Skip to main content
NCBI home page
Zoological Studies logoLink to Zoological Studies
. 2020 Jun 9;59:e17. doi: 10.6620/ZS.2020.59-17

Estimations of Riverine Distribution, Abundance, and Biomass of Anguillid Eels in Japan and Taiwan Using Environmental DNA Analysis

Hikaru Itakura 1,2, Ryoshiro Wakiya 3, Masayuki K Sakata 4, Hsiang-Yi Hsu 5, Shih-Chong Chen 5, Chih-Chao Yang 5, Yi-Cheng Huang 5, Yu-San Han 5, Satoshi Yamamoto 6, Toshifumi Minamoto 4
PMCID: PMC7688404  PMID: 33262841

Abstract

Although populations of anguillid eels have declined remarkably in recent decades, monitoring data on the spatial and temporal variation in their dynamics are often limited, particularly for tropical eel species. As there are often sympatries of multiple eel species in tropical rivers, identifying eel species based solely on morphological characteristics is challenging. Basin-scale surveys were conducted in rivers of southern Japan and northern Taiwan to investigate (1) whether the spatial distribution, abundance, and biomass of the tropical eel species, the giant mottled eel (Anguilla marmorata), can be monitored in rivers by comparing the results obtained from environmental DNA (eDNA) analysis with data from electrofishing and (2) the riverine distribution of the sympatric A. marmorata and the temperate eel species, the Japanese eel (Anguilla japonica), in this region using eDNA analysis. Although we found an much lower abundance of A. marmorata in the study region, we identified the eDNA of the species from all of the study sites (21 sites) where it was collected by electrofishing, in addition to 22 further study sites where it was not collected directly. This indicates that eDNA analysis has a greater sensitivity for detecting A. marmorata, making it a powerful tool for monitoring the spatial distribution of the species in rivers. We found a significant positive relationship between eDNA concentration and both the abundance and biomass of A. marmorata, and eDNA concentration seemed to better reflect the abundance of the species than did biomass. eDNA of both A. japonica and A. marmorata was identified from almost all rivers, indicating the sympatry of these species in this region, although the degree of sympatry differed between rivers. Though the eDNA concentration of A. japonica decreased significantly with increasing distance from the river mouth, no significant relationship was found for A. marmorata. This study is the first to demonstrate the potential usefulness of eDNA analysis for estimating the spatial distribution, abundance, and biomass of tropical eels in rivers and to further apply this method to investigate sympatry among anguillid species. eDNA analysis can help in obtaining data on the population dynamics of tropical eels, providing invaluable information for managing these species.

Keywords: Anguilla japonica, Anguilla marmorata, Environmental DNA, Monitoring, Spatial distribution

BACKGROUND

The genus Anguilla contains 16 species and 3 subspecies, which spawn in the open ocean and grow in continental waters. Populations of anguillid eels are distributed across more than 150 countries, and their ecological, commercial, and cultural importance are common globally (Jacoby et al. 2015). However, in recent decades, there have been significant declines in the stocks of some anguillid eels, and 10 of the 16 species are now listed as “Threatened” or “Nearly Threatened” in the International Union for Conservation of Nature (IUCN) Red List of Threatened Species (IUCN 2019). This critical situation among anguillid eel populations highlights that conservation and monitoring these populations is an urgent global issue.

Ideally, to enable direct comparison of results obtained from different regions or studies, a consistent protocol should be used to monitor the dynamics of a target population quantitatively and continuously throughout its distribution range. However, a quanti-tative monitoring survey necessitates extensive fieldwork and considerable effort, in addition to different sampling protocols in different environments. Anguillid eels have broad geographic ranges because they migrate between saline and freshwater environments and inhabit diverse habitats, from saline bays to upland headwaters (Moriarty 2003). They also exhibit hiding behaviors in refuges (Aoyama et al. 2005). All these factors create challenges for accurate and continuous monitoring using standardized capture-based methods throughout their range. Thus, data on spatial and temporal variation in anguillid eel population dynamics are often either sparse, patchy, or imbalanced (Jacoby et al. 2015).

Still, data on population dynamics and ecological knowledge of various life stages have accumulated for eels distributed in temperate regions (i.e., temperate eels), including the American eel (Anguilla rostrata), the European eel (Anguilla anguilla), and the Japanese eel (Anguilla japonica), which have long been commercially important species. Conversely, though most anguillid eel species (10 species) are found in the tropical Indo-Pacific regions (i.e., tropical eels) (Ege 1939; Aoyama et al. 2018), there are limited monitoring data on the population dynamics of tropical eels, and ecological studies are, as yet, incomplete (Jacoby et al. 2015), perhaps due to these eels’ relatively low commercial importance. However, data collection for tropical eels is particularly important, given a recent increase in demand for tropical eels, such as the giant mottled eel (Anguilla marmorata) and Anguilla bicolor, as a replacement for the temperate eels, especially in East Asia (Gollock et al. 2018).

Continental waters in each region usually contain a single temperate eel species because, generally, the geographic distribution of temperate eels does not overlap (with the exception of Anguilla dieffenbachii and Anguilla australis; Glova et al. 1998). Conversely, there are sympatries of temperate and tropical eels in the marginal regions of their distribution range (Shiao et al. 2003; Leander et al. 2012; Hsu et al. 2019), and sympatries of multiple eel species often occur in tropical rivers (Arai and Abdul Kadir 2017; Hagihara et al. 2018). Accurate identification based solely on morphological characteristics is difficult for some eel species in the tropical regions (Watanabe et al. 2004), which often leads to eel species being misidentified. For instance, Anguilla bengalensis bengalensis has been misidentified as A. marmorata (Arai and Abdul Kadir 2017). Genetic analysis is preferred to identify eel species in tropical regions (Arai and Abdul Kadir 2017; Hagihara et al. 2018; Nguyen et al. 2018). Moreover, recent studies indicate that different eel species can have different habitat use and riverine distribution patterns (Shiao et al. 2003; Chino and Arai 2010; Arai and Abdul Kadir 2017; Hagihara et al. 2018; Nguyen et al. 2018; Hsu et al. 2019), which makes it difficult to monitor their population dynamics. Most of these studies have examined eel distribution in tropical regions using either qualitative or non-standardized sampling protocols among different environments and have used otolith microelement analysis to estimate eels’ habitat utilization. Although these studies have obviously provided important new information on eels, other quantitative methods are required to understand their distributions precisely.

Environmental DNA (eDNA) analysis is increasing rapidly in popularity as a cost-efficient and non-lethal monitoring tool for studying and managing organisms in aquatic ecosystems (Lodge et al. 2012; Rees et al. 2014). It has been used effectively in determining the presence of aquatic species inhabiting lakes, rivers, and marine habitats (Dougherty et al. 2016; Yamamoto et al. 2016; Yamanaka and Minamoto 2016). Furthermore, as this approach can be more sensitive than conventional capture-based sampling methods for detecting the presence or absence of fish (Jerde et al. 2013; Wilcox et al. 2016; Sakata et al. 2017; Itakura et al. 2019), this is a particularly useful tool for monitoring rare and endangered species (Fukumoto et al. 2015; Sakata et al. 2017). eDNA analysis can also be used in both freshwater and marine habitats to estimate abundance and biomass (Pilliod et al. 2013; Dougherty et al. 2016; Yamamoto et al. 2016; Doi et al. 2017; Minamoto et al. 2017; Itakura et al. 2019). For anguillid eels, Itakura et al. (2019) conducted basin-scale surveys of Japanese eels across 10 rivers in Japan and detected the eDNA of the species at 92% (56 of 61 sites) of study sites from which individuals were collected by electrofishing as well as at 35 additional sites where individuals were not directly collected. This indicates that eDNA analysis has greater sensitivity for detecting the presence of Japanese eels. The study also showed that eDNA analysis can be used to estimate the abundance and biomass of Japanese eels in rivers, requiring less time and effort than electrofishing. As some tropical eels cannot be identified without using genetic characteristics, eDNA analysis is considerably helpful for monitoring and allows investigators to undertake large-scale surveys throughout the eels’ distribution ranges using a consistent method. However, eDNA analysis has not yet been applied to detect tropical eels that are often present sympatrically in rivers.

In this study, we used eDNA analysis to carry out basin-scale surveys of A. marmorata and A. japonica in 11 rivers in southern Japan and northern Taiwan. We chose these species and regions because (1) the two eel species are sympatric due to their overlapping geographic distributions (Shiao et al. 2003; Han et al. 2012a; Hsu et al. 2019), thereby providing suitable sites for applying eDNA analysis to sympatric anguillid eel species; (2) use of eDNA analysis for A. marmorata is particularly important, as misidentification of this species using only morphological characters has been reported previously (Arai and Abdul Kadir 2017); and (3) a relatively low abundance of A. marmorata is expected in some parts of these regions, as it is thought to be close to the northern limit of the distribution range of this species (Jacoby and Gollock 2014a), thus offering suitable sites for investigating the sensitivity of eDNA analysis for detecting the presence of A. marmorata. For example, the abundance of A. marmorata has been reported to be an order of magnitude lower than that of A. japonica in a river in Taiwan (Hsu et al. 2019).

We first evaluated whether eDNA analysis can be used to estimate not only the spatial distribution but also the abundance and biomass of A. marmorata in rivers. To this end, we compared the results of eDNA analysis with those of the electrofishing method to estimate the presence or absence of A. marmorata in sampling sites in three rivers in Japan and two rivers in Taiwan. Moreover, we investigated the relationship between the eDNA concentration and the abundance of A. marmorata in the Oganeku River on Amami-Oshima Island, a subtropical island of Japan where A. marmorata is known to be predominant (Wakiya et al. 2019). Lastly, we carried out further eDNA sampling in six rivers in Japan and Taiwan, and discuss the riverine distribution of the two eel species in this region based on the eDNA concentration results from all of the study rivers. Samples of A. japonica collected from two Japanese rivers (the Atsumari and Mawatari rivers) in the current study were also used in an eDNA study for this species (Itakura et al. 2019).

MATERIALS AND METHODS

Study species

Anguilla japonica spawn in the waters west of the Mariana Islands in the western North Pacific Ocean (Tsukamoto et al. 2011), and their larvae drift westward to growth habitats in East Asia including Taiwan, eastern China, Korea, and Japan. After metamorphosing into glass eels (early juvenile phase), they migrate into brackish and freshwater habitats, where they remain as growth-phase yellow eels. They grow in diverse habitats within rivers, ranging from brackish estuaries to upland headwaters, lakes, and saline bays (Kaifu et al. 2010; Wakiya et al. 2016; Itakura et al. 2018; Itakura et al. 2019). Once they are sexually mature, they metamorphose into reproductive-stage silver eels and then migrate to their spawning areas (Han et al. 2016; Chen et al. 2018; Higuchi et al. 2018). This species is a commercially important species in East Asia and, due to a notable decline in its abundance, is classified as Endangered on the IUCN Red List of Threatened Species (Jacoby and Gollock 2014b)

Anguilla marmorata is the most widespread anguillid species in the world, ranging from the western Indian Ocean, across the Indo-Pacific, to French Polynesia in the South Pacific Ocean (Ege 1939; Watanabe et al. 2004). The species has four genetically distinct populations (Minegishi et al. 2008), one of which spawns in the same region as A. japonica (Kuroki et al. 2009; Tsukamoto et al. 2011). Although A. marmorata tends to live in freshwater areas rather than brackish water or seawater (Shiao et al. 2003; Nguyen et al. 2018), the species occupies a broad range of habitats from brackish estuaries to upland headwaters (Arai and Abdul Kadir 2017; Arai and Chino 2018; Hagihara et al. 2018; Wakiya et al. 2019). A. marmorata is one of the tropical eel species for which demand is increasing, and their glass eels are being used increasingly to stock farms in East Asia (Gollock et al. 2018). This species has been classified as of Least Concern in the IUCN Red List of Threatened Species (Jacoby and Gollock 2014a).

Study area

To investigate both whether eDNA analysis can detect the presence or absence of A. marmorata in rivers and the sensitivity of the approach, we carried out eDNA sampling and conventional capture-based sampling (electrofishing) at 49 study sites from the lower to the upper reaches of rivers in Kagoshima Prefecture, Japan (the Atsumari and Mawatari rivers on the southern main island of Japan and the Oganeku River on Amami-Oshima Island) and the Fengshan and Shuang rivers in Taiwan (Table 1; Figs. 1, 2). To test whether eDNA analysis can be used to estimate the abundance and biomass of A. marmorata, we investigated the relationship between the eDNA concentration and the abundance and biomass of A. marmorata in the Oganeku River (Table 1; Fig. 1), which is approximately 0.5 km in length. We collected eels over the entire area of the Oganeku River and conducted eDNA samplings at 20 m intervals from the river mouth to the uppermost reaches of the river (25 sites in total). We did not conduct surveys at study sites between 0 to 20 m from the river mouth because these were sandy beaches with very little water flow. To further study the riverine distribution of A. marmorata and A. japonica based on their eDNA concentration, we carried out eDNA sampling at 81 sites in five rivers (Hetsuka, Hirose, Kedo, Kubota, and Mae) in Kagoshima Prefecture, Japan, and three rivers (Fengshan, Nan-ao South, and Shuang) in Taiwan (Table 1; Figs. 1, 2).

Table 1.

Characteristics of the study rivers and the samplings for the environmental DNA (eDNA) analysis and electrofishing of anguillid eels (Anguilla japonica and Anguilla marmorata)

Country River No. study site Sampling date Length (km) eDNA copies L-1
　 　 　 　 A. japonica A. marmorata
Japan Atsumari 10 (10) 23–25 September 2016 5.9 354 ± 394 1929 ± 1951
Hetsuka 11 (0) 17–18 July 2017 4.2 9 ± 0 25 ± 22
Kedo 16 (0) 27 September 2017 11.5 45 ± 33 234 ± 233
Hirose 8 (0) 17–18 July 2017 8.7 208 ± 314 84 ± 111
Kubota 8 (0) 17–18 July 2017 14.8 43 ± 42 17 ± 17
Mawatari 7 (7) 23–25 September 2016 11.5 66 ± 55 3291 ± 4392
Mae 9 (0) 17–18 July 2017 20.0 265 ± 273 38 ± 29
Oganeku 25 (25) 26–29 July 2017 0.5 - 922 ± 900
Taiwan Fengshan 11 (5) 2–6 October 2017 45.5 13 ± 11 -
Nan-ao South 12 (0) 2–6 October 2017 30.6 107 ± 109 17 ± 11
　 Shuang 13 (2) 2–6 October 2017 26.8pare 55 ± 0 31 ± 31
Open in a new tab

Number of study site for electrofishing is indicated in parentheses. eDNA copies of sites at which eDNA of anguillid eels was detected is indicated as mean ± standard deviation.

Fig. 1.

Fig. 1.

Open in a new tab

Locations of the study region and study rivers in Japan and Taiwan used for environmental DNA (eDNA) analysis and/or electrofishing of anguillid eels (Anguilla japonica and Anguilla marmorata).

Fig. 2.

Fig. 2.

Open in a new tab

Maps showing the locations of the sites in each river and the environmental DNA (eDNA) detections of anguillid eels (Anguilla japonica and Anguilla marmorata). The presence or absence of eDNA detections are shown as different colored circles. The numbers below each river name indicate the proportion of eDNA detection sites for both species (jap: A. japonica; mar: A. marmorata). (a) Kedo River, (b) Mae River, (c) Atsumari and Mawatari rivers, (d) Hirose and Kubota rivers, (e) Hetsuka River, (f) Fengshan River, (g) Nan-ao South River, and (h) Shuang River. Map scales differ among panels.

We carried out eDNA sampling and eel collection from September 2016 to October 2017 (Table 1). The length and width of study sites in the Atsumari River ranged from 20.0 to 34.0 m and 3.3 to 8.1 m, respectively, and those in the Mawatari River ranged from 13.0 to 20.0 m and from 3.0 to 8.0 m, respectively (for details, see Itakura et al. 2019), whereas we did not measure those in the Taiwanese rivers. The width of the study sites in the Oganeku River ranged from 0.5 to 13.7 m, with a mean ± standard deviation of 3.4 ± 2.5 m. We measured water velocity at the center of the downstream point at each study site of the Oganeku River in August and November 2016, and we used the mean of the three measurement points for analysis (see Data analysis). Though the Mae and Kedo rivers flow through residential areas, all the other rivers flow through agricultural and forest lands. Most study sites were located in freshwater, and some sites were in brackish water.

eDNA analysis

Water sampling

Water sampling was conducted at the downstream side of each study site. Surface water (1 L) was collected at the center of the river by submerging a bottle approximately 10 cm in depth. Following Yamanaka et al. (2017), benzalkonium chloride solution (1 mL) was added immediately to each water sample to prevent eDNA degradation. After collection, each water sample was vacuum-filtered through either one or two 47 mm GF/F glass filters (pore size c. 0.7 μm; GE Healthcare Life Science, Whatman) within an average of 2 days (maximum 5 days). Next, the filters were wrapped immediately in commercial aluminum foil and stored at -20°C until eDNA extraction. The bottles and the filtering devices (i.e., filter funnels and measuring cups used for filtration) were bleached with 0.1% sodium hypochlorite. The bottles were further washed two or more times with surface river water from each sampling site immediately before collecting the water. In study sites at which eels were also collected, water was sampled done before eel collection.

eDNA extraction

eDNA was extracted from the filters following the method of Yamamoto et al. (2016). Total eDNA was extracted from each filter using a DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany), with a minor modification to adjust for eDNA extraction. In brief, the sample filter was placed in the suspended insert within a Salivette tube (Sarstedt, Nümbrecht, Germany), and 420 μL of a solution comprising 20 μL Proteinase K, 200 μL AL buffer, and 200 μL water was poured onto the filter. Then, the tube was incubated for 30 min at 56°C, after which the liquid held in the filter was collected via centrifugation. To increase the eDNA yield, 200 μL TE buffer (pH 8.0) was poured onto the filter, and the liquid was again collected via centrifugation. Next, 200 μL AL buffer and 600 μL ethanol were added to the collected liquid, the mixture was transferred to a spin column, and the final volume of eDNA was eluted in 100 μL AE buffer, according to the manufacturer’s instructions. During the eDNA extraction procedures, we checked for cross-contamination by extracting eDNA simultaneously from pure water (extraction negative control) as one sample for every extraction procedure (i.e., there was one negative control for every 7–23 river water samples).

Real-time qPCR

We quantified the eDNA samples by real-time TaqMan qPCR using a StepOnePlus Real-Time PCR system (Life Technologies, Foster City, USA). The mitochondrial 16S ribosomal RNA (rRNA) gene fragments were amplified and quantified using species-specific primers and probes. The primers and a probe for A. japonica were developed by Watanabe et al. (2005), whereas those for A. marmorata were designed for this study (Table 2). These primers specifically amplified 153 and 171 bp fragments of 16S rRNA gene for A. japonica and A. marmorata, respectively. DNA sequences of A. marmorata and other related species (i.e., A. japonica, Anguilla bicolor pacifica, and Anguilla luzonensis) potentially living in Japan and Taiwan were collected from the National Center for Biotechnology Information (NCBI, http://www.ncbi.nlm.nih.gov/) nucleotide database. The primers and probe set for A. marmorata was designed using UGENE v1.26.1 (Unipro, Russia) and Primer Express software (Thermo Fisher Scientific, Waltham, MA, USA), checking the sequence mismatching between the target and other related species. Each 20 μL TaqMan reaction contained 2 μL extracted eDNA solution, a final concentration of 900 nM forward and reverse primers, 125 nM TaqMan probe, and 0.1 μL of AmpErase Uracil N-Glycosylase (Thermo Fisher Scientific) in 1× Environmental Master Mix 2.0 (Life Technologies). We carried out qPCR in triplicate for each eDNA sample under the following conditions: 2 min at 50°C, 10 min at 95°C, and 55 cycles of 15 s at 95°C and 1 min at 60°C.

Table 2.

Specific primers and probe for the Japanese eel (Anguilla japonica) and the giant mottled eel (Anguilla marmorata) used in this study

Target species Primer and probe Sequences (5'–3') Source
A. japonica Forward primer AATCAGTAATAAGAGGGCCCAAGC Watanabe et al. 2005
Reverse primer TGTTGGGTTAACGGTTTGTGGTA
Probe FAM-CACATGTGTAAGTCAGAACGGACCGACC-TAMRA
A. marmorata Forward primer GGACATAAATGAGCAGTTATCCTGA This study
Reverse primer TGGTTGATTTCGTATACCGACG
Probe FAM-CTCTAATGCTATTCCTAATTAC-MGB-NFQ
Open in a new tab

To quantify the number of 16S rRNA genes in a 2 μL eDNA solution sample, a dilution series of standards consisting of 3 × 101 – 3 × 104 copies of a commercially synthesized artificial DNA fragment inserted into a plasmid was analyzed in triplicate simultaneously in each round of qPCR. In addition, pure water (2 μL) was analyzed in triplicate in all rounds of qPCR as a negative control. We found that the calibration curves from all rounds of qPCR had R2 values of 0.983–0.997, slopes between -3.175 and -2.954, and intercept values of 36.253–39.712 for A. japonica, and R2 values of 0.971–0.997, slopes between -3.662 and -3.303, and intercept values of 40.098– 41.914 for A. marmorata. Based on the calibration curve of each run and the Ct value of each sample, we calculated the copy number of 16S rRNA gene fragments by averaging the three replicate values of each sample. Watanabe et al. (2005) confirmed the specificity of the primers and probe for A. japonica by using tissue-derived DNA extractions from related species. The same primers and probe also enabled us to specifically detect eDNA of A. japonica from tanks, rivers, and the open ocean (Itakura et al. 2019; Takeuchi et al. 2019a b). We confirmed the specificity of the primers and probe for A. marmorata by using real-time PCR to determine the DNA sequences of some of the amplified samples. We collected these samples from rivers in which abundant A. marmorata had been captured by electrofishing (Wakiya et al. 2019). The DNA sequencing was performed commercially (Fasmac Co., Ltd, Kanagawa, Japan). Basic Local Alignment Search Tool searches using the NCBI nucleotide database showed that all the sequenced fragments from samples had only 16S rRNA gene sequences of A. marmorata.

Eel collection

Eels were collected from the downstream to upstream direction of each study site in Japanese rivers using an electroshocker (LR-20B, Smith-Root, Inc., Vancouver, WA, USA). Captured fish were euthanized with > 10% eugenol solution (FA100; DS Pharma Animal Health Co., Ltd. Japan). Though fish from the Oganeku River were released subsequently into capture sites after measurements for other studies, fish from other rivers were stored at -20°C until measurements were made. Each specimen was identified morphologically following Watanabe et al. (2004). The growth stage of each specimen was confirmed based on the color of its body and pectoral fins following previous studies (Okamura et al. 2007; Hagihara et al. 2012). The total length and body weight of each eel were measured to the nearest 1 mm and 0.1 g, respectively. In addition, we calculated the observed abundance and biomass densities of A. marmorata at 10 m intervals in the Oganeku River by dividing either the number or the total mass of captured eels, respectively, by the area of the study site (m2). Eel sampling was conducted both under the guidance of and with the permission of the Fisheries Adjustment Rules of Kagoshima Prefectures.

In the Taiwanese rivers, eels were collected from each study site using an electroshocker (Taixing Battery Store). Captured fish were released into capture sites once the species and growth stage were determined. Eel sampling was approved by the Fishery Agency, Council of Agriculture, Executive Yuan, Taiwan

Data analysis

All statistical analyses were carried out using R 3.6.0. To assess the relationship between eDNA concentration and the abundance and biomass of A. marmorata in the Oganeku River, we used two different models: i) a Type I regression model and ii) a Type II regression model with the standardized major axis method (sma in the package smatr). Type II regression models can treat two variables (x and y) with an equal magnitude of random variation. These models included eDNA concentrations as a response variable and the abundance or biomass of A. marmorata as an explanatory variable. eDNA concentrations were log-transformed. Using the two models, we first tested the extent to which eDNA concentration reflected the abundance of eels in the river. To accomplish this, we calculated the abundances of eels between each water sampling point and over different distances by varying the distance from 10 to 150 m with 10 m increments. Furthermore, as the transport distance of eDNA from the source organisms varies with water velocity (Deiner and Altermatt 2014; Pilliod et al. 2014; Jane et al. 2015; Wilcox et al. 2016), the watershed was divided into two river basins (i.e., downstream and upstream) according to water velocity characteristics. To identify major changes in water velocity throughout the river, we carried out a changepoint analysis based on a likelihood ratio test (cpt. mean in the package changepoint). The findings showed that the mean water velocity of the river changed drastically at 150 and 160 m from the river mouth in August and November 2016, respectively (Fig. S1). Therefore, we calculated the abundances of eels between each water sampling point and the various consecutive distances for the two basins separated at the point where the water velocity changed dramatically (i.e., 160 m from the river mouth). Lastly, we used Akaike’s information criterion (AIC) and R2 values to evaluate 225 candidate models for Type I and II regression models.

As studies have reported that the abundance of anguillid eels declines with increasing distance from the river mouth (e.g., Yokouchi et al. 2008; Kaifu et al. 2010), we also examined the spatial distribution of eDNA concentration. In the Oganeku River, we used a linear regression model (Type I regression model) with eDNA concentration of A. marmorata as a response variable and distance from the river mouth as an explanatory variable. In the other rivers, we used a linear mixed model (LMM: lmer in the package lme4) with eDNA concentration of A. japonica or A. marmorata as a response variable, distance from the river mouth as an explanatory variable, and the river as a random effect. We log-transformed eDNA concentrations by adding 1 to the variable. In addition, we compared the proportion of eDNA detection sites between the two eel species for each river using Fisher’s Exact Test.

RESULTS

Number of collected eels

In all, 135 individuals of the two eel species were collected in this study (Table 3); 46 Anguilla japonica and 9 A. marmorata were collected from the Atsumari and Mawatari rivers on the southern main island of Japan (Fig. 3), 60 A. marmorata were collected from the Oganeku River on Amami-Oshima Island, Japan, and 18 A. japonica and 2 A. marmorata were collected from the two Taiwanese rivers (Fig. 3). We collected no A. japonica from the Oganeku and Shuang rivers and no A. marmorata from the Fengshan River. All of the eels captured were in their growth phase, apart from one A. marmorata glass eel collected from the Atsumari River. No anguillid species other than A. japonica and A. marmorata were collected in this study. The total length of the collected A. marmorata in the Oganeku River rangedfrom78to600mm,withamean±SDof303± 156 mm.

Fig. 3.

Fig. 3.

Open in a new tab

Species composition of anguillid eels (Anguilla japonica and Anguilla marmorata) collected in the Atsumari and Mawatari rivers in Japan and the Fengshan and Shuang rivers in Taiwan. The numbers in each bar chart indicate the number of A. japonica or A. marmorata captured at each study site. The blank indicates that no sampling was conducted.

Although both species were collected from the lower to upper reaches of the Atsumari and Mawatari rivers, higher numbers of A. japonica were collected from the lower reaches (Fig. 3). In the Fengshan River, A. japonica were collected only from the lower reaches. In the Shuang River, A. marmorata were collected only from the upper reaches because we did not sample the lower reaches of the river (Fig. 3).

Table 3.

Summary of the results of the environmental DNA (eDNA) analysis and the electrofishing surveys of anguillid eels (Anguilla japonica and A. marmorata)

Country River No. of captured eels and total length (mm)
 Eel capture sites (%)
 eDNA detection sites (%)
 eDNA detection sites where eels were captured (%)
 eDNA detection sites where no eels were captured (%)
A. japonica
 A. marmorata
No. TL No. TL A. marmorata
Japan Atsumari 24 422 ± 158 (154–780) 5 279 ± 300 (57–805) 4/10 (40) 10/10 (100) 4/4 (100) 6/6 (100)
Mawatari 22 240 ± 93 (140–470) 4 456 ± 460 (67–1120) 4/7 (57) 7/7 (100) 4/4 (100) 3/3 (100)
Oganeku 0 - 60 345 ± 160 (70–656) 12/25 (48) 25/25 (100) 12/12 (100) 13/13 (100)
Taiwan Fengshan 18 - 0 - 0/5 (0) 0/11 (0) 0/0 (0) 0/11 (0)
Shuang 0 - 2 - 1/2 (50) 9/13 (69) 1/1 (100) 0/1 (0)
Total 　 64 - 71 - 21/49 (43) 51/66 (77) 21/21 (100) 22/29 (76)
Open in a new tab

TL of eels is indicated as mean ± standard deviation (range).

Comparison of presence or absence of anguillid eels using eDNA and electrofishing

A. marmorata were collected from 21 of the 49 (43%) study sites via electrofishing in the Atsumari, Mawatari, Oganeku, Fengshan, and Shuang rivers, whereas eDNA of the species was detected not only at all sites where eels were collected but also at 22 sites where the species was not collected directly (Table 3). We identified eDNA of A. marmorata from all sites in these Japanese rivers and from nine of the 24 sites in the Taiwanese rivers (51 of 66, 77% in total). We did not find A. marmorata in the Fengshan River, either by electrofishing or eDNA analysis. In Taiwan, we collected A. japonica by electro-fishing at three of the five study sites in the Fengshan River. Conversely, we identified the eDNA of the species from one of three sites (33%) where the eels were collected and from two sites where the species was not collected directly (results not shown).

Relationships between eDNA concentration and abundance and biomass of A. marmorata

We investigated whether eDNA concentration reflected the abundance and biomass of A. marmorata in the Oganeku River, Japan. We found a significant positive relationship between eDNA concentration and the abundance of A. marmorata in all candidate models with different abundances of eels at each sampling site (Type I or II regression model, p < 0.05). The AIC and R2 values of both regression models considerably decreased or increased, respectively, when the abundance of eels was calculated more than 50 m from the water sampling point in the lower reaches of the river and more than 70 m from the water sampling point in the upper reaches (respectively) (Fig. S2). Using these distances, the models revealed that the eDNA concentration of A. marmorata was significantly positively related to both the abundance (Type I: coefficient ± SE = 1.416 ± 0.237, t = 5.965, p < 0.001, R2 = 0.590, AIC = 76.65; Type II: , R2 = 0.607, p < 0.001, AIC = 121.88) and biomass (Type I: coefficient ± SE = 0.859 ± 0.284, t = 3.029, p = 0.006, R2 = 0.254, AIC = 91.63; Type II: , R2 = 0.285, p < 0.006, AIC = 141.88) of eels (Fig. 4).

Fig. 4.

Fig. 4.

Open in a new tab

Relationships between the environmental DNA (eDNA) concentrations for the giant mottled eel (Anguilla marmorata) in the surface waters of the Oganeku River in Amami-Oshima Island, Japan, and their (a) abundance and (b) biomass. The green and red lines and shaded areas indicate the best regression slope and 95% confidence intervals (CI) of the Types I and II regression models, respectively.

Spatial distribution of anguillid eels inferred by eDNA analysis

In this study, we identified eDNA of both eel species in all the rivers apart from the Fengshan River in Taiwan—we did not identify eDNA of A. marmorata from any study site in this river (Table 3; Fig. 2). However, the proportion of eDNA detection sites among the two eel species varied between rivers. Though the proportion of eDNA detection sites of A. marmorata was significantly higher in the Hetsuka and Shuang rivers than was that of A. japonica (Fisher’s Exact Test, p < 0.05; Fig. 2), in other rivers, the proportions of eel species did not differ significantly (Fisher’s Exact Test, p > 0.05; Fig. 2). In particular, eDNA of both species was detected at most study sites in the Atsumari, Kedo, Mae, and Mawatari rivers.

We found the eDNA concentration of A. japonica to be higher in downstream sites than in more upstream sites in each river, and it declined significantly with increasing distance from the river mouth (LMM: coefficient ± SE = -0.079 ± 0.025, t = -3.212, p = 0.002; Fig. 5). Conversely, we did not find a significant relationship between the eDNA concentration of A. marmorata and distance from the river mouth (LMM: coefficient ± SE = -0.043 ± 0.026, t = -1.629, p = 0.107; Fig. 5). However, the eDNA concentration of A. marmorata in the Oganeku River was higher in downstream sites than in more upstream sites, and it declined significantly with increasing distance from the river mouth (Type I: coefficient ± SE = -0.009 ± 0.001, t = -6.678, p < 0.001).

Fig. 5.

Fig. 5.

Open in a new tab

Environmental DNA (eDNA) concentrations of anguillid eels (Anguilla japonica and A. marmorata) along distance from the river mouth of each study river. The species are shown as different colored circles and lines. The scales of the x and y axes differ among rivers.

DISCUSSION

Effectiveness of eDNA analysis for estimating the distribution of Anguilla marmorata

In this study, we carried out basin-scale surveys of Anguilla marmorata from near the river mouths to the upstream reaches of five rivers in southern Japan and northern Taiwan, and we compared the results of eDNA analysis and direct collection of fish by electrofishing. As anticipated, we collected only 11 A. marmorata (apart from the Oganeku River on Amami-Oshima Island, from which we collected 60 individuals), suggesting that the abundance of A. marmorata in the study region was extremely low compared to that of A. japonica (64 individuals collected). Still, we identified A. marmorata eDNA not only from all the study sites (21 sites) from which species was collected by electrofishing but also from a further 22 study sites where the species was not collected directly. The sites were located mainly in the upper reaches of the rivers, where eel densities are generally low (Itakura et al. unpublished data). Thus, eDNA analysis seems to be more sensitive than conventional survey techniques for detecting the presence of A. marmorata, as reported previously for A. japonica (Itakura et al. 2019) and other fish (Jerde et al. 2013; Wilcox et al. 2016; Sakata et al. 2017), and it is therefore likely a powerful tool for monitoring the spatial distribution of A. marmorata in rivers.

In addition, irrespective of model type, we found a significant positive relationship between the eDNA concentration and the abundance and biomass of A. marmorata in the Oganeku River, indicating that eDNA analysis can facilitate estimation of the abundance and biomass of this species in rivers. Although we found a significant positive relationship between eDNA concentration and the abundance of eels in all candidate models with different abundances of eels at each sampling site, this relationship increased in strength when we calculated the abundance of eels more than 50 and 70 m ahead of water sampling points in the lower and upper reaches of the river, respectively. The eDNA concentration appears to reflect the abundance of eels inhabiting areas more distant from the water sampling points in the upper reaches of the river than in the lower reaches. This is likely due to the higher water velocity in the upper reaches and the distance that the eDNA travels from the source being larger at higher water velocity (Jane et al. 2015). Although the distance that the eDNA migrates in rivers ranges from several tens of meters to kilometers (Deiner and Altermatt 2014; Pilliod et al. 2014; Jane et al. 2015; Wilcox et al. 2016), our findings indicate that the eDNA concentration would reflect the abundance of this species at more than 50 and 70 m ahead of water sampling points in the lower and upper reaches of the river, respectively. Itakura et al. (2019) did not verify that the distances that eDNA travels reflects eel abundance; thus, our findings provide new information that will help in understanding the result of eDNA concentration of anguillid eels.

The AIC and R2 values of the models indicate that the eDNA concentration of A. marmorata reflects their abundance better than it does their biomass. Although this has been reported for A. japonica in rivers (Itakura et al. 2019), this finding does not correspond with an aquarium experiment for other fish, which found that eDNA analysis reflects fish biomass better than it does fish density (Maruyama et al. 2014). This may be due to differential eDNA release at different developmental stages of eels and biased distribution of body size and individual density of eels in rivers. The rate of release of eDNA per weight decreases as the life history stages develop; thus, the release rate is higher in juvenile than in adult fish (Maruyama et al. 2014; Takeuchi et al. 2019a), possibly resulting from an ontogenetic decrease in the metabolic rate. This may lead to eDNA analysis overestimating the biomass of a population that is dominated by juveniles (Maruyama et al. 2014). In fact, we identified relatively high concentrations of eDNA in the lower reaches (sampling sites 100–200 m from the river mouth) of the Oganeku River, which included low eel biomass sites because of the dominance of small eels (< 200 mm in TL) (Itakura et al. unpublished data). Conversely, relatively moderate or low concentrations of eDNA were identified more downstream or upstream of the river, which included high or moderate eel biomass sites due to the dominance of larger eels (Itakura et al. unpublished data). This mismatch between high eDNA concentration and the low biomass of A. marmorata in the lower reach of the river could result in the model for biomass performing lower than that for abundance observed in this study. As the body size and individual density of anguillid eels tend to vary depending on the distance from the river mouth (Smogor et al. 1995; Glova et al. 1998; Goodwin and Angermeier 2003; Lasne and Laffaille 2008; Yokouchi et al. 2008; Kaifu et al. 2010), eDNA concentration can reflect the abundance of anguillid eels in rivers better than it can reflect their biomass.

The abundance of A. marmorata has been shown to decline with increasing distance from the river mouth when the species is highly dominant throughout the river (Robinet et al. 2007). Similarly, our results showed that the eDNA concentration of A. marmorata decreased significantly with increasing distance from the mouth of the Oganeku River. As eDNA does not accumulate in downstream reaches, being removed from there through processes such as settling or destruction (Jane et al. 2015; Laramie et al. 2015), our findings suggest that eDNA analysis can be used to identify such habitat utilization of A. marmorata.

We did not identify eDNA of A. marmorata in any study sites in the Fengshan River in Taiwan from where the species was not collected directly using electrofishing. However, A. marmorata has been collected previously in the upper reaches of the river using electrofishing (Hsu et al. 2019). This failure to identify eDNA of A. marmorata was likely due to an exceptionally low eDNA concentration of A. marmorata arising from lower density of the species and the larger scale of the river compared with other study rivers from which A. marmorata was collected. Hsu et al. (2019) collected 240 A. japonica and 24 A. marmorata individuals in the Fengshan River over 4 years. This ten-fold difference in the relative abundances of A. japonica and A. marmorata is higher than that observed in the Atsumari and Mawatari rivers in this study (one-fifth), suggesting that the density of A. marmorata in the Fengshan River is relatively low. Moreover, as the Fengshan River (45.5 km) is longer than the Shuang (26.8 km), Nan-ao South (30.6 km), Atsumari (5.9 km), and Mawatari (11.5 km) rivers, there should be greater attenuation of A. marmorata eDNA concentration in the Fengshan River than in the other rivers. For A. japonica study, Itakura et al. (2019) detected the eDNA of A. japonica from nearly all study sites where the species was collected directly; however, the study was conducted in relatively small-scale rivers (3.0–18.0 km). Together, the low density of A. marmorata and the large size of the river may have resulted in an exceptionally low eDNA concentration of the species that is below the detectable concentration from the 1 L river water samples collected from each study site. Therefore, it is likely that either using more than 1 L per sample of river water or increasing the sampling numbers is required (Santas et al. 2013; Moyer et al. 2014) to identify the presence of A. marmorata in rivers with a very low concentration of eDNA.

Spatial distribution of anguillid eel eDNA in rivers that multiple eel species inhabit sympatrically

We identified both A. japonica and A. marmorata eDNA from nearly all rivers, suggesting that they are sympatric in the same watershed in the study region. However, the degree of sympatry of these species varied between rivers; though A. japonica and A. marmorata were identified at either an equal or similar proportion of eDNA detection sites in some rivers, either species was dominant in other rivers. This finding is consistent with previous studies that observed a difference in percentage composition of A. marmorata and A. japonica in rivers in Taiwan (Shiao et al. 2003). Although both A. marmorata and A. japonica can potentially recruit to the study region, the relative recruitments of these species can vary even in geographically close regions (Han et al. 2012a). As these species are transported passively by oceanic currents to their growth habitats (Kimura et al. 1999; Han et al. 2012b 2016), the positional relationship between the mouths of each river and the oceanic currents may be an important factor determining whether they can recruit to the growth habitats. As such, differences in recruitment between rivers might lead to differences in the relative abundance of these species between the study rivers.

Another potential reason for the difference in predominant eel species between rivers is the difference in their environmental preferences of growth habitat after recruitment (Shiao et al. 2003). Small and steep streams and gravel and rocky substrates dominate the upper reaches of rivers, whereas gradual flow and sandy and muddy substrates dominate the lower reaches. A. japonica tends to inhabit the lower reaches of rivers in Taiwan, whereas A. marmorata tends to inhabit the upper reaches (Shiao et al. 2003; Hsu et al. 2019). Conversely, in the subtropical regions, only A. marmorata is found in some small and steep rivers, where the environments of the lower reaches and upper reaches are quite similar to each other (Shiao et al. 2003; Wakiya et al. 2019). Thus, if glass eels of both species recruit to rivers, one of them might not be able to survive to the juvenile or growth phase because of unsuitable environmental conditions; however, it should be noted that we did not study the environments in each river. Studying the recruitment and environmental preferences of juvenile eels in growth habitats will help in elucidating their spatial distribution.

A. japonica are generally predominant in the lower reaches of any river, whereas the riverine distribution of A. marmorata appears to be completely different. This study found that there was a tendency for the eDNA concentration of A. japonica to be higher at downstream sites in each river and to decline significantly with increasing distance from the river mouth. This agrees with previous studies showing that A. japonica tends to inhabit lower reaches (Shiao et al. 2003; Hagihara et al. 2018; Hsu et al. 2019) and with findings from this study using electrofishing, in addition to findings showing declines in abundance of A. japonica with increasing distance from the river mouth in rivers where only this species is present (Yokouchi et al. 2008; Kaifu et al. 2010; Itakura et al. 2019). Conversely, there was no significant (negative) relationship between the eDNA concentration of A. marmorata and the distance from the river mouth, which is not consistent with previous studies that demonstrate a negative relationship between the abundance of A. marmorata and the distance from the river mouth in rivers where this species is predominant (Robinet et al. 2007; Itakura et al. unpublished data). A. marmorata tends to be dominant in the upper reaches of rivers in which A. japonica is dominant in the lower reaches (Shiao et al. 2003; Hsu et al. 2019), but this species is also found in varying habitat types, ranging from estuary to headstream waters (Chino and Arai 2010; Arai and Abdul Kadir 2017; Arai and Chino 2018; Hagihara et al. 2018; Nguyen et al. 2018). Therefore, the variability of eDNA concentrations of A. marmorata found in the present study might reflect the plasticity of their pattern of habitat use in rivers.

We caught more A. japonica (n = 46) than A. marmorata (n = 9) in the Atsumari and Mawatari rivers (Table 3); however, we detected higher eDNA concentrations for A. marmorata than for A. japonica (Fig. 5; Table 1). This may be related to differences in eDNA release and microhabitat preferences among species. A. marmorata is heavier than A. japonica at similar lengths (Shiao et al. 2003), which indicates that, per individual, eDNA release is higher for A. marmorata than for A. japonica. Moreover, A. marmorata is one of the largest anguillid species, and the largest total length of A. marmorata (1,120 mm) collected in the rivers was considerably higher than that of A. japonica (780 mm) (Table 3). This larger body size of A. marmorata may result in an increase in eDNA release, which may contribute to the higher concentration of eDNA of A. marmorata compared with that of A. japonica observed in the rivers. In rivers where the two species are present sympatrically, A. marmorata appears to prefer deep pools, whereas A. japonica favors sandy and muddy substrates (Shiao et al. 2003). This suggests that there may be a greater abundance of A. marmorata residing in deeper parts of the rivers or upstream waters, none of which could be sampled by electrofishing in this study. It should be noted that there was no such deep water in the Oganeku River, where the relationship between eDNA concentration and abundance and biomass of A. marmorata was investigated, and only A. marmorata was collected in the river. Although understanding of the riverine distribution pattern of anguillid eels is still poor, it may be affected by both environmental conditions and interspecific competition related to either species composition or relative abundance (Arai and Abdul Kadir 2017; Hagihara et al. 2018), which may, in turn, be influenced by recruitment. Quantitative approaches for studying anguillid eels throughout their ranges, such as eDNA analysis, combined with measurement of detailed environmental conditions, are required to elucidate their patterns of riverine distribution and the mechanism underlying these patterns.

CONCLUSIONS

The present study is the first to demonstrate the potential usefulness of eDNA analysis for estimating the spatial distribution, abundance, and biomass of the tropical eel A. marmorata in rivers. As Itakura et al. (2019) demonstrated similar results for the temperate eel A. japonica, eDNA analysis could be applied to anguillid eels in general to monitor their riverine populations. We further applied this method to basin-scale surveys to investigate the riverine distribution of sympatric A. japonica and A. marmorata, and our findings indicated that eDNA analysis could reflect their spatial distributions. eDNA analysis involves less time and fewer human resources than does electrofishing (Itakura et al. 2019) and enables eel populations to be monitored over large spatial and temporal scales using a consistent protocol. This approach could be particularly beneficial in tropical regions, where identifying eel species based solely on morphological characteristics is difficult, as several eel species often co-inhabit rivers in these regions (Watanabe et al. 2004). Therefore, eDNA analysis can help in collecting data not only on the population dynamics of these anguillid species but also on their ecology, such as their spatial distribution, knowledge of both of which are limited considerably in tropical regions (Jacoby et al. 2015), thereby providing invaluable information for managing these eels.

Supplementary materials

Fig. S1.

Water velocity plotted against distance from the river mouth in the Oganeku River in the Amami-Oshima Island, Japan. The observation values on (a) August and (b) November 2016. The red line indicates the mean values of water velocity at each section that were separated by changepoint method based on likelihood ratio test. (download)

zoolstud-59-017-s001.docx (65.1KB, docx)
Fig. S2.

Relationships between Akaike’s information criterion (AIC) or R2 values and distances in lower and upper reaches that abundance of A. marmorata was calculated in the Oganeku River in the Amami-Oshima Island, Japan. AIC and R2 values were calculated by a generalised linear model (GLM) and Type II regression model. (download)

Acknowledgments

We are deeply grateful to Y.F. Chang, K. Kitamura, and B. Hsiao for their help with the field sampling and T. Hirae for his help with the laboratory work. We thank K. Arai for his constructive comments on the manuscript. H.I. was financially supported by the Research Fellowship for Young Scientists and the Postdoctoral Fellowship for Research Abroad from the Japan Society for Promotion of Science. This study was supported by the River Fund of the River Foundation, Japan, the Sasakawa Scientific Research Grant from the Japan Science Society and the CREST Programme from the Japan Science and Technology Agency. The authors thank the Ministry of Science and Technology, Executive Yuan, Taiwan (NSC 102-2628-B-002 -023 -MY3, MOST 105-2313-B-002-030, MOST 106-2313-B-002 -036 -MY3), and the Council of Agriculture, Executive Yuan, Taiwan (106AS-11.3.4-FA-F1) for funding this project.

List of abbreviations

AIC

Akaike’s information criterion.

CI

Confidence intervals.

IUCN

International Union for Conservation of Nature.

LMM

Linear mixed model.

NCBI

National Center for Biotechnology Information.

Footnotes

Authors’ contributions: HI, RW conceived the study and designed the methodology; SCC, YCH, YSH, HYH, HI, RW and CSY collected the data; HI, TM, MKS and SY analyzed the eDNA samples; HI analyzed the data; HI led the writing of the article. All authors contributed critically to the drafts and gave final approval for publication.

Competing interests: The authors declare that they have no conflict of interests.

Availability of data and materials: The data that support the findings of this study are not currently available because we are drafting another manuscript using some parts of these data.

Consent for publication: Not applicable.

Ethics approval consent to participate: Captured fish were euthanized with > 10% eugenol solution to euthanize them to satisfy both national and institutional standards.

References

  1. Aoyama J, Shinoda A, Sasai S, Miller MJ, Tsukamoto K. 2005. First observations of the burrows of Anguilla japonica. J Fish Biol 67:1534–1543. doi:10.1111/j.1095-8649.2005.00860.x.
  2. Aoyama J, Wouthuyzen S, Miller MJ, Sugeha HY, Kuroki M, Watanabe S, Syahailatua A, Tantu FY, Hagihara S, Triyanto, Otake T, Tsukamoto K. 2018. Reproductive ecology and biodiversity of freshwater eels around Sulawesi island Indonesia. Zool Stud 57:30. doi:10.6620/ZS.2018.57-30. [DOI] [PMC free article] [PubMed]
  3. Arai T, Abdul Kadir SR. 2017. Diversity, distribution and different habitat use among the tropical freshwater eels of genus Anguilla. Sci Rep 7:7593. doi:10.1038/s41598-017-07837-x. [DOI] [PMC free article] [PubMed]
  4. Arai T, Chino N. 2018. Opportunistic migration and habitat use of the giant mottled eel Anguilla marmorata (Teleostei: Elopomorpha). Sci Rep 8:5666. doi:10.1038/s41598-018-24011-z. [DOI] [PMC free article] [PubMed]
  5. 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. [DOI] [PMC free article] [PubMed]
  6. Chino N, Arai T. 2010. Migratory history of the giant mottled eel (Anguilla marmorata) in the Bonin Islands of Japan. Ecol Freshw Fish 19:19–25. doi:10.1111/j.1600-0633.2009.00385.x.
  7. Deiner K, Altermatt F. 2014. Transport distance of invertebrate environmental DNA in a natural river. PLoS ONE 9(2):e88786. doi:10.1371/journal.pone.0088786. [DOI] [PMC free article] [PubMed]
  8. Doi H, Inui R, Akamatsu Y, Kanno K, Yamanaka H, Takahara T, Minamoto T. 2017. Environmental DNA analysis for estimating the abundance and biomass of stream fish. Freshw Biol 62:30– 39. doi:10.1111/fwb.12846.
  9. Dougherty MM, Larson ER, Renshaw MA, Gantz CA, Egan SP, Erickson DM, Lodge DM. 2016. Environmental DNA (eDNA) detects the invasive rusty crayfish Orconectes rusticus at low abundances. J Appl Ecol 53:722–732. doi:10.1111/1365-2664.12621. [DOI] [PMC free article] [PubMed]
  10. Ege V. 1939. A revision of the genus Anguilla Shaw: a systematic, phylogenetic and geographical study. Dana Rep 16:1–256.
  11. Fukumoto S, Ushimaru A, Minamoto T. 2015. A basin-scale application of environmental DNA assessment for rare endemic species and closely related exotic species in rivers: A case study of giant salamanders in Japan. J Appl Ecol 52:358–365. doi:10.1111/1365-2664.12392.
  12. Glova GJ, Jellyman DJ, Bonnett ML. 1998. Factors associated with the distribution and habitat of eels (Anguilla spp.) in three New Zealand lowland streams. New Zeal J Mar Freshw Res 32:255– 269.
  13. Gollock M, Shiraishi H, Carrizo S, Crook V, Levy E. 2018. Status of non-CITES listed anguillid eels. CITES AC30 Document 18.1 Annex 2.
  14. Goodwin KR, Angermeier PL. 2003. Demographic characteristics of American eel in the Potomac River drainage, Virginia. Trans Am Fish Soc 132:524–535. doi:10.1577/1548-8659(2003)132<0524:DCOAET>2.0.CO;2.
  15. Hagihara S, Aoyama J, Limbong D, Tsukamoto K. 2012. Morphological and physiological changes of female tropical eels, Anguilla celebesensis and Anguilla marmorata, in relation to downstream migration. J Fish Biol 81:408–426. doi:10.1111/j.1095-8649.2012.03332.x. [DOI] [PubMed]
  16. Hagihara S, Aoyama J, Limbong D, Tsukamoto K. 2018. Interspecific and sexual differences in riverine distribution of tropical eels Anguilla spp. J Fish Biol 93:21–29. doi:10.1111/jfb.13666. [DOI] [PubMed]
  17. Han Y, Yambot A V, Zhang H, Hung C. 2012a. Sympatric Spawning but Allopatric Distribution of Anguilla japonica and Anguilla marmorata: Temperature-and Oceanic Current-Dependent Sieving. PLoS ONE 7(6):e37484. doi:10.1371/journal. pone.0037484. [DOI] [PMC free article] [PubMed]
  18. Han YS, Zhang H, Tseng YH, Shen ML. 2012b. Larval Japanese eel (Anguilla japonica) as sub-surface current bio-tracers on the East Asia continental shelf. Fish Oceanogr 21:281–290. doi:10.1111/j.1365-2419.2012.00624.x.
  19. Han YS, Wu CR, Iizuka Y. 2016. Batch-like arrival waves of glass eels of Anguilla japonica in offshore waters of Taiwan. Zool Stud 55:36. doi:10.6620/ZS.2016.55-36. [DOI] [PMC free article] [PubMed]
  20. Higuchi T, Watanabe S, Manabe R, Kaku T, Okamura A, Yamada Y, Miller MJ, Tsukamoto K. 2018. Tracking Anguilla japonica silver eels along the West Marina Ridge using pop-up archival transmitting tags. Zool Stud 57:24. doi:10.6620/ZS.2018.57-24. [DOI] [PMC free article] [PubMed]
  21. Hsu H, Chen H, Han Y. 2019. Habitat partitioning and its possible genetic background between two sympatrically distributed eel species in Taiwan. Zool Stud 58:27. doi:10.6620/ZS.2019.58-27. [DOI] [PMC free article] [PubMed]
  22. Itakura H, Arai K, Kaifu K, Shirai K, Yoneta A, Miyake Y, Secor DH, Kimura S. 2018. Distribution of wild and stocked Japanese eels in the lower reaches of the Tone River catchment revealed by otolith stable-isotope ratios. J Fish Biol 93:803–813. doi:10.1111/jfb.13782. [DOI] [PubMed]
  23. Itakura H, Wakiya R, Yamamoto S, Kaifu K, Sato T, Minamoto T. 2019. Environmental DNA analysis reveals the spatial distribution, abundance, and biomass of Japanese eels at the river‐basin scale. Aquat Conserv Mar Freshw Ecosyst 29:361– 373. doi:10.1002/aqc.3058.
  24. IUCN 2019. The IUCN Red List of Threatened Species. Version 2019.2.
  25. Jacoby D, Gollock M. 2014a. Anguilla marmorata. IUCN Red List Threat. Species 2014e.T166189A45832585.
  26. Jacoby DMP, Casselman JM, Crook V, DeLucia MB, Ahn H, Kaifu K, Kurwie T, Sasal P, Silfvergrip AM, Smith KG, Uchida K, Walker AM, Gollock MJ. 2015. Synergistic patterns of threat and the challenges facing global anguillid eel conservation. Glob Ecol Conserv 4:321–333. doi:10.1016/j.gecco.2015.07.009.
  27. Jacoby D, Gollock M. 2014b. Anguilla japonica. IUCN Red List Threat. Species 2014e.T166184A1117791.
  28. Jane SF, Wilcox TM, Mckelvey KS, Young MK, Schwartz MK, Lowe WH, Letcher BH, Whiteley AR. 2015. Distance, flow and PCR inhibition: eDNA dynamics in two headwater streams. Mol Ecol Resour 15:216–227. doi:10.1111/1755-0998.12285. [DOI] [PubMed]
  29. Jerde CL, Chadderton WL, Mahon AR, Renshaw MA, Corush J, Budny ML, Mysorekar S, Lodge DM. 2013. Detection of Asian carp DNA as part of a Great Lakes basin-wide surveillance program. Can J Fish Aquat Sci 70:522–526. doi:10.1139/cjfas-2012-0478.
  30. Kaifu K, Tamura M, Aoyama J, Tsukamoto K. 2010. Dispersal of yellow phase Japanese eels Anguilla japonica after recruitment in the Kojima Bay-Asahi river system, Japan. Environ Biol Fishes 88:273–282. doi:10.1007/s10641-010-9640-y.
  31. Kimura S, Döös K, Coward AC. 1999. Numerical simulation to resolve the issue of downstream migration of the Japanese eel. Mar Ecol Prog Ser 186:303–306.
  32. Kuroki M, Aoyama J, Miller MJ, Yoshinaga T, Shinoda A, Hagihara S, Tsukamoto K. 2009. Sympatric spawning of Anguilla marmorata and Anguilla japonica in the western North Pacific Ocean. J Fish Biol 74:1853–1865. doi:10.1111/j.1095-8649.2009.02299.x. [DOI] [PubMed]
  33. Laramie MB, Pilliod DS, Goldberg CS. 2015. Characterizing the distribution of an endangered salmonid using environmental DNA analysis. Biol Conserv 183:29–37. doi:10.1016/j.biocon.2014.11.025.
  34. Lasne E, Laffaille P. 2008. Analysis of distribution patterns of yellow European eels in the Loire catchment using logistic models based on presence-absence of different size-classes. Ecol Freshw Fish 17:30–37. doi:10.1111/j.1600-0633.2007.00253.x.
  35. Leander NJ, Shen KN, Chen RT, Tzeng WN. 2012. Species composition and seasonal occurrence of recruiting glass eels (Anguilla spp.) in the Hsiukuluan River, Eastern Taiwan. Zool Stud 51:59–71.
  36. Lodge DM, Turner CR, Jerde CL, Barnes MA, Chadderton L, Egan SP, Feder JL, Mahon AR, Pfrender ME. 2012. Conservation in a cup of water: estimating biodiversity and population abundance from environmental DNA. Mol Ecol 21:2555–2558. doi:10.1111/j.1365-294X.2012.05600.x. [DOI] [PMC free article] [PubMed]
  37. Maruyama A, Nakamura K, Yamanaka H, Kondoh M, Minamoto T. 2014. The release rate of environmental DNA from juvenile and adult fish. PLoS ONE 9(12):e114639. doi:10.1371/journal. pone.0114639. [DOI] [PMC free article] [PubMed]
  38. Minamoto T, Fukuda M, Katsuhara KR, Fujiwara A, Hidaka S, Yamamoto S, Takahashi K, Masuda R. 2017. Environmental DNA reflects spatial and temporal jellyfish distribution. PLoS ONE 12(2):e0173073. doi:10.1371/journal.pone.0173073. [DOI] [PMC free article] [PubMed]
  39. Minegishi Y, Aoyama J, Tsukamoto K. 2008. Multiple population structure of the giant mottled eel, Anguilla marmorata. Mol Ecol 17:3109–3122. doi:10.1111/j.1365-294X.2008.03822.x. [DOI] [PubMed]
  40. Moriarty C. 2003. The yellow eel. In: Aida K, Tsukamoto K, Yamauchi K (eds) Eel Biol. Springer, Tokyo, pp. 89–105.
  41. Moyer GR, Díaz-Ferguson E, Hill JE, Shea C. 2014. Assessing environmental DNA detection in controlled lentic systems. PLoS ONE 9(7):e103767. doi:10.1371/journal.pone.0103767. [DOI] [PMC free article] [PubMed]
  42. Nguyen AT, Tsukamoto K, Lokman PM. 2018. Composition and distribution of freshwater eels Anguilla spp. in Vietnam. Fish Sci 84:987–994. doi:10.1007/s12562-018-1239-9.
  43. Okamura A, Yamada Y, Yokouchi K, Horie N, Mikawa N, Utoh T, Tanaka S, Tsukamoto K. 2007. A silvering index for the Japanese eel Anguilla japonica. Environ Biol Fishes 80:77–89.
  44. Pilliod DS, Goldberg CS, Arkle RS, Waits LP. 2013. Estimating occupancy and abundance of stream amphibians using environmental DNA from filtered water samples. Can J Fish Aquat Sci 70:1123–1130. doi:10.1139/cjfas-2013-0047.
  45. Pilliod DS, Goldberg CS, Arkle RS, Waits LP. 2014. Factors influencing detection of eDNA from a stream-dwelling amphibian. Mol Ecol Resour 14:109–116. doi:10.1111/1755-0998.12159. [DOI] [PubMed]
  46. Rees HC, Maddison BC, Middleditch DJ, Patmore JR, Gough KC. 2014. REVIEW: The detection of aquatic animal species using environmental DNA -a review of eDNA as a survey tool in ecology. J Appl Ecol 51:1450–1459. doi:10.1111/1365-2664.12306.
  47. Robinet T, Feunteun E, Keith P, Marquet G, Olivier JM, Réveillac E, Valade P. 2007. Eel community structure, fluvial recruitment of Anguilla marmorata and indication for a weak local production of spawners from rivers of Réunion and Mauritius islands. Environ Biol Fishes 78:93–105. doi:10.1007/s10641-006-9042-3.
  48. Sakata MK, Maki N, Sugiyama H, Minamoto T. 2017. Identifying a breeding habitat of a critically endangered fish, Acheilognathus typus, in a natural river in Japan. Sci Nat 104:100. doi:10.1007/s00114-017-1521-1. [DOI] [PubMed]
  49. Santas AJ, Persaud T, Wolfe BA, Bauman JM. 2013. Noninvasive method for a statewide survey of eastern Hellbenders Cryptobranchus alleganiensis using environmental DNA. Int J Zool 2013:174056. doi:10.1155/2013/174056.
  50. Shiao JC, Iizuka Y, Chang CW, Tzeng WN. 2003. Disparities in habitat use and migratory behavior between tropical eel Anguilla marmorata and temperate eel A. japonica in four Taiwanese rivers. Mar Ecol Prog Ser 261:233–242. doi:10.3354/meps261233.
  51. Smogor RA, Angermeier PL, Gaylord CK. 1995. Distribution and abundance of American eels in Virginia streams: tests of null models across spatial scales. Trans Am Fish Soc 124:789–803.
  52. Takeuchi A, Iijima T, Kakuzen W, Watanabe S, Yamada Y, Okamura A, Horie N, Mikawa N, Miller MJ, Kojima T, Tsukamoto K. 2019a. Release of eDNA by different life history stages and during spawning activities of laboratory-reared Japanese eels for interpretation of oceanic survey data. Sci Rep 9:6074. doi:10.1038/s41598-019-42641-9. [DOI] [PMC free article] [PubMed]
  53. Takeuchi A, Watanabe S, Yamamoto S, Miller MJ, Fukuba T, Miwa T, Okino T, Minamoto T, Tsukamoto K. 2019b. First use of oceanic environmental DNA to study the spawning ecology of the Japanese eel Anguilla japonica. Mar Ecol Prog Ser 609:187– 196.
  54. Tsukamoto K, Chow S, Otake T, Kurogi H, Mochioka N, Miller MJ, Aoyama J, Kimura S, Watanabe S, Yoshinaga T, Shinoda A, Kuroki M, Oya M, Watanabe T, Hata K, Ijiri S, Kazeto Y, Nomura K, Tanaka H. 2011. Oceanic spawning ecology of freshwater eels in the western North Pacific. Nat Commun 2:179. doi:10.1038/ncomms1174. [DOI] [PMC free article] [PubMed]
  55. Wakiya R, Itakura H, Kaifu K. 2019. Age, growth, and sex ratios of the giant mottled eel, Anguilla marmorata, in freshwater habitats near its northern geographic limit: a comparison to tropical regions. Zool Stud 58:34. doi:10.6620/ZS.2019.58-34. [DOI] [PMC free article] [PubMed]
  56. Wakiya R, Kaifu K, Mochioka N. 2016. Growth conditions after recruitment determine residence-emigration tactics of female Japanese eels Anguilla japonica. Fish Sci 82:729–736. doi:10.1007/s12562-016-1006-8.
  57. Watanabe S, Aoyama J, Tsukamoto K. 2004. Reexamination of Eges (1939) Use of Taxonomic Characters of the Genus Anguilla. Bull Mar Sci 74:337–351.
  58. Watanabe S, Minegishi Y, Yoshinaga T, Aoyama J, Tsukamoto K. 2005. A Quick Method for Species Identification of Japanese Eel (Anguilla japonica) Using Real-Time PCR: An Onboard Application for Use During Sampling Surveys. Mar Biotechnol 6:566–574. doi:10.1007/s10126-004-1000-5. [DOI] [PubMed]
  59. Wilcox TM, McKelvey KS, Young MK, Sepulveda AJ, Shepard BB, Jane SF, Whiteley AR, Lowe WH, Schwartz MK. 2016. Understanding environmental DNA detection probabilities: a case study using a stream-dwelling char (Salvelinus fontinalis). Biol Conserv 194:209–216. doi:10.1016/j.biocon.2015.12.023.
  60. Yamamoto S, Minami K, Fukaya K, Takahashi K, Sawada H, Murakami H, Tsuji S, Hashizume H, Kubonaga S, Horiuchi T, Hongo M, Nishida J, Okugawa Y, Fujiwara A, Fukuda M, Hidaka S, Suzuki KW, Miya M, Araki H, Yamanaka H, Maruyama A, Miyashita K, Masuda R, Minamoto T, Kondoh M. 2016. Environmental DNA as a ‘Snapshot’ of Fish Distribution: A Case Study of Japanese Jack Mackerel in Maizuru Bay , Sea of Japan. PLoS ONE 11(3):e0149786. doi:10.1371/journal. pone.0149786.
  61. Yamanaka H, Minamoto T. 2016. The use of environmental DNA of fishes as an efficient method of determining habitat connectivity. Ecol Indic 62:147–153. doi:10.1016/j.ecolind.2015.11.022.
  62. Yamanaka H, Minamoto T, Matsuura J, Sakurai S, Tsuji S, Motozawa H, Hongo M, Sogo Y, Kakimi N, Teramura I, Sugita M, Baba M, Kondo A. 2017. A simple method for preserving environmental DNA in water samples at ambient temperature by addition of cationic surfactant. Limnology 18:233–241. doi:10.1007/s10201-016-0508-5.
  63. Yokouchi K, Aoyama J, Oka HP, Tsukamoto K. 2008. Variation in the demographic characteristics of yellow-phase Japanese eels in different habitats of the Hamana Lake system, Japan. Ecol Freshw Fish 17:639–652. doi:10.1111/j.1600-0633.2008.00315.

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Fig. S1.

Water velocity plotted against distance from the river mouth in the Oganeku River in the Amami-Oshima Island, Japan. The observation values on (a) August and (b) November 2016. The red line indicates the mean values of water velocity at each section that were separated by changepoint method based on likelihood ratio test. (download)

zoolstud-59-017-s001.docx (65.1KB, docx)
Fig. S2.

Relationships between Akaike’s information criterion (AIC) or R2 values and distances in lower and upper reaches that abundance of A. marmorata was calculated in the Oganeku River in the Amami-Oshima Island, Japan. AIC and R2 values were calculated by a generalised linear model (GLM) and Type II regression model. (download)

zoolstud-59-017-s002.docx (228KB, docx)

Articles from Zoological Studies are provided here courtesy of Biodiversity Research Center, Academia Sinica

RESOURCES