Skip booklet index and go to page content

COSEWIC assessment and status report on the American Eel in Canada

Limiting factors and threats

Because of its extended lifespan, semelparous reproductive system, and long migrations, the American eel faces a long list of natural and anthropogenic mortality factors.


Natural Limiting Factors

Global climate change is thought to be generating a northward deviation of the Gulf Stream system (Castonguay et al. 1994b; Knights 2003), and to be reducing oceanic productivity (Dekker 1998). Weakening currents could interfere with leptocephalus transport and survival either by starvation or by unfavourable transport patterns that extend the duration of oceanic migration (Knights 2003), both leading to reduced recruitment.

Continental-phase eels may die from natural causes, and they may also metamorphose into silver eels and leave for the spawning grounds. These two factors are often difficult to distinguish. The finite instantaneous instream mortality rate (M) for elvers entering the East River (Chester) has been evaluated at 0.0612, based on trap count data (Jessop 2000b). This rate is higher than those reported for European elvers (0.0107 and 0.0233; Berg and Jørgensen 1994 cited in Jessop 2000a) and may be a direct result of the toxic effects of low pH of the river (Jessop 2000a).

Annual disappearance rate estimates include both natural mortality and emigration and are based on the assumption that recruitment is stable through time. On the Sud-Ouest River, Verreault (2002) estimated an annual disappearance rate of 26.4% (instantaneous rate = 0.307) for emigrating eels aged 9 to 17. A population model indicates that 27% of eels that enter Lake Ontario survive to reach the open Gulf of St. Lawrence as pre-spawning silver eels. Abundance of escaping eels declined by 23% annually between 2000 and 2004 (J.M. Casselman, pers. obs.). Estimates show good agreement, given that recruitment during the period has virtually ceased. In the Hudson River, the disappearance rate was estimated at 15% (instantaneous rate = 0.16 ± 0.06), and no significant difference was found between freshwater and brackish water sites (Morrison and Secor 2003). A stochastic life table model estimated a disappearance rate of 22.9% (instantaneous rate = 0.26) per year in an unexploited eel component in Prince Edward Island (ICES 2001).


Anthropogenic Threats

Habitat Modifications and Dams

Dams and other barriers generate habitat loss and fragmentation for upstream migrants, and produce turbine mortality for downstream migrants.

There are two hydro complexes on the main stem of the St. Lawrence River below Lake Ontario.  The Moses-Saunders dam was completed in 1959.  The Beauharnois complex was begun in the late 1920 and completed in 1961 (Verdon and Desrochers 2003, R. Verdon, Hydro-Québec, pers. comm.). Shipping locks at these two dams offered upstream passage opportunities, but upstream passage has been provided by permanent eel ladders at the Moses-Saunders dam since 1974 and at the Beauharnois dam by a passage facility that operated in 1994-1995 and from 1998 to the present. Ladders have been available to give access to Lake Champlain via the Richelieu River at Chambly since 1997 and at St-Ours since 2001. The Ottawa River is blocked by 12 hydro-dams, none of which is equipped with an eel ladder.

There are few dams on rivers draining into the southern Gulf of St. Lawrence, except on Prince Edward Island, where about 800 low head dams are erected, and where most streams are blocked at one to several places (MacFarlane 1999). Nevertheless, since eels of all stages have the ability to ascend salmonid fish ladders, and small eels are able to ascend low dams that lack any passage facilities (Cairns et al. 2004; Lamson et al. submitted), eels are generally common in the ponds above these dams. There are numerous hydro-dams in the Scotia-Fundy region, including the Mactaquac dam which blocks the upper part of the Saint John River in New Brunswick. The only monitored eel ladder in the Maritime Provinces is at the Morgan Falls hydro-facility on the LaHave River on the south shore of Nova Scotia. The ladder has operated since 2002 (R. Bradford, pers. comm.).

Verreault et al. (2004) examined obstacles to eel recruitment in the St. Lawrence River drainage area in Quebec and Ontario and the United States. In the St. Lawrence River watershed, over 8,000 dams (at least 2.5 m high) prevent, restrict or delay access to more than 12,000 km² (Table 6) of freshwater habitat (10 m or less deep; Verreault et al. 2004). Hydro-dams (n = 151) were present on many major tributaries in the upper St. Lawrence except on the Richelieu River (FEA1). Based on data analysis from three tributaries in the St. Lawrence River watershed, restricted access could have reduced potential annual escapement substantially, by more than 800,000 eels (Table 6; Verreault et al. 2004), primarily highly fecund females (Casselman 2003; Verreault et al. 2003; Tremblay 2004). Historically, the Ottawa River habitat would have potentially contributed 255,000 female silver eels annually (Table 6; Verreault et al. 2004).  

To enhance potential annual production and escapement, these habitats should be made more accessible and downstream migrants should be protected from turbine mortality. The extended potential recruitment to these habitats, however, is unknown and actual carrying capacity is unquantified.


Table 6.  Surface areas of freshwater growth habitat that is upstream of restrictive dams in the St. Lawrence watershed, and estimated potential annual escapement (modified from Verreault et al. 2004).
Site (subwatershed)Estimated growth
habitat above dams (km2)
Potential annual
unutilized production
Upper St. Lawrence River - Lake Ontario5,800399,700 A
Ottawa River3,700255,000
RichelieuRiver - Lake Champlain1,20082,700 B

A: Access re-opened in 1974 at Moses-Saunders Dam and in 2002 at Beauharnois Dam; B: Access re-opened between 1997 (Chambly Dam) and 2001 (St-Ours Dam)

Silver eels descending rivers equipped with hydro-dams may be killed or injured as they pass through turbines. Turbine mortality is positively correlated to eel length and inversely proportional to blade spacing. It also varies with turbine type (Francis, Kaplan and propeller), turbine size, and operating conditions such as flow and generating efficiency (Montén 1985; Larinier and Dartiguelongue 1989; Travade and Larinier 1992). Since eels of the upper St. Lawrence River – Lake Ontario system (FEA1) are the longest in North America (Verreault et al. 2003), they are at greatest risk of turbine mortality. Mortality rate of emigrating eels with mean length of 88 cm has been estimated at 16% for a Francis turbine and at 24% for a propeller turbine at the Beauharnois dam (Desrochers 1995). Eels with a mean length of 102 cm passing a propeller turbine in the Moses-Saunders dam suffered an estimated mortality of 26.4% (Normandeau Associates and Skalski 2000). Migrants leaving Lake Ontario encounter both of these dams during their spawning migration, and are subjected to an accumulated turbine mortality of 40% (Verreault and Dumont 2003). This additive turbine mortality at Moses-Saunders dam and Beauharnois dam contributes to almost 75% of the anthropogenic mortality during downstream migration and reduces the annual spawning escapement by 40%. This analysis refers to the St. Lawrence system above the Beauharnois dam only.  Turbine mortality figures should be regarded as a minima because undetected sublethal injuries could further reduce the number of females that reach and successfully spawn in the Sargasso Sea (Couillard et al. 1997).

In the southern Gulf of St. Lawrence few dams are used for water power, and turbine mortality is not a significant issue. Numerous rivers in Scotia-Fundy region, Newfoundland, and Labrador are dammed but no data are available on number of dams or amount of habitat loss.


All eel fisheries target pre-spawners (Richkus and Whalen 1999). All continental stages are subject to commercial exploitation in Canada, but the stages that are exploited vary geographically, and much of the Canadian range is unfished. Fisheries for elvers and silver eels occur during narrow time windows. The yellow eel stage may last many years, so fisheries that target this stage may produce high cumulative mortality even if fishing mortality rate per year is low.

Fishery landings have limited value as indicators of abundance. Landings may be influenced by regulations, price per kg, alternative opportunities in fishing and other employment, and by changing gear efficiencies.

Total annual harvest data from commercial fisheries in Canada are presented in Figure 18 and detailed harvest for Quebec fisheries by sector is presented in Figure 11. Canadian commercial catches decreased substantially in the 1990s despite an increase in price per kg (Casselman 2003). The unreported eel catch is believed less than 5% in Lake Ontario and 8% in the St. Lawrence estuarine fishery (ICES 2001). In the Scotia-Fundy region, reported landings are closely correlated with effort and are not considered to reflect abundance (R. Bradford, pers. comm.).

Fishing mortality rate is poorly known for yellow and silver American eels. Instantaneous fishing mortality on mostly yellow eels in exploited waters of Prince Edward Island was estimated at 0.5 per year (ICES 2001). The great majority of eels taken in this fishery are yellow, and are exposed to fishing mortality over several years. Assuming no density dependence in survival rates, the model estimated that fishing in exploited Prince Edward Island waters reduced spawner escapement by 90% below what it would have been in the absence of fishing. In the St. Lawrence estuary silver eel fishery, mark-recapture experiments yielded estimates of exploitation rates of 19% in 1996 and 24% in 1997 (Caron et al. 2003).

Eel fishing effort is unevenly distributed within the Canadian range of the American eel. In some regions, there are intensive fisheries while in other regions eels are unexploited. The stage targeted by fisheries (glass eel, elver, yellow eel, silver eel) also varies geographically. In Ontario, the major yellow eel fishery in Lake Ontario and the upper St. Lawrence River was closed in 2004. In Quebec, there are major fisheries in the upper St. Lawrence River and estuary targeting mainly silver eels (> 75%; Caron et al., submitted). Eels originating in FEA2 are not exploited since Quebec fisheries target eels from FEA1. In the southern Gulf of St. Lawrence (FEA3), commercial fisheries target yellow eels in tidal waters. Yellow eels are fished extensively in coastal waters and estuaries of the Gulf New Brunswick and Prince Edward Island. There is little eel fishing effort in the Gulf of Nova Scotia, and none in most fresh waters of the southern Gulf of St. Lawrence. Winter recreational spear fisheries also contribute to anthropogenic mortality of yellow eels in the Southern Gulf of St. Lawrence. In the Scotia-Fundy region, eel fishing occurs in both fresh and marine waters, but many rivers and coastal habitats are unfished. The only documented elver fishery in Canada occurs in Scotia-Fundy. In Newfoundland (FEA4) and Labrador (FEA5), yellow and silver eels are fished principally in rivers, but there are many rivers which are not exploited. Landings for Labrador have been reported only in 1985 (4.3 tonnes) and in 1993 (0.1 tonne), and it is unknown whether this irregular pattern is related to abundance; however, landings are not large. Although seven exploratory elver licences were issued in 2004 in Newfoundland, fishing and effort data are not available.

Yellow and silver eel catches in FEA1 (Figure 18) have steadily declined since the early 1980s under relatively constant fishing effort before 1996. As a result of dramatic resource declines, commercial eel fisheries were closed in the RichelieuRiver in 1998 and in Lake Ontario and in Ontario waters of the upper St. Lawrence River in 2004. Between 1950 and 1999, catches from the estuarine fishery in the lower St. Lawrence River (FEA1 origin) were almost fivefold greater and more consistent than catches in the commercial harvest in Ontario (Casselman 2003).

Lake St. Francis is located between Moses-Saunders dam and Beauharnois dam. Catches in Lake St. Francis are now a significant part of total harvest in FEA1 (Figure 11). Harvest (in Quebec waters only) rose from 5.2 tonnes in 1988 to 29.0 tonnes in 2004, in contrast to the general abundance trend in FEA1. This fishery was opened in 1986 and relies on a consistent fishing effort. Favourable access infrastructure developed at the Beauharnois dam since 1994 by means of a trap and by a permanent ladder since 2002, could have had contributed to a rising trend in total catch (P. Dumont, MNRF, Secteur Faune Québec, pers. comm.).

Reported commercial catches in the southern Gulf of St. Lawrence (FEA3) were low from the beginning of the time series in 1917 to the 1960s, when catches increased following the introduction of new fishing methods and the development of markets (Figure 18). Therefore, catch trends prior to 1970 are probably unrelated to abundance. After a peak around 1970, reported catches dipped, peaked again in the late 1980s, dipped again, and have gradually increased since the late 1990s. Heavy exploitation of eels over the minimum size (46.0 cm up to 1997, 50.8 cm in 1998 and beyond) could have reduced the size and age distribution (Cairns et al. 2004).

The Canadian elver fishery targets arriving glass eels and elvers as they ascend estuaries. Jessop (2000b) estimated that elver fishers took 31 to 52% of arriving elvers in the East River, Chester, Nova Scotia. Some elver fishing sites (including East River, Chester) are located at the mouths of rivers that are impacted by acid rain. These rivers, or substantial portions of these rivers, may be unsuitable for eel growth and survival. Hence, the elvers that are removed by the fishery do not necessarily represent a loss to the reproductive capacity of the species, because many of them would probably die due to low pH if they had not been caught (Jessop 2000a).

There has been a trend towards increasingly restrictive fishing regulations in the last several decades, especially in the Atlantic Provinces, and especially since 2000 (e.g. Cairns et al. submitted).  Changes include shortening of seasons, increases of minimum size, caps on the number of gears that can be deployed, and freezes on the development of any new fisheries.

Chemical and Biological Contamination

Contaminants may impact eels by reducing survival and impairing reproduction (Castonguay et al. 1994a; Hodson et al. 1994; Couillard et al. 1997). In polluted waters, eels are heavy bioaccumulators since they are long-lived benthic species with a high fat content that accumulates lipophilic contaminants such as PCBs (polychlorinated biphenyls), pesticides (DDT), dioxins and furans. High summer mortalities of silver eels in the freshwater part of the St. Lawrence in the early 1970s were attributed to acute toxicity from environmental contaminant levels (Dutil 1984).

Couillard et al. (1997) used the relationship between tissue mirex concentration and body mass to identify the origin of migrating silver eels taken in the St. Lawrence estuary. Those with high loadings were presumed to have come from the upper St. Lawrence River - Lake Ontario area. The authors observed a relationship between chemical contamination and pathological lesions, and suspected a relation between organochlorine contamination and oocyte diameter that may lead to reproductive failure at a later stage of maturation (Couillard et al., unpublished data, cited in Castonguay et al. 1994a). PCBs have deleterious effects on eel fertility by impairing egg quality and embryonic development. Since migrating females are fasting (Pankhurst and Sorensen 1984), contaminants recirculate into the blood system, and chemical levels in the eggs could be even higher at hatching, increasing the likelihood of toxicity to the larvae (Hodson et al. 1994; Robinet and Feunteun 2002). Above 0.2 pg/TEQ/g[1], production of vital offspring is affected (Anonymous 2005). Van Ginneken et al. (2005) observed lower oxygen consumption rates in swimming and resting eels loaded with PCBs in comparison to a control group. Results corroborate the general depressing effect of PCBs on protein synthesis (G. Thillart, Leiden University, pers. comm.).

Contaminant levels in Lake Ontario have decreased significantly from 1970s’ levels (Luckey et al., in press), and there is little evidence to suggest that human-related contaminants [PCBs, DDT, mirex, dieldrin (insecticide), dioxins, furans, mercury] are currently impacting natural reproduction and health of Lake Ontario benthos, plankton or fish on a lakewide basis. According to a monitoring program targeting coho salmon, total PCB concentrations have decreased threefold and mirex has reduced twofold since 1970 (Luckey et al., in press). Eels from the St. Lawrence estuary tributaries have been reported to be less contaminated with mirex than those from Lake Ontario (Hodson et al. 1994). However, Renaud et al. (1995) found greater concentration of mirex during the 1990s than between 1947 and 1950 in polluted tributaries of the St. Lawrence River (St‑François and Sainte-Anne rivers). Therefore, deterioration of habitat quality could affect eel survival throughout its range depending on pollution level.

One of the most important factors influencing contaminant dynamics in Lake Ontario is the increasing proliferation of exotic nuisance species since they alter both fish community composition and food web energy flows (Luckey et al., in press). Thus, subsequent changes to pathways and fate of contaminants have resulted in altered bioaccumulation rates in portions of fish communities as evidenced by recent spikes in contaminant burdens. Alterations to the forage base of fish communities have resulted in diet shifts and in some cases, the consumption of a more contaminated prey, which produces elevated body burdens of contaminants (Luckey et al., in press).

Exotic zebra and quagga mussels have substantially altered water quality and trophic relations in Lake Ontario (Mills 2005).  However, the rapid proliferation of these mussels occurred in the early 1990s, well after the major decline in the Moses-Saunders had been completed.

Many rivers in the southern uplands area of Nova Scotia have low pH due to acid precipitation (Marcogliese and Cone 1996). Acidic conditions in these rivers may limit the survivorship of American eels (Jessop 2000a).

 Introduced Parasite: Anguillicola crassus

The swim bladder nematode parasite, Anguillicola crassus, spread into Europe from shipments of Japanese eels (Anguilla japonica) from Asia to aquaculture facilities in Germany in 1982 (KØie 1991, cited in Barse and Secor 1999). In North America, the parasite was originally discovered in a single eel captured in Winyah Bay (South Carolina) in 1995 (Fries et al. 1996). Since then, the parasite has been detected in eels in the Hudson River and Chesapeake Bay (Barse and Secor 1999; Morrison and Secor 2003). The following information on A. crassus in New England is provided by K. Oliveira, University of Massachusetts, pers. comm. Eels in the Paskamensett River, Massachusetts, showed infestation rates higher than 90%. All rivers in Rhode Island and Massachusetts that were examined had the parasite with some degree of variation in intensity. The parasite has also been documented in Maine waters. It appears that the parasite is at least past mid-coast of Maine and may be as far east as the East Machias River, which is about 40 km from the New Brunswick border.

A significant (p < 0.05) positive relationship between mean intensity of infection and eel size was found (Moser et al. 2001). Heavy infections can lead to hemorrhagic lesions, swim bladder fibrosis or collapse, skin ulceration, decreased appetite, and reduced swimming performance (Barse and Secor 1999). Van Ginneken et al. (2005) found that swim bladder parasites cause bladder shrinkage, leading to higher cost of swimming and reduced migration capacity.

Over the past five years, almost 1,200 eels have been examined in the upper St. Lawrence – Lake Ontario system and no swim bladder parasite was found (J.M. Casselman, OMNR, pers. obs.). Since 2002, many hundreds of silver eels in downstream migration in the lower St. Lawrence River have been sampled annually for swim bladder examination but no parasites have been detected (G. Verreault, MNRF, Secteur Faune Québec, pers. comm.). During the first attempt of elver stocking in the upper St. Lawrence River, from Bay of Fundy to the upper Richelieu River and Lake Champlain, none of the 128 elvers examined carried this parasite (Dumont et al. 2005). The subsample examined in the 2005 program was also free of A. crassus (P. Dumont, MNRF, Secteur Faune Québec, pers. comm.). Moreover, a Nova Scotian study found an absence of A. crassus in eels (Barker 1997).  Although A. crassus has not been detected in Canada, its advancement northward on the US coastline, and its current presence in Maine close to the Canadian border, suggests that the arrival of this parasite in Canada may be imminent. 


Stocking may represent a way to reduce the sharp decline and recruitment failure of the American eel in the upper St. Lawrence River. As a trial of this concept, 40,000 elvers (mean length 60.3 ± 3.0 mm) were captured in the Bay of Fundy (New Brunswick), marked with tetracycline, and released in Lake Morin (4 km²) on the south shore of the St. Lawrence estuary (Verreault et al., submitted). Elver stocking in Lake Champlain was proposed in 2003 by the Association des pêcheurs d’anguilles et de poissons d’eau douce du Québec (Quebec St. Lawrence Estuary commercial fishermen) in collaboration with MNRF, Secteur Faune Québec (Dumont et al. 2005). In May 2005, 600,000 elvers were translocated from Bay of Fundy to the upper Richelieu River (P. Dumont, MNRF, Secteur Faune Québec, pers. comm.). Deserted habitats not impeded with hydro-dams in the St. Lawrence watershed could be utilized as growth habitats for eels to increase the overall eel abundance (Verreault et al., submitted) in Canadian waters.

However, there is much uncertainty regarding the benefits of stocking to American eel conservation. The density of stocked eel is a consideration in stocking programs since high densities can lead to male-dominated sex ratios (Krueger and Oliveira 1999). A stocking density of 100 elvers per ha in Lake Morin (FEA2) resulted in 27.2% males after four years of growth (Verreault et al., submitted). Therefore, to maintain a high proportion of females within the St. Lawrence watershed stocking needs to be achieved on a low density basis. Another point of concern is whether silver eels originating from distant sources will find appropriate migration pathways to reach the Sargasso Sea and spawn successfully. European studies on this matter give conflicting results (Westin 1990; Moriarty and Dekker 1997; Dekker 2004). Limburg et al. (2003) reported that stocked eels in the Baltic sea showed evidence of maturity at the silver stage and migrated back to the Sargasso Sea. Nevertheless, homing to the spawning grounds may be based on the memory of migration during the early life stages, with magnetic cells in the jaws as a navigation tool (Feunteun 2002). Finally, it is not known if the survival and subsequent escapement of translocated elvers is greater than or less than their survival and subsequent escapement if the elvers had been left to grow in their natural habitat. The question is further complicated by the possibility that translocation might alter sex ratio and size at maturity, which will affect potential egg deposition per exiting silver eel.

[1]PCBs units: picogram/ toxicity equivalency factor/ gram