Psammoecus hiranoi Yoshida and Hirowatari 2013

© 2018 Academia Sinica, Taiwan Open Access Seaward Migration Routes of Indigenous Eels, Anguilla japonica, A. marmorata, and A. bicolor pacifica, via Satellite Tags Shih-Chong Chen1, Ching-Rong Chang1, and Yu-San Han1,* 1Institute of Fisheries Science, College of Life Science, National Taiwan University, Taiwan. E-mail: f05b4501[email protected] (Chen); [email protected] (Chang) (Received 16 November 2017; Accepted 11 April 2018; Published 24 May 2018; Communicated by Hin-Kiu Mok) Citation: Chen SC, Chang CR, Han YS. 2018. Seaward migration routes of Indigenous Eels, Anguilla japonica, A. marmorata, and A. bicolor pacifica, via satellite tags. Zool Stud 57:21. doi:10.6620/ZS.2018.57-21. Shih-Chong Chen, Ching-Rong Chang, and Yu-San Han (2018) The spawning migration of the Japanese eel, Anguilla japonica, along the western Mariana Ridge is considered one of the longest animal migrations. To establish an appropriate conservation policy for this species, a complete life cycle study is required, but the migratory routes of anguillid eels remain poorly documented. Anguilla japonica, A. marmorata, and A. bicolor pacifica are three common eel species in Taiwan. The present study used pop-up archival transmitting tags to track their movements during the oceanic phase of their migration. Twenty silver eels (farmed and wild) were tagged and released near the Gueishan and Liuchiu Islands. All three species in the Gueishan Island group were entrained in the Kuroshio Current, and pop-up tag data showed that, after 23 days, the farthest distance travelled was approximately 1,500 km, to the southern waters of Yakushima, Japan. In comparison, eels released near Liuchiu Island moved in the opposite direction, which suggests that they were influenced by the recirculation in the southwestern waters of Taiwan. The migration speed of eels was estimated to be between 12.9 km/d and 65.2 km/d. Several eels exhibited diel vertical migration behavior, moving into deeper water during the daytime and ascending into shallower waters at night. The farmed eels exhibited similar patterns of diel vertical migration to wild eels. The results of this study suggest that eels from Taiwan travel through the Kuroshio Current during their early migration, irrespective of species and source. Of importance, farmed eels appeared to behave similar to wild ones following artificial release. This information provides important rationale for releasing cultivated eels for resource enrichment programs. Key words: Eel, Spawning migration, Pop-up archival transmitting tag, Diel vertical migration, Kuroshio Current. *Correspondence: Shih-Chong Chen and Ching-Rong Chang contributed equally to this work. E-mail: [email protected] BACKGROUND There are 19 species and subspecies of eel (family Anguillidae, genus Anguilla) worldwide. Most species are distributed in the Indo-Pacific Ocean area. Four species of Anguilla eel have been found in Taiwan: A. japonica (Japanese eel), A. marmorata (Giant mottled eel), A. bicolor pacifica (Shortfin eel), and A. luzonensis (Luzon mottled eel) (Tzeng 1982 1983; Tzeng and Tabeta 1983; Teng et al. 2009; Watanabe et al. 2009). A. japonica and A. marmorata are the most common eel species in Taiwan (Han 2010). The anguillid eel is a catadromous species. A. japonica is a temperate eel species that was listed in the IUCN Red List of Threatened Species as an endangered species in 2014. It is characterized by long-distance oceanic migration from continental freshwater habitats to its spawning areas along the western Mariana Ridge, near 14-16°N, 141-142°E (Tsukamoto 2006 2009), which is over 2,000 km away from the Asian continent. A. marmorata and A. Zoological Studies 57: 21 (2018) doi:10.6620/ZS.2018.57-21 1 © 2018 Academia Sinica, Taiwan bicolor pacifica are tropical eel species, and their spawning areas are assumed to be in the North Equatorial Current (NEC) region, near the Mariana Ridge. Leptocephali of A. marmorata have been collected throughout the NEC region to the west of 142°E, but not to the east of it. The latitudinal range (12-18°N) of the leptocephali is wider than the determined spawning area of A. japonica (Miller et al. 2002). A. bicolor pacifica is distributed throughout the western Pacific Ocean region. The spawning site of A. bicolor pacifica has not been determined; however, one study suggested that the spawning area is located in the NEC region. The leptocephali of A. bicolor pacifica have been caught in the waters of western New Guinea and to the west of the Mariana Ridge (Aoyama 2009). Thus, establishing the seaward migration of silver eels would provide information on the final stages of their life cycle. Anguillid eels transform from yellow to silver once they have sufficiently matured for their seawater migration; however, the mechanism triggering this stage remains unclear. When the eels initiate the long oceanic migration, various morphological and physiological changes occur. Various adaptations include the silvering of the skin (Pankhurst and Lythgoe 1982; Han et al. 2003), the enlargement of eyes and pectoral fins (Pankhurst 1982; Han et al. 2003), the degeneration of the alimentary tract (Pankhurst and Sorensen 1984; Han et al. 2003), a shift in retinal sensitivity (Andjus et al. 1998; Zhang et al. 2000), changes to musculature (Egginton 1986) and fat content (Larsson et al. 1990), modifications to the swim bladder (Kleckner 1980; Yamada et al. 2001), and an increase in the chloride cells of the gills (Fontaine et al. 1995; Sasai et al. 1998). These changes appear to improve the swimming ability of eels and reduce the likelihood of predation in the ocean, which, in turn, provide them with a greater chance of reaching the spawning areas. Previous studies have shown that restocked eels might be an effective way to recover natural eel resources in Europe. In 2006, tagged restocked and wild eels were documented migrating toward the outlets of the Baltic Sea, which included both restocked and wild eels (Sjöberg et al. 2008). One study estimated that 26.7% of all coastal eels at the outlet of the Baltic Sea came from restocked sources (Limburg et al. 2003), while another study estimated 21.3% (Clevestam and Wickström 2008). In Taiwan, the Fisheries Research Institute has been releasing cultured Japanese eels since 1976 (Liao et al. 1994). The eels are released into rivers or estuaries after being injected with microchips. However, the efficiency of releasing cultured eels has been questioned because the eels are moved from aquafarms, where they have been maintained in an artificial environment for years. It is also questionable whether there are significant differences in the migratory behaviors of wild and cultured eels, including their choices of seaward migration routes. Several tracking studies have used acoustic tags to examine the migratory behavior of silver eels in the ocean (Tesch 1978 1989; Aoyama et al. 1999, 2002; McCleave and Arnold 1999; Chow et al. 2015). However, the tracking duration was too short to cover the entire migratory period and the ocean depth studied was limited. The development of pop-up satellite archival transmitters has allowed data to be collected for longer periods. Studies of released tagged European eels in the Mediterranean and Baltic Seas have been undertaken for years. In general, released European eels tend to travel to subtropical areas, and then to the Sargasso Sea (Wysujack et al. 2015; Amilhat et al. 2016; Righton et al. 2016). American eels were successfully tracked from the continental shelf and shown to cross the Gulf Stream and swim toward their spawning area in the Sargasso Sea (Béguer-Pon et al. 2015). One tagged A. marmorata eel traveled 843 km toward the South Equatorial Current from Vanuatu (Schabetsberger et al. 2013). Only a few tracking studies of Japanese eels have been carried out, and these studies have been designed differently in terms of eel-release positions, methods of tag attachment, and eel life stage (Tsukamoto 2009; Manabe et al. 2011; Chow et al. 2015). Tagged Japanese eels have been recorded swimming in the Kuroshio Current (Manabe et al. 2011). All anguillid eels exhibit diel vertical migration (DVM) behavior during their oceanic migration, with migrating eels preferring shallower water (100-300 m) at night and deeper water (500700 m) during the daytime. Eels released in Taiwan have not been previously evaluated. Thus, this study aimed to: (1) determine the seaward migration route of A. japonica, A. marmorata, and A. bicolor pacifica from Taiwan and (2) compare the seaward migration behaviors of wild and farmed eels. Our results provide information on the similarities and differences between farmed and wild eels during migration and suggest the value of release programs using farmed eels. page 2 of 13Zoological Studies 57: 21 (2018) © 2018 Academia Sinica, Taiwan MATERIALS AND METHODS Eels Eels were obtained from the fish market in Ilan and Pingtung Counties, Taiwan. Four farmed eels (two A. marmorata and two A. japonica) and six wild eels (one A. marmorata and five A. bicolor pacifica) were selected for tagging and release in 2015, while 10 farmed A. japonica were selected for tagging and release in 2016. Eels were selected based on morphological characters that distinguished them as having completed their yellow life stage: black dorsal color, bigger eyes, larger pectoral fins, and more melanin (Han et al. 2003). The eels were maintained at an ambient temperature (approximately 25°C) in a tank, and the salinity of the water was stepwise increased to 30‰ by adding salt over 10 days. The water was replaced with seawater after the tank reached seawater conditions. The temperature was then controlled using an electric system at a constant 23°C. To assess the condition and stage of the eels, measurements of total length, body weight, eye diameter, and pectoral fin length were made. The ocular index and pectoral fin index was also calculated (Pankhurst 1982; Han et al. 2003) (Table 1). Release sites In total, 20 eels were released between 2015 and 2016 at 10 eels per year. In 2015, tagged eels were released in three batches from three different locations around Gueishan Island, Ilan County. One eel was released on August 11, 2015 at 24°51'8.00"N, 121°53'40.00"E (R1, Fig. 1), four on September 18, 2015 at 24°51'50.90"N, 121°59'59.11"E (R2, Fig. 1), and five on November 3, 2015 at 24°51'38.24"N, 121°59'35.10"E (R3, Fig. 1). In 2016, 10 eels were separated into two groups and released around Gueishan Island, Ilan County and Liuchiu Island, Pingtung County. Four eels were released on October 11, 2016 at 24°52'3.00"N, 121°58'7.00"E (R4, Fig. 1) and on November 1, 2016 at 22°23'39.92"N, 120°24'12.77"E (R5, Fig. 1). Release locations ranged from 300 to 400 m in depth. Types of tags Two different types of pop-up archival transmitting tags were attached to the eels: the MiniPAT in 2015 and the mark report PAT (mrPAT) in 2016 (purchased from Wildlife Computers Inc.). MiniPAT: The MiniPAT tag is 124 mm in length, has a maximal diameter of 38 mm, and weighs 60 g in air. Its range of depth is 0 to 1700 m and range of temperature sensor is -40 to 60°C. The MiniPAT is equipped with depth, temperature, and light sensors. A mortality detection mechanism is set so that the depth variance is recorded within 3 m for 10 days. Once the mortality event is determined, the release pin burns through, which separates the tag from its tether. The tag then floats to the surface and begins transmitting data to Argos satellites. Mark report PAT: The mrPAT is 121 mm in length, has a maximal diameter of 23 mm, and weighs 26 g in air. It is equipped with only temperature and light sensors, and has no mortality detection mechanism. Sensor data are collected during deployment and archived in its internal memory. Then, on a pre-set interval, the tag is released from its host animal and a summary of the archived data is uploaded to Argos satellites. Both pop-up tags can obtain light-dependent geolocation data via the Argos satellite system. The tags were programed to pop-up from 1 to 4 months at each release site (Table 1). The sampling intervals were programmed with 75, 150, 300, and 450 s according to different scheduled days (31, 61, 91, 121). Tag attachment To attach the pop-up archival transmitting tags to the eels, nylon wire (105 lb.), PVC pads, polyethylene foam pads, and oval sleeves were used to construct an intrusive strap in front of the dorsal fin by surgery. Eels were anesthetized with 0.2‰ eugenol before surgery. A needle (3.5 in) holding one end of the monofilament nylon wire was inserted from the lateral musculature through the body dorsally. A loop was made for the insertion of the tag. Then, one strap was inserted from the dorsal musculature beside the previous wound for 1.5 cm on the same side. Another strap was constructed the same way, and the pads were used to prevent the wound from worsening at the insertion points (Fig. 2). Povidone-iodine solution was applied to prevent infection at the wound sites. All surgery was scheduled 3 days before release to ensure the eels were in good enough condition for release after tag straps were attached. page 3 of 13Zoological Studies 57: 21 (2018) © 2018 Academia Sinica, Taiwan Data treatment The pop-up locations were recorded to map out the approximate seaward migration route of anguillid eels from Taiwan. To increase the accuracy of the migration route, the swimming depths and temperature data received from the Argos system were also documented. The first pop-up location transmitted from the MiniPAT was related to the location where the tag was released. Data from the MiniPATs were subject to three procedures in order to reconstruct the migration Fig. 1. Release position of tagged eels. R1 (24°51'8.00"N, 121°53'40.00"E), R2 (24°51'50.90"N, 121°59'59.11"E), and R3 (24°51'38.24"N, 121°59'35.10"E) are the positions at which the tagged eels were released in 2015 and R4 (24°52'3.00"N, 121°58'7.00"E) and R5 (22°23'39.92"N, 120°24'12.77"E) are the 2016 release positions. The scale bar can be used in the largest map only. (The original figure was obtained from Google Earth) Fig. 2. Schematic diagram of tag attachment with (a) cross-section and (b) lateral views. (a) (b) page 4 of 13Zoological Studies 57: 21 (2018) © 2018 Academia Sinica, Taiwan routes: (1) confirm the longitudinal position, (2) compare the position with the seabed depth, and (3) compare the temperatures of (1) and (2) at different depths. The depths the eels migrate at (> 200 m) were too high for the internal daylight geolocation system of the tags to provide meaningful data for light-based estimate locations. The punctual DVM behavior, whereby eels descended through the water column at dawn and ascended at sunset, was used to determine sunrise and sunset times, with this type of migration utilized to study the horizontal movement as well. A previous study showed that migrating Japanese eels rapidly dive to deeper water 50 min before the sun rises (Chow et al. 2015). The data from the sunrisesunsetmap website (http://sunrisesunsetmap.com/) was used to cross-reference the real time data obtained from the released eels. The migration route analysis also used environmental oceanic data obtained from the Mercator Ocean website (http://www.mercatorocean.fr/) Daily Global Physical Bulletin at 1/12°. The maximum depths of migrating eels were restricted by water depth; therefore, the seafloor depth was used to eliminate any locations that were shallower than the recorded maximum depths. The temperature data were used to eliminate locations that did not match the records of migrating eels, and were obtained from the Mercator Ocean website (http://www.mercatorocean.fr/). The oceanic temperatures at 100 and 300 m were applied to check the possible locations of migrating eels. The warm Kuroshio Current flows from the low latitudinal areas of Taiwan to Japan. The annual average temperature of the Kuroshio Current is 24-26°C, 18-24°C in winter and 22-30°C in summer. Cooler water was measured around the continental shelf, at approximately 15-18°C. These differences in temperature were used to speculate as to whether the migrating eels left the warm Kuroshio Current. Time-related depth data were not collected from the mrPATs because they did not have the premature pop-up detection mechanism. If the pop-up function of the tag occurred earlier due to shedding, the recorded location at the set time would be further away from the real location of the tagged eel. The released tag would float and keep recording the temperature in the Kuroshio Current. Data on the speed of the current were obtained from the Mercator Ocean website (http:// www.mercator-ocean.fr/). These data were used to estimate the approximate location where the tag was shed. However, the migration routes from the release site to the shedding location could only be determined by process of elimination based on temperature records. RESULTS Migration duration and distance In total, 20 eels were tagged and released in two locations (Table 1). A total of 14 tag signals were successfully transmitted (10 MiniPATs and 4 mrPATs). The 10 reporting MiniPATs transmitted on average 54% (range: 30-71%) of their archival data, with recording days from 8 to 132 days and the real active duration from 2 to 32 days. However, 3 of the 10 MiniPAT tag data were defective, due to a high overall loss rate (above 65%), and were excluded from the analysis (Table 2). The mrPATs popped-up at the scheduled times due to no mortality being detected. The duration of the real active time ranged from 10 to 21 days. Seven pieces of reporting data in 2015 were used to reconstruct possible migration routes, with distances from 81 to 1,500 km (Fig. 3a). In the 2016 reporting data, because the tags lacked several environmental factor detectors, we could only calculate the migration distance by assuming that all of the eels swam along the Kuroshio axis. #164778 and #164780 tags produced raw distances of 1,769 and 1,949 km from the released site, respectively (R4). After elimination and reconstruction, the distances of possible routes were 409 and 667 km, respectively. Tags #164782 and #164785 showed distances of 197 and 466 km from release at site R5, respectively (Table 2). Migratory speed and pop-up sites The pop-up locations showed that the eels released in 2015 migrated north-eastward in the Kuroshio Current (Fig. 3a). Five tags popped up in the Okinawa Trough, which is located on the edge of the continental shelf. Two released eels migrated further. Tag #150539 popped up in the southern waters of Yakushima, Japan and tag #150540, which had the longest migratory distance, was geolocated to the eastern waters of Miyazaki. The eels’ migration speeds were reported to be 12.9 to 65.2 km/d (minimum and maximum, respectively) (Table 2). For the mrPATs released in 2016, tags page 5 of 13Zoological Studies 57: 21 (2018) © 2018 Academia Sinica, Taiwan #164778 and #164780 (Fig. 3b) showed similar routes as those in 2015, with mean migratory speeds of 37.2 and 44.5 km/d, respectively (Table 2). Tags #164782 and #164785 popped up in the eastern waters of Lan Island and the more westward waters near Guangdong Province, China, respectively (Fig. 3b). These two tags might have been influenced by the recirculation of the southwestern waters of Taiwan, which made it difficult to estimate the migrating speed of the tagged eels. Fig. 3. Pop-up locations and the estimated migratory routes of the released tags in (a) 2015 and (b) 2016. All estimated routes were cross-checked with the timing of dawn and sunset and the timing of decent and ascent. Maximum seafloor depths were used on the route corrections as well. Because of the recirculation located at the southwestern water of Taiwan, the eels or shedding tags probably recirculated in this region for days. Two individuals left the recirculation: #164782 entrained in Kuroshio and #164785 pop-upped at more western waters. Table 1. Profile of all eels included in the study No. Species TL (mm) BW (g) OI FI Release site Scheduled Days Type of ending Type 150536 AJ 725 657 3.9 6.1 R3, IL 61 Premature Farmed 150537 ABP 779 1133 7.3 5.0 R2, IL 61 Premature Wild 150538 ABP 771 1096 3.7 5.2 R3, IL 61 Premature Wild 150539 AM 998 4000 7.9 5.3 R2, IL 61 Premature Farmed 150540 ABP 714 861 7.9 5.7 R2, IL 31 Premature Wild 150541 ABP 725 1061 5.3 4.8 R3, IL 121 Premature Wild 150542 AM 1000 5000 - - R1, IL 31 Premature Farmed 150543 ABP 765 960 6.2 4.2 R2, IL 91 Premature Wild 150544 AM 1162 4000 9.3 4.9 R3, IL 91 Premature Wild 150545 AJ 755 717 3.7 5.3 R3, IL 31 Premature Farmed 164777 AJ 683 532 6.0 5.8 R4, IL 31 Lost Farmed 164778 AJ 667 583 6.7 6.7 R4, IL 61 Premature Farmed 164779 AJ 725 777 4.6 5.0 R5, PT 91 Lost Farmed 164780 AJ 672 502 6.3 5.1 R4, IL 91 Premature Farmed 164781 AJ 646 455 4.7 5.0 R5, PT 31 Lost Farmed 164782 AJ 682 818 5.2 5.1 R5, PT 61 Premature Farmed 164783 AJ 801 889 4.5 4.7 R5, PT 121 Lost Farmed 164784 AJ 710 598 7.9 5.8 R4, IL 121 Lost Farmed 164785 AJ 672 518 6.3 5.7 R5, PT 91 Premature Farmed 164786 AJ 761 1120 5.1 5.4 R5, PT 121 Lost Farmed AJ: A. japonica; ABP: A. bicolor pacifica; AM: A. marmorata; TL: total length; BW: body weight; OI: ocular index; FI: pectoral fin index. (a) (b) page 6 of 13Zoological Studies 57: 21 (2018) © 2018 Academia Sinica, Taiwan Fig. 4. Typical DVM behavior. Eels ascended to shallow water at night and descended to deeper water in daytime. Blue line is depth; orange line is temperature. Migration behaviors Twelve A. japonica eels were released in the present study (two in 2015 with MiniPAT tags and 10 in 2016 with mrPAT tags). Six mrPAT tags in 2016 did not transmit, and only tag (#150545) exhibited DVM behavior. Data recorded from 11:00 to 13:00 (daytime) and 23:00 to 01:00 (nighttime) were selected to calculate the mean depth and temperature during day and night. Five tags exhibited typical DVM behavior (Fig. 4) and their average temperatures and depths are recorded in table 3. Three A. marmorata were released during 2015 in the present study with MiniPAT tags (Table 4). Five A. bicolor pacifica were equipped with the MiniPAT tags in 2015 (Table 5). Their average temperatures are recorded in table 5. Behavior during migration of released eels The released eels tended to swim in water that was greater than 200 m deep if there were no limitations to seafloor topography. They exhibited more DVM behavior in deeper water (> 200 m) than in shallower water. #150537, #150539, #150543, #150544 and #150545, according to their occupied depth and temperature between day and night, are thought to have exhibited DVM Table 2. Migratory data on all released eels PTT No. Recording days Duration of activity Distance (km) Migration speed (km*day-1) 150536-f 95 uncertain - - 150537 62 27 666 24.7 150538 63 uncertain - - 150539-f 43 32 1089 34.0 150540 36 23 1500 65.2 150541 132 uncertain - - 150542-f 8 2 81.2 40.6 150543 94 25 461 18.4 150544 25 17 816 48.0 150545-f 24 15 194 12.9 164778-f 61 11 409 37.2 164780-f 91 15 667 44.5 164782-f 61 21 197 164785-f 91 10 466 “-f”: farmed eels. page 7 of 13Zoological Studies 57: 21 (2018) © 2018 Academia Sinica, Taiwan behavior. Predation Typical DVM behavior was interrupted for tags #150540 and #150545 by a sudden change in the records of both temperature and depth, which suggests that predation events had occurred. During these times, the two individuals spent a great amount of time at the surface (depth of 0 m), and the typical DVM curve in the depth data also disappeared. The temperature recorded by the tags rose from 10-24°C to greater than 25°C (Fig. 5). Light sensitive response Based on the Chinese lunar calendar, September 27, 2015 was a full moon day and October 13, 2015 was a new moon day. On these days, eels presented a light sensitive response, i.e., they occupied deeper water depths (approximately 200 m) when the moonlight was strong and the shallower ones (approximately 50-100 m) during the new moon (no moonlight) (Fig. 6). DISCUSSION The spawning site of the Japanese eel was previously identified as west of Mariana Ridge, over 2,000 km from Taiwan (141-142°E, 12-16°N) (Tsukamoto 1992). The spawning sites of the A. marmorata and A. bicolor pacific might overlap with those of the Japanese eel or be nearby (Miller et al. 2002; Aoyama 2009). Therefore, the three species of eel might have similar seawater migration routes from Taiwan. Although the numbers of tagged eels Table 3. Habitat occupation of A. japonica. (n = 6) (average ± standard deviation) Tag Temperature (°C) Depth (m) Day Night Day Night 150536 - - - 75 ± 12.7 150545 10.2 ± 0.8 20.8 ± 3.2 438 ± 36.1 75 ± 53.9 164778 9.8 ± 4.0 23.6 ± 2.7 - - 164780 8.2 ± 3.2 21.3 ± 2.7 - - 164782 6.8 ± 3.2 23.1 ± 3.2 - - 164785 8.3 ± 4.0 24.7 ± 3.0 - - “-”: data deficient. The mrPATs used in 2016 did not provide depth records during the oceanic migration. Table 4. Habitat occupation of A. marmorata. (n = 3) (average ± standard deviation) Tag Temperature (°C) Depth (m) Day Night Day Night 150539 8.1 ± 1.6 25.2 ± 2.7 573 ± 64.8 189 ± 80.6 150542 23.0 ± 2.3 25.3 ± 2.2 23 ± 15.7 23 ± 17.5 150544 7.1 ± 1.1 22.1 ± 2.4 597 ± 57.8 84 ± 72.5 Table 5. Habitat occupation of A. bicolor pacifica. (n = 5) (average ± standard deviation) Tag Temperature (°C) Depth (m) Day Night Day Night 150537 7.5 ± 1.1 20.0 ± 3.8 602 ± 77.3 165 ± 75.6 150538 12.8 ± 2.0 18.7 ± 4.1 - 101 ± 31.1 150540 14.3 ± 6.7 19.5 ± 4.1 295 ± 155.8 109 ± 84.5 150541 9.8 ± 3.9 18.4 ± 4.4 - 136 ± 84.8 150543 8.3 ± 1.1 23.5 ± 4.3 603 ± 88.0 130 ± 73.8 page 8 of 13Zoological Studies 57: 21 (2018) © 2018 Academia Sinica, Taiwan were low in the present study, the results clearly showed that all three species migrate in a similar direction along the Kuroshio Current. Several theories have been proposed regarding the migratory routes of silver Japanese eels. One study suggested an “anti-Kuroshio route,” whereby Japanese eels migrate upstream of the Kuroshio Current, and then eastward in the NEC (Yokose 2008). Another theory is an “along-Kuroshio route,” which states that the silver Japanese eels swim with the Kuroshio Current during their east-northward migration, and then swim southward in recirculation branches of the subtropical gyre along the Izu-Bonin Arc and down to the west of the Mariana Ridge where they spawn (Tsukamoto 2009; Chang et al. 2016). Another possibility is that silver Japanese eels migrate along the most direct and shortest route of the estuary to the spawning site. Results from the present study support the “along-Kuroshio route” theory for silver Japanese eels. However, these long migration distances require energyconsuming swimming behavior, and the silver eels have to store most of their energy for gonadal development. Therefore, this migration route appears to be too far for the silver eels. In 2015, a tracking study of American eels recorded one migration route to the spawning area in the Sargasso Sea (Béguer-Pon et al. 2015). In this study, the released eels migrated along the current during the early phase of migration, and then changed direction to reach the spawning area. Fig. 5. Predation event on (a) #150540 and (b) #150545. Two types of pattern were displayed in the transmitted data. (a) (b) page 9 of 13Zoological Studies 57: 21 (2018)