First Record of the Large-Scale Loach Paramisgurnus dabryanus (Cobitidae) in the United States

The introduction of exotic (i.e., nonnative) fish species is a topic of concern throughout North America, where approximately 46% of the described native freshwater fish taxa are considered imperiled (Jelks et al. 2008). The decline of native fish species is largely attributed to habitat degradation coupled with the introduction of exotic fish species (Richter et al. 1997; Jelks et al. 2008). In general, exotic species can negatively affect native fish through multiple mechanisms including habitat alteration, competition, predation, hybridization, and pathogen transfer (Douglas et al. 1994; Mooney and Cleland 2001; Dunham et al. 2004). The impacts of exotic fish on a fish assemblage is often mediated at local and landscape scales by the suitability of habitat, magnitude and extent of introduction, structure of the existing assemblage, and the implementation of management or control actions (Courtenay and Robins 1989; Moyle and Light 1996; Marchetti et al. 2004). Although the United States has a multitude of regulations to prevent the introduction and establishment of exotic fishes (Lodge et al. 2006), more than 138 exotic species have been introduced and become established in the United States through development of gamefish or baitfish stocks, disposal of aquarium fish, release of baitfish, or escape from aquaculture facilities (Nico and Fuller 1999; Pimentel et al. 2000; Fuller 2003).

One group of fishes with great invasive potential is the family Cobitidae, which includes approximately 177 freshwater loach species from 26 genera (Nelson 2006; Kottelat 2012). The most widely introduced species, the Oriental Weatherfish Misgurnus anguillicaudatus, has become established in several areas in the United States including Southern California, Florida, Hawaii, Idaho, Illinois, Michigan, New York, Oregon, and Washington (Courtenay et al. 1987; Tabor et al. 2001; Simon et al. 2006) and has been largely introduced by aquaculture escapes or releases by aquarists or fisherman (Courtenay and Stauffer 1990; Chang et al. 2009). The potential negative impacts of invasive loaches such as the Oriental Weatherfish on native fishes include predation of eggs or larvae, increased competition for macroinvertebrate or algal prey, introduction of foreign parasites, and the elevation of ammonia, nitrate, and turbidity levels within the environment (Logan et al. 1996; Keller and Lake 2007; Lintermans et al. 2007).

On October 6, 2014, members of the U.S. Fish and Wildlife Service (USFWS) collected a previously unknown exotic loach in a disconnected pool during a fish assemblage survey on the San Joaquin River immediately upstream from the Chowchilla Bifurcation Structure in Madera County, California. They sacrificed, measured, photographed, and discarded the specimen. Following Moyle (2002) and Nelson (2006), a member of the USFWS later identified the specimen as an Oriental Weatherfish using the morphological characteristics visible in the photographs (e.g., veliform body shape, subterminal mouth, five barbels present on each side of the jaw; C. Castle, personal communication). The detection of Oriental Weatherfish in the San Joaquin River would constitute a range expansion of the species within California. However, misidentification often occurs among cobitids when relying solely on external morphological characteristics (Bohlen et al. 2005). Further, Oriental Weatherfish are known to hybridize with Large-Scale Loach Paramisgurnus dabryanus and the hybrids exhibit morphological similarities to that of the Oriental Weatherfish (You et al. 2009). As a result, the true identity of the exotic loach was unknown. The objectives of this study were to collect additional loach specimens near the Chowchilla Bifurcation Structure in the San Joaquin River and validate the species identification using meristics and genetics.

The San Joaquin River basin has a Mediterranean–montane climate (i.e., wet–cool winters and dry–hot summers; Null and Viers 2013) where natural flows are largely generated during the spring from snowmelt runoff. In general, the San Joaquin River downstream of Millerton Lake (river kilometer [rkm] 430.5 of the San Joaquin River measuring from its confluence with the Sacramento River) is subjected to artificial flow regimes, agricultural and municipal contaminants, excessive groundwater pumping, channel confinement, and surface water diversion and export (Galloway and Riley 1999; Traum et al. 2014). As a result, the San Joaquin River is currently dominated by losing or effluent reaches where flows become intermittent and habitat becomes disconnected between Gravelly Ford (rkm 370.0) and the Merced River confluence (rkm 187.6) during nonflood conditions with the exception of where irrigation water from the federal Central Valley Project is conveyed. We sampled for the exotic loach in all disconnected pools occurring within 1 km of the Chowchilla Bifurcation Structure (rkm 347.8; Figure 1).

On November 12, 2014, USFWS and California Department of Fish and Wildlife staff members sampled four disconnected pools. Before sampling the pools, we collected dominant substrate and water-quality data along the shoreline of all pools but one. We used a YSI 85 or YSI PRO 2030 meter to measure water temperature to the nearest 0.1°C, specific conductance to the nearest 0.1 μS/cm, and dissolved oxygen to the nearest 0.1 mg/L. We measured turbidity using a HACH 2100Q turbidity meter to the nearest 0.1 nephelometric turbidity unit. We determined dominant substrate type visually and broadly classified it on the basis of particle size diameter as mud (< 0.5 mm) or sand (0.5–5.0 mm). We also broadly classified pool size as either small or large. Small pools were, on average, shallow (i.e., < 2 m in depth) and < 15 m wide or long. Conversely, large pools were, on average, deep (i.e., at least 2 m in depth) and ≥ 15 m wide or long.

We sampled fish in each disconnected pool during daylight hours using two Smith-Root LR 24 pulsed direct-current backpack electrofishers operating between 0.1 and 0.2 amperes. We generally sampled fishes using single-pass electrofishing throughout the wadeable portions of each pool starting at one end of the disconnected pool and moving to the opposite end of the pool. In large pools, we only sampled the shoreline habitats. We recorded the total shock time for each disconnected pool. We also sampled fish in two of the pools using a 15.0 × 1.2 m beach seine with 3-mm² Delta mesh in combination with the backpack electrofishers. In these pools, we deployed the beach seine at one end of each disconnected pool following behind the electrofishers to collect stunned fish that we could not detect visually. In large pools, we used a block net with 3-mm² Delta mesh to partition the pool into smaller units to allow for more effective sampling given the gear and methods used.

We identified all fish in the field that were ≥ 25 mm total length to species following Moyle (2002). All the loach-like fishes collected, we measured for total length, euthanized, and preserved them in a 95% ethanol solution for genetic and meristic analyses. We sent preserved specimens to the University of California–Davis for deoxyribonucleic acid (DNA) barcoding and to the Natural History Museum of Los Angeles County (LACM) Section of Ichthyology for radiograph-based identification.

We extracted DNA from dried barbel clips from the loach specimens using a PureGene DNA extraction kit. We used the Barcode of Life (BoL) Data System (BOLD; 2017) to identify appropriate universal primers to amplify and sequence the cytochrome c oxidase I (COI) gene in teleost species. We used polymerase chain reaction (PCR) primers VF2_t1 and FR1d_t1 and sequencing primers M13F and M13R for DNA barcoding (Table 1; Messing 1983; Ivanova et al. 2007; Geiger et al. 2014). We performed the PCR amplifications using the thermal profile described in Geiger et al. (2014). We outsourced Sanger sequencing to QuintaraBio in Richmond, California. We trimmed poor-quality ends from the COI sequences in the program Sequencher, leaving high-quality sequences ranging in length from 662 to 689 base pairs. We made preliminary species identification on the basis of sequence similarity between the unknown samples and species-specific sequences in the BoL and National Center for Biotechnology Information (NCBI) nucleotide database (U.S. National Library of Medicine, 2017.

Next, we downloaded COI sequences from BoL and NCBI for the closest genetic matches to the unknown specimens, the Large-Scale Loach (N = 5) and Fine-Scale Loach Misgurnus mizolepsis (N = 3), along with two commonly cultured loach species, the Oriental Weatherfish (N = 16) and European Weatherfish Misgurnus fossilis (N = 13). We trimmed the COI sequences from all 41 individuals to 607 base pairs and aligned them using the Clustal W approach with default settings in the program MEGA7 (Kumar et al. 2016). We created five separate groups, one for each species and one for the unknowns, and calculated the between-group mean pairwise distances in MEGA7 using all nucleotide substitutions (transitions and transversions, coding and noncoding), assuming uniform evolution among lineages and sites, and using a gap treatment of complete deletion. We performed a thousand bootstrap replicates to estimate variance.

A subset of preserved specimens was catalogued, photographed, and radiographed at the LACM (Sabaj 2016). We counted meristics on each specimen including dorsal, anal, pectoral, and pelvic fin rays, vertebrae, and barbels. We also assessed fin morphology, the presence of spots on the body or peduncle, and the presence and type of adipose crests on the caudal peduncle. Thereafter, we compared each specimen with Oriental Weatherfish and identified following Kottelat (2001, 2012), Kottelat and Freyhof (2007), and Froese and Pauly (2017).

The disconnected pools sampled (n = 4) represented a narrow range of habitat conditions. In general, all of the pools were all dominated by mud substrate and contained nonflowing (0 m/s), cool (16.2–18.5°C), well-oxygenated (8.4–9.8 mg/L), and clear (7.3 nephelometric turbidity units) water with little conductance (39.3–47.7 μS/cm). On average, we sampled each disconnected pool using the backpack electrofishers for 35.2 min (range = 21.9–49.5). We detected a total of 417 fish during the study, which represented a total of 14 known fish species (Table 2). One Sacramento Sucker Catostomus occidentalis represented the only native fish detected.

We captured a total of six loach specimens (93 to 131 mm total length) in two of the four disconnected pools in the study area (Table 2). We collected five loach specimens from the pool immediately upstream of the Chowchilla Bifurcation Structure where the original loach was captured on October 6, 2014. We also collected one loach specimen from the disconnected pool approximately 1 km upstream of the Chowchilla Bifurcation Structure.

All unknown loaches exhibited an identical haplotype at 607 base pairs of COI sequence. Although we anticipated an identification of Oriental Weatherfish, our preliminary analysis of sequence similarity found that the closest matches were the Large-Scale Loach (BoL and NCBI; 99% similarity) and Fine-Scale Loach (NCBI; 99% similarity), a taxon likely conspecific to Large-Scale Loach (see Discussion). Our analysis of pairwise distance between groups revealed that the unknown specimens were most genetically similar to Large-Scale Loach (Table 3). Interestingly, the unknowns and Large-Scale Loach samples were nearly equally divergent from the Fine-Scale Loach (0.062 vs. 0.063; Table 3). Greater sequence divergence was observed in comparisons involving Oriental Weatherfish and European Weatherfish (Table 3).

We selected one specimen to send to the LACM for meristic identification. We retained the other specimens, which were in poor physical condition, for additional genetic analysis. The specimen was identified as Large-Scale Loach (catalog number LACM 58237-1). The radiograph revealed the meristic 8 dorsal rays, 7 anal rays, 11–12 pectoral rays, about 8 pelvic rays, and 46 vertebrae (Figure 2). Physical characteristics included a pair of rostral, maxillary, and mandibular barbels, with two pairs of barbel-like mental lobes on the lower lip. Scales covered the body of the specimen. The tail was not forked, somewhat pointed, and the pectoral fin was longer than the head, with the second ray thickened. A small pouch was present on the inside of each pectoral fin but did not contain lamina circularis. The caudal peduncle had adipose crests that extended to the dorsal and anal fins, and there were spots on the side of the body but not a distinct spot on the upper caudal base. These characters are consistent with Large-Scale Loach or Fine-Scale Loach (Kottelat 2001, 2012; Froese and Pauly 2017). However, the specimen was identified as a Large-Scale Loach following Kottelat and Freyhof (2007) on the basis of the presence of four barbel-like mental lobes on the lower lip, the absence of a midlateral stripe from the eye to the caudal base, the absence of a narrow stripe from the opercle to the pelvic origin, and the presence of high adipose crests on the caudal peduncle. An Oriental Weatherfish (catalog number LACM 36986-2) was radiographed for comparative purposes and was found to have only 10 pectoral rays and 45 vertebrae, a pectoral fin that was shorter than the head (although this may be due to it being a female), a rounded tail, and a dense black spot on the upper caudal base.

We captured a total of six loach specimens in the San Joaquin River near the Chowchilla Bifurcation Structure. Using genetic techniques, we identified all as Large-Scale Loach, and the LACM meristic analysis of one specimen corroborated the genetic results. Although our initial queries of BoL and NCBI revealed strong genetic similarity between our specimens and both Large-Scale and Fine-Scale Loach, the subsequent pairwise distance analysis supported identification of the unknowns as Large-Scale Loach. Genetic similarity between Large-Scale and Fine-Scale Loach was not unexpected given the findings of phylogenetic studies of the family Cobitidae. In two recent studies using both nuclear (RAG-1) and mitochondrial (cytochrome b) gene sequence data, Large-Scale and Fine-Scale Loach resolved into a single lineage with low genetic divergence within it, supporting the hypothesis posed by several authors that the two taxa are very genetically similar (Perdices et al. 2016) or conspecific (Vasil'eva 2001; Kottelat 2012; Perdices et al. 2012). Morphological similarity of some Fine-Scale Loach specimens to Oriental Weatherfish have led to recommendations that those taxa are conspecific (Kottelat 2012), but this is not supported by genetic data (Perdices et al. 2012, 2016; this study). Our results are the first to confirm the occurrence of Large-Scale Loach in the United States after a review of the literature and the U.S. Geological Survey Nonindigenous Aquatic Species Database (U.S. Geological Survey 2017). The specimens are currently retained by the University of California–Davis (catalog numbers WFB 3240a through WFB 3240d), California Academy of Sciences (catalog number CAS 243916), and LACM (catalog number LACM 58237-1) as vouchers for future study.

The native distribution of Large-Scale Loach includes the Yangtze River basin in China and the inland waters of Taiwan (Kottelat 2012) where water temperatures can range from 10°C to 30°C (Wang and Li 2005; Zhang et al. 2012). They are found in lentic and lotic habitats containing muddy substrate, low gradient, slack water, and high turbidity (Gao et al. 2010; Sato et al. 2011; Huang et al. 2013). Additionally, Large-Scale Loach can occur in irrigation or diversion canals, ponds, and rice fields (Kanou et al. 2007; Greshishchev et al. 2015); they are able to breathe air in oxygen-depleted waters and burrow into mud to avoid desiccation or thermal extremes (Wang and Li 2005; Zhang et al. 2012). These habitat preferences and characteristics make the Large-Scale Loach suited to the San Joaquin River with its labyrinth of agricultural diversion canals and managed waterfowl habitat, muddy substrate, and thermal extremes (Galloway and Riley 1999; Traum et al. 2014).

We speculate that the introduction to the San Joaquin River may have been the result of one or more aquarium releases within 1 km of the Chowchilla Bifurcation Structure or farther upstream when the river was connected. The Large-Scale Loach is considered to have nutritional, medicinal, and ornamental value outside of the species' native distribution (Freyhof and Korte 2005; Wang and Li 2005; Kanou et al. 2007) and the species is commonly exported from eastern Asia for the aquarium or aquaculture trade (Freyhof and Korte 2005; Kanou et al. 2007; Chu et al. 2012). It has been introduced to Europe and Japan by means of aquaculture escape and aquarium release (Kanou et al. 2007; Tang et al. 2008). In the United States, cobitids are commonly sold as aquarium fish in retail stores (Courtenay and Stauffer 1990; Rixon et al. 2005), including those within the San Francisco Estuary watershed (Chang et al. 2009), providing a source of introduction.

The ecological impact of the Large-Scale Loach population in the San Joaquin River is unknown. The species has high reproductive potential and can become sexually mature at 1 or 2 y of age when fish range in length from 100 to 200 mm (Chu et al. 2012; Zhang et al. 2012). The Large-Scale Loach is an omnivorous benthic feeder and its prey consists of zooplankton, macroinvertebrates, and algae (Wang and Li 2005; Kanou et al. 2007). High population densities of Large-Scale Loach could potentially alter the invertebrate and algal community within invaded waters. Further, Large-Scale Loach could potentially interfere with other benthic feeders (Mills et al. 2004; Hazelton and Grossman 2009), although interference competition has not been documented between Large-Scale Loach and other species possessing diet overlap within Japan (Kanou et al. 2007). Additionally, the introduction of Large-Scale Loach has the potential to introduce foreign fish parasites. Introductions of Oriental Weatherfish have resulted in the transport and introductions of pathogens including tapeworms and nematodes (Sohn et al. 1993; Ernst and Dove 1998; Lintermans et al. 2007), which could affect both fish and humans.

For all these reasons, we believe that the establishment of Large-Scale Loach in the San Joaquin River could influence the aquatic ecosystems and the success of large-scale river restoration projects being implemented in the basin including the San Joaquin River Restoration Program. Therefore, we recommend further research to assess their level of establishment in the San Francisco Estuary watershed. Additional field sampling is needed throughout the San Joaquin River downstream of Friant Dam including the San Francisco Estuary, adjacent diversion canals, and managed waterfowl habitat to document the distribution and abundance of the species. This information can be used to facilitate an assessment of impact, guide further research, and inform natural resource management decisions concerning the implementation of appropriate management actions (Courtenay and Robins 1989; Lodge et al. 2006). Although the ecological impacts of introduced Large-Scale Loach are not well understood, bans on the import of Large-Scale Loach and eradication strategies may be warranted to help prevent establishment and further introductions within the United States.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Data S1. All fish capture and habitat data collected from the San Joaquin River, California near the Chowchilla Bifurcation Structure on November 12, 2014 are contained in the xls file titled Data_S1. (37 KB).

Found at DOI: http://dx.doi.org/10.3996//012017-JFWM-008.S1 (90 KB CSV).

Reference S1. Galloway DL, Riley FS. 1999. San Joaquin Valley, California—largest human alteration of the Earth's surface. Pages 23–34 in Galloway DL, Jones DR, Ingebritsen SE, editors. Land Subsidence in the United States. U.S. Geological Survey Circular 1182. Found at DOI: http://dx.doi.org/10.3996//012017-JFWM-008.S2 (5,471 KB PDF); also available at https://pubs.usgs.gov/circ/1999/1182/report.pdf (5,340 KB PDF).

Reference S2. Traum JA, Phillips SP, Bennett GL, Zamora C, and Metzger LF. 2014. Documentation of a groundwater flow model (SJRRPGW) for the San Joaquin River Restoration Program Study Area, California. U.S. Geological Survey Scientific Investigations Report 2014–5148. Found at DOI: http://dx.doi.org/10.3996//012017-JFWM-008.S3 (16,985 KB PDF); also available at https://pubs.usgs.gov/sir/2014/5148/pdf/sir2014-5148.pdf (16,500 KB PDF).

This work was jointly supported by the U.S. Fish and Wildlife Service's Aquatic Invasive Species and Anadromous Fish Restoration programs and the California Department of Fish and Wildlife. We thank D. Barnard, K. Benn, C. Castle, N. Cullen, P. Ferguson, G. Giannetta, M. Grill, and L. McMartin for assistance in field sampling. The manuscript was improved with suggestions from the Associate Editor, H. Jelks, and one anonymous reviewer. We are also grateful to Dr J. Freyhof (German Centre for Integrative Biodiversity Research) for his assistance with fish identification and J. Fisher (National Conservation Training Center Library) for his assistance acquiring international literature.

Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.