1 Introduction
The muskrat (Ondatra zibethicus ) is seen as a fixture of wetlands across North America, being one of the most common and widely distributed furbearer species on the continent. They are found primarily in marshes, but also in ponds, sloughs, lakes, ditches, rivers and streams from the east coast to the west, and from the Mackenzie Delta in Canada’s north to northern Mexico in the south (Boutin and Birkenholz, 1987). The muskrat also ranks as the most heavily harvested wild furbearer in North America of the 20th century (Obbard et al., 1987) and has contributed more than any other animal to the combined income of North America’s trappers over the past 150 years (OFMF, 2019). In the early 1900s, millions of muskrats were trapped and sold across North America and, though harvest numbers are lower today, muskrats remain a major source of income for fur trappers and are still among the most prevalent species trapped for fur (Fur Institute of Canada, 2019).
While the muskrat played a major role in the early fur trade and colonization of North America by Europeans (White et al., 2015), the species has been of cultural significance as a traditional clothing and food item and as a spiritual symbol among Indigenous people long before European explorers arrived on the continent. For example, in an Anishinaabe story of creation, the muskrat (Wa-zhushk) comes to the rescue to help re-build the Earth after a great flood and decimation of life, and is said to embody humility, courage, and determination (MacGregor, 2013).
As a wetland obligate and dominant herbivore, the muskrat plays several important roles in wetland ecosystems. Their foraging, travel, and house-building activities create numerous small openings in marsh vegetation, thereby increasing the interspersion of open water and emergent vegetation which often results in increased structural diversity and plant species richness in wetlands (Nyman et al., 1993; Connors et al., 2000; Keddy, 2010). Different biotic communities are known to respond positively to such enhancements in habitat diversity (Wilcox and Meeker, 1992); in particular, a greater density and diversity of marsh birds and waterfowl has been found in wetlands with an equal ratio of open water to emergent vegetation (Weller and Fredrickson, 1973; Kaminski and Prince, 1981; McDonnell, 1983). Muskrat houses (both new and old) can also create important loafing and nesting sites for marsh birds. For example, the black tern (Chlidonias niger ) will use muskrat houses and feeding platforms as nesting substrate (Hickey and Malecki, 1997) and it is quite common to find Canada geese (Branta canadensis ) and sometimes trumpeter swans (Cygnus buccinator ) loafing or nesting on old muskrat structures. Less recognized are the benefits that muskrats provide to snakes and turtles, many of which are species at risk. In an Illinois study, a large number of spotted turtle (Clemmys guttata ) captures occurred in deep open-water pools associated with muskrat lodges where muskrat grazing decreased the vegetative cover, and it was believed that these pools served as refugia for the turtles during periods of high temperature and/or drought (Litzgus and Brooks, 2000). Numerous vertebrate species, including more than ten herptile species, have been observed using muskrat houses, burrows, cleared pathways and other features for thermoregulation, nesting, cover and ease of travel (S. Gillingwater, unpublished data; Kiviat, 1978). Furthermore, muskrat activity can influence mussel abundance (Diggins and Stewart, 2000), invertebrate communities (De Szalay and Cassidy, 2001; Nummi et al., 2006), microbial activity (Wainscott et al., 1990) and nutrient cycling (Connors et al., 2000). Muskrats are also an important food item for many predators, such as red foxes (Vulpes vulpes ), coyotes (Canis latrans ), raccoons (Procyon lotor ), raptors and especially mink (Neovison vison ) (McDonnell, 1983).
Early research on muskrats by Paul Errington and others illustrated the density-dependent nature of muskrat populations and a variety of abundance cycles (Errington, 1951; Clark and Kroeker, 1993; Erb et al., 2000). In their analysis of almost one hundred individual time series of muskrat harvest data from the Hudson’s Bay Company in Canada, Erb et al. (2000) found that the mean period length of muskrat population cycles differed between ecozones, ranging from 3.7 to 8.6 years, with the shorter periods tending to occur at higher latitudes and in eastern regions. In some cases, however, the time series did not exhibit any periodicity. Evidence exists for both population-intrinsic factors (e.g., social factors) and extrinsic factors (e.g., disease, environmental variability and trophic interactions) explaining the observed patterns of these cycles (Errington, 1963 and Errington et al., 1963; Bulmer, 1974; Weller and Fredrickson, 1973). Despite some commonalities these patterns lack consistency across geographic areas, and in many cases muskrat population dynamics and their mechanisms of regulation remain unclear.
The muskrat is a prolific species; females typically have 2-3 litters per year and an average litter size of 6.5 kits (Boutin and Birkenholz, 1987). As well, most muskrats are able to breed the same year they are born and have high dispersal capabilities. These demographic characteristics make muskrats relatively resilient to harvest and other population pressures (e.g., disease, predation) as a small number of individuals can quickly multiply and enable population recovery. They are also reasonably flexible in their habitat requirements and, as noted by Errington (1951), muskrats often demonstrate a remarkable ability to stay alive up to the very edges of what may be considered habitable range.
Today, however, this marsh denizen may not be thriving like it once was. There is a growing body of literature suggesting declines in muskrat populations from myriad locations across North America over the past 10-20 years. Most recently Gregory et al. (2019) reported that muskrat harvest on Prince Edward Island declined by more than half when comparing the average harvest from 1977-1988 to 1988-2016. More broadly, Ahlers and Heske (2017) analyzed harvest data from 1970 to 2012 across the United States and, after controlling for pelt prices, found strong evidence that muskrat populations declined during this time period. Prior to that study, analyses by Roberts and Crimmins (2010) revealed a 75% decline in muskrat harvest across the northeastern United States and eastern Canada from 1986 to 2006. This decline was thought to be indicative of regional declines in muskrat abundance as the study’s authors found that a previously strong correlation between harvest and pelt prices had weakened in the latter years of their analysis. In other words, they felt it was likely that recent changes in muskrat harvests were reflecting underlying population change (and not simply changes in harvest effort) because they did not find the strong relationship between harvest levels and pelt prices which had previously defined the harvest dynamics of muskrats and other furbearer populations (Scognamillo and Chamberlain, 2006; Bailey, 1981). As well, in the same study the authors reported a lack of periodic fluctuations in the modern muskrat harvest data (as compared to the historic data which exhibited mild periodicity), providing further support for widespread population decline (Roberts and Crimmins, 2010).
In Ontario, similar to trends across Canada, provincial fur harvest records show a clear decline in muskrat harvest over the past 100 years, most notably over the past 30 years (Fig. 1). In fact, the average annual number of muskrats harvested in the past 30 years has declined by more than 90% from the mean in the previous 20-year period (late 1960s to late 1980s). While some of this decline can be explained by changes in fur harvest reporting structure that occurred in Ontario in the late 1980s, as well as economic factors and cultural shifts in trapping, the magnitude and time span of the decline seems too great to be explained by these factors alone. For example, the spatial pattern of change in muskrat harvests from 1972 to 2004 suggests an actual decline in muskrat population may have occurred because the pattern did not conform to the expectation arising from spatially homogeneous decline in fur price and therefore trapper effort (Gorman 2007). Furthermore, the time span of the recent period of low muskrat harvest numbers (30 years) far exceeds that of the maximum population cycle length reported for muskrats (Erb et al., 2000). As well, the low annual numbers harvested in recent years are well below the lower limit of published historic muskrat population fluctuations (Statistics Canada, 2011).
Since many furbearers (including muskrats) are difficult and costly to census (Erb and Perry, 2003), harvest data are often the only type of information related to species’ population trends that are available for a long time series or across a large geographic area (White et al., 2015). However, there are problems with relying on harvest data to infer wildlife population trends, and these have been noted by many of the same authors that have analyzed harvest trends of furbearers (e.g., Gregory, 2019; Roberts and Crimmins, 2010). For example, harvest data can encompass trapping seasons of different lengths, can be missing data for some years, or can be deficient due to lack of full reporting. More importantly, harvest data can be biased by trapping effort which can be affected by the number of trappers or by changing economic factors (e.g., pelt price) that may encourage or discourage trappers (Landholt and Genoways, 2000).
Although long-term census data on muskrat populations are generally rare, a few researchers have published evidence of declines in local muskrat populations from direct count surveys. For example, Benoit and Askins (1999) reported that counts of muskrat houses from marshes of the Quinnipiac and Connecticut Rivers decreased dramatically (78% and 100% respectively) between 1965 and 1990. As well, Ward and Gorelick (2018) analyzed muskrat house count records from 1970 to 2016 for the Peace-Athabasca Delta in Canada and found a significant decline in population density over this time frame.
Other recent studies, while not long-term in nature, have reported finding lower than expected muskrat densities in areas of seemingly suitable habitat. For example, Toner et al. (2010) reported a mean muskrat density of 0.71 houses/ha in six coastal wetlands along the Upper St. Lawrence River surveyed multiple times from 2001 to 2006, and Greenhorn et al. (2017) found a mean muskrat density of 0.27 houses/ha in 43 cattail-dominated marshes surveyed along the north shore of Lake Ontario in 2014. In 2009, Gregory et al. (2019) intensively searched four marshes on Prince Edward Island for muskrats and reported a mean density of just 0.07 houses/ha. All of these results are much lower than the typical densities of 2.1 – 3.6 muskrat houses/ha reported previously in the literature for cattail-dominated marshes in Canada (Proulx and Gilbert, 1984; Messier and Virgil, 1992).
Anecdotal and interview-based reports of muskrat decline over the past few decades have also been reported by trappers (Gregory et al., 2019), fur managers (Roberts and Crimmins, 2010) and Indigenous people across Canada (Brietzke, 2015; Straka et al., 2018; S. Mallany, personal communication, December 2016), and have also been received from a variety of sources by us. The underlying theme among these reports is that trappers and other long-term land users are simply not finding muskrats in the numbers they used to, despite efforts to do so.
Collectively these recent studies and reports point to potential widespread declines in muskrat populations. However, robust empirical data on long-term trends in muskrat populations based on field observations are generally scarce in the literature. Thus, more information is needed to confirm that declines in muskrat harvest correspond to real declines in muskrat abundance (Ahlers and Heske, 2017). Fortunately, we learned of annual muskrat field surveys having been conducted at multiple locations in Ontario, Canada between 1950 and 1990 and sought to exploit this untapped source of historic muskrat population data.
Our specific objectives in this study were to locate and revisit sites where historic muskrat survey data exists for Ontario, to replicate the historic survey methods as closely as possible, and to then compare contemporary survey results with the historic data to determine if there have been empirical changes in the muskrat populations in these areas. We hypothesized that declines observed in muskrat harvest are due at least in part to real declines in muskrat populations. Therefore, we expected to observe evidence of fewer muskrats in contemporary versus historic muskrat surveys.