the bottom, to ensure water movement around the embryos (flow ~ 0.25 L minÀ1). (d) Design of the larval rearing system with 380-litre capacity, includes (1) a 150-litre sump containing Kaldnes media for biological filtration, an in-sump protein skimmer, and a Danner 1800 Mag-Drive pump for recirculation. Plumbing includes (2) recirculating flow lines fitted with ¼-inch quarter turn labcock valves to control inflow rates, (3) ½-inch PVC return lines from rearing tanks back to sump, (4) ½-inch PVC alternate waste lines to floor drains, and (5) T-barbs connected to each centre drain to control the water level in each tank. (e) Customized 8-cm (shown) or 5-cm-diameter centre standpipe drains kept larvae in the tanks while allowing adequate water exchange. Bubble rings (blue) were fed from airlines above the larval tanks and kept larvae and copepods away from the standpipe drain.

Isochrysis galbana (CCMP 1324) (Florida Aqua 8–12 9 106 cells mLÀ1 were reached in 5–7 days Farms Inc., Dade City, FL, USA) was added twice in algae media. Cultures were started in 1 L Roux daily to densities of 1–2 9 105 cells mLÀ1, and flasks with autoclaved media, transferred to 20 L water changes of 50% were conducted three times glass carboys and finished in 100 L polycarbonate per week. Copepods were harvested using 40-lm tanks (20 and 100 L media was pasteurized to mesh screen, while size-sorted nauplii for larval 85°C). Algae media was formulated following the feeding were harvested by filtering the culture manufacturer’s instructions (F/2 Algae Food, Fritz through a 75-lm screen filter, then through a 40- Aquatics, Inc., Texas, USA). lm screen filter. Culturing of I. galbana was accomplished follow- Larval feeding ing the protocols outlined by Creswell (2010). All pieces of microalgae culture equipment were auto- Copepods (Parvocalanus sp.) were cultured in L. claved or sterilized with bleach. Isochrysis galbana dalli rearing tanks as described above prior to the were cultured at ambient temperature (20°C) with first feeding of larvae. Isochrysis densities were sterilized salt water (salinity of 35) in sealed cul- maintained daily (8–12 9 106 cells mLÀ1) ture vessels. Each culture was aerated through a through microalgae drip lines as well as direct 0.2-lm in-line filter. Optimal densities of addition to the tank. Early larval gobies were able

4 © 2015 John Wiley & Sons Ltd, Aquaculture Research, 1–18 Aquaculture Research, 2015, 1–18 Larviculture and development of the Bluebanded goby S Archambeault et al. to consume only the smaller copepod nauplii, and either before or during overnight preservation in were therefore reliant on reproducing copepods. 4% paraformaldehyde at 4°C. Embryos were stored Late stage larvae and early-settled juveniles in 100% methanol at À20°C. Whole mount in situ were fed Artemia nauplii (Brine Shrimp Direct, hybridizations were carried out as described by Odgen, UT, USA) in addition to copepod culture Archambeault, Taylor and Crow (2014). maintenance. Culturing of the copepods through the addition of microalgae was halted once all the Results larvae settled out of the water column. Juveniles were weaned onto a diet of Otohime pellets Fertilization, hatching rates, developmental timing (100–400 lm diameter) at 2 weeks post settle- and survivorship ment. Fertilization and egg deposition of L. dalli occurred most frequently in the 4 hours following first light, Developmental staging of Lythrypnus dalli 8:00–12:00 hours. (Fig. 2). Average nest size at Developmental observations of L. dalli embryos 21.6°C was 835 eggs (n = 9, range 396–1055), and larvae were conducted at San Francisco State and often represented multiple spawning events. University. The individuals used for this study Males guarded the nests until hatching, and filial were housed according to the protocol above, cannibalism was occasionally observed on clutches except that adults were kept in a closed, recirculat- that were unfertilized or experienced fungal ing sea water system maintained at 15.5°C growth (personal observations). Removal of (14.4–18.3°C). Developmental observations were clutches from broodstock tanks significantly made at two temperatures, 14.4 and 17.0°C. increased the frequency of egg laying: when Photos of embryos and larvae were taken at 43 clutches were removed, new clutches were fertil- different stages of development from 33 different ized every 9.4 days (SE = 0.86, n = 86) compared clutches of eggs. Embryonic and early larval obser- with every 11.0 days (SE = 0.85, n = 63; Mann– vations and photos were taken of live individuals Whitney test: UA=3378, p = 0.0102) when clutches using a Zeiss DiscoveryV.8 stereomicroscope and remained with adults for paternal care. However, an AxioCam ICc3 camera with AxioVision soft- hatching rates for L. dalli were more consistent with ware (Carl Zeiss Microscopy GmbH, Jena, Ger- parental care, ~95% with male care at temperatures many). If necessary, larvae were anaesthetized in of 18 – 22°C. Survivorship to sub-adulthood was a 1:25 dilution in seawater of a 0.4% MS-222 variable, between 5–40%. solution, pH 7 (Tricaine methanesulfonate; Sigma- The yolks of L. dalli embryos are orange, making Aldrich, Co., St. Louis, MO, USA). Stages of embry- new clutches easy to detect against the white PVC onic development are described for L. dalli based tubes. Like other goby eggs, L. dalli eggs are on staging criteria for D. rerio (Kimmel et al. oblong in shape and are attached to the substrate 1995) for the following periods: zygote, cleavage, by adhesive threads, in contrast to the negatively blastula, gastrula, segmentation, pharyngula, buoyant but free moving spherical eggs of D. rerio hatching and early larval. Observations of late lar- (Fig. 3). Tavolga (1950) showed in Bathygobious val development were made on individuals pre- that eggs stripped from the ovary are equipped served overnight in 4% paraformaldehyde at 4°C, with the adhesive threads, which are then acti- then transferred to 100% methanol and stored at vated by contact with water and attach to the À20°C. The proposed staging of late larval L. dalli substrate. However, in L. dalli adhesive threads are was made comparable to D. rerio following Pari- absent from the eggs while still in the ovary, and chy, Elizondo, Mills, Gordon and Engeszer (2009). thus must be added at the time of female egg deposition (Grober, unpublished data). The female external genitalia in L. dalli have two internal In situ hybridization chambers, the most ventral of which is most likely Whole mount in situ hybridization was done to used to produce the adhesive threads during egg test the feasibility of the procedure and to observe laying (Schuppe et al., in preparation). expression patterns of the posterior Hox gene, Unlike some gobies, L. dalli eggs lack oil globules hoxa13a, during multiple stages of L. dalli embry- (Yokoi & Hosoya 2006; Wittenrich, Turingan & onic development. Embryos were dechorionated Creswell 2007). Chorion dimensions are uniform

50 Lights on at 7 am Lights off at 9 pm

10 Number of clutches fertilized Number 0 12 am–4 am 4 am–8 am 8 am–12 pm 12 pm–4 pm 4 pm–8 pm 8 pm–12 am Time of day

Figure 2 Eggs are fertilized predominantly in the morning. A distribution of the timing of egg fertilization shows that eggs were most commonly fertilized in the 4 hours following ﬁrst light. Timing was determined by staging embryos and estimating their age upon discovery. Age was then subtracted from the time they were found to estimate the time of day that the eggs were fertilized. at 1.88 mm by 0.46 mm (length x = 1.878, position during early cleavage. The cells formed SE = 0.014; width x = 0.464, SE = 0.002; n = 15). by early cleavages in D. rerio are positioned on Embryonic L. dalli grew in size from 0.57 mm top of the yolk whereas the cells of the early (SE = 0.005) diameter at the zygotic stage to blastodisc of L. dalli wrap around the yolk, and 2.72 mm (SE = 0.05) in length at hatching. reach towards the vegetal pole (Figs 3b–c, 5a–e). Lythrypnus dalli embryos hatched in 10.3 days A third cleavage plane, perpendicular to the ori- at 14°C, 9.1 days at 17°C and 4 days at 21.6°C, ginal two planes, is established by the 64-cell which is slightly faster than other gobiids reared stage in both L. dalli and D. rerio embryos, result- at similar temperatures (Fig. 4), but slower than ing in multiple layers of cells (Fig. 5f). By the D. rerio at its common rearing temperature, seventh cleavage in L. dalli the blastodisc is posi- 28.5°C. To directly compare the development rate tioned on top of the yolk (Fig. 5f). During early of L. dalli to D. rerio, relative hours of development epiboly, the yolk of L. dalli becomes relatively was plotted vs. developmental stage (Fig. 4). Both oblong, mimicking the shape of the chorion L. dalli and D. rerio display a linear relationship (Fig. 5h–i). By 90%, epiboly the embryo is spheri- between time and developmental stage, thereby cal again (Fig. 5j). making it possible to estimate developmental stage based on time and temperature for L. dalli Segmentation (y = 0.0189*temp*hpf, R2 = 0.914, R2 = 0.932 for 17°C and 14°C respectively). Interestingly, L. By the 6-somite stage, both D. rerio and L. dalli dalli takes 2.3–3 times longer than D. rerio to have a distinct head bulge and tail bud, both of reach the protruding mouth stage, depending on which are more pronounced and relatively bigger the temperature. This relationship between in L. dalli (Fig. 5k). This difference contributes to developmental stage and time is robust to temper- an overall oblong shape of L. dalli, mirroring the ature, therefore, the following staging scheme for shape of the chorion. This difference in shape L. dalli is consistent at a range of developmental appears to be associated with earlier detachment temperatures. of the tail and head buds from the yolk in L. dalli (Table 1). Thus, egg shape may play a role in shaping developmental progress in some Cleavage, Blastula and Gastrula features. The early periods of development are very similar During the early segmentation period, the optic between D. rerio and L. dalli (Fig. 5a-j), however, placode and somites of L. dalli are often obscured there is an observable difference in blastodisc by a grainy texture, likely due to cell movement

(6-somite stage, Fig. 5k). Nonetheless, the optic (a) (b) placode is visible by the 6-somite stage, the otic placode is visible by the 9-somite stage, and the otic vesicle hollows and contains two otoliths by the 12-somite stage (Fig. 5k–m). Also by the 12- somite stage, the lens is present and most somites (c) exhibit the expected chevron-shape. Differentiation of the brain also occurs earlier in L. dalli than D. rerio. By the 19-somite stage, the telencephalon, diencephalon and mesenceph- alon are evident in the fore- and midbrain of L. dalli, and are separated from the hindbrain by a distinct cerebellum (Fig. 5n). In contrast, the cerebellar primordium does not become prominent in D. rerio until the 26-somite stage (Table 1).

(d) Pharyngula

The pharyngula period begins with the posterior advancement of the lateral line primordium over the trunk somites (e.g. prim-5, prim-15). The primordium is a difficult trait to monitor, and requires differential interference contrast optics to visualize. We took advantage of the large and visible neuromasts of L. dalli to characterize their developmental stage categorically. For example, when a neuromast was visible on the ninth somite, but not yet on the 14th somite, we categorized the embryo into the prim 9–14 stage. By the end of the pharyngula Figure 3 The eggs of L. dalli are oblong and attached period, neuromasts were visible on the 5, 9, 14, 23 to the substrate via adhesive threads. (a) The chorions and 29 somites of L. dalli, therefore, we suggest of 4-day post fertilization L. dalli embryos remain subdividing the pharyngula period into six stages oblong and attached perpendicularly to the substrate (prim 1–5, prim 5–9, prim 9–14, prim 14–23, prim via adhesive threads (arrows in a and b). (b) The L. 23–29 and high pec). dalli embryo is smaller than the D. rerio embryo (c). Pigmentation begins in the early prim stages The L. dalli blastomere wraps around the yolk towards the vegetal pole (marked by arrowheads in b) through for L. dalli.Byprim-9-14,L. dalli has scattered the 8-cell stage. (c) The D. rerio embryo is free-floating, black and red pigment cells along the ventral line has a spherical chorion, and the blastomere is perched of the body (Fig. 5o). Additional pigmentation on top of the yolk (arrowheads). (d) Ventral aspect of a appears on a dorsal region of somites 14–23 by cross-section through the body of a female Lythrypnus the prim 23–29 stage (Fig. 5p). Eye pigmentation dalli, noting the position of the ovaries (Ov), hypaxial develops with trunk melanocytes in the prim body musculature (M) and genital papilla (GP). The 23–29 stage, which is later than in D. rerio female GP has two lumens. The dorsal lumen is the (Table 1). putative oviduct by which the egg is communicated Muscle movement, including a visible heart- from the ovary to the genital papilla for adhesion to beat, begins in L. dalli at the prim 9–14 stage. A the substrate. The ventral lumen is the proposed adhe- small pectoral fin bud emerges by this stage as sive thread producing structure within the GP and is noted with an arrow. Tissue was stained with Haemat- well. The pectoral fin bud develops the apical oxylin and Eosin. Scale bars in b and c equal ectodermal ridge (AER) and reaches a 1:1 height 0.50 mm, and in d equals 200 lm. Vegetal poles (in b to width ratio at the “high-pec” stage (Fig. 5q). and c) and ventral surface of the fish (in d) are at the By this stage, many visible changes have bottom of the images. occurred in L. dalli including heavily pigmented

(a) 3.90 (b) 25 Leucopsarion Protruding mouth petersii Rhinogobius sp. BI 3.40 20 Elecatinus figaro Protruding mouth 2.90 Gobiosoma evelynae 15 Zosterissessor 2.40 ophiocephalus Gobius paganellus 1.90 10 Gobius cobitis

1.40 Eviota abax

Mean days to hatching 5

Embryonic length (mm) 14-somite L. dalli Eviota storthynx 0.90 6-somite D. rerio Priolepis nocturna 0 0.40 Lythrypnus dalli 0 25 50 75 100 125 150 175 200 10 15 20 25 30 Hours post fertilization

pec fin - 60

long pec - 50 high pec - 40 prim 25 -

prim 15 - 30 at 28.5°C 25 somite - Developmental stage 18 somite - 20 6 somite - 75% epiboly - 10 germ ring - 64 cell - 0 zebrafish Hours post fertilization for 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 Hours post fertilization

Figure 4 Lythrypnus dalli embryos are smaller and develop slower than Danio rerio. (a) Embryonic length (EL) as a function of hpf for L. dalli at 17°C and D. rerio at 28.5°C(D. rerio data taken from Kimmel et al. 1995). Danio rerio is longer than L. dalli at all stages of development observed, except the 14-somite stage. This increase in EL of L. dalli is due to the lifting of the tail bud from the yolk. The tail and head buds of L. dalli lift off the yolk sooner than in D. rerio which results in an increase in EL (6-somite stage in L. dalli and 14-somite stage in D. rerio). This figure highlights the longer time to hatching of L. dalli compared to D. rerio. (b) Days to hatching as a function of water temperature in 11 species of Gobiidae. Three species were reared at more than one temperature and display an inverse relationship between temperature and time to hatching. The other eight species were only reared at one temperature, however, taken together, Gobiidae species show an overall inverse relationship between time to hatching and temperature. Data for all species except L. dalli were gathered from the literature (Sunobe & Nakazono 1987; Gil et al. 1997; Privileggi et al. 1997; Arakawa et al. 1999; Borges et al. 2003; Olivotto et al. 2005; Yokoi & Hosoya 2006; Wittenrich et al. 2007; Meirelles et al. 2009). (c) To directly compare the development rate of L. dalli to D. rerio, hours post fertilization (hpf) were plotted on the vertical right axis against hpf on the horizontal axis for D. rerio (solid black diamonds). The vertical right axis (hpf) was then translated into developmental stages for D. rerio on the vertical left axis. Developmen- tal stages of Lythrypnus dalli (vertical left axis) were determined and plotted for various time points (in hpf on the horizontal axis) at 21°C, 17°C and 14°C. Intercepts of the linear trend lines were set to 0, representing fertilization and the onset of development. Both L. dalli and D. rerio display a linear relationship between time and developmental stage, thereby making it possible to estimate developmental stage based on time and temperature for L. dalli (y = 0.0189*temp*hpf, R2 = 0.914, R2 = 0.932 for 17°C and 14°C respectively). L. dalli takes 2.3–3 times longer than D. rerio to reach the protruding mouth stage, depending on the temperature. eyes, the appearance of the swim bladder just reaching a height to width ratio of 2:1. The fin posterior to the yolk, and a constriction in the buds arch posteriorly, defining the “long-pec” posterior region of the gut. stage. At this stage, blood circulation is noticeably strong, and the eyes contain iridescent cells (Fig. 6a). Hatching The ‘pec-fin’ stage is characterized by the expan- The beginning of the hatching period is character- sion of the apical ectodermal ridge in the pectoral ized by the elongation of the pectoral fin buds fins into the paddle shaped apical fold. The blood

Table 1 Heterochronic developmental characteristics of One intriguing difference between L. dalli and D. Danio rerio and Lythrypnus dalli rerio at this stage of lateral line development is the relative size of the cupula. The cupula contain kino- Stage Stage cilia that project distally from hair cells at the base L. dalli D. rerio Characteristic in in of each neuromast (Metcalfe, Kimmel & Schabtach Tail detaches from yolk 9-somite 15-somite 1985). While the neuromasts emerge during the Head detaches from yolk 19-somite Prim-20 pharyngula period, the cupula become visible Optic placode 6-somite 5-somite around 4 days after hatching in D. rerio. In con- Otic placode 9-somite 14-somite trast, within one day of hatching in most L. dalli lar- Otic vesicle with otoliths 12-somite 20-somite Lens 12-somite 20-somite vae, four pairs of cupula visibly protrude from the Chevron-shaped somites 12-somite 14-somite trunk lateral line. Up to eight pairs of cupula are vis- Distinct cerebellum 19-somite 26-somite ible along the head and trunk lateral lines of L. dalli Eye pigmentation Prim23–29 Prim-15 by two days post hatching (Fig. 8). Interestingly, Trunk melanocytes Prim23–29 Prim-15 the cupula of L. dalli are 2–7 times longer than Heartbeat Prim9–14 Prim-15 Muscle movement Prim9–14 26-somite those of D. rerio, although the larvae themselves are Swim bladder High-pec High-pec smaller. Finally, the cupula of L. dalli are visible on Pectoral fin bud Prim9–14 Prim-25 a stereoscope with the correct lighting, making Iridescent eye pigment Long-pec Long-pec them significantly easier to observe. Visible cupula 1 dph 4 dph By one-day post hatching, the lower and upper Grey boxes indicate the species in which the trait appears at jaws protrude equally anterior (Fig. 6e). The devel- an earlier stage. opment of the jaws suggests readiness for feeding. cells have a distinct red pigmentation, and muscle The gape size of L. dalli at hatching is roughly l l twitches are strong (Fig. 6b). 110 m tall by 90 m wide. The yolk is largely The protruding mouth stage is defined by the depleted by 4-5 days post hatching at 21°C, extension of the lower jaw anterior to the eyes. at which time food encounters are crucial for The head and jaws experience major rearrange- survival. ments, which are evident in the changing shape of the head (Figs 6a–d, 7). The nares appear just dor- Larval morphogenesis and settling sal to the upper jaw at this stage. The pectoral fins have expanded to cover most of the reduced yolk, Flexion occurs between 18 and 25 days post and these paddle-like fins are flapped during bouts hatching, after which caudal fin rays develop, sim- of swimming of hatched larvae. The larvae begin ilar to D. rerio (Parichy et al. 2009). Mesenchymal to hatch during the protruding mouth stage in condensations and formation of the fin rays within both L. dalli and D. rerio. Just after hatching, L. the median fin fold quickly follow to form the anal dalli larvae are under 3.0 mm long (x = 2.80 mm, and second dorsal fins (SL = 3.96 mm, SE = 0.09) SE = 0.097). (Fig. 9a). The most noticeable difference between D. rerio The paired pelvic buds emerge after the fin rays and L. dalli during the hatching period is that in L. have formed in the medial fins including the sec- dalli, the inflation of the swim bladder causes a ond dorsal, anal and caudal fins (SL = 5.68 mm, – noticeable bend in the neural tube. This bend SE = 0.48) (Fig. 9b, e f). At this stage, several becomes more prominent during the larval period rows of neuromasts are clearly visible on the cau- (Figs 6c–e, 7). dal fin, which does not occur in D. rerio. Lythrypnus dalli has two dorsal fins, and the first dorsal fin emerges from the dorsal midline anterior Post-hatching and Early Larval Period to the midline fin soon after the emergence of the By the early larval period, significant changes in pelvic fin buds (SL = 7.35 mm, SE = 0.11) the neurocranium and jaw structures result in a (Fig. 9c), similar to what has been described in dramatic reshaping of the head. The jaws appear Rhinogobius sp. BI (Yokoi & Hosoya 2006). Most to gape in most embryos, and the antero-dorsal acanthomorphs have two dorsal fins, but the slope of the head slants forward to the gaping molecular basis of the differentiated first dorsal fin mouth (Figs 6e, 7). has not been studied.

(a) (b) (c) (d) (e)

(f) (g) (h) (i) (j)

(k) (l) (m) (n)

(o)

(p)

(q)

Figure 5 Early embryogenesis of Lythrypnus dalli. The zygotic (a), cleavage (b–e), blastula (f–g), gastrula (h–j), segmentation (k-n) and pharyngula (o–q) periods. (a) In the 1-cell stage, the blastomere of L. dalli wraps around the yolk. This continues through the (b) 2-cell and (c) 4-cell stages. Although the cells no longer wrap around the yolk, they remain in a single layer through the (d) 16-cell and (e) 32-cell stages. Cells are clearly perched on top of the yolk from (f) the 128-cell stage through (g) the high-cell stage. Cell movement begins during early epiboly. The layer of cells is uniform in thickness through (h) 50% epiboly. (i) After the formation of the germ ring, there is a clear asymmetry to the embryo. This asymmetry in cell thickness continues throughout epiboly, including (j) 90% epiboly. (k) The optic placode is visible by the 6-somite stage, despite the grainy texture of the embryo. (l) The tail bud lifts off the yolk and the anterior somites are chevron-shaped by the 9-somite stage. (m) The lens and otic pla- codes are visible by the 12-somite stage. (n) The grainy texture of the embryo is gone by the 19-somite stage, the head has begun to lift off the yolk, and the cerebellum demarcating the division of the mid- and hind-brain is evident (arrowhead). (o) Pigmentation appears on the ventral side of the trunk by the prim 9–14 stage. (p) The pectoral fin buds emerge (arrowhead) and pigmentation occurs in the eyes and dorsal trunk by the prim23–29 stage. (q) The pectoral fin buds develop an apical ectodermal ridge, the swim bladder appears and a neuromast develops on the 29 somite (arrow) by the high-pec stage. All scale bars are 0.50 mm. The first vertical scale bar applies to a–j. The second vertical scale bar applies to k–n. The horizontal scale bar applies to o–q.

Fin rays are evident in the pelvic fins by SL = anterior edges of the pelvic fins fuse together to 7.67 mm (SE = 0.15) (Fig. 9c, g), but the pelvic complete the sucking disc structure. A similar pro- fins do not fuse to form the adult sucking disc cess is pictured in Rhinogobius sp. BI (Yokoi & Ho- until approximately 11.06 mm SL. The fusion of soya 2006). the pelvic fins occurs in two stages. First, the pos- Juvenile L. dalli began to settle out of the water col- terior portions of the two fins fuse together umn starting around 40 dph and concurrent with (Fig. 9g), likely as a result of the outgrowth of the the onset of adult colouration, after the development fins as a single structure at later stages, although of the pelvic fin sucking disc (Fig. 10), as has been this has not been observed directly. Next, the described for other Gobiidae with pelagic larvae

(a)

Figure 6 Hatching period. (a) By (b) the long-pec stage, the hindgut pinches off from the midgut (arrowhead), the eyes develop some iridescence, and the ratio of pectoral fin height to width is (c) about 2:1. (b) The pectoral fin takes on a paddle shape by the pec-fin stage. (c) The lower jaw protrudes past the anterior edge of the eye and the neural tube begins to bend as the swim bladder (d) inflates (arrow) by the protruding mouth stage. (d) The mouth gapes open and most larvae have hatched by the early larval stage. (e) Both jaws extend anteriorly, (e) and the angle of the bend in the neural tube decreases noticeably in the larval stage. Scale bar is 0.50 mm and all photos are to scale.

(a) In situ hybridization

We successfully performed whole mount in situ hybridizations on embryonic L. dalli and D. rerio (b) for hoxa13a. The two species display strikingly similar patterns of expression of hoxa13a in the trunk, tail bud, gut and pectoral ﬁn buds during the stages of embryonic development examined (Fig. 11).

Discussion

Lythrypnus dalli is interesting to aquarists and evolutionary developmental biologists for its bright colouration, bidirectional sex change, modified pelvic fins, size and patterning of neuromasts Figure 7 Two different angle measurements proved and alternative male breeding morphs. In order helpful in staging L. dalli. a) The head-trunk angle to promote research and rearing of this derived (HTA), shown in red, and the forehead slope, shown in teleost for comparative studies of vertebrate blue. (b) The HTA of L. dalli increases in a linear fash- development and to facilitate large-scale produc- ion similar to D. rerio, except at an earlier stage. At the tion to supply the aquarium trade, an optimized end of the pharyngula period, the jaws of L. dalli pro- rearing protocol and developmental staging trude anteriorly causing an increase in the slope of the scheme were developed. Using this rearing proto- forehead. col, 300–500 juvenile L. dalli can be reared in each of several tanks at once, allowing for the (Yokoi & Hosoya 2006; Kondo, Maeda, Yamasaki & production of large numbers of individuals for Tachihara 2012). There was considerable divergence research or aquarium use. in settlement time with some larvae taking Some of the difficulties breeders have faced in 90–100 days to settle. The juvenile period lasted rearing diminuitive species such as L. dalli and roughly 1–2 months, at which time most individuals Priolepis nocturna include determining optimal were sexually mature. This timing is similar to D. breeding and rearing temperatures, and culturing rerio, which reach sexual maturity in 10–12 weeks appropriately small foods for onset of first feeding (Westerfield 2000). (Wittenrich et al. 2007). We successfully reared

(a)

(b) (c)

Figure 8 The neuromasts and cupula of the head and trunk lateral lines of L. dalli are prominent in recently- hatched larvae. (a) Eight pairs of neuromasts are marked with dots on a 2 day post-hatching larvae. Three pairs are located anterior to the pectoral ﬁns on the head lateral line (blue dots), and ﬁve pairs are on the posterior lateral line (red dots). (b) The neuromasts are clearly visible bumps along the posterior lateral line from a dorsal view (arrowheads). (c) The cupula containing the kinocilia extend at least a body width perpendicular to either side of the trunk (arrows).

(a) (e)

(b) (f)

(d)

Figure 9 Larval development post flexion. (a) Flexion is concurrent with development of the caudal, anal and second dorsal fin rays (arrowheads) within mesenchymal condensations of the midline fin. (b) Three rows of neuromasts appear on the caudal fin prior to the development of the first dorsal fin. Two pelvic fin buds emerge ventral to the pectoral fins (arrow). (c) Fin rays appear in the pelvic fin buds (arrow) and the first dorsal fin emerges anterior to the second dorsal fin (arrowhead). (d) The pelvic fin rays elongate posteriorly while the caudal fin rays branch. (e) Scanning electron micrograph of the ventral anterior side of a larva with emerging pelvic fin buds. Note the thoracic placement of the pelvic fins ventral to the pectoral fins. (f) Enlargement of the pelvic fin buds in (e). The pelvic fins emerge as two separate buds that (g) first fuse posteriorly, and later anteriorly (not shown). Photos are of preserved larvae. Anterior is to the left in all photos. Dorsal is up in a–d, while ventral is up in e–g. Standard length (SL) of individuals pictured are given for a–d, while average SL for stage pictured are given for f–g. Scale bars are 1.0 mm. Magnifications are:(e) 144X, (f) 728X and (g) 549X. three generations of L. dalli at 18–21°C, near the nauplii in, while allowing for water and waste top of their reported natural temperature range exchange. This protocol is readily adaptable for (18–22°C, FishBase 2013). We also observed rearing other gobies or small marine ornamental successful reproduction and development through species. early larval stages at a range of temperatures, A clear understanding of the evolutionary his- from 14°C–22°C, suggesting that this species can tory of ontogenetic characters is crucial to the be bred and reared at a range of temperatures. interpretation of results from developmental The relatively small gape size of L. dalli at hatch- research (Metscher & Ahlberg 1999). To build a ing and at exogenous feeding have posed a chal- phylogenetic context for multiple ontogenetic char- lenge, as common larval marine fish foods such acters, developmental studies that include a vari- as artemia and rotifers (>120 lm) are too large ety of vertebrate taxa such as derived teleosts are to consume at first feeding. We therefore devel- essential. Studying development within a phyloge- oped and optimized the protocol for the co-cul- netic context identifies interesting developmental turing of copepod nauplii and L. dalli larvae in and evolutionary novelties and the appropriate tanks designed to keep the larvae and copepod organisms in which to study them. Our work is

(a)

(b)

Figure 10 Pigment pattern devel- (c) opment in L. dalli. (a) Pigmentation prior to settling is minimal ensur- ing that larva are relatively transparent. (b) Once the pelvic ﬁns are well developed and larvae begin to settle, pigmentation begins to ﬁll. (c) The adult pigmentation pattern becomes evident soon after settling occurs.

(a) (b)

Figure 11 Expression of hoxa13a in the hindgut and tail of Lythrypnus dalli and Danio rerio. Whole mount in situ hybridization reveals hoxa13a expression in the posterior gut (arrowheads) and tail bud of (a) prim 9–14 stage L. dalli embryos and (b) prim-10 stage D. rerio embryos. Scale bar is 0.50 mm and all photos are to scale. the ﬁrst description of development of a member of Our comparative analysis has highlighted novel- the genus Lythrypnus, and strengthens the phylo- ties in L. dalli, such as the blastodisc shape and genetic context for ontogenetic characters within large larval cupula. The blastodisc shape of L. dalli gobiid species (Sunobe & Nakazono 1987; Gil, may be due to the oblong shape of the chorion. Goncalves, Faria, Almada, Baptista & Carreiro However, this unique blastodisc shape has not been 1997; Privileggi, Ota & Ferrero 1997; Arakawa, mentioned in other gobiid species, and importantly Kanno, Akiyama, Kitano, Nakatsuji & Nakatsuji not in Priolepis nocturna, a member of a closely 1999; Borges, Faria, Gil, Goncßalves Emanuel & Al- related clade to Lythrypnus (Wittenrich et al. mada 2003; Olivotto, Zenobi, Rollo, Migliarini, Av- 2007), despite the conserved fusiform shape of the ella & Carnevali 2005; Yokoi & Hosoya 2006; chorion. Interestingly, the third cleavage in Wittenrich et al. 2007; Meirelles et al. 2009; another gobiid species, Leucopsarion petersii, occurs Kondo et al. 2012). In addition, our ability to in the horizontal plane, which results in two layers show comparable expression of hoxa13a during of cells at the 8-cell stage (Arakawa et al. 1999). the formation of the gut and trunk in D. rerio and In comparison, the third cleavage plane is used L. dalli using in situ hybridization suggests that L. during the sixth cleavage in D. rerio and L. dalli dalli is also tractable for studies of gene expression. (Kimmel et al. 1995). These cells are therefore

14 © 2015 John Wiley & Sons Ltd, Aquaculture Research, 1–18 Aquaculture Research, 2015, 1–18 Larviculture and development of the Bluebanded goby S Archambeault et al. perched on top of the yolk, which may be an alter- shared feature of gobiid embryos, again perhaps native adaptation to the fusiform chorions of the associated with the shape of the chorion. Gobiidae. Other gobiids may share or have addi- The pelvic fins of L. dalli are of interest for sev- tional strategies to deal with the size and shape eral reasons. First, they articulate in the thoracic limitations of their chorions. region as is common in many acanthomorphs, rel- Temperature plays a critical role in develop- ative to the abdominal pelvic fins of D. rerio. Sec- mental timing, hatching success and survival of ond, the pelvic fins of L. dalli are fused on both the fish embryos (Schirone & Gross 1968; Kimmel posterior and anterior edges, as are the basiptery- et al. 1995; Kucharczyk, Luczynski, Kujawa & gia, the internal bony pelvic girdle supporting the Czerkies 1997; Bermudes & Ritar 1999; Arenzon, pelvic fins. These modifications to the pelvic fins Lemos & Bohrer 2002; Yang & Chen 2005; Ahn, form a sucking disc – a unique morphological Yamada, Okamura, Horie, Mikawa, Tanaka & character of many gobies – in a similar manner as Tsukamoto 2012). Higher temperatures accelerate described in Rhinogobius sp. BI (Yokoi & Hosoya development and reduces time to hatching in 2006). These pelvic fin modifications make mem- many species, including D. rerio (Kimmel et al. bers of the Gobiidae, including L. dalli, interesting 1995). At temperatures outside this range, higher taxa in which to study the evolution of fin modifi- rates of malformation and lower rates of survival cations and body plan diversity. are common (Kimmel et al. 1995). We observed An interesting developmental trait of L. dalli is shorter times to hatching at higher temperatures the pronounced neuromasts and cupula of the for Lythrypnus dalli in the range tested (14.4°C– hatching-stage embryos and early larvae. We 21.6°C), and did not observe higher rates of mal- found eight pairs of neuromasts and corresponding formation or death, suggesting that L. dalli devel- cupula on newly hatched larvae, which extended opment progresses normally within this up to a body width out of the neuromast. Compar- temperature range. Danio rerio occur in warmer atively, the neuromasts of D. rerio are much smal- waters than L. dalli, and therefore are kept in ler and difficult to find. Similarly pronounced, and warmer water, develop faster and hatch sooner comparable numbers of cupula have been reported than L. dalli in the lab. in two Eviota species, E. abax and E. storthynx Interestingly, despite the temperature and over- (Sunobe & Nakazono 1987), but were specifically all time to hatching differences in L. dalli and D. not found in a more closely related species to Lyt- rerio, we found that both species progress through hrypnus, Priolepis nocturna (Wittenrich et al. 2007). larval and juvenile stages in a linear fashion. We found no other reports of these cupula in lar- While neither species allocates more time to any val gobiids, however, preservation in 5% formalin particular developmental stage based on the stag- may destroy them (Sunobe & Nakazono 1987). It ing criteria proposed here, and following Kimmel would be interesting to further examine the preva- et al. (1995), there were several variations in the lence of these large cupula and neuromasts in gob- timing of specific characters, which appeared at iid species as well as in a variety of pelagic and different developmental stages between the taxa demersal larval fishes. (Table 1). For example, the otic placode, otoliths Several distinct lineages within the Gobiidae and lens develop at earlier developmental stages in family are bidirectional sex changers (St. Mary L. dalli, and eye pigmentation occurs earlier in D. 1993; Sunobe & Nakazono 1993). This is a rare rerio, indicating variation in timing of individual characteristic in teleosts and represents an extreme organs with respect to developmental stage. Inter- in sexual plasticity. Little is known about the estingly, some characteristics of the segmentation development of the gonads of these species, but period develop at similar stages in the ice goby, the limited data suggest similar molecular path- but at later stages in D. rerio (Kimmel et al. 1995; ways are involved in initial gonad development Arakawa et al. 1999), which suggests an evolu- and the reallocation process of gonadal tissue dur- tionary component to these differences in develop- ing sex change (Miyake, Sakai & Kuniyoshi 2012). mental timing. In addition, L. dalli are smaller In L. dalli, preliminary data indicate that juveniles than D. rerio at all embryonic stages and times, develop initially into a bipotential phase where with the exception of the 14-somite stage, likely gonads contain both eggs and sperm and genitalia due to the lifting of the head and tail bud off the are ambiguous (Lorenzi and Grober, unpublished yolk at an earlier stage, which appears to be a data), followed by social regulation of initial sexual

© 2015 John Wiley & Sons Ltd, Aquaculture Research, 1–18 15 Larviculture and development of the Bluebanded goby S Archambeault et al. Aquaculture Research, 2015, 1–18 development that is responsive to social status of the annual fish Cynopoecilus melanotaenia (Cyprin- (Solomon-Lane, et al., in preparation) in the same odontiformes, Rivulidae). Brazilian Journal of Biology way that has been shown for adults (Rodgers et al. 62, 743–747. 2007). Bermudes M. & Ritar A.J. (1999) Effects of temperature on the embryonic development of the striped trumpeter Lythrypnus dalli is a small ornamental fish (Latris lineata Bloch and Schneider, 1801). Aquaculture that is easily maintained in a laboratory setting. 176, 245–255. Furthermore, adults readily lay eggs in captivity, Black M.P., Reavis R.H. & Grober M.S. (2004) Socially and are of interest for studies of sexual plasticity, induced sex change regulates forebrain isotocin in Lyt- sex change and alternative reproductive strategies. hrypnus dalli. NeuroReport 15, 185–189. This taxon is tractable and evolutionarily interest- Black M.P., Balthazart J., Baillien M. & Grober M.S. ing for studies of developmental traits and novel- (2005) Socially induced and rapid increases in aggres- ties, such as the early neuromasts of the lateral sion are inversely related to brain aromatase activity line and modified pelvic fins. This work provides a in a sex-changing fish, Lythrypnus dalli. Proceedings useful staging scheme that can be used to study Biological Sciences/The Royal Society 272, 2435–2440. the development of several novel features in a Black M.P., Balthazart J., Baillien M. & Grober M.S. (2011) Rapid increase in aggressive behavior precedes representative derived teleost. the decrease in brain aromatase activity during socially mediated sex change in Lythrypnus dalli. Gen- eral and Comparative Endocrinology 170, 119–124. Acknowledgments Borges R., Faria C., Gil F., Goncßalves Emanuel J. & Al- We thank Darren Gewant for his support and mada V.C. (2003) Embryonic and larval development expertise. Funding for this research was provided of Gobius paganellus (Pisces: Gobiidae). Journal of the by the National Science Foundation (proposal num- Marine Biological Association of the UK (JMBA) 83, – ber 1022509 and grant number DUE-0728279 1151 1156. Collazo A. & Bolker J.A. (1994) A phylogenetic perspective awarded to Dr Karen Crow and San Francisco State on teleost gastrulation. American Naturalist 144, 133. University, respectively; # IOB0548567 to Mat- Creswell L (2010) Phytoplankton Culture for Aquaculture thew Grober and a Doctoral Dissertation Improve- Feed, Southern Regional Aquaculture Center, SRAC ment Grant (DDIG) #1311303 to Tessa Solomon- publication No. 5004. September 2010. Lane) and a Sigma Xi Grant-in-Aid of Research to Drilling C. & Grober M. (2005) An initial description of Tessa Solomon-Lane. Funding was also provided by alternative male reproductive phenotypes in the blue- the Institute of Museums and Library Sciences for banded goby, Lythrypnus dalli (Teleostei, Gobiidae). the work carried out at Roger Williams University. Environmental Biology of Fishes 72, 361–372. The authors declare no competing interests. Eschmeyer W. & Fricke R. (2012) Catalog of Fishes. Cali- fornia Academy of Sciences, San Francisco. FishBase (2013) World Wide Web electronic publication, Available at: http://www.fishbase.org, version References (accessed 06/2013). Ahn H., Yamada Y., Okamura A., Horie N., Mikawa N., Gil M.F., Goncalves E.J., Faria C., Almada V.C., Baptista Tanaka S. & Tsukamoto K. (2012) Effect of water tem- C. & Carreiro H. (1997) Embryonic and larval develop- perature on embryonic development and hatching time ment of the giant goby Gobius cobitis (Pisces: Gobii- of the Japanese eel Anguilla japonica. Aquaculture dae). Journal of Natural History 31, 799–804. 330–333, 100–105. Kendall A.W., Ahlstrom E.H. & Moser H.G. (1984) Early Andrews C (1990) The ornamental fish trade and fish life history stages of fishes and their characters. In: conservation. Journal of Fish Biology 37,53–59 Ontogeny and Systematics of Fishes: Based on an Inter- Arakawa T., Kanno Y., Akiyama N., Kitano T., Nakatsuji national Symposium Dedicated to the Memory of Elbert N. & Nakatsuji T. (1999) Stages of Embryonic Develop- Halvor Ahlstrom/Sponsored by the National Marine Fish- ment of the Ice Goby (Shiro-uo), Leucopsarion petersii. eries Service, National Oceanic and Atmospheric Adminis- Zoological Science 16, 761–773. tration (ed. by H.G. Moser, W.J. Richards, D.M. Archambeault S., Taylor J. & Crow K. (2014) HoxA and Cohen, M.P. Fahay, A.W. Jr Kendall & S.L. Richard- HoxD expression in a variety of vertebrate body plan son), pp. 11–22. United States Dept. of Commerce. features reveals an ancient origin for the distal Hox American Society of Ichthyologists and Herpetologists, program. EvoDevo 5, 44. New York, USA. Arenzon A., Lemos C.A. & Bohrer M.B.C. (2002) The Kimmel C.B., Ballard W.W., Kimmel S.R., Ullmann B. & influence of temperature on the embryonic development Schilling T.F. (1995) Stages of embryonic development

of the zebrafish. Developmental Dynamics 203, 253– Figla and Dmrt1, in the protogynous hermaphroditic 310. fish, halichoeres poecilopterus. Zoological Science 29, Kondo M., Maeda K., Yamasaki N. & Tachihara K. 690–701. (2012) Spawning habitat and early development of Moe M.A. Jr (2008)Culture of Marine Ornamentals: for Luciogobius ryukyuensis (Gobiidae). Environmental Love, for Money, and for Science In: Marine Ornamental Biology of Fishes 95, 291–300. Species: Collection, Culture and Conservation (ed. by JC Kucharczyk D., Luczynski M., Kujawa R. & Czerkies P. Cato & CL Brown), pp. 11–28. Wiley Blackwell, Ames, (1997) Effect of temperature on embryonic and larval Iowa, USA. development of bream (Abramis brama L.). Aquatic Sci- Moyle P.B. & Cech J.J. Jr (2004) Fishes: An Introduction to ence 59, 214–224. Ichthyology. Pearson Benjamin Cummings, San Francisco. Lorenzi V. & Grober M.S. (2012) Immunohistochemical Nakatsuji T., Kitano T., Akiyama N. & Nakatsuji N. localization of serotonin in the brain during natural sex (1997) Ice goby (shiro-uo), Leucopsarion petersii, may change in the hermaphroditic goby Lythrypnus be a useful material for studying teleostean embryo- dalli. General and Comparative Endocrinology 175, genesis. Zoological Science 14, 443–448. 527–536. Near T.J., Eytan R.I., Dornburg A., Kuhn K.L., Moore Lorenzi V., Earley R.L. & Grober M.S. (2006) Preventing J.A., Davis M.P., Wainwright P.C., Friedman M. & Behavioural Interactions with a Male Facilitates Sex Smith W.L. (2012) Resolution of ray-finned fish phylo- Change in Female Bluebanded Gobies, Lythrypnus geny and timing of diversification. Proceedings of the dalli. Behavioral Ecology and Sociobiology 59, 715–722. National Academy of Sciences 109, 13698–13703. Lorenzi V., Earley R.L., Rodgers E.W., Pepper D.R. & Nelson J.S. (2006) Fishes of the World. John Wiley & Sons Grober M.S. (2008) Diurnal patterns and sex differ- Inc, Hoboken, New Jersey. ences in cortisol, 11-ketotestosterone, testosterone, Olivotto I., Zenobi A., Rollo A., Migliarini B., Avella M. & and 17beta-estradiol in the bluebanded goby (Lyt- Carnevali O. (2005) Breeding, rearing and feeding hrypnus dalli). General and Comparative Endocrinology studies in the cleaner goby Gobiosoma evelynae. Aqua- – 155,438 446. culture 250, 175–182. Lorenzi V., Carpenter R.E., Summers C.H., Earley R.L. & Olivotto I., Holt S.A., Carnevali O. & Holt G.J. (2006) Grober M.S. (2009) Serotonin, social status and sex Spawning, early development, and first feeding in the change in the bluebanded goby Lythrypnus dalli. Phys- lemonpeel angelfish Centropyge flavissimus. Aquacul- – iology & Behavior 97, 476 483. ture 253, 270–278. Lorenzi V., Earley R.L. & Grober M.S. (2012) Differential Olivotto I., Planas M., Simoes N., Holt G.J., Avella M.A. responses of brain, gonad and muscle steroid levels to & Calado R. (2011) Advances in breeding and rearing changes in social status and sex in a sequential and marine ornamentals. Journal of the World Aquaculture bidirectional hermaphroditic fish. PLoS ONE 7, e51158. Society 42, 135–166. Maxfield J.M., Van Tassell J.L., St Mary C.M., Joyeux J.C. Parichy D.M., Elizondo M.R., Mills M.G., Gordon T.N. & & Crow K.D. (2012) Extreme gender flexibility: using a Engeszer R.E. (2009) Normal table of postembryonic phylogenetic framework to infer the evolution of varia- zebrafish development: staging by externally visible tion in sex allocation, phylogeography, and speciation anatomy of the living fish. Developmental Dynamics: An in a genus of bidirectional sex changing fishes (Lyt- Official Publication of the American Association of Anato- hrypnus, Gobiidae). Molecular Phylogenetics and Evolu- mists 238, 2975–3015. tion 64, 416–427. Patzner R, Van Tassell J.L., Kovacic M & Kapoor B.G. Meirelles M.E., Tsuzuki M.Y., Ribeiro F.F., Medeiros R.C. (2011) The Biology of Gobies. In: The Biology of Gobies & Silva I.D. (2009) Reproduction, early development (ed. by R. Patzner, J.L. van Tassell, M. Kovacic & B.G. and larviculture of the barber goby, Elacatinus figaro – (Sazima, Moura & Rosa 1997). Aquaculture Research Kapoor), pp. 1 650. Science Publishers, Enfield, New 41,11–18. Hampshire, USA. Metcalfe W.K., Kimmel C.B. & Schabtach E. (1985) Anat- Privileggi N., Ota D. & Ferrero E.A. (1997) Embryonic omy of the posterior lateral line system in young lar- and larval development of the grass goby Zosterisessor vae of the zebrafish. The Journal of Comparative ophiocephalus (Teleostei, Gobiidae). Italian Journal of – Neurology 233, 377–389. Zoology 64, 201 207. Metscher B.D. & Ahlberg P.E. (1999) Zebrafish in Con- Rodgers E.W., Drane S. & Grober M.S. (2005) Sex text: uses of a Laboratory Model in Comparative reversal in Pairs of Lythrypnus dalli: behavioral and Studies. Developmental Biology 210,1–14. morphological changes. Biological Bulletin 208, 120– Miller D.J.L.R. (1972) Guide to Coastal Marine Fishes of 126. California. California Department of Fish and Game, Rodgers E.W., Earley R.L. & Grober M.S. (2006) Elevated Fish Bulletin. 11-ketotestosterone during paternal behavior in the Miyake Y., Sakai Y. & Kuniyoshi H. (2012) Molecular Bluebanded goby (Lythrypnus dalli). Hormones and cloning and expression profile of sex-specific genes, Behavior 49, 610–614.

Rodgers E.W., Earley R.L. & Grober M.S. (2007) Social Stiassny M., Wiley E., Johnson G. & Carvalho Md (2004) status determines sexual phenotype in the bi-directional Gnathostome fishes. Oxford Univ Press, New York. sex changing bluebanded goby Lythrypnus dalli. Journal Sunobe T. & Nakazono A. (1987) Embryonic-develop- of Fish Biology 70,1660–1668. ment and larvae of genus Eviota (Pisces, Gobiidae).1. Scaggiante M., Grober M.S., Lorenzi V. & Rasotto M.B. eviota-abax and eviota-storthynx. Journal of the (2004) Changes along the male reproductive axis in Faculty of Agriculture Kyushu University 31, 287– response to social context in a gonochoristic gobiid, Zoster- 295. isessor ophiocephalus (Teleostei, Gobiidae), with alterna- Sunobe T. & Nakazono A. (1993) Sex change in both tive mating tactics. Hormones and Behavior 46,607–617. directions by alteration of social dominance in Schirone R.C. & Gross L. (1968) Effect of temperature on trimma okinawae (Pisces: Gobiidae). Ethology 94, early embryological development of the zebra fish, 339–345. Brachydanio rerio. Journal of Experimental Zoology 169, Sweet T (2014) Captive-breeding: state of the Art 2014. 43–52. In: CORAL (ed. by J.M. Lawrence & D. Knop). Dart- St. Mary C (1993) Novel sexual patterns in two simulta- mouth Printing, Hanover, New Hampshire. neously hermaphroditic gobies, Lythrypnus dalli and Tavolga W.N. (1950) Development of the gobiid fish, Lythrypnus zebra. Copeia, 1993, 1062–1072. Bathygobius soporator. Journal of Morphology 87, 467– St. Mary C (1994) Sex allocation in a simultaneous her- 492. maphrodite, the blue-banded goby (Lythrypnus dalli): Van Tassell J, Lea R & Bearez P (2010) IUCN Red List of the effects of body size and behavioral gender and the Threatened Species; Lythrypnus dalli. consequences for reproduction. Behavioral Ecology 5, Westerfield M. (2000) The Zebrafish Book. A Guide for the 304–313. Laboratory use of Zebrafish (Danio Rerio). University of St. Mary C (1996) Sex allocation in a simultaneous her- Oregon Press, Eugene. maphrodite, the zebra goby Lythrypnus zebra insights Wittenrich M.L., Turingan R.G. & Creswell R.L. (2007) gained through a comparison with its sympatric con- Spawning, early development and first feeding in the gob- gener, Lythrypnus dalli. Environmental Biology of Fishes iid fish Priolepis nocturna. Aquaculture 270, 132–141. 45, 177–190. Yang Z. & Chen Y. (2005) Effect of temperature on St. Mary C (1998) Characteristic gonad structure in the incubation period and hatching success of obscure gobiid Genus Lythrypnus with comparisons to other puffer Takifugu obscurus (Abe) eggs. Aquaculture hermaphroditic gobies. Copeia 1998, 720–724. 246,173–179. St. Mary C (2000) Sex allocation in Lythrypnus (Gobii- Yokoi K-i & Hosoya K. (2006) Early development of dae): variations on a hermaphroditic theme. Environ- the endangered freshwater goby, Rhinogobius sp. BI mental Biology of Fishes 58, 321–333. (Gobiidae). Ichthyological Research 53, 160–165.