Warning This Web page has been archived on the Web.

Archived Content

Information identified as archived on the Web is for reference, research or recordkeeping purposes. It has not been altered or updated after the date of archiving. Web pages that are archived on the Web are not subject to the Government of Canada Web Standards, as per the Policy on Communications and Federal Identity.

Skip booklet index and go to page content

COSEWIC assessment and status report on the Atlantic Salmon (Inner Bay of Fundy populations) in Canada

Species Information

Name and Classification
ClassOsteichthyes / Actinopterygii
Latin binomialSalmo salar L.
Designatable UnitInner Bay of Fundy Populations
Common species names:  
 English – Atlantic salmon
 French – Saumon atlantique
 Local Aboriginal – polam (Maliseet, Francis and Leavitt 2004); plamu (Mi'kmaq, National Recovery Team 2002). 
 Other common names exist for various forms and life history stages of the species (e.g., see Froese and Pauly 2004)


Morphological Description

To date, no overt morphological distinctions have been described for the Inner Bay of Fundy (iBoF) Atlantic salmon.  As a species, the Atlantic salmon (Figure 1) is a medium-sized trout-like fish.  Salmon mature can mature either in fresh water or in the ocean.  Throughout the species range, females typically grow to maturity in the ocean and return to fresh water to reproduce (i.e., they are anadromous).  Although males are often anadromous as well, the incidence of maturation in fresh water, in the absence of a seaward migration, exceeds 50% in many Canadian populations (Myers et al. 1986).  Features that distinguish Atlantic salmon from other salmonids include the maxillary ending below the centre of the eye, 2-3 large spots on the gill cover, and 11 dorsal fin rays (Scott and Crossman 1998). Ocean colour is silvery on the sides, darker on the top, and lighter underneath; spawning salmon in fresh water are a bronzed dark brown, typically with reddish spots on head and body.  Males have a pronounced hook (kype) on the lower jaw.  After spawning, the sexes darken and re-enter the ocean.  Juveniles generally have 8-11 parr marks on their sides with a single red spot between each parr mark along the lateral line.

Figure 1.  Line drawing of Atlantic salmon, Salmo salar

Figure 1.  Line drawing of Atlantic salmon, Salmo salar.From Amiro (2003).


Genetic Description

Atlantic salmon are naturally structured into genetically differentiated populations (Stahl 1987, Verspoor 1997, Nielsen 1998, National Research Council 2002).  This results from adult homing to natal rivers, juvenile rearing within rivers, and the spatial isolation of river systems.  These barriers to gene flow promote reproductive isolation and divergence through natural selection and drift.  Genetic divergence among populations is generally hierarchical, with groups of populations that are more similar nested within a gradient of difference that generally increases with geographic scale.  Groups often differ in life history and other phenotypic traits, such as egg size, body size, and adult age, reflecting adaptations to their local environment (e.g., Taylor 1991, Hutchings and Jones 1998).

Atlantic salmon originated in Europe and colonized North America prior to the last glaciation. North American populations are thought to have been pushed into refugia, from which they later recolonized the continent during glacial retreat (Bernatchez and Wilson 1998).  Verspoor (2005) provides an extensive analysis of current genetic differentiation across Canadian populations (Figure 2).  Based primarily on variation at allozyme loci, he places these populations into six genetically distinct regional groups: (1) Labrador/Ungava; (2) Gulf of Saint Lawrence; (3) Newfoundland (excluding Gulf rivers); (4) Atlantic Shore/Southern Uplands of Nova Scotia; (5) the outer Bay of Fundy (oBoF); and (6) the inner Bay of Fundy (iBoF).  The average genetic divergence is strong: FST values indicate that approximately 25% of all variation is associated with differences among the regional groups.  Figure 2 illustrates the oBoF rivers as overlapping the Gulf of St. Lawrence group, but additional mtDNA (Verspoor et al. 2002, Verspoor et al. 2004) and microsatellite data (McConnell et al. 1997) support their distinction as a separate group (Verspoor 2004).  Populations in Maine likely represent an additional extant North American regional group (King et al. 2001, National Research Council 2002, Spidle et al. 2003).


Designatable Unit: InnerBayof Fundy (iBoF) Populations

The iBoF encompasses 48 rivers, of which 32-40 may historically have supported Atlantic salmon (National Recovery Team 2002) (Figure 3).  The DU includes two major regions: Chignecto Bay (CB) in the west, and Minas Basin (MB) in the east. The best information on iBoF salmon comes from two index rivers, the Big Salmon River (CB) and the Stewiacke River (MB). Chignecto Bay is adjacent to the oBoF, while the Minas Basin is the more insular region of the DU.

Figure 2.Allozyme variation in Canadian Atlantic salmon populations.

Figure 2.  Allozyme variation in Canadian Atlantic salmon populations.

(a)  Map showing locations of 53 rivers in eastern Canadathat were included in a recent multilocus allozyme study (Verspoor 2005).  (b)  Multidimensional scaling (MDS) plot of allozyme variation for 48 of these 53 rivers (indicated by number).  Four regions cluster with relatively little overlapping: Labrador/Ungava, Newfoundland, Gulf of Saint Lawrence and Nova Scotia/Bay of Fundy.  Exclusion of the Hammond River (50, a tributary of the Saint John River in the outer Bay of Fundy) further resolves the Nova Scotia/Bay of Fundy region into three distinct clusters: the inner Bay of Fundy (48-49), the outer Bay of Fundy (50-53), and the Atlantic Shore/Southern Uplands of Nova Scotia (45-47), as also supported by neighbour joining (NJ) analysis (not shown).  Slightly modified from Verspoor (2005).

Figure 3.  Map showing the locations of 48 inner Bay of Fundy(iBoF) rivers. Recreational fishery data suggest that at least 32 rivers (identified by asterisks) supported self-sustaining Atlantic salmon populations.

Figure 3.  Map showing the locations of 48 inner Bay of Fundy(iBoF) rivers. Recreational fishery data suggest that at least 32 rivers (identified by asterisks) supported self-sustaining Atlantic salmon populations. However, at least an additional 8 rivers likely contained salmon. Map from Gibson et al. (2003a).

Below we review the evidence that iBoF Atlantic salmon populations are distinct from neighbouring regional groupings, especially the oBoF and Scotian Coast, by examining what is known about their genetics, phylogeographic history, local selection, life history, behaviour, demography, and stocking history.


Measures of genetic divergence between iBoF salmon and other regions are available from allelic enzymes (allozymes), microsatellite DNA, and mitochondrial DNA (mtDNA).  Each marker has its own strengths and limitations for resolving population structure (see for example Frankham et al. 2002, Avise 2004).  We therefore provide an interpretation based on the combined results.

Allozymes.  Recent analyses of 53 salmon populations across 24 allozyme loci suggest that salmon in the Minas Basin region of the iBoF, as represented by the Stewiacke and Gaspereau rivers, are genetically distinct from other Canadian salmon (Verspoor 2005).  However, the genetic data and statistical analyses are not unequivocal.  Multidimensional scaling analysis (Figure 2) sorts Stewiacke and Gaspereau salmon into their own regional group only after exclusion of the Hammond River (a tributary to the Saint John River of the oBoF, which appears to be a distinct population).  Neighbour joining statistics group the Stewiacke and Gaspereau together, but give less obvious regional clustering.  Furthermore, the majority of iBoF rivers have yet to be assessed with allozymes, and the rivers sampled to date may not reflect the situation throughout the iBoF, especially since the Chignecto Bay region was not included. Thus, the results from allozyme studies lend qualified support to the DU status of the iBoF, but do not alone justify its designation.

Microsatellites. Microsatellite studies consistently show that iBoF rivers contain salmon populations with a high degree of reproductive isolation and genetic structure. However, no single study provides compelling evidence that the populations within the iBoF, as a whole, form a distinct group.

1.     McConnell et al. (1995): The first microsatellite study, using 3 loci, reveals strong division between the Stewiacke (iBoF:MB; n = 45) population and two Scotian Coast populations, the Salmon (n = 20) and Gold Rivers (n = 35) (Figure 4).

Figure 4.  UPGMA dendrogram of Nei’s unbiased genetic distance based on microsatellite variation at 3 loci for five Atlantic salmon populations.


Figure 4.  UPGMA dendrogram of Nei’s unbiased genetic distance based on microsatellite variation at 3 loci for five Atlantic salmon populations. Redrawn from McConnell et al. (1995).

2.     McConnell et al. (1997):  A larger microsatellite study using 8 loci for 16 populations (507 individuals overall), including 3 from the Minas Basin of the iBoF (St. Croix Wells/Hants (n = 34), Stewiacke (n = 34), and Gaspereau (n = 30) rivers).  Multidimensional scaling and a UPGMA phenogram of Roger’s distance distinguish the three Minas Basin populations from a group of 10 Scotian Coast populations, two Gulf of St. Lawrence populations and one Newfoundland population.  However, neighbour joining networks under a variety of methods cluster populations without any clear regional pattern. 

3.     Jones (2001): This PhD thesis used fewer loci (5) but populations from both Chignecto Bay (Point Wolfe (n = 673), Upper Salmon (n = 459), Big Salmon (= 450), Petitcodiac (n = 56)) and Minas Basin (Stewiacke (n = 226), Gaspereau (n = 78)), as well as one oBoF (Hammond, n = 113) and one Gulf of St Lawrence population (Margaree, n = 168).  It suggests that, rather than distinct regional clusters, salmon within each river have some degree of reproductive isolation correlated with distance. 

4.     King et al. (2001):  An extensive microsatellite study using 12 loci and 27 Atlantic salmon populations (1,682 individuals overall) from across their global distribution, including the Stewiacke (iBoF:MB; n = 56) and the Saint John (oBoF; n = 66) finds the Stewiacke to be clearly distinct from all other North American populations in both multidimensional scaling and a neighbour joining network of genetic distance.  Genetic distances correlated with geographic distances, as in Jones (2001), but the Stewiacke is a unique exception.  It is strongly differentiated from salmon in the two nearest regions: the Saint John River of the oBoF, and the Gold River of the Scotian Coast.

5.     Spidle et al. (2003): A microsatellite study with 11 loci and 23 populations (3,863 individuals overall) includes the Stewiacke (iBoF:MB; n = 56), the Saint John (oBoF; n = 66), the St. Croix (oBoF, NB and Maine border; n = 63), and the Gold (Scotian Coast; n = 54) rivers.  Both multidimensional scaling and neighbour joining statistics distinguish the three regions (MB, oBoF, Scotian Coast) from each other.  Furthermore, using maximum likelihood statistics, individual fish are assignable to their river of origin: Stewiacke 77% (4th highest assignment success of the 23 populations), St Croix 73%, Saint John 37%, Gold 80% (2nd highest of 23).  The mean assignment success for the 23 rivers is 56% (random assignment success is 4%).

6.     O’Reilly (In Preparation): This multilocus microsatellite study with samples from 24 rivers throughout the iBoF, oBoF and the Scotian Coast, is still in progress.  A UPGMA dendrogram of pairwise FST differences distinguishes two major groupings: Bay of Fundy rivers (MB: Stewiacke (n = 822), Gaspereau (n = 50), Economy (n = 53), Great Village (n = 47); CB: Big Salmon (n = 797), Upper Salmon (n = 54), Point Wolfe (n = 103); oBoF:  Saint John (n = 152)) and eight Scotian Coast rivers.  It does not distinguish between the seven iBoF rivers and the Saint John River of the oBoF.  The data suggest that the Big Salmon River (iBoF; CB) is more related to the Saint John River (oBoF) than to any other iBoF river. 

Mitochondrial DNA (mtDNA). Recent studies of variation at mtDNA loci provide strong evidence of phylogenetic distinctiveness of salmon within the iBoF.

1.     Verspoor et al. (2002):  Analysis of sequence variation in two regions of the mitochondrial ND1 gene, encompassing 710 base pairs, reveals a unique haplotype variant (detectable via RFLP analysis using Alu I) that is present at moderate frequencies (mean 0.35, range 0.17-0.75) in six of nine iBoF rivers, but absent from both the Saint John River (oBoF) and the Narraguagus River in Maine.  The distribution of this variant within the iBoF is strongly geographically structured, occurring in each of six Minas Basin sites while absent from three rivers surveyed in Chignecto Bay (Big Salmon, Irish, Black). 

2.     Verspoor et al. (2004): This study expands that undertaken by Verspoor et al. (2002) to 94 rivers, including 20 iBoF rivers and 74 non-iBoF rivers, across the eastern coastline of North America from Maine to Ungava Bay. The ND1 variant referred to above was absent from the 74 non-iBoF rivers, but detected at moderate frequencies in 11 of 20 iBoF rivers, including 10 of 11 Minas Basin rivers and one of nine Chignecto Bay rivers (Gardner Creek) (Figure 5).  Furthermore, across all 35 haplotypes studied, cluster analyses group several rivers from Chignecto Bay and the Minas Basin.  Although the Minas Basin was most differentiated from the oBoF, the haplotype frequencies also show highly significant regional differentiation between Chignecto Bay (iBoF) and the oBoF. 

3.     O’Reilly (In Preparation): This study uses single nucleotide polymorphism (SNP) analyses at the same region of the ND1 gene as Verspoor et al. (2002, 2004).  It reveals the presence of the unique haplotype variant in three additional Chignecto Bay rivers (Upper Salmon, Point Wolfe, Big Salmon).  This brings the total number of rivers with the variant to four of nine rivers sampled in Chignecto Bay.  The frequency in Chignecto Bay rivers is lower than in Minas Basin, but the variant is clearly present in the iBoF while absent outside the iBoF. 

A caveat in these mtDNA studies is that mtDNA is matrilineally inherited, and thus reflects the phylogeny through females but not through males.  The microsatellite analyses of O’Reilly (In Preparation) suggest gene flow between the Big Salmon (iBoF: CB) and Saint John (oBoF) rivers.  If males are more likely to disperse or stray than females, the iBoF and oBoF regions may not be as strongly differentiated in their nuclear DNA as in their mtDNA.

Nevertheless, based on these mtDNA studies, three distinctive evolutionary lineages appear to exist within the Bay of Fundy: the Minas Basin, Chignecto Bay, and the outer Bay.  The distinctiveness of the Chignecto Bay region is not as clear, but it seems partially shared with the Minas Basin while different from that of the oBoF. 

Figure 5.  North American distribution and frequencies of two regionally unique mitochondrial DNA haplotypes

Figure 5.  North American distribution and frequencies of two regionally unique mitochondrial DNA haplotypes.

One is restricted to the iBoF (black wedges) and one is restricted to the Scotian Shelf (grey wedges).  Numbers indicate sample sizes.  From Verspoor et al. (2004).  Asterices indicate 3 Chignecto Bay rivers where the iBoF variant has since been found at low frequencies (O’Reilly, In Preparation).

Phylogeographic History

Another approach to understanding the distinctiveness of iBoF salmon is the reconstruction of their phylogeographic history (Bernatchez and Wilson 1998, Avise 2000).  The Bay of Fundy region was ice-covered during the last glacial maximum of 18,000 years BP (Pielou 1991).  Several glacial refugia, now submerged in the ocean, are believed to have existed on eastern islands and shelves near the Bay of Fundy.  Additional refugia are believed to have been in rivers south of what is now Long Island.  Detailed reconstruction suggests that parts of the iBoF were the first sections of the coast to be free of glacial ice (Pielou 1991, Shaw et al. 2002) and thus the first sites available to salmon colonization.  Verspoor et al. (2002) suggest that salmon colonization began in the Chignecto Bay of the iBoF, and that established Chignecto populations then colonized the Minas Basin as the ice receded.  Furthermore, for thousands of years afterward, the Minas Basin is thought to have been a single watershed, which could explain the close genetic relationship among its current populations.  IBoF salmon may therefore have a different history of colonization than oBoF salmon.

Local Selection

The iBoF ecosystem is physically and hydrologically unique because of its extreme tides (Greenberg 1987).  The tide levels increase as one moves into the iBoF: the average difference between high and low water in the Minas Basin is about 12 metres; the flow of water into Minas Basin is some 2,000 times the discharge of the St. Lawrence River (Thurston 1990). This extreme movement causes extensive erosion to the sand and clay shoreline and substrate of the iBoF, suspending silt in the water breathed by migrating and resident salmon. The river entries are dynamic: tidal bores surge up many of the rivers and on retreat leave shallow reaches.  These extreme conditions will almost certainly generate selection pressures that differ from those experienced by Atlantic salmon elsewhere, and may contribute to genetic isolation by excluding salmon that are not adapted.

Life History, Behavioural and Demographic Characteristics

If iBoF salmon are a distinctive group from those of the oBoF and Scotian Coast, we might expect to find differences among these groups in life histories, behaviour and demography, but few studies have been conducted on these characteristics. The strongest life history data concern age and sex ratio differences at maturity in the Big Salmon (CB) and Stewiacke (MB) rivers of the iBoF, and the Saint John River of the oBoF.  A greater proportion of Big Salmon and Stewiacke adults mature after one sea-winter (SW), and a greater proportion of these are females.  In the iBoF, the mean proportion of 1SW returns is about 93%, and about 72% of these fish are female (Stewiacke = 94.4% 1SW, 72.2% female, n = 298; Big Salmon = 91.8% 1SW, 70.8% female, n = 3,334;) (Amiro and McNeill 1986, Amiro 2003). In the oBoF, about 55% are 1SW and 14% of these are female (Hutchings and Jones 1998).  IBoF populations also tend to exhibit more repeat spawning (Amiro 2003).

Migration behaviour of iBoF salmon may also differ from that of oBoF and other Maritime Canada salmon, as iBoF smolts have been hypothesized to grow to adulthood within local waters, including the Bay of Fundy and northern Gulf of Maine (Jessop 1976, Ritter 1989, Amiro 2003), whereas other Canadian populations typically migrate north to the oceans off Labrador, Newfoundland and Greenland (Hutchings and Jones 1998).  Of more than 30,000 tagged, hatchery-grown smolt from the Big Salmon and Stewiacke rivers released in the iBoF between 1985 and 1990, zero were recaptured in the Newfoundland and Greenland fisheries (Amiro 2003), but nor were any recaptured returning to rivers.  In addition, the earlier maturity of iBoF salmon and the greater amount of repeat spawning are consistent with a hypothesis of local migration.  However, in at least one iBoF population (the Gaspereau River), some salmon have been recovered in Newfoundland and Greenland, and there might therefore be different migratory patterns among rivers within the iBoF.  Migration patterns might also differ within rivers: a recent tracking study of Big Salmon River post-smolts revealed that some (approximately 40%) moved rapidly out of the Inner Bay and did not return, whereas the majority (approximately 60%) initially left the Inner Bay but later displayed resident behaviour (i.e., were detected within the Inner Bay) for at least two months (Lacroix et al. 2005).  Residency near coastal habitat is not necessarily anomalous or unique to Bay of Fundy salmon populations: post-smolts have been detected near shore in late summer in the northern Gulf of St. Lawrence and in the Baltic Sea (Lacroix et al. 2005).  A recent unpublished meta-analysis of recapture data by Dadswell (2004) found that migration patterns of Big Salmon River (iBoF) and Saint John River (oBoF) releases were not statistically different.  The author points out that a lack of historical fisheries within the Bay of Fundy or Gulf of Maine, a lack of tag returns from these areas between October and early May (despite commercial fishery operation for other species using gear that would catch salmon), and similarity in coastal migration patterns for iBoF and oBoF salmon all argue against the hypothesis of local migration.  Dadswell (2004) proposes that the distance for departure from the Bay (up to 300 km), strong tidal currents, possible retention in the mid-Bay gyre, and temporal fishing patterns may underlie the seeming anomaly in iBoF tag recaptures.  There is therefore a lack of consensus regarding migration patterns of iBoF salmon.

There is some indication that iBoF populations are demographically disassociated from at least the Scotian Coast populations (Amiro 2003); however, detailed demographic analyses are needed.  Thus, there is some support for the iBoF DU hypothesis in the limited life history, behavioural and demographic information.

Stocking History

There has been extensive stocking of iBoF rivers by management agencies and public groups: over 40 million salmon have been released since 1900 (Gibson et al. 2003b).   Most of these stocked fish are from outside the iBoF, which raises the concern that introgression of non-local genes may have destroyed any local population structure and adaptations of iBoF populations.  Roughly 80% of these releases were fry, and occurred before the 1960s.  Stocking then changed to parr and later to smolt.  A similar history of extensive fry stocking characterizes the Maritimes in general, and also Maine, where over 96 million salmon have been stocked since 1870 (Baum 1997).  Despite this prolific stocking, however, there is little evidence that it has contributed to significant loss of genetic structure in either the Maritimes or Maine.  Elson (1957, cited in Gibson et al. 2003b) concluded that fry stocking had negligible influence on adult recruitment in the Maritimes.  Baum (1997) concluded from experimental fry and parr marking in the 1940s and 1950s that adult returns ranged from “insignificant to nil”.  Microsatellite studies by King et al. (2001) and O’Reilly (In Preparation) reveal very high genetic diversity both between and within river populations despite the stocking.  A recent in-depth assessment of stocking impacts by the National Research Council (2002) concluded that wild Atlantic salmon populations can retain their genetic uniqueness even under heavy stocking.  Thus, while stocking issues are a legitimate concern, the integrity of the iBoF DU may have remained largely intact, or at least not seriously compromised.

Summary of DU Designation

The cumulative evidence from genetics, phylogeography, local selection, life history, behaviour and demography suggests that there are several extant lineages of salmon in the Bay of Fundy.  There is strong mtDNA evidence that the iBoF is unique from any other region, although microsatellite DNA suggests that there may be some gene flow from the oBoF.  It should be noted that the current “border” between the iBoF and oBoF regions recognizes the different biological characteristics between two major index rivers (Big Salmon in Chignecto Bay of the ‘inner’ Bay and Saint John in the ‘outer’ Bay), but there is no obvious physical basis underlying the border.  Phylogeographic reconstruction provides insight into the historical potential for the biological divergence within the Bay of Fundy.  Local selection and spatial isolation through the extreme hydrodynamics and geography of the Bay could also contribute to the evolution of local life history, behavioural and demographic differences.  In summary, there is compelling yet qualified evidence that the iBoF region constitutes a COSEWIC DU for Atlantic salmon.