Hydrothermal vent animals were kept and displayed at atmospheric pressure with an artificial hydrothermal vent system in Enoshima Aquarium, Fujisawa, Kanagawa, Japan. The artificial hydrothermal vent system was composed of a main rearing tank, a heating tank, a hot water outlet with added Na2S as the source of H2S, and added CO2 for chemosynthetic bacteria and pH regulation. When the need arises, a dissolved oxygen control unit and submersible heaters can be attached. We are now rearing hydrothermal vent crabs (Austinograea yunohana), hydrothermal vent galatheid crabs (Shinkaia crosnieri), vestimentiferan tubeworms (Lamellibrachia satsuma), hydrothermal vent shrimp (Opaepele spp.), hydrothermal vent barnacles (Ashinkailepas seepiophilia and Neoverruca sp.), and tonguefish (Symphurus sp.). In the artificial hydrothermal vent tank, shrimp and crabs have been observed to cluster close to the artificial hydrothermal vent. In particular, large (adult) crabs needed a heat source to live in the aquarium over a long term. Additionally, some species (A. yunohana, S. crosnieri, Opaepele sp. and Symphurus sp.) have spawned and hatched in captivity. It is likely higher water temperatures are needed for egg and larval development compared with temperatures for adult requirements.
The study of hydrothermal vent animals has been progressing since the discovery of hydrothermal vents in 1977 using manned submersibles and ROVs (Corliss & Ballard 1977; Corliss et al. 1979). The study of deep-sea animals including hydrothermal vent and cold seep animals has been conducted in a similar fashion as shallow environment studies but with the use of special tools and/or concepts to deal with high water pressure (Gage & Tyler 1991; Herring 2002). However, there have been a few long-term observations through the rearing of deep-sea animals. Havenhand et al. (2006) kept a predatory tunicate in an aquarium and observed larval development. Miyake & Lindsay (2003) reported on the sampling and rearing methods of deep-sea hydroids. Miyake (2005) reported on the rearing of mid-water plankton. As for cold seep and hydrothermal vent organisms, Miyake et al. (2005) attempted to rear deep-sea giant white clams, Calyptogena soyoae and C. solidissima using an MT push core which was developed to keep Calyptogena in good condition and to re-create their natural environment as closely as possible (except for water pressure). Miura et al. (1997) reported keeping vestimentiferan tubeworms, Lamellibrachia satsuma, in a tank with only regulated temperature and added sodium sulfide. Kádár et al. (2005) reported on relationships between vent mussels Bathymodiolus azoricus and their endosymbionts under cultivation. Tsuchida et al. (1998) observed molting of the vent crab Austinograea yunohana in a long-term rearing period, over 6 years. Watanabe et al. (2004) reared vent barnacles, Neoverruca sp., and observed larval development and intermolting time.
One of the important steps in deep-sea hydrothermal vent biology research is the observation and study of living animals in the laboratory. To accomplish this, deep-sea animals need to be kept available so that researchers can conduct experiments at any time without deep-sea diving. Deep-sea chemosynthetic ecosystem organisms select strictly suitable environments based on a gradient of physico-chemical factors (Gaill & Hunt 1991; Hesseler & Kaharl 1995; Barry et al. 1997; Olu et al. 1997). The rearing of deep-sea hydrothermal vent and cold seep animals is difficult because of problems in maintaining high pressure, low pH, H2S concentration, high CO2 concentration, low dissolved oxygen (DO), no light and low temperature conditions. These conditions are just the opposite of the normal rearing conditions of fishes in aquaria. Here we report on a new method of rearing hydrothermal vent animals at atmospheric pressure and discuss the ecological aspects of hydrothermal vent animals reared in a tank with an artificial hydrothermal vent system. This rearing system represents an important breakthrough in the study of the biology of deep-sea hydrothermal vent animals.
Material and Methods
Hydrothermal vent animals were collected using manipulators and a slurp gun with single or six canisters set on the ROV Hyper-dolphin of JAMSTEC. Hydrothermal vent galatheid crabs, Shinkaia crosnieri (Baba & Williams 1998) and hydrothermal vent shrimp, Opaepele spp. were collected at Izena Hole (27°16′ N, 127°05′ E, 1400 m) and Hatoma Knoll (24°51′ N, 123°05′ E, 1500 m) in the Okinawa Trough. Vestimentiferan tubeworms Lamellibrachia satsuma (Hashimoto et al. 1993; Miura et al. 1997) were collected at the Nikko Seamount (23°05′ N, 142°20′ E, 430 m) on the Izu-Ogasawara Ridge. Hydrothermal vent shrimp, Opaepele sp. the hydrothermal vent crab Austinograea yunohana, and the tonguefish Symphurus sp. (Tsuchida et al. 2001) were collected at Nikko Seamount and Kaikata Seamount (26°42′ N, 141°05′ E, 450 m) on the Izu-Ogasawara Ridge. Hydrothermal vent barnacles Ashinkailepas seepiophilia (Watanabe et al. 2003) and Neoverruca sp. (Watanabe et al. 2003, 2004) were collected at Hatoma Knoll and Izena Knoll on the Okinawa Trough and at Myojin Knoll (32°06′ N, 139°52′ E, 1300 m) on the Izu-Ogasawara Ridge.
Sampled animals were transferred to shipboard tanks immediately after recovery of the ROV. The water temperatures in the tanks were set at 4 °C (S. crosnieri, Opaepele spp., A. seepiophilia, and Neoverruca sp.) and 12 °C (L. sastuma, Opaepelesp., A. yunohana, and Symphurus sp.) to match each animal's respective habitat. Canister filters (EHEIM) were used for water filtration. Rearing water was changed when it became cloudy, i.e. when biological filtration was not working properly at low water temperatures. Subsequently, these animals were transported to rearing tanks at the Enoshima Aquarium by car or air.
The rearing tank included an artificial hydrothermal vent system (Fig. 1). This system is composed of a main rearing tank (Fig. 1A), a filtration tank (Fig. 1B), a heating tank (Fig. 1C), a hot water inlet with added Na2S (Fig. 1D) as a source of H2S, and added CO2 (Fig. 1E) for chemosynthetic bacteria and pH regulation. When the need arises, submersible heaters or a DO control unit (DOCCS) are attached to the heating tank (Miyake et al. 2005). We made two artificial hydrothermal vent tanks. The volume of the main tank was 1000 l for the animals living at 12 °C, and 300 l for the animals living at 4 °C. Austinograea yunohana, Opaepele sp., L. satsuma and Symphurus sp. were kept together in the large artificial hydrothermal vent tank. Shinkaia crosnieri, Opaepele spp, A. seepiophilia, and Neoverruca sp. were kept together in the small artificial hydrothermal vent tank. The main circulation of the rearing water is represented by the bold black line in Fig. 1. The rearing water overflowed to the filtration tank. The filtration material was coral sand. Filtered water was then pumped up to the main tank through the brine cooler to chill the rearing water. Some of the filtered rearing water was branched out to the heating tank. Natural seawater drawn by the Enoshima Aquarium from the sea in front of the aquarium was also poured into the heating tank (refer to the dash line in Fig. 1). The temperature of the added natural seawater varied between 14 °C in winter and 27 °C in summer. Water circulation in the heating tank is represented by the bold orange line in Fig. 1. The filtration material was placed in the heating tank as biological filtration is more efficient in hot water than in chilled water. A small portion of the hot water was pumped up to the main rearing tank. Extra hot water in the heating tank overflowed to the filtration tank. The temperature of the hot water was 20–35 °C. The fluctuation of its temperature depended on the balance of the water volume of the inlet and outlet to the heating tank. Next, 10 ml of a solution of sodium sulfide (500 g Na2S/20 laq) was mixed in the hot water every 10 min using an electromagnetic metering pump (EHN-YT; Rei-Sea Co., Ltd). Carbon dioxide was added by bubbling from carbon dioxide diffusers (Misty 30; Jaqno). The inlet of hot water was covered by fragments of chimney collected at a hydrothermal vent field. Red LED ramps were used for the lighting of the tank during the business hours of the aquarium. There was no light during the closed hours. Sometimes a halogen lamp or white LED lamp was used to observe in detail the condition of the animals.
As for the vent crabs, we kept a Kaikata population of vent crabs Austinograea yunhana in the normal rearing tank in the aquarium with no heat sources. Some female individuals which brooded eggs on the hairs on their abdomen were separated into other tanks with a heat source. The temperatures of these nursery tanks were set at 12 °C and 18–19 °C.
Temperature, salinity, and pH were measured every day using a pH meter (Model PH81; Yokogawa) and a salinity meter (Ecoscan Salt 6; Eutech Instruments). Hydrogen sulfide was measured by the Methylene Blue Method using a DR2400 portable spectrophotometer (Hach Company) as needed. Feeding for vent crabs, shrimp and tonguefish occurred in the mornings on Mondays, Wednesdays and Fridays; the animals were fed filleted fish, defrosted krill, and defrosted mysids.
Observations of the animals (survival, release of larvae or eggs, feeding conditions in vent crabs and shrimp, molting in crustaceans, and mating behavior) were conducted daily. Observations of eggs, embryos or larvae were made under the stereomicroscope. As for the tonguefish, the incubation temperatures were 12, 20 and 26 °C, respectively.
The temperature, salinity, pH, and DO in the rearing tanks (the large artificial hydrothermal vent tank, the small artificial hydrothermal vent tank, and the normal tank without any heat source or artificial hydrothermal system) are given in Table 1.
Table 1. Environmental factors in rearing tanks.
large hydrothermal vent tank
small hydrothermal vent tank
normal tank with no heat source
Dissolved oxygen (ml l−1)
In the artificial hydrothermal vent tank, vent crabs Austinograea yunohana and vent shrimp Opaepele spp. were observed to cluster close to the artificial hydrothermal vent (Fig. 2A and B). In particular, A. yunohana moved their backs into direct contact with submersible heaters and spread themselves over the hot-water vent. The survival rate of A. yunohana kept in the tank without any heat source was 28.6% (n = 18) and the average size of the dead individuals and surviving individuals was 31.1 mm (range 23.2–40.7 mm) and 20.7 mm (range 13.7–26.7 mm), respectively. On the other hand, the survival rate of A. yunohana in the artificial hydrothermal vent tank was 81.1% (n = 38) and the average size of the dead individuals and surviving individuals was 36.7 mm (range 30.0–42.0 mm) and 29.6 mm (range 17.5–47.4 mm), respectively. The dead individuals were of adult size. The surviving individuals lived for more than a year. Females were guarded by the males before molting. After molting, copulation was observed (Fig. 2C). On the other hand, males that had just molted were cannibalized by the other individuals in the aquarium. Some gravid females were collected and some females spawned in the rearing tank. Gravid crabs maintained the eggs on their abdomen at the hot water vent. Larvae hatched from the collected individuals in the 18–19 °C tank (Fig. 2D), but not in the 12 °C tank. Adult crabs had no eyes, however, the hatched larvae had eyes.
Vent shrimp (Opaepele spp. from Izena Hole, Hatoma Knoll, and Nikko Seamount) clustered at the inlets of hot water and grazed on bacterial filaments and mats attached around the inlet of hot water when A. yunohana was not around the vent. However, the vent shrimp did not approach the inlet of hot water when A. yunohana occupied the hot water vent and instead perched on tubeworms L. satsuma. Some shrimp were brooding eggs when originally collected and continued to brood in the tank. Eggs, which came off female pleopods, sank to the bottom. The eggs were orange in color in the early stages of development. Well-developed eggs changed to a whitish-orange color. Hatched larvae in the aquarium swam upward in an upside-down posture (Fig. 2E). However, shrimp larvae could not be reared successfully. Adult shrimp had no eyes, however, the hatched larvae had eyes.
Tonguefish Symphurus sp. (Fig. 2F) were observed in dense colonies around the shimmering hot water at both Kaikata Seamount and Nikko Seamount. This species is the dominant species at these sites, as is A. yunohana. In the artificial hydrothermal vent tank, Symphurus sp. was usually either on the sand and hidden in the sand, or sometimes on rocks or walls of the tank. They fed on any food which was dropped in front of them. Symphurus sp. spawned in the tank. The eggs were buoyant and were observed on the surface of the water. Egg size was about the 0.9 mm. Live eggs were buoyant, but dead eggs sank to the bottom. The higher the incubation temperature rose, the faster was the developmental speed. Larval fish hatched in 1 day at 26 °C, in 3 days at 20 °C and in 14 days at 12 °C. Recently hatched larvae had a yolk sac and did not have a completely developed digestive canal or eyes (Fig. 2G). The digestive canal extended to the anus in 3-day-old larvae. Seven-day-old larvae had completely formed eyes and mouth (Fig. 2H) and ate live rotifers.
Hydrothermal vent galatheid crabs Shinkaia crosnieri (Fig. 2I) had bushy white hair with dense bacterial filaments on their ventral side at the time of sampling, because hot vent water in situ provided a suitable environment for bacteria (i.e. enough hydrogen sulfide and carbon dioxide and low DO concentration). However, a few days after sampling, these dense bacterial filaments had disappeared in the tank with added Na2S and without CO2 or hot water on board. In the artificial hydrothermal vent system, bacterial filaments or mats on the white hairs on the ventral side of S. crosnieri increased. It has been observed that S. crosnieri graze on these bacteria using their mouth-parts. Shinkaia crosnieri did not cluster around the artificial vent. But S. crosnieri was fatally weakened when carbon dioxide was added accidentally in a large volume and the pH of the tank decreased to 6.0. Some S. crosnieri were brooding eggs when they were collected. The size of the eggs was about 2.9 mm in length and 2.2 mm in width. Eggs from the female pleopods were buoyant and rose to the water surface. The eggs of one individual hatched. The size of the larvae was about 6.6 mm and larvae were white in color (Fig. 2J). Larvae were buoyant as were the eggs, and became trapped on the water surface. Adult individuals had no eyes, however the hatched larvae had eyes.
Additionally, hydrothermal vent barnacles, Ashinkailepas seepiophilia (Fig. 2K) and Neoverruca sp. (Fig. 2L) had the cirris feeding legs covered with bacteria just after sampling from their habitat. A few days after sampling, bacteria on their cirris disappeared in an H2S-only environment. In the artificial hydrothermal vent tank, bacteria increased on the cirris as seen in situ. The growth of both species could be observed clearly in the Myojin population samples, as both species of the Myojin population had black pigments deposited on their bodies just after collection. Additionally, the newly grown parts of the bodies in the artificial tank were white.
The rearing of deep-sea animals had previously been carried out using pressure cylindrical aquaria for a span of a few days to a few months (Gaill et al. 1997; Goffredi et al. 1997; Shillito et al. 1997, 1999; Holden 1998; Girguis et al. 2000). Soon, Koyama et al. (2002) developed the Deep-Aquarium to capture and rear animals at their normal in situ pressure. However, these systems use expensive and sophisticated components to make high-pressure environments. In our system, these components are not needed at atmospheric pressure. Our system is much less expensive, allowing researchers to make systems quickly as needed.
In the previous reports regarding the rearing of hydrothermal vent animals at atmospheric pressure, sodium sulfide was added to the rearing tanks for endosymbiotic bacteria, epibiotic bacteria or free-living chemosynthetic bacteria (Hashimoto et al. 1993; Miura et al. 1997; Watanabe et al. 2004). However, carbon dioxide was not added in the previous reports, despite being an important carbon source. The appearance of bacterial mats on the hairs of S. crosnieri and cirri of Neoverruca sp. was faster and bacteria were more abundant in the tanks to which carbon dioxide was added rather than in tanks with added hydrogen sulfide, but not carbon dioxide. The addition of carbon dioxide and hydrogen sulfide appears to be essential for rearing animals in chemosynthetic ecosystems.
Furthermore, adding hot water along with hydrogen sulfide and carbon dioxide is required to successfully rear hydrothermal vent animals. In particular, Austinograea yunohana needed a heat source to survive in the tanks. Adult A. yunohana died in tanks without any heat source despite the juvenile individuals surviving in such tanks. It may be that the juvenile vent crabs do not need a heat source. Bathograeid crabs in situ release their larvae on bare rocks away from hydrothermal vent areas (Van Dover 2000). This behavior may be employed to escape predation of larvae. Vent crabs have voracious appetites in situ and in aquaria. Dropped eggs from the abdomens of gravid crabs were eaten by other vent crabs in the aquaria. Juvenile crabs may grow to a size large enough to escape cannibalism outside of the hydrothermal vent areas. Hydrothermal vent crabs need to constantly move and re-colonize other hydrothermal vents that are distributed in patches on the deep-sea floor, but dispersed larvae do not always reach new hydrothermal vents. In this case, megalope or juvenile crabs may be able to survive in non-vent environments. If the juveniles are dispersed near hydrothermal vents, they may be able reach new vents before growth to adult stages, which requires a heat source. This difference between the adult and juvenile vent crabs may be an adaptive strategy.
On the other hand, Shinkaia crosnieri did not cluster around the hot vents. The hot vent in the tank was at a very low temperature compared with the hydrothermal vents in S. crosnieri’s habitat. There were Paralvinella hessleri in the habitat of S. crosnieri. It appears that S. crosnieri may actually need higher temperature vents to cluster around the vent in the aquarium. However, S. crosnieri were kept for over 1 year; and thus a minimum condition to keep this species long term is making an environment that is sufficient for epibionts to grow on the ventral hairs at 4–5 °C.
Larvae of S. crosnieri were buoyant. Similarly, the larval development of tonguefish Symphurus sp. was faster at high temperatures and the eggs and larvae were buoyant. In situ, this buoyancy takes the larvae or eggs up to the mid-water range where the temperature is higher compared with the temperature at the seafloor. Similarly, the larvae of A. yunohana hatched in our high-temperature tank. Heat sources or high temperatures may be essential for the development of embryos and the growth of larvae of many hydrothermal vent organisms.
The artificial hydrothermal vent system in this study allows us to observe the behavior and modes of the life of hydrothermal vent animals as if they were in situ. This system is an important breakthrough in hydrothermal vent biology, allowing for long-term observation in aquaria.
We sincerely thank the captain and crew of the R/V Natsushima and the commander, pilots, and operations team of the ROV Hyper-Dolphin for their dedicated efforts. We also thank Dr James D. Reimer of JAMSTEC for his editing assistance and valuable comments. Finally, we appreciate the useful comments regarding the manuscript from the editors and anonymous reviewers.