field surveys on the indonesian coelacanth, …4seikei education and research center for sustainable...

8
INTRODUCTION The first living coelacanth, Latimeria chalumnae, was discovered in South Africa in the Indian Ocean in 1938 (SMITH, 1939). The first observation of the living coelacanth habitats using submersibles was reported in Comoros (FRICKE et al., 1987). Latimeria chalumnae hide in submarine caves along steep slopes between 100 m and 200 m depth during daytime and come out and drift near the ocean floor for hunting prey at night (FRICKE et al., 1991). However, since the record of juvenile or small (below 80 cm long) individuals is virtually absent, the growth, development and breeding biology of L. chalumnae remain poorly known. Our knowledge of the reproduction biology of the coelacanths is almost restricted to an ovoviviparity strategy since some caught large females carried unfertilized eggs or embryos in their oviduct (SMITH et al., 1975; CUPELLO et al., 2015). The Indonesian coelacanth, L. menadoensis, was first discovered in Manado, Sulawesi Island, Indonesia, in 1997. The first specimen was captured in 1998 (ERDMANN et al ., 1998). According to a genetic study, the specimen differed from L. chalumnae and a new species was erected (POUYAUD et al ., 1999). In 1999 two individuals were observed by submersible in a location 360 km away from Manado (FRICKE et al., 2000). Today 8 individuals were officially caught in the Indonesian seas. However, the details of their habitat have not Bull. Kitakyushu Mus. Nat. Hist. Hum. Hist., Ser. A, 17: 49–56, March 31, 2019 Field surveys on the Indonesian coelacanth, Latimeria menadoensis using remotely operated vehicles from 2005 to 2015 ABSTRACT Habitats of the Indonesian coelacanth, Latimeria menadoensis, were investigated by Remotely Operated Vehicles (ROVs) surveys in the northern coast of Sulawesi Island and southern coast of Biak Island by collaboration of Aquamarine Fukushima (Japan), and Indonesian Institute of Sciences and Sam Ratulangi University (Indonesia) from 2005 to 2015. The Remotely Operated Vehicles operations were conducted 1173 times and coelacanths were encountered 30 times. A total of 30 different individuals were observed at a depth range from 115.6 m to 218.9 m deep. The water temperature was between 12.4 to 21.5 ºC. Most of the individuals were found alone, however, schools of two, three and six individuals were also observed. The Indonesian coelacanth, L. menadoensis, was observed at a similar depth of the African coelacanth, Latimeria chalumnae, and the temperature range also seems to be very similar. However, Latimeria menadoensis was sometimes observed besides big rocks or the steep wall. It seems to be less sensitive to daylight than L. chalumnae. Here we report also, for the first time in the world, a juvenile coelacanth was observed in a small crack at 165 to 171 m depth during these surveys. KEY WORDS: Indonesian coelacanth, underwater survey, Latimeria menadoensis, Sulawesi, Biak, ROV survey. Masamitsu IWATA 1 , Yoshitaka YABUMOTO 2 , Toshiro SARUWATARI 3,4 , Shinya YAMAUCHI 1 , Kenichi FUJII 1 , Rintaro ISHII 1 , Toshiaki MORI 1 , Frensly D. HUKOM 5 , DIRHAMSYAH 5 , Teguh PERISTIWADY 6 , Augy SYAHAILATUA 7 , Kawilarang W. A. MASENGI 8 , Ixchel F. MANDAGI 8 , Fransisco PANGALILA 8 and Yoshitaka ABE 1 1 Aquamarine Fukushima, Marine Science Museum, 50 Tatsumi-cho, Onahama, Iwaki, Fukushima, 971-8101, Japan 2 Kitakyushu Museum of Natural History and Human History, 2-4-1 Higashida, Yahata Higashi-ku, Kitakyushu, Fukuoka, 805-0071, Japan 3 Atmosphere and Ocean Research Institute, The University of Tokyo, 5-1-5 Kashimanohara, Kashiwa-shi, Chiba, 277-8564, Japan 4 Seikei Education and Research Center for Sustainable Development, Seikei Gakuen, 3-3-1 Kichijoji-Kitamachi, Musashino-shi, Tokyo, 180-8633, Japan 5 Research Center for Oceanography, Indonesian Institute of Science, Jl. Pasir Putih I, Ancol Timur, Jakarta, 14430, Indonesia 6 Technical Implementation Unit Marine Biota Conservation Bitung, Indonesian Institute of Sciences, Jl. Colombo, Kec, Maesa, Bitung Tengah, Maesa, Kota Bitung, Sulawesi Utara, 95511, Indonesia 7 Research Center for Deep Sea, Indonesian Institute of Science, Jl. Y. Syaranamual, Guru-Guru, Poka, Ambon, 97233, Indonesia 8 Faculty of Fisheries and Marine Science, Sam Ratulangi University, Jl. Kampus UNSRAT Bahu, Manado, Sulawesi Utara, 95115, Indonesia (Received August 24, 2018; accepted November 9, 2018)

Upload: others

Post on 10-Jan-2020

0 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Field surveys on the Indonesian coelacanth, …4Seikei Education and Research Center for Sustainable Development, Seikei Gakuen, 3-3-1 Kichijoji-Kitamachi ... Guru-Guru, Poka, Ambon,

Allometric growth in the extant coelacanth lung during ontogenetic development. Nature Communications, 6: 8222. DOI: 10.1038/ncomms9222

DE VOS, L. and OYUGI, D. 2002. First capture of a coelacanth, Latimeria chalumnae SMITH, 1939 (Pisces, Latimeriidae), off Kenya. South African Journal of Science, 98: 345–347.

ERDMANN, M. V., CALDWELL, R. L. and MOOSA, M. K. 1998. Indonesian ‘king of the sea’ discovered. Nature, 395: 335.

FRICKE, H., REINICKE, O., HOFER, H. and NACHTIGALL, W. 1987. Locomotion of the coelacanth Latimeria chalumnae in its natural environment. Nature, 329: 331–333.

FRICKE, H., HISSMANN, K., SCHAUER, J., REINICKE, O., KASANG, L. and PLANTE, R. 1991. Habitat and population size of coelacanth Latimeria chalumnae at Grand Comoro. Environmental Biology of Fishes, 32: 287–300.

FRICKE, H., HISSMANN, K., SCHAUER, J., ERDMANN, M., MOOSA, M. K. and PLANTE, R. 2000. Biogeography of the Indonesian coelacanth. Nature, 403: 38.

HEEMSTRA, P. C., FREEMAN, A. L., WONG, H. Y., HENSLEY, D. A. and RABESANDRATANA, H. D. 1996. First authentic

capture of a coelacanth off Madagascar. South African Journal of Science 92: 150–151.

POUYAUD, L., WIRJOATMODJO, S., RACHMATIKA, I., TJAKRAWIDJAJA, A., HADIATY, R. and HADIE, W. 1999. A new species of coelacanth. Genetic and morphologic proof. C. R. Academy of Science, 322: 261–267.

SMITH, J. L. B. 1939. The living coelacanth fish from South Africa. Nature, 143: 748–750.

SMITH, J. L. B. 1953. The second coelacanth. Nature, 171: 99–101.

SMITH, C. L., RAND, C. S., SCHAEFFER, B. and ATZ, J. W. 1975. Latimeria, the living coelacanth, is ovoviviparous. Science, 190: 1105–1106.

VENTER, P., TIMM, P., GUNN, G., LE ROUX, E., SERFONTEIN, E., SMITH, P., SMITH, E., BENSCH, M., HARDING, D. and HEEMSTRA, P. 2000. Discovery of a viable population of coelacanths (Latimeria chalumnae SMITH, 1939) at Sodwana Bay, South Africa. South African Journal of Science, 96: 567–568.

INTRODUCTION

The first living coelacanth, Latimeria chalumnae, was discovered in South Africa in the Indian Ocean in 1938 (SMITH, 1939). The first observation of the living coelacanth habitats using submersibles was reported in Comoros (FRICKE et al., 1987). Latimeria chalumnae hide in submarine caves along steep slopes between 100 m and 200 m depth during daytime and come out and drift near the ocean floor for hunting prey at night (FRICKE et al., 1991). However, since the record of juvenile or small (below 80 cm long) individuals is virtually absent, the growth, development and breeding biology of L. chalumnae remain poorly known. Our knowledge

of the reproduction biology of the coelacanths is almost restricted to an ovoviviparity strategy since some caught large females carried unfertilized eggs or embryos in their oviduct (SMITH et al., 1975; CUPELLO et al., 2015).

The Indonesian coelacanth, L. menadoensis, was first discovered in Manado, Sulawesi Island, Indonesia, in 1997. The first specimen was captured in 1998 (ERDMANN et al., 1998). According to a genetic study, the specimen differed from L. chalumnae and a new species was erected (POUYAUD et al., 1999). In 1999 two individuals were observed by submersible in a location 360 km away from Manado (FRICKE et al., 2000). Today 8 individuals were officially caught in the Indonesian seas. However, the details of their habitat have not

been described. The aim of this study is to clear their habitat and distribution.

METHODS

Two remotely operated vehicles (ROVs) (Kowa; VEGA300) were used for the surveys. The first ROV was replaced with the second one in 2007. Our ROVs are able to dive up to 300 m depth. These had two vertical, two horizontal and two right-left propellers and these were controlled from a boat on surface through a 400 m long tether cable. The underwater operations were visualized on a screen with information including directions of ROVs, depth, water temperature, date and time. All data were directly recorded on video-tapes on the boat. Since the 2007 survey, water temperature and depth were independently recorded with more accuracy by external measuring memories (Alec Electronics, later named JFE Alec; MDS-MKV/T, MDS-MKV/D). In addition two laser beam irradiators, that provide line lasers of 20-cm distance, were attached to the second ROV and have been used to register the size of encountered animals. The surveys were conducted off northern parts of Sulawesi Island and along the Biak Island located in northwestern New Guinea Island, Indonesia. The areas and time periods of each survey are shown in Table 1 and Fig. 1.

RESULTS

The field survey of the Indonesian coelacanth, Latimeria menadoensis, was conducted 14 times from 2005 to 2015 with 1173 underwater operations around Sulawesi and Biak islands in Indonesia (Fig. 1, Table 1). No coelacanth was encountered during the first surveys in 2005 although 452 underwater operations were conducted around Manado-tua Island and other close Indonesian islands from the 17th to 30th of April (Table 1, R1). The second field survey was held along the northern coast of Sulawesi from the 6th to 19th May in 2006, but once again despite of 107 underwater operations no coelacanth was recorded.

The first individual of Indonesian coelacanth (Table 2, ID 1) was recorded during the third field survey, in the morning of the 30th May 2006 (Table 1, R3). The coelacanth was observed in a cave of 165 m depth (Table 2, Boul-1) for 10 minutes (Table 2, E1). The adult individual was in stationary position in a cave (Buol-1) slowly moving its pectoral, pelvic, second dorsal and anal fins. The day after, another individual (ID 2) and an unidentified one (UN1) were found in the same cave (Buol-1). The unidentified one immediately hid deeper in the cave, where the ROV could not access, and only its lateral side of the body was recorded. The individual (ID 2) was observed for one minute before it hid inside the cave (Table 2: E2). When a new individual could be identified by the white

spots pattern on its body, an ID number was allocated. The individual which could not be distinguished was allocated an UN (unknown) + number.

In the end of the morning of the same day, a third individual (Table 2: E3, ID 3) was found in another cave (Buol-2), larger and located 20 m deeper (Buol-2) than the first cave (Buol-1). After 29 minutes of recording, this individual hid inside the cave. Although the water temperature increased of more than five degrees, from 14.8 to 20.4 ºC, it stayed at the same place, not moving away. Despite it disappeared from the camera frame we kept setting the ROV there in front of the cave until the next morning, however it did not appear again.

On the 4th June in 2006, three individuals, including ID 3 which was observed on the 31th May at Buol-2 and two new individuals, were observed in a vertical crack which was several hundred meters far from Buol-2 (Buol-3) for 2 hours and 14 minutes (Fig. 2C; Table 2: E4, ID 3, 4, 5). During this long period of time all individuals stayed near the rock wall always keeping their head against the water flow. The water

Bull. Kitakyushu Mus. Nat. Hist. Hum. Hist., Ser. A, 17: 49–56, March 31, 2019

Field surveys on the Indonesian coelacanth, Latimeria menadoensis using remotely operated vehicles from 2005 to 2015

ABSTRACT − Habitats of the Indonesian coelacanth, Latimeria menadoensis, were investigated by Remotely Operated Vehicles (ROVs) surveys in the northern coast of Sulawesi Island and southern coast of Biak Island by collaboration of Aquamarine Fukushima (Japan), and Indonesian Institute of Sciences and Sam Ratulangi University (Indonesia) from 2005 to 2015. The Remotely Operated Vehicles operations were conducted 1173 times and coelacanths were encountered 30 times. A total of 30 different individuals were observed at a depth range from 115.6 m to 218.9 m deep. The water temperature was between 12.4 to 21.5 ºC. Most of the individuals were found alone, however, schools of two, three and six individuals were also observed. The Indonesian coelacanth, L. menadoensis, was observed at a similar depth of the African coelacanth, Latimeria chalumnae, and the temperature range also seems to be very similar. However, Latimeria menadoensis was sometimes observed besides big rocks or the steep wall. It seems to be less sensitive to daylight than L. chalumnae. Here we report also, for the first time in the world, a juvenile coelacanth was observed in a small crack at 165 to 171 m depth during these surveys.

KEY WORDS: Indonesian coelacanth, underwater survey, Latimeria menadoensis, Sulawesi, Biak, ROV survey.

Masamitsu IWATA1, Yoshitaka YABUMOTO2, Toshiro SARUWATARI3,4, Shinya YAMAUCHI1, Kenichi FUJII1, Rintaro ISHII1, Toshiaki MORI1, Frensly D. HUKOM5, DIRHAMSYAH5, Teguh PERISTIWADY6, Augy SYAHAILATUA7,

Kawilarang W. A. MASENGI8, Ixchel F. MANDAGI8, Fransisco PANGALILA8 and Yoshitaka ABE1

1Aquamarine Fukushima, Marine Science Museum, 50 Tatsumi-cho, Onahama, Iwaki, Fukushima, 971-8101, Japan2Kitakyushu Museum of Natural History and Human History, 2-4-1 Higashida, Yahata Higashi-ku, Kitakyushu, Fukuoka,

805-0071, Japan3Atmosphere and Ocean Research Institute, The University of Tokyo, 5-1-5 Kashimanohara, Kashiwa-shi, Chiba, 277-8564, Japan

4Seikei Education and Research Center for Sustainable Development, Seikei Gakuen, 3-3-1 Kichijoji-Kitamachi, Musashino-shi, Tokyo, 180-8633, Japan

5Research Center for Oceanography, Indonesian Institute of Science, Jl. Pasir Putih I, Ancol Timur, Jakarta, 14430, Indonesia6Technical Implementation Unit Marine Biota Conservation Bitung, Indonesian Institute of Sciences, Jl. Colombo, Kec, Maesa,

Bitung Tengah, Maesa, Kota Bitung, Sulawesi Utara, 95511, Indonesia7Research Center for Deep Sea, Indonesian Institute of Science, Jl. Y. Syaranamual, Guru-Guru, Poka, Ambon, 97233, Indonesia

8Faculty of Fisheries and Marine Science, Sam Ratulangi University, Jl. Kampus UNSRAT Bahu, Manado, Sulawesi Utara, 95115, Indonesia

(Received August 24, 2018; accepted November 9, 2018)

temperature changed up and down by five degrees during the observation due to water flow reversals. When the water flow was reversed some times, all individuals always changed their directions against the water flow.

During the field survey carried on December of the same year, in the same Buol area (Table 1: R4) the sixth individual (Table 2: E5, ID 6) was recognized at 14:34 on the 12th December alongside a large rock. This individual stayed in the beginning of the observation and then moved ahead to deeper slowly. Eventually it reached 12 m deeper during one hour and 42 minutes.

On the 14th December, the individual ID 3 previously recorded at Buol-2 in a cave and at Buol-3 in a crack, was encountered again (Table 2: Buol-5) near a vertical wall at 145 m depth (Table 2: E6). After a while, this individual (ID 3) moved upward about 18 m and stayed behind a large rock (Fig. 3). The observation time was the longest of all surveys with four hours and 16 minutes of recording. The temperature changed up and down several times (Fig. 4). Its difference

between the maximum and the minimum was 6.7 °C, which was the largest range of all surveys.

On the next day, on the 15th December, the ID 3 was observed again under an overhang (Table 2: E7, Buol-6) near other sites where it was previously found at Buol-2 , Buol-3 and Buol-5. Its posture was upside down with its ventral side facing the ceiling of the overhang. Such posture is also known in coelacanths from the Comorian Archipelago (FRICKE et al., 1987). ID 3 was found totally four times in this area in two surveys during the seven months and seems to live in this area.

During the 2007 survey (Table 1: R5) a coelacanth (Table 2: ID 7) was recorded off Manado by 195 m depth, with a temperature of 12.4 °C, the coolest one registered during all surveys.

The field survey in 2009 was carried out off Manado and around Talise and Bangka islands off the northern coast of Sulawesi Island from the 12th September to the 9th October (Table 1: R8). On the 24th September 2009, six individuals (Table 2: E11, ID 9, 10, 11, 12, UN 2, 3) were found under an

overhang off Talise Island between 144 and 150 m deep. There was almost no water current in this area. All individuals were close to each other under the same overhang and most of them stayed still with their head facing downward. Four days later, ID 9 was observed again, alone on a steep slope at 172.9 m depth (Table 2: E12), which was located within several dozen meters far from Talise-2 and 30 m deeper.

Four new coelacanths were recorded from the 29th September to the 6th October 2009 in Talise and Banggka Islands and off Manado (Table 2: E13 and 14, ID 13, 14, 15). ID 15 is the only juvenile coelacanth filmed so far. It was in a small crack between 164.6 and 170.9 m deep in a temperature between 14.5 and 15 °C.

The shallowest record of a coelacanth was at 115.6 m depth off the southern part of Manado (Table 2: E16, Manado-3) on the 9th October in 2009. ID 16 (Table 2) was found at the end of the morning, alone alongside a large rock on the edge of a shelf. The water temperature was about 20.0 ºC.

From the 6th to the 16th November in 2010, the field survey was conducted around Biak Island (Table 1: R9). Adult coelacanths were found at two different sites (Table 2: E17–19). Two individuals (ID 17 and 18) were found under an overhang between 212.5 and 218.9 m deep (Table 2: Biak-1) on the 11th November. Three other ones (ID 19, 20, 21) were registered under an overhang between 193.2 and 195.9 m deep (Table 2: Biak-2) on the 13th November. Two days later, Biak-2 was observed again and only ID 20 stayed there.

Two new coelacanths were recorded off Manado in December 2010 (Table 1: R10; Table 2: E 20 and 21; ID 22 and 23).

On the 4th May 2012, an individual (Table 2: E22, ID 24) was found on a steep slope at 169.4 m depth off Manado (Table 2: Manado-6). It swam up and down 5.4 m between 168.1 and 173.5 m deep for 17 minutes. Thirteen months later (the 9th June 2003), this individual (ID 24) was observed again near a cliff at 152.3 m depth (Table 2: E24, Manado-8), where located on the same slope of Manado-6.

Another new coelacanth (Table 2: E23, ID 25) was recorded alone on a Manado site (Manado-7).

Field surveys were conducted from the 19th to 30th May 2015 off Lolak Island and from the 30th October to the 16th November 2015 off Lolak and Bitung Islands (Table 1; R13 and R14). Three new individuals (Table 2: ID 26–28) were recognized at different places off Lolak Island (Lolak 1-4). All individuals were alone. ID 26 was observed in May and once

again in October 2015 on a steep slope (Table 2: E25, E27). The coelacanth ID 28 was found on the 1st November along one of the large rocks scattered on a gentle slope in a bay of the Lolak Island by 125 m depth.

The encountered Indonesian coelacanths during these surveys were in caves, alongside large rocks, under overhangs or on steep slopes (Table 2; Fig. 2).

Most individuals were found under overhangs or alongside large rocks. Some individuals were encountered not to hide in any shade and stayed just on rocky slopes. In overhangs and cracks, all individuals had their ventral side alongside and close to a rock wall, but without touching it.

Estimated total lengths are shown in Table 3. Most individuals are more than 1 m in total length. ID 15 is a juvenile individual of 31.5 cm and ID 23 is 90 cm in total length. No individual beyond 140 cm has been recorded in Indonesia so far.

DISCUSSION

During the field surveys by underwater ROV recording for Indonesian coelacanths, Latimeria menadoensis, from 2005 to 2015, 30 different individuals were identified and three others were unidentified. Among those, six individuals were observed more than twice: ID 3 at Buol-2, 3, 5, 6 (Table 2: E3, E4, E6 and E7); ID 8 at Talise-1 (E9 and E10), ID 9 at Talise 2 and 3 (E11 and E12), ID 20 at Biak-2 (E18 and E19), ID 24 at Manado-6 and 8 (E22 and E24), ID 26 at Lolak-1 and 3 (E25 and E27). ID 8 was observed on the 14th September 2009 and was encountered again at the same place (Talise-1) the day after. ID 24 was observed on the 4th May 2012 and then 13 months later, on the 9th June 2013 at the same locations of Manado area within several kilometers.

Al l individuals observed more than twice were encountered in the same area, at locations close from each other within hundreds of meters. However, in the Comorian Archipelago, same individuals of L. chalumnae were observed, within a two week period, in several caves distributed in a 8 kilometers wide area (FRICKE et al., 1991).

When individuals of L. menadoensis were encountered, some of them stayed stationary at same place but some individuals such as ID 6 on the 12th December 2006 (E5) and ID 24 on the 4th May 2012 (E22) swam away. It seemed to avoid and escape from the brightness of the light or the sounds

of the propellers of the ROV.Temperature and depth data for each Indonesian

coelacanth record are shown in Fig. 3. Though an average of water temperature on the ocean surface was about 30 °C, the temperature often dropped into about 11 °C at 300 m water depth. The temperatures changed widely even though the changes of depth were small at several places. All Indonesian coelacanths, L. menadoensis, were encountered during day time at a depth between 115.6 and 218.9 m and a temperature between 12.4 and 21.5 ºC. Indonesian coelacanths have exceptionally been recorded in an environment with a temperature exceeding 20 ºC, regardless of the depth. The African coelacanth, L. chalumnae, was observed by submarine vehicles in steep slopes between 150 m and 253 m (FRICKE et al., 1991). Divers recorded the presence of this species at 104 m depth at Sodwana Bay, South Africa (VENTER et al., 2000). Latimeria chalumnae seem to prefer a temperature range of 15–20 °C, where they tend to choose their daytime habitats (FRICKE et al., 1991). The suitable temperature for L. menadoensis seems to be almost the same as L. chalumnae.

Water temperature can increase or drop for more than 5 °C in a very short period of time (about 30 min according to our recordings). The coelacanths however often stayed there without moving away or significantly modifying their behavior. These water currents of different temperatures are caused by the movement of the marine thermocline. During our surveys we observed in some area that as the thermocline

level changed, the water flow was reversed. Coelacanths always kept their head against the water flow, like other observed actinopterygian fishes did. It means that as the direction of the water flow changed, the swimming direction of each coelacanth individual changed accordingly.

On the 24th September 2009 (Table 2: E11) six individuals were recognized under an overhang (Talise-2). The all had about the same size. There was almost no water flow, and each individual was head downward. They were close to each other but no direct contacts or interactions between coelacanths were observed. No evidence of a social behavior has been recognized in the Comorian populations of coelacanths (FRICKE et al., 1991).

According to the observations of the African coelacanths, the species L. chalumnae is nocturnal and hides in cave in daytime but comes out at night, supposedly in search of food (FRICKE et al., 1991). The living geomorphologic environment of L. menadoensis is similar to that of L. chalumnae, with rocky steep slopes with undermarine caves or overhangs. However, Latimeria menadoensis was also observed alongside large rocks or along vertical wall in daytime. Some of these individuals were encountered above 200 m deep, in an environment that remains sl ightly bright . Latimeria menadoensis seems to be less sensitive to the daylight than L. chalumnae.

These first observations are potentially crucial since the habitats of the Indonesian coelacanths could be possibly not as

restricted as those of the African coelacanths, and so a broader distribution of the Indonesian coelacanth could be expected. Occurrences of coelacanths along the eastern coast of the African continent were recorded in South Africa (SMITH, 1939), Comoros (SMITH, 1953), Madagascar (HEEMSTRA et al., 1996), Kenya (DE VOS and OYUGI, 2002), Tanzania (BENNO et al., 2006), and Mozambique (BRUTON et al., 1992). In Indonesia, coelacanths have only been found off the northern coast of the Sulawesi Island and the southern coast of the Biak Island in northern New Guinea.

Field surveys of the Indonesian coelacanths should be continued in all area of Indonesia and expanded to all Southeast Asia in order to increase our knowledge of the distribution of this rare species and to define, in a close future, the most relevant conservation policies of the different known populations.

ACKNOWLEDGMENTS

We are grateful to Dr. Teruya UYENO for his valuable advice, guidance and encouragement throughout the present study. Many thanks go to colleagues of the Faculty of Fisheries and Marine Sciences of the Sam Ratulangi University at Manado, Indonesia, for their generous supports to conduct expeditions. Also we express our thanks to Dr. SUHARSONO, Dr. Zainal ARIFIN, Dr. DIRHAMSYAH and colleagues of the Indonesian Institute of Science for their advice and administrative work to obtain our research permit. Other thanks go to Mr. OPO, Mr. REFRY and colleagues at Murex Dive Resort, and DAUD, AL,

and other staffs for our work on the research boat. Dr. BATUNA and the Murex staffs supported our daily life during the expeditions. We would like to express our sincere thanks to colleagues at Aquamarine Fukushima, Japan, for their understanding and supports. We would like to thank Dr. Gaël CLÉMENT of Muséum national d’Histoire naturelle and Dr. Camila CUPELLO of Universidade do Estado do Rio de Janeiro as a referee for their critical reading of the manuscript and their comments.

Unfortunately one of our authors, Dr. Djoko Hadi KUNARSO has passed away during the process of this contribution. We pray for the response of his soul.

This study was supported in part by funding from Interdisciplinary Collaborative Research program of Atmosphere and Ocean Research Institute, the University of Tokyo.

REFERENCES

BENNO, B., VERHEIJI, E., STAPLEY, J., RUMISHA, C., NGATUNGA, B., ABDALLAH, A. and KALOMBO, H. 2006. Coelacanth (Latimeria chalumnae SMITH, 1939) discoveries and conservation in Tanzania. South African Journal of Science, 102: 486–490.

BRUTON, M. N., CABRAL, A. J. P. and FRICKE, H. 1992. First capture of a coelacanth, Latimeria chalumnae (Pisces, Latimeriidae), off Mozambique. South African Journal of Science, 88: 225–227.

CUPELLO, C., BRITO, P. M., HERBIN, M., MEUNIER, F. J., JA N V I E R, P . , DU T E L, H. and CL É M E N T, G. 2015.

Page 2: Field surveys on the Indonesian coelacanth, …4Seikei Education and Research Center for Sustainable Development, Seikei Gakuen, 3-3-1 Kichijoji-Kitamachi ... Guru-Guru, Poka, Ambon,

Allometric growth in the extant coelacanth lung during ontogenetic development. Nature Communications, 6: 8222. DOI: 10.1038/ncomms9222

DE VOS, L. and OYUGI, D. 2002. First capture of a coelacanth, Latimeria chalumnae SMITH, 1939 (Pisces, Latimeriidae), off Kenya. South African Journal of Science, 98: 345–347.

ERDMANN, M. V., CALDWELL, R. L. and MOOSA, M. K. 1998. Indonesian ‘king of the sea’ discovered. Nature, 395: 335.

FRICKE, H., REINICKE, O., HOFER, H. and NACHTIGALL, W. 1987. Locomotion of the coelacanth Latimeria chalumnae in its natural environment. Nature, 329: 331–333.

FRICKE, H., HISSMANN, K., SCHAUER, J., REINICKE, O., KASANG, L. and PLANTE, R. 1991. Habitat and population size of coelacanth Latimeria chalumnae at Grand Comoro. Environmental Biology of Fishes, 32: 287–300.

FRICKE, H., HISSMANN, K., SCHAUER, J., ERDMANN, M., MOOSA, M. K. and PLANTE, R. 2000. Biogeography of the Indonesian coelacanth. Nature, 403: 38.

HEEMSTRA, P. C., FREEMAN, A. L., WONG, H. Y., HENSLEY, D. A. and RABESANDRATANA, H. D. 1996. First authentic

capture of a coelacanth off Madagascar. South African Journal of Science 92: 150–151.

POUYAUD, L., WIRJOATMODJO, S., RACHMATIKA, I., TJAKRAWIDJAJA, A., HADIATY, R. and HADIE, W. 1999. A new species of coelacanth. Genetic and morphologic proof. C. R. Academy of Science, 322: 261–267.

SMITH, J. L. B. 1939. The living coelacanth fish from South Africa. Nature, 143: 748–750.

SMITH, J. L. B. 1953. The second coelacanth. Nature, 171: 99–101.

SMITH, C. L., RAND, C. S., SCHAEFFER, B. and ATZ, J. W. 1975. Latimeria, the living coelacanth, is ovoviviparous. Science, 190: 1105–1106.

VENTER, P., TIMM, P., GUNN, G., LE ROUX, E., SERFONTEIN, E., SMITH, P., SMITH, E., BENSCH, M., HARDING, D. and HEEMSTRA, P. 2000. Discovery of a viable population of coelacanths (Latimeria chalumnae SMITH, 1939) at Sodwana Bay, South Africa. South African Journal of Science, 96: 567–568.

50 Field surveys on Latimeria menadoensis using ROV from 2005 to 2015 51

INTRODUCTION

The first living coelacanth, Latimeria chalumnae, was discovered in South Africa in the Indian Ocean in 1938 (SMITH, 1939). The first observation of the living coelacanth habitats using submersibles was reported in Comoros (FRICKE et al., 1987). Latimeria chalumnae hide in submarine caves along steep slopes between 100 m and 200 m depth during daytime and come out and drift near the ocean floor for hunting prey at night (FRICKE et al., 1991). However, since the record of juvenile or small (below 80 cm long) individuals is virtually absent, the growth, development and breeding biology of L. chalumnae remain poorly known. Our knowledge

of the reproduction biology of the coelacanths is almost restricted to an ovoviviparity strategy since some caught large females carried unfertilized eggs or embryos in their oviduct (SMITH et al., 1975; CUPELLO et al., 2015).

The Indonesian coelacanth, L. menadoensis, was first discovered in Manado, Sulawesi Island, Indonesia, in 1997. The first specimen was captured in 1998 (ERDMANN et al., 1998). According to a genetic study, the specimen differed from L. chalumnae and a new species was erected (POUYAUD et al., 1999). In 1999 two individuals were observed by submersible in a location 360 km away from Manado (FRICKE et al., 2000). Today 8 individuals were officially caught in the Indonesian seas. However, the details of their habitat have not

been described. The aim of this study is to clear their habitat and distribution.

METHODS

Two remotely operated vehicles (ROVs) (Kowa; VEGA300) were used for the surveys. The first ROV was replaced with the second one in 2007. Our ROVs are able to dive up to 300 m depth. These had two vertical, two horizontal and two right-left propellers and these were controlled from a boat on surface through a 400 m long tether cable. The underwater operations were visualized on a screen with information including directions of ROVs, depth, water temperature, date and time. All data were directly recorded on video-tapes on the boat. Since the 2007 survey, water temperature and depth were independently recorded with more accuracy by external measuring memories (Alec Electronics, later named JFE Alec; MDS-MKV/T, MDS-MKV/D). In addition two laser beam irradiators, that provide line lasers of 20-cm distance, were attached to the second ROV and have been used to register the size of encountered animals. The surveys were conducted off northern parts of Sulawesi Island and along the Biak Island located in northwestern New Guinea Island, Indonesia. The areas and time periods of each survey are shown in Table 1 and Fig. 1.

RESULTS

The field survey of the Indonesian coelacanth, Latimeria menadoensis, was conducted 14 times from 2005 to 2015 with 1173 underwater operations around Sulawesi and Biak islands in Indonesia (Fig. 1, Table 1). No coelacanth was encountered during the first surveys in 2005 although 452 underwater operations were conducted around Manado-tua Island and other close Indonesian islands from the 17th to 30th of April (Table 1, R1). The second field survey was held along the northern coast of Sulawesi from the 6th to 19th May in 2006, but once again despite of 107 underwater operations no coelacanth was recorded.

The first individual of Indonesian coelacanth (Table 2, ID 1) was recorded during the third field survey, in the morning of the 30th May 2006 (Table 1, R3). The coelacanth was observed in a cave of 165 m depth (Table 2, Boul-1) for 10 minutes (Table 2, E1). The adult individual was in stationary position in a cave (Buol-1) slowly moving its pectoral, pelvic, second dorsal and anal fins. The day after, another individual (ID 2) and an unidentified one (UN1) were found in the same cave (Buol-1). The unidentified one immediately hid deeper in the cave, where the ROV could not access, and only its lateral side of the body was recorded. The individual (ID 2) was observed for one minute before it hid inside the cave (Table 2: E2). When a new individual could be identified by the white

spots pattern on its body, an ID number was allocated. The individual which could not be distinguished was allocated an UN (unknown) + number.

In the end of the morning of the same day, a third individual (Table 2: E3, ID 3) was found in another cave (Buol-2), larger and located 20 m deeper (Buol-2) than the first cave (Buol-1). After 29 minutes of recording, this individual hid inside the cave. Although the water temperature increased of more than five degrees, from 14.8 to 20.4 ºC, it stayed at the same place, not moving away. Despite it disappeared from the camera frame we kept setting the ROV there in front of the cave until the next morning, however it did not appear again.

On the 4th June in 2006, three individuals, including ID 3 which was observed on the 31th May at Buol-2 and two new individuals, were observed in a vertical crack which was several hundred meters far from Buol-2 (Buol-3) for 2 hours and 14 minutes (Fig. 2C; Table 2: E4, ID 3, 4, 5). During this long period of time all individuals stayed near the rock wall always keeping their head against the water flow. The water

temperature changed up and down by five degrees during the observation due to water flow reversals. When the water flow was reversed some times, all individuals always changed their directions against the water flow.

During the field survey carried on December of the same year, in the same Buol area (Table 1: R4) the sixth individual (Table 2: E5, ID 6) was recognized at 14:34 on the 12th December alongside a large rock. This individual stayed in the beginning of the observation and then moved ahead to deeper slowly. Eventually it reached 12 m deeper during one hour and 42 minutes.

On the 14th December, the individual ID 3 previously recorded at Buol-2 in a cave and at Buol-3 in a crack, was encountered again (Table 2: Buol-5) near a vertical wall at 145 m depth (Table 2: E6). After a while, this individual (ID 3) moved upward about 18 m and stayed behind a large rock (Fig. 3). The observation time was the longest of all surveys with four hours and 16 minutes of recording. The temperature changed up and down several times (Fig. 4). Its difference

between the maximum and the minimum was 6.7 °C, which was the largest range of all surveys.

On the next day, on the 15th December, the ID 3 was observed again under an overhang (Table 2: E7, Buol-6) near other sites where it was previously found at Buol-2 , Buol-3 and Buol-5. Its posture was upside down with its ventral side facing the ceiling of the overhang. Such posture is also known in coelacanths from the Comorian Archipelago (FRICKE et al., 1987). ID 3 was found totally four times in this area in two surveys during the seven months and seems to live in this area.

During the 2007 survey (Table 1: R5) a coelacanth (Table 2: ID 7) was recorded off Manado by 195 m depth, with a temperature of 12.4 °C, the coolest one registered during all surveys.

The field survey in 2009 was carried out off Manado and around Talise and Bangka islands off the northern coast of Sulawesi Island from the 12th September to the 9th October (Table 1: R8). On the 24th September 2009, six individuals (Table 2: E11, ID 9, 10, 11, 12, UN 2, 3) were found under an

overhang off Talise Island between 144 and 150 m deep. There was almost no water current in this area. All individuals were close to each other under the same overhang and most of them stayed still with their head facing downward. Four days later, ID 9 was observed again, alone on a steep slope at 172.9 m depth (Table 2: E12), which was located within several dozen meters far from Talise-2 and 30 m deeper.

Four new coelacanths were recorded from the 29th September to the 6th October 2009 in Talise and Banggka Islands and off Manado (Table 2: E13 and 14, ID 13, 14, 15). ID 15 is the only juvenile coelacanth filmed so far. It was in a small crack between 164.6 and 170.9 m deep in a temperature between 14.5 and 15 °C.

The shallowest record of a coelacanth was at 115.6 m depth off the southern part of Manado (Table 2: E16, Manado-3) on the 9th October in 2009. ID 16 (Table 2) was found at the end of the morning, alone alongside a large rock on the edge of a shelf. The water temperature was about 20.0 ºC.

From the 6th to the 16th November in 2010, the field survey was conducted around Biak Island (Table 1: R9). Adult coelacanths were found at two different sites (Table 2: E17–19). Two individuals (ID 17 and 18) were found under an overhang between 212.5 and 218.9 m deep (Table 2: Biak-1) on the 11th November. Three other ones (ID 19, 20, 21) were registered under an overhang between 193.2 and 195.9 m deep (Table 2: Biak-2) on the 13th November. Two days later, Biak-2 was observed again and only ID 20 stayed there.

Two new coelacanths were recorded off Manado in December 2010 (Table 1: R10; Table 2: E 20 and 21; ID 22 and 23).

On the 4th May 2012, an individual (Table 2: E22, ID 24) was found on a steep slope at 169.4 m depth off Manado (Table 2: Manado-6). It swam up and down 5.4 m between 168.1 and 173.5 m deep for 17 minutes. Thirteen months later (the 9th June 2003), this individual (ID 24) was observed again near a cliff at 152.3 m depth (Table 2: E24, Manado-8), where located on the same slope of Manado-6.

Another new coelacanth (Table 2: E23, ID 25) was recorded alone on a Manado site (Manado-7).

Field surveys were conducted from the 19th to 30th May 2015 off Lolak Island and from the 30th October to the 16th November 2015 off Lolak and Bitung Islands (Table 1; R13 and R14). Three new individuals (Table 2: ID 26–28) were recognized at different places off Lolak Island (Lolak 1-4). All individuals were alone. ID 26 was observed in May and once

again in October 2015 on a steep slope (Table 2: E25, E27). The coelacanth ID 28 was found on the 1st November along one of the large rocks scattered on a gentle slope in a bay of the Lolak Island by 125 m depth.

The encountered Indonesian coelacanths during these surveys were in caves, alongside large rocks, under overhangs or on steep slopes (Table 2; Fig. 2).

Most individuals were found under overhangs or alongside large rocks. Some individuals were encountered not to hide in any shade and stayed just on rocky slopes. In overhangs and cracks, all individuals had their ventral side alongside and close to a rock wall, but without touching it.

Estimated total lengths are shown in Table 3. Most individuals are more than 1 m in total length. ID 15 is a juvenile individual of 31.5 cm and ID 23 is 90 cm in total length. No individual beyond 140 cm has been recorded in Indonesia so far.

DISCUSSION

During the field surveys by underwater ROV recording for Indonesian coelacanths, Latimeria menadoensis, from 2005 to 2015, 30 different individuals were identified and three others were unidentified. Among those, six individuals were observed more than twice: ID 3 at Buol-2, 3, 5, 6 (Table 2: E3, E4, E6 and E7); ID 8 at Talise-1 (E9 and E10), ID 9 at Talise 2 and 3 (E11 and E12), ID 20 at Biak-2 (E18 and E19), ID 24 at Manado-6 and 8 (E22 and E24), ID 26 at Lolak-1 and 3 (E25 and E27). ID 8 was observed on the 14th September 2009 and was encountered again at the same place (Talise-1) the day after. ID 24 was observed on the 4th May 2012 and then 13 months later, on the 9th June 2013 at the same locations of Manado area within several kilometers.

Al l individuals observed more than twice were encountered in the same area, at locations close from each other within hundreds of meters. However, in the Comorian Archipelago, same individuals of L. chalumnae were observed, within a two week period, in several caves distributed in a 8 kilometers wide area (FRICKE et al., 1991).

When individuals of L. menadoensis were encountered, some of them stayed stationary at same place but some individuals such as ID 6 on the 12th December 2006 (E5) and ID 24 on the 4th May 2012 (E22) swam away. It seemed to avoid and escape from the brightness of the light or the sounds

of the propellers of the ROV.Temperature and depth data for each Indonesian

coelacanth record are shown in Fig. 3. Though an average of water temperature on the ocean surface was about 30 °C, the temperature often dropped into about 11 °C at 300 m water depth. The temperatures changed widely even though the changes of depth were small at several places. All Indonesian coelacanths, L. menadoensis, were encountered during day time at a depth between 115.6 and 218.9 m and a temperature between 12.4 and 21.5 ºC. Indonesian coelacanths have exceptionally been recorded in an environment with a temperature exceeding 20 ºC, regardless of the depth. The African coelacanth, L. chalumnae, was observed by submarine vehicles in steep slopes between 150 m and 253 m (FRICKE et al., 1991). Divers recorded the presence of this species at 104 m depth at Sodwana Bay, South Africa (VENTER et al., 2000). Latimeria chalumnae seem to prefer a temperature range of 15–20 °C, where they tend to choose their daytime habitats (FRICKE et al., 1991). The suitable temperature for L. menadoensis seems to be almost the same as L. chalumnae.

Water temperature can increase or drop for more than 5 °C in a very short period of time (about 30 min according to our recordings). The coelacanths however often stayed there without moving away or significantly modifying their behavior. These water currents of different temperatures are caused by the movement of the marine thermocline. During our surveys we observed in some area that as the thermocline

level changed, the water flow was reversed. Coelacanths always kept their head against the water flow, like other observed actinopterygian fishes did. It means that as the direction of the water flow changed, the swimming direction of each coelacanth individual changed accordingly.

On the 24th September 2009 (Table 2: E11) six individuals were recognized under an overhang (Talise-2). The all had about the same size. There was almost no water flow, and each individual was head downward. They were close to each other but no direct contacts or interactions between coelacanths were observed. No evidence of a social behavior has been recognized in the Comorian populations of coelacanths (FRICKE et al., 1991).

According to the observations of the African coelacanths, the species L. chalumnae is nocturnal and hides in cave in daytime but comes out at night, supposedly in search of food (FRICKE et al., 1991). The living geomorphologic environment of L. menadoensis is similar to that of L. chalumnae, with rocky steep slopes with undermarine caves or overhangs. However, Latimeria menadoensis was also observed alongside large rocks or along vertical wall in daytime. Some of these individuals were encountered above 200 m deep, in an environment that remains sl ightly bright . Latimeria menadoensis seems to be less sensitive to the daylight than L. chalumnae.

These first observations are potentially crucial since the habitats of the Indonesian coelacanths could be possibly not as

restricted as those of the African coelacanths, and so a broader distribution of the Indonesian coelacanth could be expected. Occurrences of coelacanths along the eastern coast of the African continent were recorded in South Africa (SMITH, 1939), Comoros (SMITH, 1953), Madagascar (HEEMSTRA et al., 1996), Kenya (DE VOS and OYUGI, 2002), Tanzania (BENNO et al., 2006), and Mozambique (BRUTON et al., 1992). In Indonesia, coelacanths have only been found off the northern coast of the Sulawesi Island and the southern coast of the Biak Island in northern New Guinea.

Field surveys of the Indonesian coelacanths should be continued in all area of Indonesia and expanded to all Southeast Asia in order to increase our knowledge of the distribution of this rare species and to define, in a close future, the most relevant conservation policies of the different known populations.

ACKNOWLEDGMENTS

We are grateful to Dr. Teruya UYENO for his valuable advice, guidance and encouragement throughout the present study. Many thanks go to colleagues of the Faculty of Fisheries and Marine Sciences of the Sam Ratulangi University at Manado, Indonesia, for their generous supports to conduct expeditions. Also we express our thanks to Dr. SUHARSONO, Dr. Zainal ARIFIN, Dr. DIRHAMSYAH and colleagues of the Indonesian Institute of Science for their advice and administrative work to obtain our research permit. Other thanks go to Mr. OPO, Mr. REFRY and colleagues at Murex Dive Resort, and DAUD, AL,

and other staffs for our work on the research boat. Dr. BATUNA and the Murex staffs supported our daily life during the expeditions. We would like to express our sincere thanks to colleagues at Aquamarine Fukushima, Japan, for their understanding and supports. We would like to thank Dr. Gaël CLÉMENT of Muséum national d’Histoire naturelle and Dr. Camila CUPELLO of Universidade do Estado do Rio de Janeiro as a referee for their critical reading of the manuscript and their comments.

Unfortunately one of our authors, Dr. Djoko Hadi KUNARSO has passed away during the process of this contribution. We pray for the response of his soul.

This study was supported in part by funding from Interdisciplinary Collaborative Research program of Atmosphere and Ocean Research Institute, the University of Tokyo.

REFERENCES

BENNO, B., VERHEIJI, E., STAPLEY, J., RUMISHA, C., NGATUNGA, B., ABDALLAH, A. and KALOMBO, H. 2006. Coelacanth (Latimeria chalumnae SMITH, 1939) discoveries and conservation in Tanzania. South African Journal of Science, 102: 486–490.

BRUTON, M. N., CABRAL, A. J. P. and FRICKE, H. 1992. First capture of a coelacanth, Latimeria chalumnae (Pisces, Latimeriidae), off Mozambique. South African Journal of Science, 88: 225–227.

CUPELLO, C., BRITO, P. M., HERBIN, M., MEUNIER, F. J., JA N V I E R, P . , DU T E L, H. and CL É M E N T, G. 2015.

IWATA et al.

E1E2E3E4E5E6E7E8E9E10

E11

E12E13E14E15E16E17E18E19E20E21E22E23E24E25E26E27E28E29E30

30. May 0631. May 0631. May 064. Jun. 06

12. Dec. 061. Dec. 0615. Dec 0627. Jun. 0714. Sep. 0915. Sep. 09

24. Sep. 09

28. Sep. 0929. Sep. 0929. Sep. 096. Oct. 099. Oct. 09

11. Nov. 1013. Nov. 1015. Nov. 105. Dec. 1016. Dec. 104. May. 1213. May, 129. Jun. 13

27. May. 1528. May. 1530. Oct. 151. Nov. 1516. Nov. 1516. Nov. 15

Date Place SiteNo. Start8:307:5811:1518:4214:3414:2114:0212:1510:338:40

14:51

8:229:2214:2611:3410:1914:2216:4814:5414:3914:3810:4514:3213:3114:4816:1815:3214:5313:3914:36

End8:407:5911:4420:5616:1618:3414:2412:4611:008:42

14:55

8:309:2614:3311:5111:0214:4617:0515:2414:5114:5911:0214:3913:4814:5516:3515:4014:5813:4114:39

Duration101291341022532231272

4

84717432417301221177197178523

Max.165.0165.0183.0156.2163.8145.0154.2195.0172.2166.8

149.9

172.9158.6212.6170.9122.2218.9195.9199.5179.5199.6173.5149.1154.0121.2150.3131.0125.6149.1156.1

Min.165.0165.0183.0153.9151.0126.5152.3190.8159.8160.6

143.7

158.9156.5211.5164.6115.6212.5193.2198.2171.9187.4168.1148.7147.1116.6142.2122.7122.5147.3151.8

Range000

2.312.818.51.94.212.46.2

6.2

142.11.16.36.66.42.71.37.612.25.40.46.94.68.18.33.11.84.3

Max.17.115.520.417.816.621.215.616.215.820.2

18.5

18.517.113.015.021.517.617.118.120.317.818.918.720.819.819.121.217.419.120.3

Min.16.415.314.812.814.214.515.212.414.319.7

18.2

17.616.712.514.518.917.216.817.818.817.616.618.716.519.216.420.715.819.120.2

Ave.16.815.417.513.715.717.915.415.015.520.1

18.4

18.016.912.714.820.017.416.917.919.617.718.018.718.619.417.520.916.219.120.2

12,UN1

33,4,5

633788

9,10,11,12UN2,UN3

913141516

17,1819,20,21

202223242524262726282930

cavecavecavecrackwallwall

over hangbeside big rock

over hangover hang

over hang

steep slope, beside big rockover hangover hangover hangover hangover hangover hangover hang

wallcrack

steep slopeslopewall

beside big rockbeside big rock

steep slopebeside big rock

steep slopesteep slope

ConfigurationIDTemperature (ºC)Depth (m)Time

BuolBuolBuolBuolBuolBuolBuol

MandoTalise Isl.Talise Isl.

Talise Isl.

Talise Isl.Talise Isl.

Banggka Isl.ManadoManado

BiakBiakBiak

ManadoManadoManadoManadoManadoLolakLolakLolakLolakBitungBitung

Buol-1Buol-1Buol-2Buol-3Buol-4Buol-5Buol-6

Mando-1Talise-1Talise-1

Talise-2

Talise-3Talise-4Bang-1

Manado-2Manado-3

Biak-1Biak-2Biak-2

Manado-4Manado-5Manado-6Manado-7Manado-8Lolak-1Lolak-2Lolak-3Lolak-4Bitung-1Bitung-2

Fig. 1. Localities of the Indonesian coelacanth survey using ROV from 2005 to 2015.

Table 1. Periods and areas of the Latimeria menadoensis survey using ROV. Total of 14 expeditions were conducted in the north of Sulawesi Island and Biak Island located in north west of New Guinea Island. The survey areas are shown in Fig. 1.

No.R1R2R3R4R5R6R7R8R9R10R11R12R13R14

Survey AreaManado-tua Isl. and other islandsNorthern coast of SulawesiBuolBuolManadoBuolTalise and Bangka Isl.Talise and Bangka Isl. ManadoBiak IslandManadoManadoManadoBitung, LolakBitung, Lolak

Date17 Apr. – 30 Apr. 20056 May – 19. May 200627 May – 5 Jun. 200611 Dec. – 20 Dec. 200627 Jun. – 5 Jul. 20079 Jul. – 12 Jul. 20072 Dec. – 8 Dec. 200812 Sep. – 9 Oct. 20096 Nov – 16 Nov. 20105 Dec. – 18 Dec. 20102 May – 13 May 20123 Jun. – 10 Jun. 201319 May – 30 May 201530 Oct. – 16 Nov. 2015

Table 2. Registers of Latimeria menadoensis. Time shows start, end and duration of each observation. When an individual could be identified by patterns of white spot on the body, an ID number was allocated. Individuals that could not been distinguished were allocated UN (unknown) + number. Individuals observed several times were identified as numbers with underbar. Duration shows long of each observation time. Range indicates vertical movement of the individual during observation and not the distance traveled horizontally. Some individuals swam along the slope.

Page 3: Field surveys on the Indonesian coelacanth, …4Seikei Education and Research Center for Sustainable Development, Seikei Gakuen, 3-3-1 Kichijoji-Kitamachi ... Guru-Guru, Poka, Ambon,

Allometric growth in the extant coelacanth lung during ontogenetic development. Nature Communications, 6: 8222. DOI: 10.1038/ncomms9222

DE VOS, L. and OYUGI, D. 2002. First capture of a coelacanth, Latimeria chalumnae SMITH, 1939 (Pisces, Latimeriidae), off Kenya. South African Journal of Science, 98: 345–347.

ERDMANN, M. V., CALDWELL, R. L. and MOOSA, M. K. 1998. Indonesian ‘king of the sea’ discovered. Nature, 395: 335.

FRICKE, H., REINICKE, O., HOFER, H. and NACHTIGALL, W. 1987. Locomotion of the coelacanth Latimeria chalumnae in its natural environment. Nature, 329: 331–333.

FRICKE, H., HISSMANN, K., SCHAUER, J., REINICKE, O., KASANG, L. and PLANTE, R. 1991. Habitat and population size of coelacanth Latimeria chalumnae at Grand Comoro. Environmental Biology of Fishes, 32: 287–300.

FRICKE, H., HISSMANN, K., SCHAUER, J., ERDMANN, M., MOOSA, M. K. and PLANTE, R. 2000. Biogeography of the Indonesian coelacanth. Nature, 403: 38.

HEEMSTRA, P. C., FREEMAN, A. L., WONG, H. Y., HENSLEY, D. A. and RABESANDRATANA, H. D. 1996. First authentic

capture of a coelacanth off Madagascar. South African Journal of Science 92: 150–151.

POUYAUD, L., WIRJOATMODJO, S., RACHMATIKA, I., TJAKRAWIDJAJA, A., HADIATY, R. and HADIE, W. 1999. A new species of coelacanth. Genetic and morphologic proof. C. R. Academy of Science, 322: 261–267.

SMITH, J. L. B. 1939. The living coelacanth fish from South Africa. Nature, 143: 748–750.

SMITH, J. L. B. 1953. The second coelacanth. Nature, 171: 99–101.

SMITH, C. L., RAND, C. S., SCHAEFFER, B. and ATZ, J. W. 1975. Latimeria, the living coelacanth, is ovoviviparous. Science, 190: 1105–1106.

VENTER, P., TIMM, P., GUNN, G., LE ROUX, E., SERFONTEIN, E., SMITH, P., SMITH, E., BENSCH, M., HARDING, D. and HEEMSTRA, P. 2000. Discovery of a viable population of coelacanths (Latimeria chalumnae SMITH, 1939) at Sodwana Bay, South Africa. South African Journal of Science, 96: 567–568.

50 Field surveys on Latimeria menadoensis using ROV from 2005 to 2015 51

INTRODUCTION

The first living coelacanth, Latimeria chalumnae, was discovered in South Africa in the Indian Ocean in 1938 (SMITH, 1939). The first observation of the living coelacanth habitats using submersibles was reported in Comoros (FRICKE et al., 1987). Latimeria chalumnae hide in submarine caves along steep slopes between 100 m and 200 m depth during daytime and come out and drift near the ocean floor for hunting prey at night (FRICKE et al., 1991). However, since the record of juvenile or small (below 80 cm long) individuals is virtually absent, the growth, development and breeding biology of L. chalumnae remain poorly known. Our knowledge

of the reproduction biology of the coelacanths is almost restricted to an ovoviviparity strategy since some caught large females carried unfertilized eggs or embryos in their oviduct (SMITH et al., 1975; CUPELLO et al., 2015).

The Indonesian coelacanth, L. menadoensis, was first discovered in Manado, Sulawesi Island, Indonesia, in 1997. The first specimen was captured in 1998 (ERDMANN et al., 1998). According to a genetic study, the specimen differed from L. chalumnae and a new species was erected (POUYAUD et al., 1999). In 1999 two individuals were observed by submersible in a location 360 km away from Manado (FRICKE et al., 2000). Today 8 individuals were officially caught in the Indonesian seas. However, the details of their habitat have not

been described. The aim of this study is to clear their habitat and distribution.

METHODS

Two remotely operated vehicles (ROVs) (Kowa; VEGA300) were used for the surveys. The first ROV was replaced with the second one in 2007. Our ROVs are able to dive up to 300 m depth. These had two vertical, two horizontal and two right-left propellers and these were controlled from a boat on surface through a 400 m long tether cable. The underwater operations were visualized on a screen with information including directions of ROVs, depth, water temperature, date and time. All data were directly recorded on video-tapes on the boat. Since the 2007 survey, water temperature and depth were independently recorded with more accuracy by external measuring memories (Alec Electronics, later named JFE Alec; MDS-MKV/T, MDS-MKV/D). In addition two laser beam irradiators, that provide line lasers of 20-cm distance, were attached to the second ROV and have been used to register the size of encountered animals. The surveys were conducted off northern parts of Sulawesi Island and along the Biak Island located in northwestern New Guinea Island, Indonesia. The areas and time periods of each survey are shown in Table 1 and Fig. 1.

RESULTS

The field survey of the Indonesian coelacanth, Latimeria menadoensis, was conducted 14 times from 2005 to 2015 with 1173 underwater operations around Sulawesi and Biak islands in Indonesia (Fig. 1, Table 1). No coelacanth was encountered during the first surveys in 2005 although 452 underwater operations were conducted around Manado-tua Island and other close Indonesian islands from the 17th to 30th of April (Table 1, R1). The second field survey was held along the northern coast of Sulawesi from the 6th to 19th May in 2006, but once again despite of 107 underwater operations no coelacanth was recorded.

The first individual of Indonesian coelacanth (Table 2, ID 1) was recorded during the third field survey, in the morning of the 30th May 2006 (Table 1, R3). The coelacanth was observed in a cave of 165 m depth (Table 2, Boul-1) for 10 minutes (Table 2, E1). The adult individual was in stationary position in a cave (Buol-1) slowly moving its pectoral, pelvic, second dorsal and anal fins. The day after, another individual (ID 2) and an unidentified one (UN1) were found in the same cave (Buol-1). The unidentified one immediately hid deeper in the cave, where the ROV could not access, and only its lateral side of the body was recorded. The individual (ID 2) was observed for one minute before it hid inside the cave (Table 2: E2). When a new individual could be identified by the white

spots pattern on its body, an ID number was allocated. The individual which could not be distinguished was allocated an UN (unknown) + number.

In the end of the morning of the same day, a third individual (Table 2: E3, ID 3) was found in another cave (Buol-2), larger and located 20 m deeper (Buol-2) than the first cave (Buol-1). After 29 minutes of recording, this individual hid inside the cave. Although the water temperature increased of more than five degrees, from 14.8 to 20.4 ºC, it stayed at the same place, not moving away. Despite it disappeared from the camera frame we kept setting the ROV there in front of the cave until the next morning, however it did not appear again.

On the 4th June in 2006, three individuals, including ID 3 which was observed on the 31th May at Buol-2 and two new individuals, were observed in a vertical crack which was several hundred meters far from Buol-2 (Buol-3) for 2 hours and 14 minutes (Fig. 2C; Table 2: E4, ID 3, 4, 5). During this long period of time all individuals stayed near the rock wall always keeping their head against the water flow. The water

temperature changed up and down by five degrees during the observation due to water flow reversals. When the water flow was reversed some times, all individuals always changed their directions against the water flow.

During the field survey carried on December of the same year, in the same Buol area (Table 1: R4) the sixth individual (Table 2: E5, ID 6) was recognized at 14:34 on the 12th December alongside a large rock. This individual stayed in the beginning of the observation and then moved ahead to deeper slowly. Eventually it reached 12 m deeper during one hour and 42 minutes.

On the 14th December, the individual ID 3 previously recorded at Buol-2 in a cave and at Buol-3 in a crack, was encountered again (Table 2: Buol-5) near a vertical wall at 145 m depth (Table 2: E6). After a while, this individual (ID 3) moved upward about 18 m and stayed behind a large rock (Fig. 3). The observation time was the longest of all surveys with four hours and 16 minutes of recording. The temperature changed up and down several times (Fig. 4). Its difference

between the maximum and the minimum was 6.7 °C, which was the largest range of all surveys.

On the next day, on the 15th December, the ID 3 was observed again under an overhang (Table 2: E7, Buol-6) near other sites where it was previously found at Buol-2 , Buol-3 and Buol-5. Its posture was upside down with its ventral side facing the ceiling of the overhang. Such posture is also known in coelacanths from the Comorian Archipelago (FRICKE et al., 1987). ID 3 was found totally four times in this area in two surveys during the seven months and seems to live in this area.

During the 2007 survey (Table 1: R5) a coelacanth (Table 2: ID 7) was recorded off Manado by 195 m depth, with a temperature of 12.4 °C, the coolest one registered during all surveys.

The field survey in 2009 was carried out off Manado and around Talise and Bangka islands off the northern coast of Sulawesi Island from the 12th September to the 9th October (Table 1: R8). On the 24th September 2009, six individuals (Table 2: E11, ID 9, 10, 11, 12, UN 2, 3) were found under an

overhang off Talise Island between 144 and 150 m deep. There was almost no water current in this area. All individuals were close to each other under the same overhang and most of them stayed still with their head facing downward. Four days later, ID 9 was observed again, alone on a steep slope at 172.9 m depth (Table 2: E12), which was located within several dozen meters far from Talise-2 and 30 m deeper.

Four new coelacanths were recorded from the 29th September to the 6th October 2009 in Talise and Banggka Islands and off Manado (Table 2: E13 and 14, ID 13, 14, 15). ID 15 is the only juvenile coelacanth filmed so far. It was in a small crack between 164.6 and 170.9 m deep in a temperature between 14.5 and 15 °C.

The shallowest record of a coelacanth was at 115.6 m depth off the southern part of Manado (Table 2: E16, Manado-3) on the 9th October in 2009. ID 16 (Table 2) was found at the end of the morning, alone alongside a large rock on the edge of a shelf. The water temperature was about 20.0 ºC.

From the 6th to the 16th November in 2010, the field survey was conducted around Biak Island (Table 1: R9). Adult coelacanths were found at two different sites (Table 2: E17–19). Two individuals (ID 17 and 18) were found under an overhang between 212.5 and 218.9 m deep (Table 2: Biak-1) on the 11th November. Three other ones (ID 19, 20, 21) were registered under an overhang between 193.2 and 195.9 m deep (Table 2: Biak-2) on the 13th November. Two days later, Biak-2 was observed again and only ID 20 stayed there.

Two new coelacanths were recorded off Manado in December 2010 (Table 1: R10; Table 2: E 20 and 21; ID 22 and 23).

On the 4th May 2012, an individual (Table 2: E22, ID 24) was found on a steep slope at 169.4 m depth off Manado (Table 2: Manado-6). It swam up and down 5.4 m between 168.1 and 173.5 m deep for 17 minutes. Thirteen months later (the 9th June 2003), this individual (ID 24) was observed again near a cliff at 152.3 m depth (Table 2: E24, Manado-8), where located on the same slope of Manado-6.

Another new coelacanth (Table 2: E23, ID 25) was recorded alone on a Manado site (Manado-7).

Field surveys were conducted from the 19th to 30th May 2015 off Lolak Island and from the 30th October to the 16th November 2015 off Lolak and Bitung Islands (Table 1; R13 and R14). Three new individuals (Table 2: ID 26–28) were recognized at different places off Lolak Island (Lolak 1-4). All individuals were alone. ID 26 was observed in May and once

again in October 2015 on a steep slope (Table 2: E25, E27). The coelacanth ID 28 was found on the 1st November along one of the large rocks scattered on a gentle slope in a bay of the Lolak Island by 125 m depth.

The encountered Indonesian coelacanths during these surveys were in caves, alongside large rocks, under overhangs or on steep slopes (Table 2; Fig. 2).

Most individuals were found under overhangs or alongside large rocks. Some individuals were encountered not to hide in any shade and stayed just on rocky slopes. In overhangs and cracks, all individuals had their ventral side alongside and close to a rock wall, but without touching it.

Estimated total lengths are shown in Table 3. Most individuals are more than 1 m in total length. ID 15 is a juvenile individual of 31.5 cm and ID 23 is 90 cm in total length. No individual beyond 140 cm has been recorded in Indonesia so far.

DISCUSSION

During the field surveys by underwater ROV recording for Indonesian coelacanths, Latimeria menadoensis, from 2005 to 2015, 30 different individuals were identified and three others were unidentified. Among those, six individuals were observed more than twice: ID 3 at Buol-2, 3, 5, 6 (Table 2: E3, E4, E6 and E7); ID 8 at Talise-1 (E9 and E10), ID 9 at Talise 2 and 3 (E11 and E12), ID 20 at Biak-2 (E18 and E19), ID 24 at Manado-6 and 8 (E22 and E24), ID 26 at Lolak-1 and 3 (E25 and E27). ID 8 was observed on the 14th September 2009 and was encountered again at the same place (Talise-1) the day after. ID 24 was observed on the 4th May 2012 and then 13 months later, on the 9th June 2013 at the same locations of Manado area within several kilometers.

Al l individuals observed more than twice were encountered in the same area, at locations close from each other within hundreds of meters. However, in the Comorian Archipelago, same individuals of L. chalumnae were observed, within a two week period, in several caves distributed in a 8 kilometers wide area (FRICKE et al., 1991).

When individuals of L. menadoensis were encountered, some of them stayed stationary at same place but some individuals such as ID 6 on the 12th December 2006 (E5) and ID 24 on the 4th May 2012 (E22) swam away. It seemed to avoid and escape from the brightness of the light or the sounds

of the propellers of the ROV.Temperature and depth data for each Indonesian

coelacanth record are shown in Fig. 3. Though an average of water temperature on the ocean surface was about 30 °C, the temperature often dropped into about 11 °C at 300 m water depth. The temperatures changed widely even though the changes of depth were small at several places. All Indonesian coelacanths, L. menadoensis, were encountered during day time at a depth between 115.6 and 218.9 m and a temperature between 12.4 and 21.5 ºC. Indonesian coelacanths have exceptionally been recorded in an environment with a temperature exceeding 20 ºC, regardless of the depth. The African coelacanth, L. chalumnae, was observed by submarine vehicles in steep slopes between 150 m and 253 m (FRICKE et al., 1991). Divers recorded the presence of this species at 104 m depth at Sodwana Bay, South Africa (VENTER et al., 2000). Latimeria chalumnae seem to prefer a temperature range of 15–20 °C, where they tend to choose their daytime habitats (FRICKE et al., 1991). The suitable temperature for L. menadoensis seems to be almost the same as L. chalumnae.

Water temperature can increase or drop for more than 5 °C in a very short period of time (about 30 min according to our recordings). The coelacanths however often stayed there without moving away or significantly modifying their behavior. These water currents of different temperatures are caused by the movement of the marine thermocline. During our surveys we observed in some area that as the thermocline

level changed, the water flow was reversed. Coelacanths always kept their head against the water flow, like other observed actinopterygian fishes did. It means that as the direction of the water flow changed, the swimming direction of each coelacanth individual changed accordingly.

On the 24th September 2009 (Table 2: E11) six individuals were recognized under an overhang (Talise-2). The all had about the same size. There was almost no water flow, and each individual was head downward. They were close to each other but no direct contacts or interactions between coelacanths were observed. No evidence of a social behavior has been recognized in the Comorian populations of coelacanths (FRICKE et al., 1991).

According to the observations of the African coelacanths, the species L. chalumnae is nocturnal and hides in cave in daytime but comes out at night, supposedly in search of food (FRICKE et al., 1991). The living geomorphologic environment of L. menadoensis is similar to that of L. chalumnae, with rocky steep slopes with undermarine caves or overhangs. However, Latimeria menadoensis was also observed alongside large rocks or along vertical wall in daytime. Some of these individuals were encountered above 200 m deep, in an environment that remains sl ightly bright . Latimeria menadoensis seems to be less sensitive to the daylight than L. chalumnae.

These first observations are potentially crucial since the habitats of the Indonesian coelacanths could be possibly not as

restricted as those of the African coelacanths, and so a broader distribution of the Indonesian coelacanth could be expected. Occurrences of coelacanths along the eastern coast of the African continent were recorded in South Africa (SMITH, 1939), Comoros (SMITH, 1953), Madagascar (HEEMSTRA et al., 1996), Kenya (DE VOS and OYUGI, 2002), Tanzania (BENNO et al., 2006), and Mozambique (BRUTON et al., 1992). In Indonesia, coelacanths have only been found off the northern coast of the Sulawesi Island and the southern coast of the Biak Island in northern New Guinea.

Field surveys of the Indonesian coelacanths should be continued in all area of Indonesia and expanded to all Southeast Asia in order to increase our knowledge of the distribution of this rare species and to define, in a close future, the most relevant conservation policies of the different known populations.

ACKNOWLEDGMENTS

We are grateful to Dr. Teruya UYENO for his valuable advice, guidance and encouragement throughout the present study. Many thanks go to colleagues of the Faculty of Fisheries and Marine Sciences of the Sam Ratulangi University at Manado, Indonesia, for their generous supports to conduct expeditions. Also we express our thanks to Dr. SUHARSONO, Dr. Zainal ARIFIN, Dr. DIRHAMSYAH and colleagues of the Indonesian Institute of Science for their advice and administrative work to obtain our research permit. Other thanks go to Mr. OPO, Mr. REFRY and colleagues at Murex Dive Resort, and DAUD, AL,

and other staffs for our work on the research boat. Dr. BATUNA and the Murex staffs supported our daily life during the expeditions. We would like to express our sincere thanks to colleagues at Aquamarine Fukushima, Japan, for their understanding and supports. We would like to thank Dr. Gaël CLÉMENT of Muséum national d’Histoire naturelle and Dr. Camila CUPELLO of Universidade do Estado do Rio de Janeiro as a referee for their critical reading of the manuscript and their comments.

Unfortunately one of our authors, Dr. Djoko Hadi KUNARSO has passed away during the process of this contribution. We pray for the response of his soul.

This study was supported in part by funding from Interdisciplinary Collaborative Research program of Atmosphere and Ocean Research Institute, the University of Tokyo.

REFERENCES

BENNO, B., VERHEIJI, E., STAPLEY, J., RUMISHA, C., NGATUNGA, B., ABDALLAH, A. and KALOMBO, H. 2006. Coelacanth (Latimeria chalumnae SMITH, 1939) discoveries and conservation in Tanzania. South African Journal of Science, 102: 486–490.

BRUTON, M. N., CABRAL, A. J. P. and FRICKE, H. 1992. First capture of a coelacanth, Latimeria chalumnae (Pisces, Latimeriidae), off Mozambique. South African Journal of Science, 88: 225–227.

CUPELLO, C., BRITO, P. M., HERBIN, M., MEUNIER, F. J., JA N V I E R, P . , DU T E L, H. and CL É M E N T, G. 2015.

IWATA et al.

E1E2E3E4E5E6E7E8E9E10

E11

E12E13E14E15E16E17E18E19E20E21E22E23E24E25E26E27E28E29E30

30. May 0631. May 0631. May 064. Jun. 06

12. Dec. 061. Dec. 0615. Dec 0627. Jun. 0714. Sep. 0915. Sep. 09

24. Sep. 09

28. Sep. 0929. Sep. 0929. Sep. 096. Oct. 099. Oct. 09

11. Nov. 1013. Nov. 1015. Nov. 105. Dec. 1016. Dec. 104. May. 1213. May, 129. Jun. 13

27. May. 1528. May. 1530. Oct. 151. Nov. 1516. Nov. 1516. Nov. 15

Date Place SiteNo. Start8:307:5811:1518:4214:3414:2114:0212:1510:338:40

14:51

8:229:2214:2611:3410:1914:2216:4814:5414:3914:3810:4514:3213:3114:4816:1815:3214:5313:3914:36

End8:407:5911:4420:5616:1618:3414:2412:4611:008:42

14:55

8:309:2614:3311:5111:0214:4617:0515:2414:5114:5911:0214:3913:4814:5516:3515:4014:5813:4114:39

Duration101291341022532231272

4

84717432417301221177197178523

Max.165.0165.0183.0156.2163.8145.0154.2195.0172.2166.8

149.9

172.9158.6212.6170.9122.2218.9195.9199.5179.5199.6173.5149.1154.0121.2150.3131.0125.6149.1156.1

Min.165.0165.0183.0153.9151.0126.5152.3190.8159.8160.6

143.7

158.9156.5211.5164.6115.6212.5193.2198.2171.9187.4168.1148.7147.1116.6142.2122.7122.5147.3151.8

Range000

2.312.818.51.94.212.46.2

6.2

142.11.16.36.66.42.71.37.612.25.40.46.94.68.18.33.11.84.3

Max.17.115.520.417.816.621.215.616.215.820.2

18.5

18.517.113.015.021.517.617.118.120.317.818.918.720.819.819.121.217.419.120.3

Min.16.415.314.812.814.214.515.212.414.319.7

18.2

17.616.712.514.518.917.216.817.818.817.616.618.716.519.216.420.715.819.120.2

Ave.16.815.417.513.715.717.915.415.015.520.1

18.4

18.016.912.714.820.017.416.917.919.617.718.018.718.619.417.520.916.219.120.2

12,UN1

33,4,5

633788

9,10,11,12UN2,UN3

913141516

17,1819,20,21

202223242524262726282930

cavecavecavecrackwallwall

over hangbeside big rock

over hangover hang

over hang

steep slope, beside big rockover hangover hangover hangover hangover hangover hangover hang

wallcrack

steep slopeslopewall

beside big rockbeside big rock

steep slopebeside big rock

steep slopesteep slope

ConfigurationIDTemperature (ºC)Depth (m)Time

BuolBuolBuolBuolBuolBuolBuol

MandoTalise Isl.Talise Isl.

Talise Isl.

Talise Isl.Talise Isl.

Banggka Isl.ManadoManado

BiakBiakBiak

ManadoManadoManadoManadoManadoLolakLolakLolakLolakBitungBitung

Buol-1Buol-1Buol-2Buol-3Buol-4Buol-5Buol-6

Mando-1Talise-1Talise-1

Talise-2

Talise-3Talise-4Bang-1

Manado-2Manado-3

Biak-1Biak-2Biak-2

Manado-4Manado-5Manado-6Manado-7Manado-8Lolak-1Lolak-2Lolak-3Lolak-4Bitung-1Bitung-2

Fig. 1. Localities of the Indonesian coelacanth survey using ROV from 2005 to 2015.

Table 1. Periods and areas of the Latimeria menadoensis survey using ROV. Total of 14 expeditions were conducted in the north of Sulawesi Island and Biak Island located in north west of New Guinea Island. The survey areas are shown in Fig. 1.

No.R1R2R3R4R5R6R7R8R9R10R11R12R13R14

Survey AreaManado-tua Isl. and other islandsNorthern coast of SulawesiBuolBuolManadoBuolTalise and Bangka Isl.Talise and Bangka Isl. ManadoBiak IslandManadoManadoManadoBitung, LolakBitung, Lolak

Date17 Apr. – 30 Apr. 20056 May – 19. May 200627 May – 5 Jun. 200611 Dec. – 20 Dec. 200627 Jun. – 5 Jul. 20079 Jul. – 12 Jul. 20072 Dec. – 8 Dec. 200812 Sep. – 9 Oct. 20096 Nov – 16 Nov. 20105 Dec. – 18 Dec. 20102 May – 13 May 20123 Jun. – 10 Jun. 201319 May – 30 May 201530 Oct. – 16 Nov. 2015

Table 2. Registers of Latimeria menadoensis. Time shows start, end and duration of each observation. When an individual could be identified by patterns of white spot on the body, an ID number was allocated. Individuals that could not been distinguished were allocated UN (unknown) + number. Individuals observed several times were identified as numbers with underbar. Duration shows long of each observation time. Range indicates vertical movement of the individual during observation and not the distance traveled horizontally. Some individuals swam along the slope.

Page 4: Field surveys on the Indonesian coelacanth, …4Seikei Education and Research Center for Sustainable Development, Seikei Gakuen, 3-3-1 Kichijoji-Kitamachi ... Guru-Guru, Poka, Ambon,

Allometric growth in the extant coelacanth lung during ontogenetic development. Nature Communications, 6: 8222. DOI: 10.1038/ncomms9222

DE VOS, L. and OYUGI, D. 2002. First capture of a coelacanth, Latimeria chalumnae SMITH, 1939 (Pisces, Latimeriidae), off Kenya. South African Journal of Science, 98: 345–347.

ERDMANN, M. V., CALDWELL, R. L. and MOOSA, M. K. 1998. Indonesian ‘king of the sea’ discovered. Nature, 395: 335.

FRICKE, H., REINICKE, O., HOFER, H. and NACHTIGALL, W. 1987. Locomotion of the coelacanth Latimeria chalumnae in its natural environment. Nature, 329: 331–333.

FRICKE, H., HISSMANN, K., SCHAUER, J., REINICKE, O., KASANG, L. and PLANTE, R. 1991. Habitat and population size of coelacanth Latimeria chalumnae at Grand Comoro. Environmental Biology of Fishes, 32: 287–300.

FRICKE, H., HISSMANN, K., SCHAUER, J., ERDMANN, M., MOOSA, M. K. and PLANTE, R. 2000. Biogeography of the Indonesian coelacanth. Nature, 403: 38.

HEEMSTRA, P. C., FREEMAN, A. L., WONG, H. Y., HENSLEY, D. A. and RABESANDRATANA, H. D. 1996. First authentic

capture of a coelacanth off Madagascar. South African Journal of Science 92: 150–151.

POUYAUD, L., WIRJOATMODJO, S., RACHMATIKA, I., TJAKRAWIDJAJA, A., HADIATY, R. and HADIE, W. 1999. A new species of coelacanth. Genetic and morphologic proof. C. R. Academy of Science, 322: 261–267.

SMITH, J. L. B. 1939. The living coelacanth fish from South Africa. Nature, 143: 748–750.

SMITH, J. L. B. 1953. The second coelacanth. Nature, 171: 99–101.

SMITH, C. L., RAND, C. S., SCHAEFFER, B. and ATZ, J. W. 1975. Latimeria, the living coelacanth, is ovoviviparous. Science, 190: 1105–1106.

VENTER, P., TIMM, P., GUNN, G., LE ROUX, E., SERFONTEIN, E., SMITH, P., SMITH, E., BENSCH, M., HARDING, D. and HEEMSTRA, P. 2000. Discovery of a viable population of coelacanths (Latimeria chalumnae SMITH, 1939) at Sodwana Bay, South Africa. South African Journal of Science, 96: 567–568.

INTRODUCTION

The first living coelacanth, Latimeria chalumnae, was discovered in South Africa in the Indian Ocean in 1938 (SMITH, 1939). The first observation of the living coelacanth habitats using submersibles was reported in Comoros (FRICKE et al., 1987). Latimeria chalumnae hide in submarine caves along steep slopes between 100 m and 200 m depth during daytime and come out and drift near the ocean floor for hunting prey at night (FRICKE et al., 1991). However, since the record of juvenile or small (below 80 cm long) individuals is virtually absent, the growth, development and breeding biology of L. chalumnae remain poorly known. Our knowledge

of the reproduction biology of the coelacanths is almost restricted to an ovoviviparity strategy since some caught large females carried unfertilized eggs or embryos in their oviduct (SMITH et al., 1975; CUPELLO et al., 2015).

The Indonesian coelacanth, L. menadoensis, was first discovered in Manado, Sulawesi Island, Indonesia, in 1997. The first specimen was captured in 1998 (ERDMANN et al., 1998). According to a genetic study, the specimen differed from L. chalumnae and a new species was erected (POUYAUD et al., 1999). In 1999 two individuals were observed by submersible in a location 360 km away from Manado (FRICKE et al., 2000). Today 8 individuals were officially caught in the Indonesian seas. However, the details of their habitat have not

been described. The aim of this study is to clear their habitat and distribution.

METHODS

Two remotely operated vehicles (ROVs) (Kowa; VEGA300) were used for the surveys. The first ROV was replaced with the second one in 2007. Our ROVs are able to dive up to 300 m depth. These had two vertical, two horizontal and two right-left propellers and these were controlled from a boat on surface through a 400 m long tether cable. The underwater operations were visualized on a screen with information including directions of ROVs, depth, water temperature, date and time. All data were directly recorded on video-tapes on the boat. Since the 2007 survey, water temperature and depth were independently recorded with more accuracy by external measuring memories (Alec Electronics, later named JFE Alec; MDS-MKV/T, MDS-MKV/D). In addition two laser beam irradiators, that provide line lasers of 20-cm distance, were attached to the second ROV and have been used to register the size of encountered animals. The surveys were conducted off northern parts of Sulawesi Island and along the Biak Island located in northwestern New Guinea Island, Indonesia. The areas and time periods of each survey are shown in Table 1 and Fig. 1.

RESULTS

The field survey of the Indonesian coelacanth, Latimeria menadoensis, was conducted 14 times from 2005 to 2015 with 1173 underwater operations around Sulawesi and Biak islands in Indonesia (Fig. 1, Table 1). No coelacanth was encountered during the first surveys in 2005 although 452 underwater operations were conducted around Manado-tua Island and other close Indonesian islands from the 17th to 30th of April (Table 1, R1). The second field survey was held along the northern coast of Sulawesi from the 6th to 19th May in 2006, but once again despite of 107 underwater operations no coelacanth was recorded.

The first individual of Indonesian coelacanth (Table 2, ID 1) was recorded during the third field survey, in the morning of the 30th May 2006 (Table 1, R3). The coelacanth was observed in a cave of 165 m depth (Table 2, Boul-1) for 10 minutes (Table 2, E1). The adult individual was in stationary position in a cave (Buol-1) slowly moving its pectoral, pelvic, second dorsal and anal fins. The day after, another individual (ID 2) and an unidentified one (UN1) were found in the same cave (Buol-1). The unidentified one immediately hid deeper in the cave, where the ROV could not access, and only its lateral side of the body was recorded. The individual (ID 2) was observed for one minute before it hid inside the cave (Table 2: E2). When a new individual could be identified by the white

spots pattern on its body, an ID number was allocated. The individual which could not be distinguished was allocated an UN (unknown) + number.

In the end of the morning of the same day, a third individual (Table 2: E3, ID 3) was found in another cave (Buol-2), larger and located 20 m deeper (Buol-2) than the first cave (Buol-1). After 29 minutes of recording, this individual hid inside the cave. Although the water temperature increased of more than five degrees, from 14.8 to 20.4 ºC, it stayed at the same place, not moving away. Despite it disappeared from the camera frame we kept setting the ROV there in front of the cave until the next morning, however it did not appear again.

On the 4th June in 2006, three individuals, including ID 3 which was observed on the 31th May at Buol-2 and two new individuals, were observed in a vertical crack which was several hundred meters far from Buol-2 (Buol-3) for 2 hours and 14 minutes (Fig. 2C; Table 2: E4, ID 3, 4, 5). During this long period of time all individuals stayed near the rock wall always keeping their head against the water flow. The water

52 53

Fig. 2. Configuration of encounter point. A, over hang (E9); B, over hang (E17); C, crack (E4); D, crack (E21); E, beside big rock (E28); F, beside big rock (E25); G, wall (E20); H, steep slope (E12); I, steep slope, beside big rock (E13). Red triangles indicate coelacanths.

temperature changed up and down by five degrees during the observation due to water flow reversals. When the water flow was reversed some times, all individuals always changed their directions against the water flow.

During the field survey carried on December of the same year, in the same Buol area (Table 1: R4) the sixth individual (Table 2: E5, ID 6) was recognized at 14:34 on the 12th December alongside a large rock. This individual stayed in the beginning of the observation and then moved ahead to deeper slowly. Eventually it reached 12 m deeper during one hour and 42 minutes.

On the 14th December, the individual ID 3 previously recorded at Buol-2 in a cave and at Buol-3 in a crack, was encountered again (Table 2: Buol-5) near a vertical wall at 145 m depth (Table 2: E6). After a while, this individual (ID 3) moved upward about 18 m and stayed behind a large rock (Fig. 3). The observation time was the longest of all surveys with four hours and 16 minutes of recording. The temperature changed up and down several times (Fig. 4). Its difference

between the maximum and the minimum was 6.7 °C, which was the largest range of all surveys.

On the next day, on the 15th December, the ID 3 was observed again under an overhang (Table 2: E7, Buol-6) near other sites where it was previously found at Buol-2 , Buol-3 and Buol-5. Its posture was upside down with its ventral side facing the ceiling of the overhang. Such posture is also known in coelacanths from the Comorian Archipelago (FRICKE et al., 1987). ID 3 was found totally four times in this area in two surveys during the seven months and seems to live in this area.

During the 2007 survey (Table 1: R5) a coelacanth (Table 2: ID 7) was recorded off Manado by 195 m depth, with a temperature of 12.4 °C, the coolest one registered during all surveys.

The field survey in 2009 was carried out off Manado and around Talise and Bangka islands off the northern coast of Sulawesi Island from the 12th September to the 9th October (Table 1: R8). On the 24th September 2009, six individuals (Table 2: E11, ID 9, 10, 11, 12, UN 2, 3) were found under an

overhang off Talise Island between 144 and 150 m deep. There was almost no water current in this area. All individuals were close to each other under the same overhang and most of them stayed still with their head facing downward. Four days later, ID 9 was observed again, alone on a steep slope at 172.9 m depth (Table 2: E12), which was located within several dozen meters far from Talise-2 and 30 m deeper.

Four new coelacanths were recorded from the 29th September to the 6th October 2009 in Talise and Banggka Islands and off Manado (Table 2: E13 and 14, ID 13, 14, 15). ID 15 is the only juvenile coelacanth filmed so far. It was in a small crack between 164.6 and 170.9 m deep in a temperature between 14.5 and 15 °C.

The shallowest record of a coelacanth was at 115.6 m depth off the southern part of Manado (Table 2: E16, Manado-3) on the 9th October in 2009. ID 16 (Table 2) was found at the end of the morning, alone alongside a large rock on the edge of a shelf. The water temperature was about 20.0 ºC.

From the 6th to the 16th November in 2010, the field survey was conducted around Biak Island (Table 1: R9). Adult coelacanths were found at two different sites (Table 2: E17–19). Two individuals (ID 17 and 18) were found under an overhang between 212.5 and 218.9 m deep (Table 2: Biak-1) on the 11th November. Three other ones (ID 19, 20, 21) were registered under an overhang between 193.2 and 195.9 m deep (Table 2: Biak-2) on the 13th November. Two days later, Biak-2 was observed again and only ID 20 stayed there.

Two new coelacanths were recorded off Manado in December 2010 (Table 1: R10; Table 2: E 20 and 21; ID 22 and 23).

On the 4th May 2012, an individual (Table 2: E22, ID 24) was found on a steep slope at 169.4 m depth off Manado (Table 2: Manado-6). It swam up and down 5.4 m between 168.1 and 173.5 m deep for 17 minutes. Thirteen months later (the 9th June 2003), this individual (ID 24) was observed again near a cliff at 152.3 m depth (Table 2: E24, Manado-8), where located on the same slope of Manado-6.

Another new coelacanth (Table 2: E23, ID 25) was recorded alone on a Manado site (Manado-7).

Field surveys were conducted from the 19th to 30th May 2015 off Lolak Island and from the 30th October to the 16th November 2015 off Lolak and Bitung Islands (Table 1; R13 and R14). Three new individuals (Table 2: ID 26–28) were recognized at different places off Lolak Island (Lolak 1-4). All individuals were alone. ID 26 was observed in May and once

again in October 2015 on a steep slope (Table 2: E25, E27). The coelacanth ID 28 was found on the 1st November along one of the large rocks scattered on a gentle slope in a bay of the Lolak Island by 125 m depth.

The encountered Indonesian coelacanths during these surveys were in caves, alongside large rocks, under overhangs or on steep slopes (Table 2; Fig. 2).

Most individuals were found under overhangs or alongside large rocks. Some individuals were encountered not to hide in any shade and stayed just on rocky slopes. In overhangs and cracks, all individuals had their ventral side alongside and close to a rock wall, but without touching it.

Estimated total lengths are shown in Table 3. Most individuals are more than 1 m in total length. ID 15 is a juvenile individual of 31.5 cm and ID 23 is 90 cm in total length. No individual beyond 140 cm has been recorded in Indonesia so far.

DISCUSSION

During the field surveys by underwater ROV recording for Indonesian coelacanths, Latimeria menadoensis, from 2005 to 2015, 30 different individuals were identified and three others were unidentified. Among those, six individuals were observed more than twice: ID 3 at Buol-2, 3, 5, 6 (Table 2: E3, E4, E6 and E7); ID 8 at Talise-1 (E9 and E10), ID 9 at Talise 2 and 3 (E11 and E12), ID 20 at Biak-2 (E18 and E19), ID 24 at Manado-6 and 8 (E22 and E24), ID 26 at Lolak-1 and 3 (E25 and E27). ID 8 was observed on the 14th September 2009 and was encountered again at the same place (Talise-1) the day after. ID 24 was observed on the 4th May 2012 and then 13 months later, on the 9th June 2013 at the same locations of Manado area within several kilometers.

Al l individuals observed more than twice were encountered in the same area, at locations close from each other within hundreds of meters. However, in the Comorian Archipelago, same individuals of L. chalumnae were observed, within a two week period, in several caves distributed in a 8 kilometers wide area (FRICKE et al., 1991).

When individuals of L. menadoensis were encountered, some of them stayed stationary at same place but some individuals such as ID 6 on the 12th December 2006 (E5) and ID 24 on the 4th May 2012 (E22) swam away. It seemed to avoid and escape from the brightness of the light or the sounds

of the propellers of the ROV.Temperature and depth data for each Indonesian

coelacanth record are shown in Fig. 3. Though an average of water temperature on the ocean surface was about 30 °C, the temperature often dropped into about 11 °C at 300 m water depth. The temperatures changed widely even though the changes of depth were small at several places. All Indonesian coelacanths, L. menadoensis, were encountered during day time at a depth between 115.6 and 218.9 m and a temperature between 12.4 and 21.5 ºC. Indonesian coelacanths have exceptionally been recorded in an environment with a temperature exceeding 20 ºC, regardless of the depth. The African coelacanth, L. chalumnae, was observed by submarine vehicles in steep slopes between 150 m and 253 m (FRICKE et al., 1991). Divers recorded the presence of this species at 104 m depth at Sodwana Bay, South Africa (VENTER et al., 2000). Latimeria chalumnae seem to prefer a temperature range of 15–20 °C, where they tend to choose their daytime habitats (FRICKE et al., 1991). The suitable temperature for L. menadoensis seems to be almost the same as L. chalumnae.

Water temperature can increase or drop for more than 5 °C in a very short period of time (about 30 min according to our recordings). The coelacanths however often stayed there without moving away or significantly modifying their behavior. These water currents of different temperatures are caused by the movement of the marine thermocline. During our surveys we observed in some area that as the thermocline

level changed, the water flow was reversed. Coelacanths always kept their head against the water flow, like other observed actinopterygian fishes did. It means that as the direction of the water flow changed, the swimming direction of each coelacanth individual changed accordingly.

On the 24th September 2009 (Table 2: E11) six individuals were recognized under an overhang (Talise-2). The all had about the same size. There was almost no water flow, and each individual was head downward. They were close to each other but no direct contacts or interactions between coelacanths were observed. No evidence of a social behavior has been recognized in the Comorian populations of coelacanths (FRICKE et al., 1991).

According to the observations of the African coelacanths, the species L. chalumnae is nocturnal and hides in cave in daytime but comes out at night, supposedly in search of food (FRICKE et al., 1991). The living geomorphologic environment of L. menadoensis is similar to that of L. chalumnae, with rocky steep slopes with undermarine caves or overhangs. However, Latimeria menadoensis was also observed alongside large rocks or along vertical wall in daytime. Some of these individuals were encountered above 200 m deep, in an environment that remains sl ightly bright . Latimeria menadoensis seems to be less sensitive to the daylight than L. chalumnae.

These first observations are potentially crucial since the habitats of the Indonesian coelacanths could be possibly not as

restricted as those of the African coelacanths, and so a broader distribution of the Indonesian coelacanth could be expected. Occurrences of coelacanths along the eastern coast of the African continent were recorded in South Africa (SMITH, 1939), Comoros (SMITH, 1953), Madagascar (HEEMSTRA et al., 1996), Kenya (DE VOS and OYUGI, 2002), Tanzania (BENNO et al., 2006), and Mozambique (BRUTON et al., 1992). In Indonesia, coelacanths have only been found off the northern coast of the Sulawesi Island and the southern coast of the Biak Island in northern New Guinea.

Field surveys of the Indonesian coelacanths should be continued in all area of Indonesia and expanded to all Southeast Asia in order to increase our knowledge of the distribution of this rare species and to define, in a close future, the most relevant conservation policies of the different known populations.

ACKNOWLEDGMENTS

We are grateful to Dr. Teruya UYENO for his valuable advice, guidance and encouragement throughout the present study. Many thanks go to colleagues of the Faculty of Fisheries and Marine Sciences of the Sam Ratulangi University at Manado, Indonesia, for their generous supports to conduct expeditions. Also we express our thanks to Dr. SUHARSONO, Dr. Zainal ARIFIN, Dr. DIRHAMSYAH and colleagues of the Indonesian Institute of Science for their advice and administrative work to obtain our research permit. Other thanks go to Mr. OPO, Mr. REFRY and colleagues at Murex Dive Resort, and DAUD, AL,

and other staffs for our work on the research boat. Dr. BATUNA and the Murex staffs supported our daily life during the expeditions. We would like to express our sincere thanks to colleagues at Aquamarine Fukushima, Japan, for their understanding and supports. We would like to thank Dr. Gaël CLÉMENT of Muséum national d’Histoire naturelle and Dr. Camila CUPELLO of Universidade do Estado do Rio de Janeiro as a referee for their critical reading of the manuscript and their comments.

Unfortunately one of our authors, Dr. Djoko Hadi KUNARSO has passed away during the process of this contribution. We pray for the response of his soul.

This study was supported in part by funding from Interdisciplinary Collaborative Research program of Atmosphere and Ocean Research Institute, the University of Tokyo.

REFERENCES

BENNO, B., VERHEIJI, E., STAPLEY, J., RUMISHA, C., NGATUNGA, B., ABDALLAH, A. and KALOMBO, H. 2006. Coelacanth (Latimeria chalumnae SMITH, 1939) discoveries and conservation in Tanzania. South African Journal of Science, 102: 486–490.

BRUTON, M. N., CABRAL, A. J. P. and FRICKE, H. 1992. First capture of a coelacanth, Latimeria chalumnae (Pisces, Latimeriidae), off Mozambique. South African Journal of Science, 88: 225–227.

CUPELLO, C., BRITO, P. M., HERBIN, M., MEUNIER, F. J., JA N V I E R, P . , DU T E L, H. and CL É M E N T, G. 2015.

IWATA et al. Field surveys on Latimeria menadoensis using ROV from 2005 to 2015

Table 3. Total length of each coelacanth estimated with 20 cm distance parallel laser lines applied from the ROV.

ID numberEstimated TL (cm)

7123

8115

14110

1531.5

16113

17125

19114

20113

22106

2390

24137

Page 5: Field surveys on the Indonesian coelacanth, …4Seikei Education and Research Center for Sustainable Development, Seikei Gakuen, 3-3-1 Kichijoji-Kitamachi ... Guru-Guru, Poka, Ambon,

Allometric growth in the extant coelacanth lung during ontogenetic development. Nature Communications, 6: 8222. DOI: 10.1038/ncomms9222

DE VOS, L. and OYUGI, D. 2002. First capture of a coelacanth, Latimeria chalumnae SMITH, 1939 (Pisces, Latimeriidae), off Kenya. South African Journal of Science, 98: 345–347.

ERDMANN, M. V., CALDWELL, R. L. and MOOSA, M. K. 1998. Indonesian ‘king of the sea’ discovered. Nature, 395: 335.

FRICKE, H., REINICKE, O., HOFER, H. and NACHTIGALL, W. 1987. Locomotion of the coelacanth Latimeria chalumnae in its natural environment. Nature, 329: 331–333.

FRICKE, H., HISSMANN, K., SCHAUER, J., REINICKE, O., KASANG, L. and PLANTE, R. 1991. Habitat and population size of coelacanth Latimeria chalumnae at Grand Comoro. Environmental Biology of Fishes, 32: 287–300.

FRICKE, H., HISSMANN, K., SCHAUER, J., ERDMANN, M., MOOSA, M. K. and PLANTE, R. 2000. Biogeography of the Indonesian coelacanth. Nature, 403: 38.

HEEMSTRA, P. C., FREEMAN, A. L., WONG, H. Y., HENSLEY, D. A. and RABESANDRATANA, H. D. 1996. First authentic

capture of a coelacanth off Madagascar. South African Journal of Science 92: 150–151.

POUYAUD, L., WIRJOATMODJO, S., RACHMATIKA, I., TJAKRAWIDJAJA, A., HADIATY, R. and HADIE, W. 1999. A new species of coelacanth. Genetic and morphologic proof. C. R. Academy of Science, 322: 261–267.

SMITH, J. L. B. 1939. The living coelacanth fish from South Africa. Nature, 143: 748–750.

SMITH, J. L. B. 1953. The second coelacanth. Nature, 171: 99–101.

SMITH, C. L., RAND, C. S., SCHAEFFER, B. and ATZ, J. W. 1975. Latimeria, the living coelacanth, is ovoviviparous. Science, 190: 1105–1106.

VENTER, P., TIMM, P., GUNN, G., LE ROUX, E., SERFONTEIN, E., SMITH, P., SMITH, E., BENSCH, M., HARDING, D. and HEEMSTRA, P. 2000. Discovery of a viable population of coelacanths (Latimeria chalumnae SMITH, 1939) at Sodwana Bay, South Africa. South African Journal of Science, 96: 567–568.

INTRODUCTION

The first living coelacanth, Latimeria chalumnae, was discovered in South Africa in the Indian Ocean in 1938 (SMITH, 1939). The first observation of the living coelacanth habitats using submersibles was reported in Comoros (FRICKE et al., 1987). Latimeria chalumnae hide in submarine caves along steep slopes between 100 m and 200 m depth during daytime and come out and drift near the ocean floor for hunting prey at night (FRICKE et al., 1991). However, since the record of juvenile or small (below 80 cm long) individuals is virtually absent, the growth, development and breeding biology of L. chalumnae remain poorly known. Our knowledge

of the reproduction biology of the coelacanths is almost restricted to an ovoviviparity strategy since some caught large females carried unfertilized eggs or embryos in their oviduct (SMITH et al., 1975; CUPELLO et al., 2015).

The Indonesian coelacanth, L. menadoensis, was first discovered in Manado, Sulawesi Island, Indonesia, in 1997. The first specimen was captured in 1998 (ERDMANN et al., 1998). According to a genetic study, the specimen differed from L. chalumnae and a new species was erected (POUYAUD et al., 1999). In 1999 two individuals were observed by submersible in a location 360 km away from Manado (FRICKE et al., 2000). Today 8 individuals were officially caught in the Indonesian seas. However, the details of their habitat have not

been described. The aim of this study is to clear their habitat and distribution.

METHODS

Two remotely operated vehicles (ROVs) (Kowa; VEGA300) were used for the surveys. The first ROV was replaced with the second one in 2007. Our ROVs are able to dive up to 300 m depth. These had two vertical, two horizontal and two right-left propellers and these were controlled from a boat on surface through a 400 m long tether cable. The underwater operations were visualized on a screen with information including directions of ROVs, depth, water temperature, date and time. All data were directly recorded on video-tapes on the boat. Since the 2007 survey, water temperature and depth were independently recorded with more accuracy by external measuring memories (Alec Electronics, later named JFE Alec; MDS-MKV/T, MDS-MKV/D). In addition two laser beam irradiators, that provide line lasers of 20-cm distance, were attached to the second ROV and have been used to register the size of encountered animals. The surveys were conducted off northern parts of Sulawesi Island and along the Biak Island located in northwestern New Guinea Island, Indonesia. The areas and time periods of each survey are shown in Table 1 and Fig. 1.

RESULTS

The field survey of the Indonesian coelacanth, Latimeria menadoensis, was conducted 14 times from 2005 to 2015 with 1173 underwater operations around Sulawesi and Biak islands in Indonesia (Fig. 1, Table 1). No coelacanth was encountered during the first surveys in 2005 although 452 underwater operations were conducted around Manado-tua Island and other close Indonesian islands from the 17th to 30th of April (Table 1, R1). The second field survey was held along the northern coast of Sulawesi from the 6th to 19th May in 2006, but once again despite of 107 underwater operations no coelacanth was recorded.

The first individual of Indonesian coelacanth (Table 2, ID 1) was recorded during the third field survey, in the morning of the 30th May 2006 (Table 1, R3). The coelacanth was observed in a cave of 165 m depth (Table 2, Boul-1) for 10 minutes (Table 2, E1). The adult individual was in stationary position in a cave (Buol-1) slowly moving its pectoral, pelvic, second dorsal and anal fins. The day after, another individual (ID 2) and an unidentified one (UN1) were found in the same cave (Buol-1). The unidentified one immediately hid deeper in the cave, where the ROV could not access, and only its lateral side of the body was recorded. The individual (ID 2) was observed for one minute before it hid inside the cave (Table 2: E2). When a new individual could be identified by the white

spots pattern on its body, an ID number was allocated. The individual which could not be distinguished was allocated an UN (unknown) + number.

In the end of the morning of the same day, a third individual (Table 2: E3, ID 3) was found in another cave (Buol-2), larger and located 20 m deeper (Buol-2) than the first cave (Buol-1). After 29 minutes of recording, this individual hid inside the cave. Although the water temperature increased of more than five degrees, from 14.8 to 20.4 ºC, it stayed at the same place, not moving away. Despite it disappeared from the camera frame we kept setting the ROV there in front of the cave until the next morning, however it did not appear again.

On the 4th June in 2006, three individuals, including ID 3 which was observed on the 31th May at Buol-2 and two new individuals, were observed in a vertical crack which was several hundred meters far from Buol-2 (Buol-3) for 2 hours and 14 minutes (Fig. 2C; Table 2: E4, ID 3, 4, 5). During this long period of time all individuals stayed near the rock wall always keeping their head against the water flow. The water

52 53

Fig. 2. Configuration of encounter point. A, over hang (E9); B, over hang (E17); C, crack (E4); D, crack (E21); E, beside big rock (E28); F, beside big rock (E25); G, wall (E20); H, steep slope (E12); I, steep slope, beside big rock (E13). Red triangles indicate coelacanths.

temperature changed up and down by five degrees during the observation due to water flow reversals. When the water flow was reversed some times, all individuals always changed their directions against the water flow.

During the field survey carried on December of the same year, in the same Buol area (Table 1: R4) the sixth individual (Table 2: E5, ID 6) was recognized at 14:34 on the 12th December alongside a large rock. This individual stayed in the beginning of the observation and then moved ahead to deeper slowly. Eventually it reached 12 m deeper during one hour and 42 minutes.

On the 14th December, the individual ID 3 previously recorded at Buol-2 in a cave and at Buol-3 in a crack, was encountered again (Table 2: Buol-5) near a vertical wall at 145 m depth (Table 2: E6). After a while, this individual (ID 3) moved upward about 18 m and stayed behind a large rock (Fig. 3). The observation time was the longest of all surveys with four hours and 16 minutes of recording. The temperature changed up and down several times (Fig. 4). Its difference

between the maximum and the minimum was 6.7 °C, which was the largest range of all surveys.

On the next day, on the 15th December, the ID 3 was observed again under an overhang (Table 2: E7, Buol-6) near other sites where it was previously found at Buol-2 , Buol-3 and Buol-5. Its posture was upside down with its ventral side facing the ceiling of the overhang. Such posture is also known in coelacanths from the Comorian Archipelago (FRICKE et al., 1987). ID 3 was found totally four times in this area in two surveys during the seven months and seems to live in this area.

During the 2007 survey (Table 1: R5) a coelacanth (Table 2: ID 7) was recorded off Manado by 195 m depth, with a temperature of 12.4 °C, the coolest one registered during all surveys.

The field survey in 2009 was carried out off Manado and around Talise and Bangka islands off the northern coast of Sulawesi Island from the 12th September to the 9th October (Table 1: R8). On the 24th September 2009, six individuals (Table 2: E11, ID 9, 10, 11, 12, UN 2, 3) were found under an

overhang off Talise Island between 144 and 150 m deep. There was almost no water current in this area. All individuals were close to each other under the same overhang and most of them stayed still with their head facing downward. Four days later, ID 9 was observed again, alone on a steep slope at 172.9 m depth (Table 2: E12), which was located within several dozen meters far from Talise-2 and 30 m deeper.

Four new coelacanths were recorded from the 29th September to the 6th October 2009 in Talise and Banggka Islands and off Manado (Table 2: E13 and 14, ID 13, 14, 15). ID 15 is the only juvenile coelacanth filmed so far. It was in a small crack between 164.6 and 170.9 m deep in a temperature between 14.5 and 15 °C.

The shallowest record of a coelacanth was at 115.6 m depth off the southern part of Manado (Table 2: E16, Manado-3) on the 9th October in 2009. ID 16 (Table 2) was found at the end of the morning, alone alongside a large rock on the edge of a shelf. The water temperature was about 20.0 ºC.

From the 6th to the 16th November in 2010, the field survey was conducted around Biak Island (Table 1: R9). Adult coelacanths were found at two different sites (Table 2: E17–19). Two individuals (ID 17 and 18) were found under an overhang between 212.5 and 218.9 m deep (Table 2: Biak-1) on the 11th November. Three other ones (ID 19, 20, 21) were registered under an overhang between 193.2 and 195.9 m deep (Table 2: Biak-2) on the 13th November. Two days later, Biak-2 was observed again and only ID 20 stayed there.

Two new coelacanths were recorded off Manado in December 2010 (Table 1: R10; Table 2: E 20 and 21; ID 22 and 23).

On the 4th May 2012, an individual (Table 2: E22, ID 24) was found on a steep slope at 169.4 m depth off Manado (Table 2: Manado-6). It swam up and down 5.4 m between 168.1 and 173.5 m deep for 17 minutes. Thirteen months later (the 9th June 2003), this individual (ID 24) was observed again near a cliff at 152.3 m depth (Table 2: E24, Manado-8), where located on the same slope of Manado-6.

Another new coelacanth (Table 2: E23, ID 25) was recorded alone on a Manado site (Manado-7).

Field surveys were conducted from the 19th to 30th May 2015 off Lolak Island and from the 30th October to the 16th November 2015 off Lolak and Bitung Islands (Table 1; R13 and R14). Three new individuals (Table 2: ID 26–28) were recognized at different places off Lolak Island (Lolak 1-4). All individuals were alone. ID 26 was observed in May and once

again in October 2015 on a steep slope (Table 2: E25, E27). The coelacanth ID 28 was found on the 1st November along one of the large rocks scattered on a gentle slope in a bay of the Lolak Island by 125 m depth.

The encountered Indonesian coelacanths during these surveys were in caves, alongside large rocks, under overhangs or on steep slopes (Table 2; Fig. 2).

Most individuals were found under overhangs or alongside large rocks. Some individuals were encountered not to hide in any shade and stayed just on rocky slopes. In overhangs and cracks, all individuals had their ventral side alongside and close to a rock wall, but without touching it.

Estimated total lengths are shown in Table 3. Most individuals are more than 1 m in total length. ID 15 is a juvenile individual of 31.5 cm and ID 23 is 90 cm in total length. No individual beyond 140 cm has been recorded in Indonesia so far.

DISCUSSION

During the field surveys by underwater ROV recording for Indonesian coelacanths, Latimeria menadoensis, from 2005 to 2015, 30 different individuals were identified and three others were unidentified. Among those, six individuals were observed more than twice: ID 3 at Buol-2, 3, 5, 6 (Table 2: E3, E4, E6 and E7); ID 8 at Talise-1 (E9 and E10), ID 9 at Talise 2 and 3 (E11 and E12), ID 20 at Biak-2 (E18 and E19), ID 24 at Manado-6 and 8 (E22 and E24), ID 26 at Lolak-1 and 3 (E25 and E27). ID 8 was observed on the 14th September 2009 and was encountered again at the same place (Talise-1) the day after. ID 24 was observed on the 4th May 2012 and then 13 months later, on the 9th June 2013 at the same locations of Manado area within several kilometers.

Al l individuals observed more than twice were encountered in the same area, at locations close from each other within hundreds of meters. However, in the Comorian Archipelago, same individuals of L. chalumnae were observed, within a two week period, in several caves distributed in a 8 kilometers wide area (FRICKE et al., 1991).

When individuals of L. menadoensis were encountered, some of them stayed stationary at same place but some individuals such as ID 6 on the 12th December 2006 (E5) and ID 24 on the 4th May 2012 (E22) swam away. It seemed to avoid and escape from the brightness of the light or the sounds

of the propellers of the ROV.Temperature and depth data for each Indonesian

coelacanth record are shown in Fig. 3. Though an average of water temperature on the ocean surface was about 30 °C, the temperature often dropped into about 11 °C at 300 m water depth. The temperatures changed widely even though the changes of depth were small at several places. All Indonesian coelacanths, L. menadoensis, were encountered during day time at a depth between 115.6 and 218.9 m and a temperature between 12.4 and 21.5 ºC. Indonesian coelacanths have exceptionally been recorded in an environment with a temperature exceeding 20 ºC, regardless of the depth. The African coelacanth, L. chalumnae, was observed by submarine vehicles in steep slopes between 150 m and 253 m (FRICKE et al., 1991). Divers recorded the presence of this species at 104 m depth at Sodwana Bay, South Africa (VENTER et al., 2000). Latimeria chalumnae seem to prefer a temperature range of 15–20 °C, where they tend to choose their daytime habitats (FRICKE et al., 1991). The suitable temperature for L. menadoensis seems to be almost the same as L. chalumnae.

Water temperature can increase or drop for more than 5 °C in a very short period of time (about 30 min according to our recordings). The coelacanths however often stayed there without moving away or significantly modifying their behavior. These water currents of different temperatures are caused by the movement of the marine thermocline. During our surveys we observed in some area that as the thermocline

level changed, the water flow was reversed. Coelacanths always kept their head against the water flow, like other observed actinopterygian fishes did. It means that as the direction of the water flow changed, the swimming direction of each coelacanth individual changed accordingly.

On the 24th September 2009 (Table 2: E11) six individuals were recognized under an overhang (Talise-2). The all had about the same size. There was almost no water flow, and each individual was head downward. They were close to each other but no direct contacts or interactions between coelacanths were observed. No evidence of a social behavior has been recognized in the Comorian populations of coelacanths (FRICKE et al., 1991).

According to the observations of the African coelacanths, the species L. chalumnae is nocturnal and hides in cave in daytime but comes out at night, supposedly in search of food (FRICKE et al., 1991). The living geomorphologic environment of L. menadoensis is similar to that of L. chalumnae, with rocky steep slopes with undermarine caves or overhangs. However, Latimeria menadoensis was also observed alongside large rocks or along vertical wall in daytime. Some of these individuals were encountered above 200 m deep, in an environment that remains sl ightly bright . Latimeria menadoensis seems to be less sensitive to the daylight than L. chalumnae.

These first observations are potentially crucial since the habitats of the Indonesian coelacanths could be possibly not as

restricted as those of the African coelacanths, and so a broader distribution of the Indonesian coelacanth could be expected. Occurrences of coelacanths along the eastern coast of the African continent were recorded in South Africa (SMITH, 1939), Comoros (SMITH, 1953), Madagascar (HEEMSTRA et al., 1996), Kenya (DE VOS and OYUGI, 2002), Tanzania (BENNO et al., 2006), and Mozambique (BRUTON et al., 1992). In Indonesia, coelacanths have only been found off the northern coast of the Sulawesi Island and the southern coast of the Biak Island in northern New Guinea.

Field surveys of the Indonesian coelacanths should be continued in all area of Indonesia and expanded to all Southeast Asia in order to increase our knowledge of the distribution of this rare species and to define, in a close future, the most relevant conservation policies of the different known populations.

ACKNOWLEDGMENTS

We are grateful to Dr. Teruya UYENO for his valuable advice, guidance and encouragement throughout the present study. Many thanks go to colleagues of the Faculty of Fisheries and Marine Sciences of the Sam Ratulangi University at Manado, Indonesia, for their generous supports to conduct expeditions. Also we express our thanks to Dr. SUHARSONO, Dr. Zainal ARIFIN, Dr. DIRHAMSYAH and colleagues of the Indonesian Institute of Science for their advice and administrative work to obtain our research permit. Other thanks go to Mr. OPO, Mr. REFRY and colleagues at Murex Dive Resort, and DAUD, AL,

and other staffs for our work on the research boat. Dr. BATUNA and the Murex staffs supported our daily life during the expeditions. We would like to express our sincere thanks to colleagues at Aquamarine Fukushima, Japan, for their understanding and supports. We would like to thank Dr. Gaël CLÉMENT of Muséum national d’Histoire naturelle and Dr. Camila CUPELLO of Universidade do Estado do Rio de Janeiro as a referee for their critical reading of the manuscript and their comments.

Unfortunately one of our authors, Dr. Djoko Hadi KUNARSO has passed away during the process of this contribution. We pray for the response of his soul.

This study was supported in part by funding from Interdisciplinary Collaborative Research program of Atmosphere and Ocean Research Institute, the University of Tokyo.

REFERENCES

BENNO, B., VERHEIJI, E., STAPLEY, J., RUMISHA, C., NGATUNGA, B., ABDALLAH, A. and KALOMBO, H. 2006. Coelacanth (Latimeria chalumnae SMITH, 1939) discoveries and conservation in Tanzania. South African Journal of Science, 102: 486–490.

BRUTON, M. N., CABRAL, A. J. P. and FRICKE, H. 1992. First capture of a coelacanth, Latimeria chalumnae (Pisces, Latimeriidae), off Mozambique. South African Journal of Science, 88: 225–227.

CUPELLO, C., BRITO, P. M., HERBIN, M., MEUNIER, F. J., JA N V I E R, P . , DU T E L, H. and CL É M E N T, G. 2015.

IWATA et al. Field surveys on Latimeria menadoensis using ROV from 2005 to 2015

Table 3. Total length of each coelacanth estimated with 20 cm distance parallel laser lines applied from the ROV.

ID numberEstimated TL (cm)

7123

8115

14110

1531.5

16113

17125

19114

20113

22106

2390

24137

Page 6: Field surveys on the Indonesian coelacanth, …4Seikei Education and Research Center for Sustainable Development, Seikei Gakuen, 3-3-1 Kichijoji-Kitamachi ... Guru-Guru, Poka, Ambon,

Allometric growth in the extant coelacanth lung during ontogenetic development. Nature Communications, 6: 8222. DOI: 10.1038/ncomms9222

DE VOS, L. and OYUGI, D. 2002. First capture of a coelacanth, Latimeria chalumnae SMITH, 1939 (Pisces, Latimeriidae), off Kenya. South African Journal of Science, 98: 345–347.

ERDMANN, M. V., CALDWELL, R. L. and MOOSA, M. K. 1998. Indonesian ‘king of the sea’ discovered. Nature, 395: 335.

FRICKE, H., REINICKE, O., HOFER, H. and NACHTIGALL, W. 1987. Locomotion of the coelacanth Latimeria chalumnae in its natural environment. Nature, 329: 331–333.

FRICKE, H., HISSMANN, K., SCHAUER, J., REINICKE, O., KASANG, L. and PLANTE, R. 1991. Habitat and population size of coelacanth Latimeria chalumnae at Grand Comoro. Environmental Biology of Fishes, 32: 287–300.

FRICKE, H., HISSMANN, K., SCHAUER, J., ERDMANN, M., MOOSA, M. K. and PLANTE, R. 2000. Biogeography of the Indonesian coelacanth. Nature, 403: 38.

HEEMSTRA, P. C., FREEMAN, A. L., WONG, H. Y., HENSLEY, D. A. and RABESANDRATANA, H. D. 1996. First authentic

capture of a coelacanth off Madagascar. South African Journal of Science 92: 150–151.

POUYAUD, L., WIRJOATMODJO, S., RACHMATIKA, I., TJAKRAWIDJAJA, A., HADIATY, R. and HADIE, W. 1999. A new species of coelacanth. Genetic and morphologic proof. C. R. Academy of Science, 322: 261–267.

SMITH, J. L. B. 1939. The living coelacanth fish from South Africa. Nature, 143: 748–750.

SMITH, J. L. B. 1953. The second coelacanth. Nature, 171: 99–101.

SMITH, C. L., RAND, C. S., SCHAEFFER, B. and ATZ, J. W. 1975. Latimeria, the living coelacanth, is ovoviviparous. Science, 190: 1105–1106.

VENTER, P., TIMM, P., GUNN, G., LE ROUX, E., SERFONTEIN, E., SMITH, P., SMITH, E., BENSCH, M., HARDING, D. and HEEMSTRA, P. 2000. Discovery of a viable population of coelacanths (Latimeria chalumnae SMITH, 1939) at Sodwana Bay, South Africa. South African Journal of Science, 96: 567–568.

INTRODUCTION

The first living coelacanth, Latimeria chalumnae, was discovered in South Africa in the Indian Ocean in 1938 (SMITH, 1939). The first observation of the living coelacanth habitats using submersibles was reported in Comoros (FRICKE et al., 1987). Latimeria chalumnae hide in submarine caves along steep slopes between 100 m and 200 m depth during daytime and come out and drift near the ocean floor for hunting prey at night (FRICKE et al., 1991). However, since the record of juvenile or small (below 80 cm long) individuals is virtually absent, the growth, development and breeding biology of L. chalumnae remain poorly known. Our knowledge

of the reproduction biology of the coelacanths is almost restricted to an ovoviviparity strategy since some caught large females carried unfertilized eggs or embryos in their oviduct (SMITH et al., 1975; CUPELLO et al., 2015).

The Indonesian coelacanth, L. menadoensis, was first discovered in Manado, Sulawesi Island, Indonesia, in 1997. The first specimen was captured in 1998 (ERDMANN et al., 1998). According to a genetic study, the specimen differed from L. chalumnae and a new species was erected (POUYAUD et al., 1999). In 1999 two individuals were observed by submersible in a location 360 km away from Manado (FRICKE et al., 2000). Today 8 individuals were officially caught in the Indonesian seas. However, the details of their habitat have not

been described. The aim of this study is to clear their habitat and distribution.

METHODS

Two remotely operated vehicles (ROVs) (Kowa; VEGA300) were used for the surveys. The first ROV was replaced with the second one in 2007. Our ROVs are able to dive up to 300 m depth. These had two vertical, two horizontal and two right-left propellers and these were controlled from a boat on surface through a 400 m long tether cable. The underwater operations were visualized on a screen with information including directions of ROVs, depth, water temperature, date and time. All data were directly recorded on video-tapes on the boat. Since the 2007 survey, water temperature and depth were independently recorded with more accuracy by external measuring memories (Alec Electronics, later named JFE Alec; MDS-MKV/T, MDS-MKV/D). In addition two laser beam irradiators, that provide line lasers of 20-cm distance, were attached to the second ROV and have been used to register the size of encountered animals. The surveys were conducted off northern parts of Sulawesi Island and along the Biak Island located in northwestern New Guinea Island, Indonesia. The areas and time periods of each survey are shown in Table 1 and Fig. 1.

RESULTS

The field survey of the Indonesian coelacanth, Latimeria menadoensis, was conducted 14 times from 2005 to 2015 with 1173 underwater operations around Sulawesi and Biak islands in Indonesia (Fig. 1, Table 1). No coelacanth was encountered during the first surveys in 2005 although 452 underwater operations were conducted around Manado-tua Island and other close Indonesian islands from the 17th to 30th of April (Table 1, R1). The second field survey was held along the northern coast of Sulawesi from the 6th to 19th May in 2006, but once again despite of 107 underwater operations no coelacanth was recorded.

The first individual of Indonesian coelacanth (Table 2, ID 1) was recorded during the third field survey, in the morning of the 30th May 2006 (Table 1, R3). The coelacanth was observed in a cave of 165 m depth (Table 2, Boul-1) for 10 minutes (Table 2, E1). The adult individual was in stationary position in a cave (Buol-1) slowly moving its pectoral, pelvic, second dorsal and anal fins. The day after, another individual (ID 2) and an unidentified one (UN1) were found in the same cave (Buol-1). The unidentified one immediately hid deeper in the cave, where the ROV could not access, and only its lateral side of the body was recorded. The individual (ID 2) was observed for one minute before it hid inside the cave (Table 2: E2). When a new individual could be identified by the white

spots pattern on its body, an ID number was allocated. The individual which could not be distinguished was allocated an UN (unknown) + number.

In the end of the morning of the same day, a third individual (Table 2: E3, ID 3) was found in another cave (Buol-2), larger and located 20 m deeper (Buol-2) than the first cave (Buol-1). After 29 minutes of recording, this individual hid inside the cave. Although the water temperature increased of more than five degrees, from 14.8 to 20.4 ºC, it stayed at the same place, not moving away. Despite it disappeared from the camera frame we kept setting the ROV there in front of the cave until the next morning, however it did not appear again.

On the 4th June in 2006, three individuals, including ID 3 which was observed on the 31th May at Buol-2 and two new individuals, were observed in a vertical crack which was several hundred meters far from Buol-2 (Buol-3) for 2 hours and 14 minutes (Fig. 2C; Table 2: E4, ID 3, 4, 5). During this long period of time all individuals stayed near the rock wall always keeping their head against the water flow. The water

54 55

temperature changed up and down by five degrees during the observation due to water flow reversals. When the water flow was reversed some times, all individuals always changed their directions against the water flow.

During the field survey carried on December of the same year, in the same Buol area (Table 1: R4) the sixth individual (Table 2: E5, ID 6) was recognized at 14:34 on the 12th December alongside a large rock. This individual stayed in the beginning of the observation and then moved ahead to deeper slowly. Eventually it reached 12 m deeper during one hour and 42 minutes.

On the 14th December, the individual ID 3 previously recorded at Buol-2 in a cave and at Buol-3 in a crack, was encountered again (Table 2: Buol-5) near a vertical wall at 145 m depth (Table 2: E6). After a while, this individual (ID 3) moved upward about 18 m and stayed behind a large rock (Fig. 3). The observation time was the longest of all surveys with four hours and 16 minutes of recording. The temperature changed up and down several times (Fig. 4). Its difference

between the maximum and the minimum was 6.7 °C, which was the largest range of all surveys.

On the next day, on the 15th December, the ID 3 was observed again under an overhang (Table 2: E7, Buol-6) near other sites where it was previously found at Buol-2 , Buol-3 and Buol-5. Its posture was upside down with its ventral side facing the ceiling of the overhang. Such posture is also known in coelacanths from the Comorian Archipelago (FRICKE et al., 1987). ID 3 was found totally four times in this area in two surveys during the seven months and seems to live in this area.

During the 2007 survey (Table 1: R5) a coelacanth (Table 2: ID 7) was recorded off Manado by 195 m depth, with a temperature of 12.4 °C, the coolest one registered during all surveys.

The field survey in 2009 was carried out off Manado and around Talise and Bangka islands off the northern coast of Sulawesi Island from the 12th September to the 9th October (Table 1: R8). On the 24th September 2009, six individuals (Table 2: E11, ID 9, 10, 11, 12, UN 2, 3) were found under an

overhang off Talise Island between 144 and 150 m deep. There was almost no water current in this area. All individuals were close to each other under the same overhang and most of them stayed still with their head facing downward. Four days later, ID 9 was observed again, alone on a steep slope at 172.9 m depth (Table 2: E12), which was located within several dozen meters far from Talise-2 and 30 m deeper.

Four new coelacanths were recorded from the 29th September to the 6th October 2009 in Talise and Banggka Islands and off Manado (Table 2: E13 and 14, ID 13, 14, 15). ID 15 is the only juvenile coelacanth filmed so far. It was in a small crack between 164.6 and 170.9 m deep in a temperature between 14.5 and 15 °C.

The shallowest record of a coelacanth was at 115.6 m depth off the southern part of Manado (Table 2: E16, Manado-3) on the 9th October in 2009. ID 16 (Table 2) was found at the end of the morning, alone alongside a large rock on the edge of a shelf. The water temperature was about 20.0 ºC.

From the 6th to the 16th November in 2010, the field survey was conducted around Biak Island (Table 1: R9). Adult coelacanths were found at two different sites (Table 2: E17–19). Two individuals (ID 17 and 18) were found under an overhang between 212.5 and 218.9 m deep (Table 2: Biak-1) on the 11th November. Three other ones (ID 19, 20, 21) were registered under an overhang between 193.2 and 195.9 m deep (Table 2: Biak-2) on the 13th November. Two days later, Biak-2 was observed again and only ID 20 stayed there.

Two new coelacanths were recorded off Manado in December 2010 (Table 1: R10; Table 2: E 20 and 21; ID 22 and 23).

On the 4th May 2012, an individual (Table 2: E22, ID 24) was found on a steep slope at 169.4 m depth off Manado (Table 2: Manado-6). It swam up and down 5.4 m between 168.1 and 173.5 m deep for 17 minutes. Thirteen months later (the 9th June 2003), this individual (ID 24) was observed again near a cliff at 152.3 m depth (Table 2: E24, Manado-8), where located on the same slope of Manado-6.

Another new coelacanth (Table 2: E23, ID 25) was recorded alone on a Manado site (Manado-7).

Field surveys were conducted from the 19th to 30th May 2015 off Lolak Island and from the 30th October to the 16th November 2015 off Lolak and Bitung Islands (Table 1; R13 and R14). Three new individuals (Table 2: ID 26–28) were recognized at different places off Lolak Island (Lolak 1-4). All individuals were alone. ID 26 was observed in May and once

again in October 2015 on a steep slope (Table 2: E25, E27). The coelacanth ID 28 was found on the 1st November along one of the large rocks scattered on a gentle slope in a bay of the Lolak Island by 125 m depth.

The encountered Indonesian coelacanths during these surveys were in caves, alongside large rocks, under overhangs or on steep slopes (Table 2; Fig. 2).

Most individuals were found under overhangs or alongside large rocks. Some individuals were encountered not to hide in any shade and stayed just on rocky slopes. In overhangs and cracks, all individuals had their ventral side alongside and close to a rock wall, but without touching it.

Estimated total lengths are shown in Table 3. Most individuals are more than 1 m in total length. ID 15 is a juvenile individual of 31.5 cm and ID 23 is 90 cm in total length. No individual beyond 140 cm has been recorded in Indonesia so far.

DISCUSSION

During the field surveys by underwater ROV recording for Indonesian coelacanths, Latimeria menadoensis, from 2005 to 2015, 30 different individuals were identified and three others were unidentified. Among those, six individuals were observed more than twice: ID 3 at Buol-2, 3, 5, 6 (Table 2: E3, E4, E6 and E7); ID 8 at Talise-1 (E9 and E10), ID 9 at Talise 2 and 3 (E11 and E12), ID 20 at Biak-2 (E18 and E19), ID 24 at Manado-6 and 8 (E22 and E24), ID 26 at Lolak-1 and 3 (E25 and E27). ID 8 was observed on the 14th September 2009 and was encountered again at the same place (Talise-1) the day after. ID 24 was observed on the 4th May 2012 and then 13 months later, on the 9th June 2013 at the same locations of Manado area within several kilometers.

Al l individuals observed more than twice were encountered in the same area, at locations close from each other within hundreds of meters. However, in the Comorian Archipelago, same individuals of L. chalumnae were observed, within a two week period, in several caves distributed in a 8 kilometers wide area (FRICKE et al., 1991).

When individuals of L. menadoensis were encountered, some of them stayed stationary at same place but some individuals such as ID 6 on the 12th December 2006 (E5) and ID 24 on the 4th May 2012 (E22) swam away. It seemed to avoid and escape from the brightness of the light or the sounds

of the propellers of the ROV.Temperature and depth data for each Indonesian

coelacanth record are shown in Fig. 3. Though an average of water temperature on the ocean surface was about 30 °C, the temperature often dropped into about 11 °C at 300 m water depth. The temperatures changed widely even though the changes of depth were small at several places. All Indonesian coelacanths, L. menadoensis, were encountered during day time at a depth between 115.6 and 218.9 m and a temperature between 12.4 and 21.5 ºC. Indonesian coelacanths have exceptionally been recorded in an environment with a temperature exceeding 20 ºC, regardless of the depth. The African coelacanth, L. chalumnae, was observed by submarine vehicles in steep slopes between 150 m and 253 m (FRICKE et al., 1991). Divers recorded the presence of this species at 104 m depth at Sodwana Bay, South Africa (VENTER et al., 2000). Latimeria chalumnae seem to prefer a temperature range of 15–20 °C, where they tend to choose their daytime habitats (FRICKE et al., 1991). The suitable temperature for L. menadoensis seems to be almost the same as L. chalumnae.

Water temperature can increase or drop for more than 5 °C in a very short period of time (about 30 min according to our recordings). The coelacanths however often stayed there without moving away or significantly modifying their behavior. These water currents of different temperatures are caused by the movement of the marine thermocline. During our surveys we observed in some area that as the thermocline

level changed, the water flow was reversed. Coelacanths always kept their head against the water flow, like other observed actinopterygian fishes did. It means that as the direction of the water flow changed, the swimming direction of each coelacanth individual changed accordingly.

On the 24th September 2009 (Table 2: E11) six individuals were recognized under an overhang (Talise-2). The all had about the same size. There was almost no water flow, and each individual was head downward. They were close to each other but no direct contacts or interactions between coelacanths were observed. No evidence of a social behavior has been recognized in the Comorian populations of coelacanths (FRICKE et al., 1991).

According to the observations of the African coelacanths, the species L. chalumnae is nocturnal and hides in cave in daytime but comes out at night, supposedly in search of food (FRICKE et al., 1991). The living geomorphologic environment of L. menadoensis is similar to that of L. chalumnae, with rocky steep slopes with undermarine caves or overhangs. However, Latimeria menadoensis was also observed alongside large rocks or along vertical wall in daytime. Some of these individuals were encountered above 200 m deep, in an environment that remains sl ightly bright . Latimeria menadoensis seems to be less sensitive to the daylight than L. chalumnae.

These first observations are potentially crucial since the habitats of the Indonesian coelacanths could be possibly not as

restricted as those of the African coelacanths, and so a broader distribution of the Indonesian coelacanth could be expected. Occurrences of coelacanths along the eastern coast of the African continent were recorded in South Africa (SMITH, 1939), Comoros (SMITH, 1953), Madagascar (HEEMSTRA et al., 1996), Kenya (DE VOS and OYUGI, 2002), Tanzania (BENNO et al., 2006), and Mozambique (BRUTON et al., 1992). In Indonesia, coelacanths have only been found off the northern coast of the Sulawesi Island and the southern coast of the Biak Island in northern New Guinea.

Field surveys of the Indonesian coelacanths should be continued in all area of Indonesia and expanded to all Southeast Asia in order to increase our knowledge of the distribution of this rare species and to define, in a close future, the most relevant conservation policies of the different known populations.

ACKNOWLEDGMENTS

We are grateful to Dr. Teruya UYENO for his valuable advice, guidance and encouragement throughout the present study. Many thanks go to colleagues of the Faculty of Fisheries and Marine Sciences of the Sam Ratulangi University at Manado, Indonesia, for their generous supports to conduct expeditions. Also we express our thanks to Dr. SUHARSONO, Dr. Zainal ARIFIN, Dr. DIRHAMSYAH and colleagues of the Indonesian Institute of Science for their advice and administrative work to obtain our research permit. Other thanks go to Mr. OPO, Mr. REFRY and colleagues at Murex Dive Resort, and DAUD, AL,

and other staffs for our work on the research boat. Dr. BATUNA and the Murex staffs supported our daily life during the expeditions. We would like to express our sincere thanks to colleagues at Aquamarine Fukushima, Japan, for their understanding and supports. We would like to thank Dr. Gaël CLÉMENT of Muséum national d’Histoire naturelle and Dr. Camila CUPELLO of Universidade do Estado do Rio de Janeiro as a referee for their critical reading of the manuscript and their comments.

Unfortunately one of our authors, Dr. Djoko Hadi KUNARSO has passed away during the process of this contribution. We pray for the response of his soul.

This study was supported in part by funding from Interdisciplinary Collaborative Research program of Atmosphere and Ocean Research Institute, the University of Tokyo.

REFERENCES

BENNO, B., VERHEIJI, E., STAPLEY, J., RUMISHA, C., NGATUNGA, B., ABDALLAH, A. and KALOMBO, H. 2006. Coelacanth (Latimeria chalumnae SMITH, 1939) discoveries and conservation in Tanzania. South African Journal of Science, 102: 486–490.

BRUTON, M. N., CABRAL, A. J. P. and FRICKE, H. 1992. First capture of a coelacanth, Latimeria chalumnae (Pisces, Latimeriidae), off Mozambique. South African Journal of Science, 88: 225–227.

CUPELLO, C., BRITO, P. M., HERBIN, M., MEUNIER, F. J., JA N V I E R, P . , DU T E L, H. and CL É M E N T, G. 2015.

IWATA et al. Field surveys on Latimeria menadoensis using ROV from 2005 to 2015

Fig. 3. Water temperature and depth during observations of coelacanths in each locality. The orange box shows depth. The top and bottom of the box mean a minimum depth and maximum depth of coelacanths stay. The blue bar shows a maximum and minimum water temperatures. The open circle shows an average of temperature during observation. All individuals of L. menadoensis were observed in depth between 115.6 and 218.9 m and water temperature between 12.4 and 21.5 ºC.

Fig. 4. Change of the water temperature at Buol-5 off Buol of Sulawesi Island.

Page 7: Field surveys on the Indonesian coelacanth, …4Seikei Education and Research Center for Sustainable Development, Seikei Gakuen, 3-3-1 Kichijoji-Kitamachi ... Guru-Guru, Poka, Ambon,

Allometric growth in the extant coelacanth lung during ontogenetic development. Nature Communications, 6: 8222. DOI: 10.1038/ncomms9222

DE VOS, L. and OYUGI, D. 2002. First capture of a coelacanth, Latimeria chalumnae SMITH, 1939 (Pisces, Latimeriidae), off Kenya. South African Journal of Science, 98: 345–347.

ERDMANN, M. V., CALDWELL, R. L. and MOOSA, M. K. 1998. Indonesian ‘king of the sea’ discovered. Nature, 395: 335.

FRICKE, H., REINICKE, O., HOFER, H. and NACHTIGALL, W. 1987. Locomotion of the coelacanth Latimeria chalumnae in its natural environment. Nature, 329: 331–333.

FRICKE, H., HISSMANN, K., SCHAUER, J., REINICKE, O., KASANG, L. and PLANTE, R. 1991. Habitat and population size of coelacanth Latimeria chalumnae at Grand Comoro. Environmental Biology of Fishes, 32: 287–300.

FRICKE, H., HISSMANN, K., SCHAUER, J., ERDMANN, M., MOOSA, M. K. and PLANTE, R. 2000. Biogeography of the Indonesian coelacanth. Nature, 403: 38.

HEEMSTRA, P. C., FREEMAN, A. L., WONG, H. Y., HENSLEY, D. A. and RABESANDRATANA, H. D. 1996. First authentic

capture of a coelacanth off Madagascar. South African Journal of Science 92: 150–151.

POUYAUD, L., WIRJOATMODJO, S., RACHMATIKA, I., TJAKRAWIDJAJA, A., HADIATY, R. and HADIE, W. 1999. A new species of coelacanth. Genetic and morphologic proof. C. R. Academy of Science, 322: 261–267.

SMITH, J. L. B. 1939. The living coelacanth fish from South Africa. Nature, 143: 748–750.

SMITH, J. L. B. 1953. The second coelacanth. Nature, 171: 99–101.

SMITH, C. L., RAND, C. S., SCHAEFFER, B. and ATZ, J. W. 1975. Latimeria, the living coelacanth, is ovoviviparous. Science, 190: 1105–1106.

VENTER, P., TIMM, P., GUNN, G., LE ROUX, E., SERFONTEIN, E., SMITH, P., SMITH, E., BENSCH, M., HARDING, D. and HEEMSTRA, P. 2000. Discovery of a viable population of coelacanths (Latimeria chalumnae SMITH, 1939) at Sodwana Bay, South Africa. South African Journal of Science, 96: 567–568.

INTRODUCTION

The first living coelacanth, Latimeria chalumnae, was discovered in South Africa in the Indian Ocean in 1938 (SMITH, 1939). The first observation of the living coelacanth habitats using submersibles was reported in Comoros (FRICKE et al., 1987). Latimeria chalumnae hide in submarine caves along steep slopes between 100 m and 200 m depth during daytime and come out and drift near the ocean floor for hunting prey at night (FRICKE et al., 1991). However, since the record of juvenile or small (below 80 cm long) individuals is virtually absent, the growth, development and breeding biology of L. chalumnae remain poorly known. Our knowledge

of the reproduction biology of the coelacanths is almost restricted to an ovoviviparity strategy since some caught large females carried unfertilized eggs or embryos in their oviduct (SMITH et al., 1975; CUPELLO et al., 2015).

The Indonesian coelacanth, L. menadoensis, was first discovered in Manado, Sulawesi Island, Indonesia, in 1997. The first specimen was captured in 1998 (ERDMANN et al., 1998). According to a genetic study, the specimen differed from L. chalumnae and a new species was erected (POUYAUD et al., 1999). In 1999 two individuals were observed by submersible in a location 360 km away from Manado (FRICKE et al., 2000). Today 8 individuals were officially caught in the Indonesian seas. However, the details of their habitat have not

been described. The aim of this study is to clear their habitat and distribution.

METHODS

Two remotely operated vehicles (ROVs) (Kowa; VEGA300) were used for the surveys. The first ROV was replaced with the second one in 2007. Our ROVs are able to dive up to 300 m depth. These had two vertical, two horizontal and two right-left propellers and these were controlled from a boat on surface through a 400 m long tether cable. The underwater operations were visualized on a screen with information including directions of ROVs, depth, water temperature, date and time. All data were directly recorded on video-tapes on the boat. Since the 2007 survey, water temperature and depth were independently recorded with more accuracy by external measuring memories (Alec Electronics, later named JFE Alec; MDS-MKV/T, MDS-MKV/D). In addition two laser beam irradiators, that provide line lasers of 20-cm distance, were attached to the second ROV and have been used to register the size of encountered animals. The surveys were conducted off northern parts of Sulawesi Island and along the Biak Island located in northwestern New Guinea Island, Indonesia. The areas and time periods of each survey are shown in Table 1 and Fig. 1.

RESULTS

The field survey of the Indonesian coelacanth, Latimeria menadoensis, was conducted 14 times from 2005 to 2015 with 1173 underwater operations around Sulawesi and Biak islands in Indonesia (Fig. 1, Table 1). No coelacanth was encountered during the first surveys in 2005 although 452 underwater operations were conducted around Manado-tua Island and other close Indonesian islands from the 17th to 30th of April (Table 1, R1). The second field survey was held along the northern coast of Sulawesi from the 6th to 19th May in 2006, but once again despite of 107 underwater operations no coelacanth was recorded.

The first individual of Indonesian coelacanth (Table 2, ID 1) was recorded during the third field survey, in the morning of the 30th May 2006 (Table 1, R3). The coelacanth was observed in a cave of 165 m depth (Table 2, Boul-1) for 10 minutes (Table 2, E1). The adult individual was in stationary position in a cave (Buol-1) slowly moving its pectoral, pelvic, second dorsal and anal fins. The day after, another individual (ID 2) and an unidentified one (UN1) were found in the same cave (Buol-1). The unidentified one immediately hid deeper in the cave, where the ROV could not access, and only its lateral side of the body was recorded. The individual (ID 2) was observed for one minute before it hid inside the cave (Table 2: E2). When a new individual could be identified by the white

spots pattern on its body, an ID number was allocated. The individual which could not be distinguished was allocated an UN (unknown) + number.

In the end of the morning of the same day, a third individual (Table 2: E3, ID 3) was found in another cave (Buol-2), larger and located 20 m deeper (Buol-2) than the first cave (Buol-1). After 29 minutes of recording, this individual hid inside the cave. Although the water temperature increased of more than five degrees, from 14.8 to 20.4 ºC, it stayed at the same place, not moving away. Despite it disappeared from the camera frame we kept setting the ROV there in front of the cave until the next morning, however it did not appear again.

On the 4th June in 2006, three individuals, including ID 3 which was observed on the 31th May at Buol-2 and two new individuals, were observed in a vertical crack which was several hundred meters far from Buol-2 (Buol-3) for 2 hours and 14 minutes (Fig. 2C; Table 2: E4, ID 3, 4, 5). During this long period of time all individuals stayed near the rock wall always keeping their head against the water flow. The water

54 55

temperature changed up and down by five degrees during the observation due to water flow reversals. When the water flow was reversed some times, all individuals always changed their directions against the water flow.

During the field survey carried on December of the same year, in the same Buol area (Table 1: R4) the sixth individual (Table 2: E5, ID 6) was recognized at 14:34 on the 12th December alongside a large rock. This individual stayed in the beginning of the observation and then moved ahead to deeper slowly. Eventually it reached 12 m deeper during one hour and 42 minutes.

On the 14th December, the individual ID 3 previously recorded at Buol-2 in a cave and at Buol-3 in a crack, was encountered again (Table 2: Buol-5) near a vertical wall at 145 m depth (Table 2: E6). After a while, this individual (ID 3) moved upward about 18 m and stayed behind a large rock (Fig. 3). The observation time was the longest of all surveys with four hours and 16 minutes of recording. The temperature changed up and down several times (Fig. 4). Its difference

between the maximum and the minimum was 6.7 °C, which was the largest range of all surveys.

On the next day, on the 15th December, the ID 3 was observed again under an overhang (Table 2: E7, Buol-6) near other sites where it was previously found at Buol-2 , Buol-3 and Buol-5. Its posture was upside down with its ventral side facing the ceiling of the overhang. Such posture is also known in coelacanths from the Comorian Archipelago (FRICKE et al., 1987). ID 3 was found totally four times in this area in two surveys during the seven months and seems to live in this area.

During the 2007 survey (Table 1: R5) a coelacanth (Table 2: ID 7) was recorded off Manado by 195 m depth, with a temperature of 12.4 °C, the coolest one registered during all surveys.

The field survey in 2009 was carried out off Manado and around Talise and Bangka islands off the northern coast of Sulawesi Island from the 12th September to the 9th October (Table 1: R8). On the 24th September 2009, six individuals (Table 2: E11, ID 9, 10, 11, 12, UN 2, 3) were found under an

overhang off Talise Island between 144 and 150 m deep. There was almost no water current in this area. All individuals were close to each other under the same overhang and most of them stayed still with their head facing downward. Four days later, ID 9 was observed again, alone on a steep slope at 172.9 m depth (Table 2: E12), which was located within several dozen meters far from Talise-2 and 30 m deeper.

Four new coelacanths were recorded from the 29th September to the 6th October 2009 in Talise and Banggka Islands and off Manado (Table 2: E13 and 14, ID 13, 14, 15). ID 15 is the only juvenile coelacanth filmed so far. It was in a small crack between 164.6 and 170.9 m deep in a temperature between 14.5 and 15 °C.

The shallowest record of a coelacanth was at 115.6 m depth off the southern part of Manado (Table 2: E16, Manado-3) on the 9th October in 2009. ID 16 (Table 2) was found at the end of the morning, alone alongside a large rock on the edge of a shelf. The water temperature was about 20.0 ºC.

From the 6th to the 16th November in 2010, the field survey was conducted around Biak Island (Table 1: R9). Adult coelacanths were found at two different sites (Table 2: E17–19). Two individuals (ID 17 and 18) were found under an overhang between 212.5 and 218.9 m deep (Table 2: Biak-1) on the 11th November. Three other ones (ID 19, 20, 21) were registered under an overhang between 193.2 and 195.9 m deep (Table 2: Biak-2) on the 13th November. Two days later, Biak-2 was observed again and only ID 20 stayed there.

Two new coelacanths were recorded off Manado in December 2010 (Table 1: R10; Table 2: E 20 and 21; ID 22 and 23).

On the 4th May 2012, an individual (Table 2: E22, ID 24) was found on a steep slope at 169.4 m depth off Manado (Table 2: Manado-6). It swam up and down 5.4 m between 168.1 and 173.5 m deep for 17 minutes. Thirteen months later (the 9th June 2003), this individual (ID 24) was observed again near a cliff at 152.3 m depth (Table 2: E24, Manado-8), where located on the same slope of Manado-6.

Another new coelacanth (Table 2: E23, ID 25) was recorded alone on a Manado site (Manado-7).

Field surveys were conducted from the 19th to 30th May 2015 off Lolak Island and from the 30th October to the 16th November 2015 off Lolak and Bitung Islands (Table 1; R13 and R14). Three new individuals (Table 2: ID 26–28) were recognized at different places off Lolak Island (Lolak 1-4). All individuals were alone. ID 26 was observed in May and once

again in October 2015 on a steep slope (Table 2: E25, E27). The coelacanth ID 28 was found on the 1st November along one of the large rocks scattered on a gentle slope in a bay of the Lolak Island by 125 m depth.

The encountered Indonesian coelacanths during these surveys were in caves, alongside large rocks, under overhangs or on steep slopes (Table 2; Fig. 2).

Most individuals were found under overhangs or alongside large rocks. Some individuals were encountered not to hide in any shade and stayed just on rocky slopes. In overhangs and cracks, all individuals had their ventral side alongside and close to a rock wall, but without touching it.

Estimated total lengths are shown in Table 3. Most individuals are more than 1 m in total length. ID 15 is a juvenile individual of 31.5 cm and ID 23 is 90 cm in total length. No individual beyond 140 cm has been recorded in Indonesia so far.

DISCUSSION

During the field surveys by underwater ROV recording for Indonesian coelacanths, Latimeria menadoensis, from 2005 to 2015, 30 different individuals were identified and three others were unidentified. Among those, six individuals were observed more than twice: ID 3 at Buol-2, 3, 5, 6 (Table 2: E3, E4, E6 and E7); ID 8 at Talise-1 (E9 and E10), ID 9 at Talise 2 and 3 (E11 and E12), ID 20 at Biak-2 (E18 and E19), ID 24 at Manado-6 and 8 (E22 and E24), ID 26 at Lolak-1 and 3 (E25 and E27). ID 8 was observed on the 14th September 2009 and was encountered again at the same place (Talise-1) the day after. ID 24 was observed on the 4th May 2012 and then 13 months later, on the 9th June 2013 at the same locations of Manado area within several kilometers.

Al l individuals observed more than twice were encountered in the same area, at locations close from each other within hundreds of meters. However, in the Comorian Archipelago, same individuals of L. chalumnae were observed, within a two week period, in several caves distributed in a 8 kilometers wide area (FRICKE et al., 1991).

When individuals of L. menadoensis were encountered, some of them stayed stationary at same place but some individuals such as ID 6 on the 12th December 2006 (E5) and ID 24 on the 4th May 2012 (E22) swam away. It seemed to avoid and escape from the brightness of the light or the sounds

of the propellers of the ROV.Temperature and depth data for each Indonesian

coelacanth record are shown in Fig. 3. Though an average of water temperature on the ocean surface was about 30 °C, the temperature often dropped into about 11 °C at 300 m water depth. The temperatures changed widely even though the changes of depth were small at several places. All Indonesian coelacanths, L. menadoensis, were encountered during day time at a depth between 115.6 and 218.9 m and a temperature between 12.4 and 21.5 ºC. Indonesian coelacanths have exceptionally been recorded in an environment with a temperature exceeding 20 ºC, regardless of the depth. The African coelacanth, L. chalumnae, was observed by submarine vehicles in steep slopes between 150 m and 253 m (FRICKE et al., 1991). Divers recorded the presence of this species at 104 m depth at Sodwana Bay, South Africa (VENTER et al., 2000). Latimeria chalumnae seem to prefer a temperature range of 15–20 °C, where they tend to choose their daytime habitats (FRICKE et al., 1991). The suitable temperature for L. menadoensis seems to be almost the same as L. chalumnae.

Water temperature can increase or drop for more than 5 °C in a very short period of time (about 30 min according to our recordings). The coelacanths however often stayed there without moving away or significantly modifying their behavior. These water currents of different temperatures are caused by the movement of the marine thermocline. During our surveys we observed in some area that as the thermocline

level changed, the water flow was reversed. Coelacanths always kept their head against the water flow, like other observed actinopterygian fishes did. It means that as the direction of the water flow changed, the swimming direction of each coelacanth individual changed accordingly.

On the 24th September 2009 (Table 2: E11) six individuals were recognized under an overhang (Talise-2). The all had about the same size. There was almost no water flow, and each individual was head downward. They were close to each other but no direct contacts or interactions between coelacanths were observed. No evidence of a social behavior has been recognized in the Comorian populations of coelacanths (FRICKE et al., 1991).

According to the observations of the African coelacanths, the species L. chalumnae is nocturnal and hides in cave in daytime but comes out at night, supposedly in search of food (FRICKE et al., 1991). The living geomorphologic environment of L. menadoensis is similar to that of L. chalumnae, with rocky steep slopes with undermarine caves or overhangs. However, Latimeria menadoensis was also observed alongside large rocks or along vertical wall in daytime. Some of these individuals were encountered above 200 m deep, in an environment that remains sl ightly bright . Latimeria menadoensis seems to be less sensitive to the daylight than L. chalumnae.

These first observations are potentially crucial since the habitats of the Indonesian coelacanths could be possibly not as

restricted as those of the African coelacanths, and so a broader distribution of the Indonesian coelacanth could be expected. Occurrences of coelacanths along the eastern coast of the African continent were recorded in South Africa (SMITH, 1939), Comoros (SMITH, 1953), Madagascar (HEEMSTRA et al., 1996), Kenya (DE VOS and OYUGI, 2002), Tanzania (BENNO et al., 2006), and Mozambique (BRUTON et al., 1992). In Indonesia, coelacanths have only been found off the northern coast of the Sulawesi Island and the southern coast of the Biak Island in northern New Guinea.

Field surveys of the Indonesian coelacanths should be continued in all area of Indonesia and expanded to all Southeast Asia in order to increase our knowledge of the distribution of this rare species and to define, in a close future, the most relevant conservation policies of the different known populations.

ACKNOWLEDGMENTS

We are grateful to Dr. Teruya UYENO for his valuable advice, guidance and encouragement throughout the present study. Many thanks go to colleagues of the Faculty of Fisheries and Marine Sciences of the Sam Ratulangi University at Manado, Indonesia, for their generous supports to conduct expeditions. Also we express our thanks to Dr. SUHARSONO, Dr. Zainal ARIFIN, Dr. DIRHAMSYAH and colleagues of the Indonesian Institute of Science for their advice and administrative work to obtain our research permit. Other thanks go to Mr. OPO, Mr. REFRY and colleagues at Murex Dive Resort, and DAUD, AL,

and other staffs for our work on the research boat. Dr. BATUNA and the Murex staffs supported our daily life during the expeditions. We would like to express our sincere thanks to colleagues at Aquamarine Fukushima, Japan, for their understanding and supports. We would like to thank Dr. Gaël CLÉMENT of Muséum national d’Histoire naturelle and Dr. Camila CUPELLO of Universidade do Estado do Rio de Janeiro as a referee for their critical reading of the manuscript and their comments.

Unfortunately one of our authors, Dr. Djoko Hadi KUNARSO has passed away during the process of this contribution. We pray for the response of his soul.

This study was supported in part by funding from Interdisciplinary Collaborative Research program of Atmosphere and Ocean Research Institute, the University of Tokyo.

REFERENCES

BENNO, B., VERHEIJI, E., STAPLEY, J., RUMISHA, C., NGATUNGA, B., ABDALLAH, A. and KALOMBO, H. 2006. Coelacanth (Latimeria chalumnae SMITH, 1939) discoveries and conservation in Tanzania. South African Journal of Science, 102: 486–490.

BRUTON, M. N., CABRAL, A. J. P. and FRICKE, H. 1992. First capture of a coelacanth, Latimeria chalumnae (Pisces, Latimeriidae), off Mozambique. South African Journal of Science, 88: 225–227.

CUPELLO, C., BRITO, P. M., HERBIN, M., MEUNIER, F. J., JA N V I E R, P . , DU T E L, H. and CL É M E N T, G. 2015.

IWATA et al. Field surveys on Latimeria menadoensis using ROV from 2005 to 2015

Fig. 3. Water temperature and depth during observations of coelacanths in each locality. The orange box shows depth. The top and bottom of the box mean a minimum depth and maximum depth of coelacanths stay. The blue bar shows a maximum and minimum water temperatures. The open circle shows an average of temperature during observation. All individuals of L. menadoensis were observed in depth between 115.6 and 218.9 m and water temperature between 12.4 and 21.5 ºC.

Fig. 4. Change of the water temperature at Buol-5 off Buol of Sulawesi Island.

Page 8: Field surveys on the Indonesian coelacanth, …4Seikei Education and Research Center for Sustainable Development, Seikei Gakuen, 3-3-1 Kichijoji-Kitamachi ... Guru-Guru, Poka, Ambon,

ACKNOWLEDGEMENTS

We are grateful to Dr. Teruya UYENO for his valuable advice, guidance and encouragement throughout the present study. Many thanks go to colleagues of the Faculty of Fisheries and Marine Sciences of the Sam Ratulangi University at Manado for their generous supports to conduct expeditions. Also, we express our thanks to Dr. SUHARSONO, Dr. Zainal ARIFIN, and colleagues of the Indonesian Institute of Science for advices and administrable work to obtain our research permit. Thanks also to Dr. BATUNA and the Murex Dive Resort staffs. We would like to express our sincere thanks to colleagues at Aquamarine Fukushima for their understanding and supports. We thank Dr. Rik NULENS to provide information of CCC records. We would like to thank Dr. Camila CUPELLO of Departamento de Zoologia, Universidade do Estado do Rio de Janeiro as a referee for her critical reading of the manuscript

and her comments. This study was supported in part by funding from Interdisciplinary Collaborative Research Program of Atmosphere and Ocean Research Institute, The University of Tokyo.

REFERENCES

BENNO, B., VERHEIJ, E., STAPLEY, C. R., NGATUNGA, B., ABDALLAH, A. and KALOMBOM, H. 2006. Coelacanth (Latimeria chalumnae SMITH, 1939) discoveries and conservation in Tanzania. South African Journal of Science, 102: 486–490.

BRUTON, M. N., CABRAL, A. J. P. and FRICKE, H. 1992. First capture of a coelacanth, Latimeria chalumnae (Pisces, Latimeriidae), off Mozambique. South African Journal of Science, 88: 225–227.

CUPELLO, C., BRITO, P. M., HERBIN, M., MEUNIER, F. J., JA N V I E R, P . , DU T E L, H. and CL É M E N T, G. 2015.

INTRODUCTION

The first individual of the extant coelacanth, Latimeria chalumnae, was discovered in South Africa in 1938 (SMITH, 1939). The first observation of the living coelacanth habitats using submersible was reported in Comoros (FRICKE et al., 1987). Juveniles or small individuals of Latimeria are rarely obsereved or caught, and the growth and reproductive biology of extant species of Latimeria remain mostly unknown. The only knowledge of its reproduction is that Latimeria is

ovoviviparous because captured female specimens carried developing embryos (SMITH et al., 1975, WOURMS et al., 1991).

Aquamarine Fukushima has conducted field surveys for Indonesian coelacanth, Latimeria menadoensis from 2005 to 2015 (IWATA et al., 2019). During the surveys a small coelacanth was observed in 2009. This discovery was simply reported as a scientific news (HOLDEN, 2009). Heterochrony of recent coelacanths was discussed based on this discovery. However, the detail of findings was not described. In the present paper, this finding will be described in detail and the probable habitat

of juveniles of coelacanth is discussed. In addition, the detail measurements and counts of the juvenile based on video footages are made and compared to embryos of L. chalumnae, which are from a large female (CCC no. 162) caught off Mozambique water in 1991 (BRUTON et al., 1992). This morphological data, based on extant coelacanths, should provide invaluable information to the study of fossil and extant coelacanths.

MATERIALS AND METHODS

Materials are the video footage of the juvenile of Indonesian coelacanth L. menadoensis (Fig. 1) and photos of the embryos of the African coelacanth L. chalumnae, which are CCC no. 162.16 housed in the J. L. B SMITH Institution, South Africa (Fig. 2A) (SECRETARIAT AT THE J. L. B. SMITH INSTITUTE OF ICHTHYOLOGY, 1993, fig. 3), CCC no. 162.12 (Fig. 2B) and CCC no. 162.22 (Fig. 2C) housed in University of Guelph (HENSEL and BALON, 2001, fig. 2a, b). CCC stands for the Coelacanth Conservation Council. The video footages including this finding were taken by a remotely operated vehicle (ROV) (Kowa; HDTV VEGA-300) in the north of Sulawesi Island and Biak Island in Indonesia from 2005 to 2015 (IWATA et al., 2019). The video was recorded with two-line lasers placed 20 cm apart. These lines were applied by two laser beam irradiators attached to the ROV. The measurements of the juvenile were calculated based on the beams (Fig. 3; Table 1) and the fin ray counts are made from the video footages (Fig. 1). Among the embryos, the total length of CCC no. 162.12, 162.16 and 162.22 are 33.3 cm, 34.8 cm and 34.4 cm respectively (SECRETARIAT AT THE J. L. B. SMITH INSTITUTE OF ICHTHYOLOGY, 1993). The body part lengths and proportion to the total length of the three embryos are calculated by measuring the figures (Table 1). In the present study, names of fins follow that of FOREY (1998) which differ from UYENO (1991). Dorsal lobe of caudal fin, ventral lobe of caudal fin and supplementary lobe of caudal fin in FOREY (1998) are 3rd dorsal fin, 2nd anal fin and caudal fin of UYENO (1991), respectively. The terminology of fins by UYENO (1991) is given in parentheses to avoid confusion of names. Fin names are abbreviated in the tables as follows. A1: Anal fin. A2: Ventral lobe of caudal fin. C: Supplementary lobe of caudal fin. D1: First dorsal fin. D2: Second dorsal fin. D3: Dorsal lobe of caudal fin. P1: Pectoral fin. P2: Pelvic fin. Measurements are abbreviated as follows. HL: Head length. SL: Standard length. TL: Total length.

RESULTS

ROV observationsThe juvenile of Latimeria menadoensis (ID 15 in IWATA

et al., 2019) was found on the southern coast of Manado Bay of Sulawesi, Indonesia, on the 6th October in 2009 (Fig. 1). The individual stayed in a narrow overhang at 164.6 m depth (Fig. 4A). The overhang was long and progressively became deeper (Fig. 4B). The height of the overhang seemed narrower than the observed juvenile standard length (Fig. 3). The juvenile was observed for 17 minutes from 11:34. It stayed below the overhang and moved slowly, between the depth of 164.6 and 170.9 m. The water current was gentle and the water temperature during the observation was stable, from 14.5 ºC to 15.0 ºC, the average temperature was 14.8 ºC.

Coelacanth has six lobed fins (two pectoral, two pelvic, one dorsal, and one anal fins). It exhibited a unique swimming style by paddling and twisting the lobed fins (FRICKE and HISSMANN, 1992). The swimming of the juvenile was similar to those of adults. During the observation of the juvenile, one cycle of paddling was five seconds and the juvenile swam forward and backward. The paddling pattern and the cycle were the same both the forward and backward swimming.

The dorso-posterior edge of the first dorsal fin, the dorsal edge of the dorsal lobe of the caudal fin (the third dorsal fin), and the ventral edge of the ventral lobe of the caudal fin (the second anal fin) are white. The dorsal margin of the head is convex and angular, where seems to be the intracranial joint (Fig. 1B). Scales have many minute spines, which reflect the light of the ROV.

Comparative description of the juvenile of L. menadoensis and embryos of L. chalumnae

The total length of the juvenile is 31.5 cm. The standard length is 26.1 cm. It is smaller than the embryos of L. chalumnae used here for comparison, which are 30.8–35.8 cm (BRUTON et al., 1992) and it is smaller than the smallest specimen (CCC no. 94) of L. chalumnae that has been collected is 42.5 cm TL.. Drawings of the L. menadoensis juvenile and L. chalumnae embryos (CCC no. 162.16) are shown in Fig. 5. The body of the Indonesian coelacanth juvenile is slender than that of the L. chalumnae embryos (Table 1). The body depth is 6.8 cm, 26 % of the standard length in the L. menadoensis juvenile and 8.7 cm and 28 % in L. chalumnae (Table 1). The head is slightly shorter than that of L. chalumnae. It is 5.5 cm and 22 % of the standard length in L. menadoensis and 8.0 cm and 26 % in L. chalumnae. The depth of the head is slightly smaller than that of L. chalumnae. The mouth is slightly smaller than that of L. chalumnae. The gape of the mouth is probably smaller than that of L. chalumnae. The eye is smaller than that of L. chalumnae. The orbit diameter of the L. menadoensis juvenile is almost half that of L. chalumnae. The part (caudal peduncle) between the second dorsal and the anal fins and the anterior ends of the dorsal lobe of the caudal fin (the third dorsal fin) and the ventral lobe of the caudal fin (the second anal fin) is shorter and deeper than that of L. chalumnae. The distance between the posterior end of the base of the second dorsal fin peduncle and the anterior end of the

Bull. Kitakyushu Mus. Nat. Hist. Hum. Hist., Ser. A, 17: 57–65, March 31, 2019

Observation of the first juvenile Indonesian coelacanth, Latimeria menadoensis from Indonesian waters with a comparison to embryos of Latimeria chalumnae

ABSTRACT − The juvenile of Indonesian coelacanth, Latimeria menadoensis is here described for the first time in detail with comparison to embryos of Latimeria chalumnae. The juvenile was found in free swimming at 164.6 m depth off Manado, Indonesia on the 6th October in 2009. Because the total length of the juvenile is 31.5 cm, which is smaller than the embryos of L. chalumnae, it is speculated that not much time has passed from its birth. The depth at which the juvenile was found is within the range of the depth where adult L. menadoensis were observed, hidden in a narrow and long overhang where large predators could not enter. The juvenile has a more slender body, smaller orbit, shorter and deeper posterior part of the body (caudal peduncle) between the second dorsal and the anal fins and anterior ends of the dorsal and ventral lobes of the caudal fin (the third dorsal and second anal fins), longer dorsal and ventral lobes of the caudal fin (the third dorsal and second anal fins), broader peduncles of broader lobed fins, larger first dorsal fin and longer supplementary lobe of the caudal fin (caudal fin) than embryos of L. chalumnae. This indicates clear differences in the first ontogenetic stages of the two species, although adults have almost the same morphological features. Latimeria menadoensis appears to reproduce in a rather confined area, because both the juvenile and adults have been found within the same area inside of Manado Bay.

KEY WORDS: habitat, heterochrony, Indonesian coelacanth, juvenile, Latimeria menadoensis

Masamitsu IWATA1, Yoshitaka YABUMOTO2, Toshiro SARUWATARI3,4, Shinya YAMAUCHI1, Kenichi FUJII1, Rintaro ISHII1, Toshiaki MORI1, Frensly D. HUKOM5, DIRHAMSYAH5,Teguh PERISTIWADY6, Augy

SYAHAILATUA7, Kawilarang W. A. MASENGI8,9, Ixchel F. MANDAGI8,9, Fransisco PANGALILA8,9 and Yoshitaka ABE1

1Aquamarine Fukushima, Marine Science Museum, 50 Tatsumi-cho, Onahama, Iwaki, Fukushima, 971-8101, Japan2Kitakyushu Museum of Natural History and Human History, 2-4-1 Higashida, Yahata Higashi-ku,

Kitakyushu, Fukuoka, 805-0071, Japan3Atmosphere and Ocean Research Institute, The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa,

Chiba, 277-8564, Japan4Seikei Education and Research Center for Sustainable Development, 3-3-1 Kichijoji-Kitamachi,

Musashino-shi, Tokyo, 180-8633, Japan5Research Center for Oceanography, Indonesian Institute of Science, Jakarta, Indonesia

6Division of Marine Resources Research, Indonesian Institute of Sciences, Jakarta, Indonesia7Centre for Deep Sea Research,Indonesian Institute of Sciences, Jakarta, Indonesia

8International Coelacanth Research Center and Marine Museum, Sam Ratulangi University, Manado 95115, Indonesia9Faculty of Fisheries and Marine Science, Sam Ratulangi University, Manado 95115, Indonesia

(Received August 24, 2018; accepted November 9, 2018)

fins, longer fins, larger principal caudal fin (the third dorsal and second anal fins), and longer supplementary lobe of the caudal fin (caudal fin) than those of the embryos of African coelacanth.

One individual of L. menadoensis, 90 cm of total length and the second smallest one among the coelacanths observed in Indonesian water from 2005 to 2015 (IWATA et al., 2019), was observed at 187.4 m depth on the 6th December 2012 (Fig. 6: ID 23 in IWATA et al., 2019). Although ID 23 is the second smallest individual among the coelacanths observed in Indonesian water, it is considered to be an adult because its morphological feature is almost identical to other adults. Also, juvenile characters, observed in the juvenile here described, are not observed in ID 23. The body proportion of the juvenile is very different from that of the adult in L. menadoensis (Figs. 1 and 6). The morphological differences between L.

menadoensis and L. chalumnae are scarce when compared between adults. However, there are several differences between juveniles and embryos of these species (Fig. 5). This indicates possibility of the ontogenetic difference of the two species.

The juvenile of L. menadoensis has a long supplemental lobe of the caudal fin, which is seen in adult and often young specimens of some fossil species and early stages of embryos of L. chalumnae (BENNO et al., 2006, fig. 2; CUPELLO et al., 2015, fig. 1a; FOREY, 1998; YABUMOTO, 2008), indicating L. menadoensis retained more juvenile characters during its development than L. chalumnae. This indicates presence of heterochrony in the development of Latimeria, with L. menadoensis being paedomorphic during its juvenile stage.

dorsal lobe of the caudal fin (the third dorsal fin) is shorter than that of L. chalumnae (Fig. 5). Positions of fins are almost the same in both species, except for the dorsal lobe of the caudal fin (the third dorsal fin). The anterior end of the dorsal lobe of the caudal fin base is located more anteriorly than that of L. chalumnae (Fig. 5). The bases of the dorsal lobe of the caudal fin (the third dorsal fin) and the ventral lobe of the caudal fin (the second anal fin) are longer than those of L. chalumnae (Table 1). The base of the dorsal lobe of the caudal fin (the third dorsal fin) is longer than that of the ventral lobe of the caudal fin (the the second anal fin) in both species. The supplementary lobe of the caudal fin (caudal fin) is longer than that of L. chalumnae. All fins are larger and broader than those of L. chalumnae. Peduncles of all lobed fins are also broader than those of L. chalumnae. The juvenile of L. menadoensis has 8 spines of the first dorsal fin; 26 rays of the secod dorsal fin; 23 rays of the anal fin; 25 rays of the dorsal lobe of the caudal fin (the third dorsal fin); 23 rays of the ventral lobe of the caudal f in ( the second anal f in) ; 26 rays of the supplementary lobe of the caudal fin (caudal fin) ; 29 rays of

the pectoral fin and 26 rays of the pelvic fin.

DISCUSSION

Latimeria chalumnae is known to be ovoviviparous (SMITH et al., 1975). To date, no females of L. menadoensis with matured eggs or embryos has been captured. Reproductive organs of L. chalumnae and L. menadoensis are almost the same anatomically (our observation) and, consequently, L. menadoensis is here considered to be ovoviviparous as well. A large female of L. chalumnae that held juveniles from 30.8 cm to 35.8 cm inside its body was caught off Mozambique water in 1991 (BRUTON et al., 1992). Latimeria chalumnae embryo CCC no. 29.5 found inside CCC no. 29 is 32.3 cm TL with a yolk sac, and CCC no, 162.21 is a late embryo without a yolk sac of 35.6 cm TL found inside CCC no. 126 captured in 1991 (CUPELLO et al., 2015). CCC no. 29.5 is housed in Muséum national d’Histoire naturelle, Paris (France) and CCC no. 162.21 is housed in Zoologisches Staatsaammling, München (Germany). These specimens suggest that L. chalumnae delivers juveniles of around 30 cm in total length. Since the smaller individuals of L. chalumnae observed in their habitat were from 50 to 60 cm long, the ecology of juveniles remain poorly known (FRICKE et al., 2011). According to CCC records, the size of captured small coelacanths are: 50 cm for CCC no. 84 caught in Comoros in 1973; 42.5 cm TL for CCC no. 94 caught in Comoros in 1974; 60 cm TL for CCC no. 116 caught in Comoros in 1979; and 62 cm for CCC no. 160 caught in Comoros in 1989 (NULENS et al., 2011).

The underwater observation at 164.6 m depth off Manado, Indonesia on the 6th October in 2009 is the first account of a live juvenile of L. menadoensis in its natural habitat. The juvenile is 31.5 cm in total length. The estimated size at newborn of L. chalumnae is around 30 cm as mention earlier. This suggests that not much time has passed since the observed juvenile L. menadoensis was born. About 30 adult coelacanths of about 1 m long were found at the depth between 115.6 m and 218.9 m in Indonesia (IWATA et al. 2019). The depth of 164.6 m at which the juvenile was recorded is well within the range of adult habitat. The individual of 113 cm total length (ID 16 in Iwata et al., 2019), was observed at 115.6 m depth in the same area three days after the juvenile was found. Presence of both the juvenile and adults of L. menadoensis within the same area inside of Manado Bay, suggests that the species reproduce in a rather confined area.

Comparisons of the juvenile L. menadoensis and the embryos of L. chalumnae show some interesting facts about ontogeny of the congeneric species. The embryos of L. chalumnae are larger than the juvenile L. menadoensis and have almost completely absorbed their yolk reserves according to CCC newsletter (SECRETARIAT AT J. L. B. SMITH INSTITUTE OF ICHTHYOLOGY, 1993), just before delivery. The juvenile of L. menadoensis has a smaller eye, larger peduncles of lobed

Allometric growth in the extant coelacanth lung during ontogenetic development. Nature Communications, 6: 8222. DOI 10.1038/ncomms9222

FOREY, P. L. 1998. History of the Coelacanth Fishes. Chapman and Hall, London, xiii+419 pp.

FRICKE, H., REINICKE, O., HOFER, H. and NACHTIGALL, W. 1987. Locomotion of the coelacanth Latimeria chalumnae in its natural environment. Nature, 329: 331–333.

FRICKE, H. and HISSMANN, K. 1992. Locomotion, fin coordination and body form of the living coelacanth, Latimeria chalumnae. Environmental Biology of Fishes, 34: 329–356.

FRICKE, H., HISSMANN, K., FROESE, R., SCHAUER, J., PLANTE, R. and FRICKE, S. 2011. The populations biology of the living coelacanth studied over 21 years. Marine Biology, 158: 1511–1522. DOI 10.1007/s00227-011-1667-x

HOLDEN, C. 2009. Infant fossil. Science, 326: 1327.HENSEL, K. and BALON, K. 2001. The sensory canal system of

the living coelacanth, Latimeria chalumnae: a new instalment. Environmental Biology of Fishes, 61: 117–124.

IWATA, M., YABUMOTO, Y., SARUWATARI, T., YAMAUCHI, S., FUJII, K., ISHII, R., MORI, T., HUKOM, F., DIRHAMSYAH, PERISTIWADY, T., SYAHAILATUS, A., MASENGI, K. W. A., MANDAGI, I. F. and ABE, Y. 2019. Field Survey on the

Indonesian coelacanth, Latimeria menadoensis using remotely operated vehicles from 2005 to 2015. Bulletin of Kitakyushu Museum of Natural History and Human History, Series A, Natural History, 17: 49–56.

NULENS, R., SCOTT, L. and HERBIN, M. 2011. An updated inventory of all known specimens of the coelacanth Latimeria spp. South African Institute for Aquatic Biodiversity, 3: 1–52.

SECRETARIAT AT THE J. L. B. SMITH INSTITUTE OF ICHTHYOLOGY. 1993. Coelacanth conservation council newsletter no.4. Environmental Biology of Fishes, 36: 395–406.

SMITH, C. L., RAND, C. S., SCHAEFFER, B. and ATZ, J. W. 1975. Latimeria, the living coelacanth, is ovoviviparous. Science, 190: 1105–1106.

UYENO, T. 1991. Observations on locomotion and feeding of released coelacanths, Latimeria chalumnae. Environmen-tal Biology of Fishes, 32: 267–273.

WOURMS, J. P., ATZ, J. W. and STRIBLING, M. D. 1991, Viviparity and meternal-embryonic relationship in the coelacanth Latimeria chalumnae. Environmental Biology of Fishes, 32: 225–248.

YABUMOTO, Y. 2008. A new Mesozoic coelacanth from Brazil (Sarcopterygii, Actinistia). Paleontological Research, 12: 329–343.

56

Allometric growth in the extant coelacanth lung during ontogenetic development. Nature Communications, 6: 8222. DOI: 10.1038/ncomms9222

DE VOS, L. and OYUGI, D. 2002. First capture of a coelacanth, Latimeria chalumnae SMITH, 1939 (Pisces, Latimeriidae), off Kenya. South African Journal of Science, 98: 345–347.

ERDMANN, M. V., CALDWELL, R. L. and MOOSA, M. K. 1998. Indonesian ‘king of the sea’ discovered. Nature, 395: 335.

FRICKE, H., REINICKE, O., HOFER, H. and NACHTIGALL, W. 1987. Locomotion of the coelacanth Latimeria chalumnae in its natural environment. Nature, 329: 331–333.

FRICKE, H., HISSMANN, K., SCHAUER, J., REINICKE, O., KASANG, L. and PLANTE, R. 1991. Habitat and population size of coelacanth Latimeria chalumnae at Grand Comoro. Environmental Biology of Fishes, 32: 287–300.

FRICKE, H., HISSMANN, K., SCHAUER, J., ERDMANN, M., MOOSA, M. K. and PLANTE, R. 2000. Biogeography of the Indonesian coelacanth. Nature, 403: 38.

HEEMSTRA, P. C., FREEMAN, A. L., WONG, H. Y., HENSLEY, D. A. and RABESANDRATANA, H. D. 1996. First authentic

capture of a coelacanth off Madagascar. South African Journal of Science 92: 150–151.

POUYAUD, L., WIRJOATMODJO, S., RACHMATIKA, I., TJAKRAWIDJAJA, A., HADIATY, R. and HADIE, W. 1999. A new species of coelacanth. Genetic and morphologic proof. C. R. Academy of Science, 322: 261–267.

SMITH, J. L. B. 1939. The living coelacanth fish from South Africa. Nature, 143: 748–750.

SMITH, J. L. B. 1953. The second coelacanth. Nature, 171: 99–101.

SMITH, C. L., RAND, C. S., SCHAEFFER, B. and ATZ, J. W. 1975. Latimeria, the living coelacanth, is ovoviviparous. Science, 190: 1105–1106.

VENTER, P., TIMM, P., GUNN, G., LE ROUX, E., SERFONTEIN, E., SMITH, P., SMITH, E., BENSCH, M., HARDING, D. and HEEMSTRA, P. 2000. Discovery of a viable population of coelacanths (Latimeria chalumnae SMITH, 1939) at Sodwana Bay, South Africa. South African Journal of Science, 96: 567–568.

INTRODUCTION

The first living coelacanth, Latimeria chalumnae, was discovered in South Africa in the Indian Ocean in 1938 (SMITH, 1939). The first observation of the living coelacanth habitats using submersibles was reported in Comoros (FRICKE et al., 1987). Latimeria chalumnae hide in submarine caves along steep slopes between 100 m and 200 m depth during daytime and come out and drift near the ocean floor for hunting prey at night (FRICKE et al., 1991). However, since the record of juvenile or small (below 80 cm long) individuals is virtually absent, the growth, development and breeding biology of L. chalumnae remain poorly known. Our knowledge

of the reproduction biology of the coelacanths is almost restricted to an ovoviviparity strategy since some caught large females carried unfertilized eggs or embryos in their oviduct (SMITH et al., 1975; CUPELLO et al., 2015).

The Indonesian coelacanth, L. menadoensis, was first discovered in Manado, Sulawesi Island, Indonesia, in 1997. The first specimen was captured in 1998 (ERDMANN et al., 1998). According to a genetic study, the specimen differed from L. chalumnae and a new species was erected (POUYAUD et al., 1999). In 1999 two individuals were observed by submersible in a location 360 km away from Manado (FRICKE et al., 2000). Today 8 individuals were officially caught in the Indonesian seas. However, the details of their habitat have not

been described. The aim of this study is to clear their habitat and distribution.

METHODS

Two remotely operated vehicles (ROVs) (Kowa; VEGA300) were used for the surveys. The first ROV was replaced with the second one in 2007. Our ROVs are able to dive up to 300 m depth. These had two vertical, two horizontal and two right-left propellers and these were controlled from a boat on surface through a 400 m long tether cable. The underwater operations were visualized on a screen with information including directions of ROVs, depth, water temperature, date and time. All data were directly recorded on video-tapes on the boat. Since the 2007 survey, water temperature and depth were independently recorded with more accuracy by external measuring memories (Alec Electronics, later named JFE Alec; MDS-MKV/T, MDS-MKV/D). In addition two laser beam irradiators, that provide line lasers of 20-cm distance, were attached to the second ROV and have been used to register the size of encountered animals. The surveys were conducted off northern parts of Sulawesi Island and along the Biak Island located in northwestern New Guinea Island, Indonesia. The areas and time periods of each survey are shown in Table 1 and Fig. 1.

RESULTS

The field survey of the Indonesian coelacanth, Latimeria menadoensis, was conducted 14 times from 2005 to 2015 with 1173 underwater operations around Sulawesi and Biak islands in Indonesia (Fig. 1, Table 1). No coelacanth was encountered during the first surveys in 2005 although 452 underwater operations were conducted around Manado-tua Island and other close Indonesian islands from the 17th to 30th of April (Table 1, R1). The second field survey was held along the northern coast of Sulawesi from the 6th to 19th May in 2006, but once again despite of 107 underwater operations no coelacanth was recorded.

The first individual of Indonesian coelacanth (Table 2, ID 1) was recorded during the third field survey, in the morning of the 30th May 2006 (Table 1, R3). The coelacanth was observed in a cave of 165 m depth (Table 2, Boul-1) for 10 minutes (Table 2, E1). The adult individual was in stationary position in a cave (Buol-1) slowly moving its pectoral, pelvic, second dorsal and anal fins. The day after, another individual (ID 2) and an unidentified one (UN1) were found in the same cave (Buol-1). The unidentified one immediately hid deeper in the cave, where the ROV could not access, and only its lateral side of the body was recorded. The individual (ID 2) was observed for one minute before it hid inside the cave (Table 2: E2). When a new individual could be identified by the white

spots pattern on its body, an ID number was allocated. The individual which could not be distinguished was allocated an UN (unknown) + number.

In the end of the morning of the same day, a third individual (Table 2: E3, ID 3) was found in another cave (Buol-2), larger and located 20 m deeper (Buol-2) than the first cave (Buol-1). After 29 minutes of recording, this individual hid inside the cave. Although the water temperature increased of more than five degrees, from 14.8 to 20.4 ºC, it stayed at the same place, not moving away. Despite it disappeared from the camera frame we kept setting the ROV there in front of the cave until the next morning, however it did not appear again.

On the 4th June in 2006, three individuals, including ID 3 which was observed on the 31th May at Buol-2 and two new individuals, were observed in a vertical crack which was several hundred meters far from Buol-2 (Buol-3) for 2 hours and 14 minutes (Fig. 2C; Table 2: E4, ID 3, 4, 5). During this long period of time all individuals stayed near the rock wall always keeping their head against the water flow. The water

temperature changed up and down by five degrees during the observation due to water flow reversals. When the water flow was reversed some times, all individuals always changed their directions against the water flow.

During the field survey carried on December of the same year, in the same Buol area (Table 1: R4) the sixth individual (Table 2: E5, ID 6) was recognized at 14:34 on the 12th December alongside a large rock. This individual stayed in the beginning of the observation and then moved ahead to deeper slowly. Eventually it reached 12 m deeper during one hour and 42 minutes.

On the 14th December, the individual ID 3 previously recorded at Buol-2 in a cave and at Buol-3 in a crack, was encountered again (Table 2: Buol-5) near a vertical wall at 145 m depth (Table 2: E6). After a while, this individual (ID 3) moved upward about 18 m and stayed behind a large rock (Fig. 3). The observation time was the longest of all surveys with four hours and 16 minutes of recording. The temperature changed up and down several times (Fig. 4). Its difference

between the maximum and the minimum was 6.7 °C, which was the largest range of all surveys.

On the next day, on the 15th December, the ID 3 was observed again under an overhang (Table 2: E7, Buol-6) near other sites where it was previously found at Buol-2 , Buol-3 and Buol-5. Its posture was upside down with its ventral side facing the ceiling of the overhang. Such posture is also known in coelacanths from the Comorian Archipelago (FRICKE et al., 1987). ID 3 was found totally four times in this area in two surveys during the seven months and seems to live in this area.

During the 2007 survey (Table 1: R5) a coelacanth (Table 2: ID 7) was recorded off Manado by 195 m depth, with a temperature of 12.4 °C, the coolest one registered during all surveys.

The field survey in 2009 was carried out off Manado and around Talise and Bangka islands off the northern coast of Sulawesi Island from the 12th September to the 9th October (Table 1: R8). On the 24th September 2009, six individuals (Table 2: E11, ID 9, 10, 11, 12, UN 2, 3) were found under an

overhang off Talise Island between 144 and 150 m deep. There was almost no water current in this area. All individuals were close to each other under the same overhang and most of them stayed still with their head facing downward. Four days later, ID 9 was observed again, alone on a steep slope at 172.9 m depth (Table 2: E12), which was located within several dozen meters far from Talise-2 and 30 m deeper.

Four new coelacanths were recorded from the 29th September to the 6th October 2009 in Talise and Banggka Islands and off Manado (Table 2: E13 and 14, ID 13, 14, 15). ID 15 is the only juvenile coelacanth filmed so far. It was in a small crack between 164.6 and 170.9 m deep in a temperature between 14.5 and 15 °C.

The shallowest record of a coelacanth was at 115.6 m depth off the southern part of Manado (Table 2: E16, Manado-3) on the 9th October in 2009. ID 16 (Table 2) was found at the end of the morning, alone alongside a large rock on the edge of a shelf. The water temperature was about 20.0 ºC.

From the 6th to the 16th November in 2010, the field survey was conducted around Biak Island (Table 1: R9). Adult coelacanths were found at two different sites (Table 2: E17–19). Two individuals (ID 17 and 18) were found under an overhang between 212.5 and 218.9 m deep (Table 2: Biak-1) on the 11th November. Three other ones (ID 19, 20, 21) were registered under an overhang between 193.2 and 195.9 m deep (Table 2: Biak-2) on the 13th November. Two days later, Biak-2 was observed again and only ID 20 stayed there.

Two new coelacanths were recorded off Manado in December 2010 (Table 1: R10; Table 2: E 20 and 21; ID 22 and 23).

On the 4th May 2012, an individual (Table 2: E22, ID 24) was found on a steep slope at 169.4 m depth off Manado (Table 2: Manado-6). It swam up and down 5.4 m between 168.1 and 173.5 m deep for 17 minutes. Thirteen months later (the 9th June 2003), this individual (ID 24) was observed again near a cliff at 152.3 m depth (Table 2: E24, Manado-8), where located on the same slope of Manado-6.

Another new coelacanth (Table 2: E23, ID 25) was recorded alone on a Manado site (Manado-7).

Field surveys were conducted from the 19th to 30th May 2015 off Lolak Island and from the 30th October to the 16th November 2015 off Lolak and Bitung Islands (Table 1; R13 and R14). Three new individuals (Table 2: ID 26–28) were recognized at different places off Lolak Island (Lolak 1-4). All individuals were alone. ID 26 was observed in May and once

again in October 2015 on a steep slope (Table 2: E25, E27). The coelacanth ID 28 was found on the 1st November along one of the large rocks scattered on a gentle slope in a bay of the Lolak Island by 125 m depth.

The encountered Indonesian coelacanths during these surveys were in caves, alongside large rocks, under overhangs or on steep slopes (Table 2; Fig. 2).

Most individuals were found under overhangs or alongside large rocks. Some individuals were encountered not to hide in any shade and stayed just on rocky slopes. In overhangs and cracks, all individuals had their ventral side alongside and close to a rock wall, but without touching it.

Estimated total lengths are shown in Table 3. Most individuals are more than 1 m in total length. ID 15 is a juvenile individual of 31.5 cm and ID 23 is 90 cm in total length. No individual beyond 140 cm has been recorded in Indonesia so far.

DISCUSSION

During the field surveys by underwater ROV recording for Indonesian coelacanths, Latimeria menadoensis, from 2005 to 2015, 30 different individuals were identified and three others were unidentified. Among those, six individuals were observed more than twice: ID 3 at Buol-2, 3, 5, 6 (Table 2: E3, E4, E6 and E7); ID 8 at Talise-1 (E9 and E10), ID 9 at Talise 2 and 3 (E11 and E12), ID 20 at Biak-2 (E18 and E19), ID 24 at Manado-6 and 8 (E22 and E24), ID 26 at Lolak-1 and 3 (E25 and E27). ID 8 was observed on the 14th September 2009 and was encountered again at the same place (Talise-1) the day after. ID 24 was observed on the 4th May 2012 and then 13 months later, on the 9th June 2013 at the same locations of Manado area within several kilometers.

Al l individuals observed more than twice were encountered in the same area, at locations close from each other within hundreds of meters. However, in the Comorian Archipelago, same individuals of L. chalumnae were observed, within a two week period, in several caves distributed in a 8 kilometers wide area (FRICKE et al., 1991).

When individuals of L. menadoensis were encountered, some of them stayed stationary at same place but some individuals such as ID 6 on the 12th December 2006 (E5) and ID 24 on the 4th May 2012 (E22) swam away. It seemed to avoid and escape from the brightness of the light or the sounds

of the propellers of the ROV.Temperature and depth data for each Indonesian

coelacanth record are shown in Fig. 3. Though an average of water temperature on the ocean surface was about 30 °C, the temperature often dropped into about 11 °C at 300 m water depth. The temperatures changed widely even though the changes of depth were small at several places. All Indonesian coelacanths, L. menadoensis, were encountered during day time at a depth between 115.6 and 218.9 m and a temperature between 12.4 and 21.5 ºC. Indonesian coelacanths have exceptionally been recorded in an environment with a temperature exceeding 20 ºC, regardless of the depth. The African coelacanth, L. chalumnae, was observed by submarine vehicles in steep slopes between 150 m and 253 m (FRICKE et al., 1991). Divers recorded the presence of this species at 104 m depth at Sodwana Bay, South Africa (VENTER et al., 2000). Latimeria chalumnae seem to prefer a temperature range of 15–20 °C, where they tend to choose their daytime habitats (FRICKE et al., 1991). The suitable temperature for L. menadoensis seems to be almost the same as L. chalumnae.

Water temperature can increase or drop for more than 5 °C in a very short period of time (about 30 min according to our recordings). The coelacanths however often stayed there without moving away or significantly modifying their behavior. These water currents of different temperatures are caused by the movement of the marine thermocline. During our surveys we observed in some area that as the thermocline

level changed, the water flow was reversed. Coelacanths always kept their head against the water flow, like other observed actinopterygian fishes did. It means that as the direction of the water flow changed, the swimming direction of each coelacanth individual changed accordingly.

On the 24th September 2009 (Table 2: E11) six individuals were recognized under an overhang (Talise-2). The all had about the same size. There was almost no water flow, and each individual was head downward. They were close to each other but no direct contacts or interactions between coelacanths were observed. No evidence of a social behavior has been recognized in the Comorian populations of coelacanths (FRICKE et al., 1991).

According to the observations of the African coelacanths, the species L. chalumnae is nocturnal and hides in cave in daytime but comes out at night, supposedly in search of food (FRICKE et al., 1991). The living geomorphologic environment of L. menadoensis is similar to that of L. chalumnae, with rocky steep slopes with undermarine caves or overhangs. However, Latimeria menadoensis was also observed alongside large rocks or along vertical wall in daytime. Some of these individuals were encountered above 200 m deep, in an environment that remains sl ightly bright . Latimeria menadoensis seems to be less sensitive to the daylight than L. chalumnae.

These first observations are potentially crucial since the habitats of the Indonesian coelacanths could be possibly not as

restricted as those of the African coelacanths, and so a broader distribution of the Indonesian coelacanth could be expected. Occurrences of coelacanths along the eastern coast of the African continent were recorded in South Africa (SMITH, 1939), Comoros (SMITH, 1953), Madagascar (HEEMSTRA et al., 1996), Kenya (DE VOS and OYUGI, 2002), Tanzania (BENNO et al., 2006), and Mozambique (BRUTON et al., 1992). In Indonesia, coelacanths have only been found off the northern coast of the Sulawesi Island and the southern coast of the Biak Island in northern New Guinea.

Field surveys of the Indonesian coelacanths should be continued in all area of Indonesia and expanded to all Southeast Asia in order to increase our knowledge of the distribution of this rare species and to define, in a close future, the most relevant conservation policies of the different known populations.

ACKNOWLEDGMENTS

We are grateful to Dr. Teruya UYENO for his valuable advice, guidance and encouragement throughout the present study. Many thanks go to colleagues of the Faculty of Fisheries and Marine Sciences of the Sam Ratulangi University at Manado, Indonesia, for their generous supports to conduct expeditions. Also we express our thanks to Dr. SUHARSONO, Dr. Zainal ARIFIN, Dr. DIRHAMSYAH and colleagues of the Indonesian Institute of Science for their advice and administrative work to obtain our research permit. Other thanks go to Mr. OPO, Mr. REFRY and colleagues at Murex Dive Resort, and DAUD, AL,

and other staffs for our work on the research boat. Dr. BATUNA and the Murex staffs supported our daily life during the expeditions. We would like to express our sincere thanks to colleagues at Aquamarine Fukushima, Japan, for their understanding and supports. We would like to thank Dr. Gaël CLÉMENT of Muséum national d’Histoire naturelle and Dr. Camila CUPELLO of Universidade do Estado do Rio de Janeiro as a referee for their critical reading of the manuscript and their comments.

Unfortunately one of our authors, Dr. Djoko Hadi KUNARSO has passed away during the process of this contribution. We pray for the response of his soul.

This study was supported in part by funding from Interdisciplinary Collaborative Research program of Atmosphere and Ocean Research Institute, the University of Tokyo.

REFERENCES

BENNO, B., VERHEIJI, E., STAPLEY, J., RUMISHA, C., NGATUNGA, B., ABDALLAH, A. and KALOMBO, H. 2006. Coelacanth (Latimeria chalumnae SMITH, 1939) discoveries and conservation in Tanzania. South African Journal of Science, 102: 486–490.

BRUTON, M. N., CABRAL, A. J. P. and FRICKE, H. 1992. First capture of a coelacanth, Latimeria chalumnae (Pisces, Latimeriidae), off Mozambique. South African Journal of Science, 88: 225–227.

CUPELLO, C., BRITO, P. M., HERBIN, M., MEUNIER, F. J., JA N V I E R, P . , DU T E L, H. and CL É M E N T, G. 2015.

IWATA et al.