Population Structure of Atlantic Salmon in Maine with Reference to Populations from Atlantic Canada

—Anadromous Atlantic salmon Salmo salar from 12 rivers in Maine, 3 rivers in New Brunswick, and 2 rivers each in Nova Scotia, Quebec, Newfoundland, and Labrador as well as 2 landlocked strains in Maine (N 5 3,863) were genotyped at 11 microsatellite loci. Fish in the drainages of Maine’s Kennebec and Penobscot rivers were genetically similar to those sampled from the 8 rivers recently listed as containing an endangered distinct population segment under the United States’ Endangered Species Act. Genetic distance estimates confirm that Maine’s Atlantic salmon, both landlocked and anadromous, represent a discrete population unit, genetically as independent from any Canadian population as the Canadian populations are from each other. Within Maine, the anadromous and landlocked populations were statistically distinct. Anadromous Atlantic salmon were more genetically similar among year-classes within rivers than among rivers, as would be expected if the river is the unit of population. The effective number of breeders estimated within each river is larger than the number of adults estimated from samples and redd counts over the 10-year period from 1991 to 2000. The abundant Atlantic salmon Salmo salar in North American rivers represented an important * Corresponding author: [email protected] Received February 14, 2002; accepted September 2, 2002 subsistence and commercial fishery resource for both native North Americans and European settlers through the beginning of the industrial revolution. The Atlantic salmon was nearly extirpated from New England in the 19th century, when habitat loss resulting from dam construction and logging 197 POPULATION STRUCTURE OF MAINE ATLANTIC SALMON TABLE 1.—Adult Atlantic salmon returns to distinct population segment rivers (either 95% confidence interval from a model based on redd and adult counts or count from a trap) as estimated by the United States Atlantic Salmon Assessment Committee’s technical advisory committee (USASAC 2001). Year Sheepscot River Ducktrap River Narraguagus River Pleasant River Machias River East Machias River Dennys River 1991 1992 1993 1994 1995 19–54 17–48 19–55 3–8 23–66 11–32 12–35 18–51 10–29 95 73 94 51 56 20–58 11–31 13–37 7–19 38–110 30–87 20–58 22–63 12–36 3–8 11–31 12–34 30–85 25–73 14–40 10–29 21–62 1996 1997 1998 1999 2000 Harmonic mean 9–25 7–19 4–12 12–36 10–29 7.6–21.4 20–58 3–8 7–21 15–44 3–8 7.7–21.4 64 37 22 33 23 42.7 19–55 2–5 7–21 3 5.8–11.0 34–99 24–70 28–80 21–59 13–38 21.3–61.9 19–55 8–24 28–80 13–39 8–22 8.7–24.2 16–45 17–50 16–47 13–38 2 9.5–14.5 rendered former levels of harvest unsustainable (Beland and Bielak 2002). By the 1890s, 70–90% of the United States’ commercial catch of Atlantic salmon was restricted to Maine’s Penobscot River (USASAC 2001). Nonetheless, anadromous fish continued to spawn in several Maine rivers, in abundances too low to support commercial harvest (Baum 1997). Beginning in the 1870s, the Penobscot River was the primary focus of efforts to maintain and improve salmon production in Maine. Through 1947, Penobscot River production was primarily supplemented with progeny from broodstock of Penobscot origin (Baum 1997, summarized in Spidle et al. 2001). From the 1940s through the 1960s, returns to the Penobscot River were extremely low (on the order of tens of animals per year) despite the increase in stocking efforts. During the period from 1948 to 1967, hatchery releases originated primarily from broodstock taken from the Miramichi River in New Brunswick. From 1968 to 1971, most hatchery releases originated from the Machias and Narraguagus rivers in Maine (Baum 1997; Spidle et al. 2001), and adult returns to the Penobscot River rebounded. These returns resumed their former role as the primary broodstock for the Penobscot River restoration program in 1974 (Baum 1997; Spidle et al. 2001). From 1972 to 1992, fish of Penobscot River origin were used to supplement the limited wild returns in most other Maine rivers (Baum 1997). These last stocking efforts, consisting entirely of broodstock collected in Maine, would be expected to have made the greatest contribution to the current population of Atlantic salmon in Maine. Prompted by the diminishing spawning runs and low juvenile densities indicated by adult surveys and redd counts (Table 1), U.S. resource managers have recently designated the Atlantic salmon in eight Maine rivers as an endangered distinct population segment (DPS; included are the Sheepscot, Ducktrap, Narraguagus, Pleasant, Machias, East Machias, and Dennys rivers and Cove Brook of the Penobscot River drainage) under the U.S. Endangered Species Act (ESA). At the time of listing, the Kennebec River and its tributaries were excluded because of a consensus that its native salmon had been extirpated owing to habitat loss and that the river had been repopulated by strays of unknown origin. The main-stem Penobscot River was also excluded because it was known to have received substantial supplemental stocking from Canadian rivers (see Baum 1997 and Spidle et al. 2001), even though the extent to which such stocking had contributed to the populations in the river was not clearly understood. River-specific broodstocks were established for six of the eight populations within the Maine DPS designation as early as 1992 (the Sheepscot, Narraguagus, Pleasant, Machias, East Machias, and Dennys rivers; the Penobscot River population of Atlantic salmon has also been maintained as a river-specific strain since the 1970s but was not included in the DPS). It is hoped that the separate river-specific strains will maintain overall genetic diversity and, ideally, river-specific traits in the face of low effective population size. Rare alleles are expected to be lost in some populations and fixed in others, resulting in the overall preservation of genetic diversity. In contrast, under panmixia, alleles would be fixed or lost throughout the entire population, although the expectation of loss and fixation would be less. Because all of the riverspecific strains have remained small, there has been concern that genetic drift and demographic stochasticity may overwhelm natural selection, po-