Distribution of a boreal rodent linked to a lobe of the Wisconsinan glaciation.
Knowledge of the geographic range of a species is fundamental to understanding its ecology (MacArthur, 1972; Gaston, 2009; Heads, 2015), speciation and adaptation (Mayr, 1963; Geber, 2011), management (Soule, 1986), and responses to climate change (Parmesan, 2001; Walther et al., 2002; Post, 2013). Many distributions reflect the clear imprint of the late Wisconsinan glaciation (Smith, 1957), such as the relictual distributions of the southern bog lemming (Synaptomys cooperi) (Wetzel, 1955; Connor, 1959) and the persistence of the boreal red-backed vole (Myodes gapperi) on mountaintops in New Mexico (Bailey, 1932).
The rock vole (yellow-nosed vole, Microtus chrotorrhinus) has a broad geographic range (Fig. 1) from Labrador to Tennessee and west across Canada to northeastern Minnesota (Kirkland and Jannett, 1982; Lansing, 2005) and is most commonly associated with talus. It is generally considered to be a boreal species (e.g., Hoffmann and Koeppl, 1985), and its known localities in Tennessee, Virginia, and West Virginia are in cool damp higher elevational habitats of the Appalachian Mountains. The species is generally considered rare or uncommon in much of its range (Kirkland and Jannett, 1982) and until 1996 was accorded "Special Concern" status in Minnesota (Coffin and Pfannmuller, 1988). Its presence in Minnesota has been enigmatic. After the first specimen was captured near Burntside Lake, the exact site unknown, in northern St. Louis County in 1921 (Swanson et al, 1945; Handley, 1954), the species was not documented again until the 1970s, when specimens were taken at two localities in Cook County in extreme northeastern Minnesota (Timm, 1974; Buech et al., 1977). The paucity of distributional records continued despite extensive field work in Cook County in the 1970s (Timm, 1975) and unsuccessful attempts to document its presence at the original locality (Buech et al, 1977; Nordquist and Birney, 1988; J. Daniels, K. Rusterholtz, pers. comm.). Then, in the early 1980s specimens were collected at 59 sites in Cook Co. and one in Lake Co. by Christian and Daniels (1985), and in one other area in Cook Co. by Etnier (1989). The only additional sites (two) since then are the westernmost localities, found in northern St. Louis Co. in Voyageurs National Park (VNP) (Jannett et al., 2007). Long-term monitoring (1983-ongoing) of rock vole and other small mammal populations by one of us (FJJ) at some of these sites and at other new sites reported here has indicated relatively constant population numbers over multiple years (Jannett, 1990) but with occasional declines to extremely low numbers and complete "absences" of variable duration (FJJ, pers. obs.)
This pattern of variable presence and abundance has complicated understanding the distribution and range limits of the species in Minnesota. Here we present results to further define the western boundary of the range of Microtus chrotorrhinus, document greater range of habitats, and suggest that the distribution in Minnesota is concordant with the extent of the boulder- and rock-rich Rainy lobe of the Wisconsinan glaciation.
Museum Special break-back traps (Woodstream Corp., Lititz, Pennsylvania) baited with peanut butter, apple, and rolled oats, or with rolled oats alone, were used in all inventorying and monitoring. Additionally, other types of traps or bait were deployed (Table 1). Voucher specimens were deposited in the collections of the Biology Department of the University of Minnesota-Duluth, Bell Museum of Natural History at the University of Minnesota-Twin Cities, Museum of Texas Tech University (MoTTU), and Science Museum of Minnesota (for which the designated repository is MoTTU). Where there may not have been voucher specimens prepared (four localities), a list of those localities with rock voles is archived at MoTTU and in the Natural Heritage database of the Minnesota Department of Natural Resources. Localities for sites within VNP where no rock voles were caught were reported to Park staff. Field methods were approved by the University of Minnesota Institutional Animal Care and Use Committee (most recent approval 008A59761 to FJJ in 2001) and conform with the American Society of Mammalogists guidelines for use of mammals in research (Sikes et al., 2011).
Extended sampling periods with Museum Special traps for up to 19 d and with small nonfolding aluminum Sherman live-traps (H. B. Sherman, DeLand, FL) for up to 4 d have shown rock voles enter traps of both types mostly in the first few days (Jannett, pers. obs.); therefore, the two-day sampling in some of our efforts is likely sufficient to ascertain presence of rock voles. Trapping was conducted for 2 d if rock voles were to be monitored at a new site or if they were not secured in the first day at an inventory site; the effort may have been less if the site was merely to be inventoried and rock voles were taken before the end of 2 d. The following efforts were for 6 d unless abbreviated due to disturbance by carnivores: monitored (M) sites in VNP (Table 2), and sites inventoried for eastern heather vole (Phenacomys ungava) and northern bog lemming (Synaptomys borealis) (see below). Some sites were trapped over 5, 6, or up to 33 y with permanent trap stations to ascertain if populations persisted in certain habitats (e.g., eskers and clearcuts) and other inventory sites were trapped repeatedly in VNP (Table 2) despite catching no rock voles. We noted parous and first-pregnancy females which we summarize as "adult females" below in eskers and clearcuts monitored for several years.
Trapping in 1986 where rock voles had been present in 1982 (Christian and Daniels, 1985) showed their continued presence at all 17 sites (Christian, pers. obs.), and annual (1984-ongoing) monitoring and census attempts at another 19 to 32 sites where they were previously reported and at some new sites herein yielded almost annual occurrences at all sites through 2006 (Jannett, pers. obs.). Therefore, we were undertaking our survey for new sites with rock voles (1982-2001) while rock vole populations were detectable where they were known to occur based on our prior and contemporary field work on them in northeastern Minnesota.
Because the first specimen of the eastern heather vole (Phenacomys ungava) in Minnesota was for long identified as a rock vole (Handley, 1954), we examined the diagnostic dental patterns of several hundred rock voles and the first 100 heather voles encountered. All had initially been correctly identified, so we are confident in identifications of rock voles.
We selected 66 sites from Cook, Lake, St. Louis and Lake of the Woods Counties in Minnesota (Fig. 1) to inventory for rock voles. Except for two outlier sites in Lake of the Woods County, our sites were bounded on the west and north by 47.586314, -93.087333 in extreme northwestern St. Louis County, on the south at Carleton Peak in Lake County (47.583338, -90.860168), and on the east in Cook County (47.973023, -90.327476). We selected sites based on (1) our field experience with rock voles in Minnesota and New York (FJJ); (2) the report of rock voles on clearcuts by Kirkland (1977); and (3) our suspicion that eskers are appropriate rock vole habitat. Most sites we trapped were appropriate for rock voles with rock or exposed outcrops, a forb layer, and, occasionally, proximity to water (Kirkland and Jannett, 1982). We did not trap on or near boulders fields such as those already reported to be habitat of rock voles in Minnesota (Timm et al., 1977; Christian and Daniels, 1985). Only one site (in VNP) contained a small talus field, typical habitat in eastern United States (Kirkland and Jannett, 1982). We included other habitats in Cook Co. such as recent and older clearcuts, forested mountain slopes, and eskers, and, in all four counties (Cook, Lake, St. Louis, and Lake of the Woods) forested areas with little or no exposed rock but often with vegetation-covered rock. We trapped on clearcuts without prior knowledge of rock abundance or exposed outcrops.
Sites were chosen to be inventoried in forested areas by reconnoitering, especially further west than where rock voles were known to be in the Superior National Forest (SNF) and VNP. Clearcuts, their respective ages, and notations of barrel scarification (cf., Larson and Hallman, 1980) were found using maps provided by SNF staff. Eskers were located according to Sharp (1953). For clearcuts and eskers, we concentrated on Cook Co. for logistic reasons.
We also report rock vole presence and absence on two additional suites of sites inventoried for other species. For eastern heather voles, there were 25 dry sites dominated by ericaceous shrubs, of which only 15 had some scattered boulders or bedrock outcrops. For northern bog lemmings, we trapped on 25 lowland bogs, without detectable rocks, that were probably not appropriate for rock roles but many were bounded by habitat that was, and some vole species have been found in numbers in habitat perhaps unusual for them, e.g., Phenacomys intermedins in fen and bog-like habitat (Pearson, 1999). Sites appropriate for these two species were selected on the bases of SNF maps of vegetation, consultation with SNF staff, and reconnoitering.
Our new sites were discretely different habitats if <1 km from site(s) known to have rock voles (Christian and Daniels, 1985). New sites inventoried included some proximate to each other, e.g., six within 7.4 km and three within 1.6 km, if they were separated by habitat deemed inappropriate for rock voles, e.g, lowland black spruce (Picea mariana), wetland, or open water. The exceptions were three pair of recent clearcuts, in each of which the two sites were across the respective single lane unpaved road. In addition, two new sites in the heather vole inventory were not near any site known to have rock voles but were ~100 m from each other on different bedrock outcrops.
Canopy was generally mixed deciduous-coniferous with dominant species of paper birch (Betula papyrifera), aspen (Populus tremuloides), black spruce, balsam fir (Abies balsamea), jack pine (Pinus banksiana), red pine (P. resinosa), and white pine (P. strobus). Alder [Alnus sp(p).], willow [Salix sp(p).], blueberry [Vaccinium sp(p).], and Labrador-tea (Ledum groenlandicum) were locally common. Most dry sites with ericaceous species dominant in the shrub layer had canopies of jack pine and/or black spruce. Lowland bogs were dominated by black spruce and larch (Larix laricina), with less commonly white cedar (Thuja occidentalis) and bog birch (B. glandulifera). Within each habitat type, except for recent clearcuts and bogs, degree of canopy closure varied but tended toward closure.
Efforts to secure rock voles at new sites are summarized in Table 1, wherein one trap-night (TN) is one trap left out for approximately 24 h. Trap stations were either spaced at varying intervals in specific habitat patches or with longer regular intervals to cover more area between stations. Trap stations were permanent at each site sampled in multiple years.
Table 2 summarizes, in greater detail, efforts in VNP in extreme northwest St. Louis Co., an area of particular interest at the western margin of the range. There were 10 sites deemed suitable habitat for rock voles west of where they were previously reported (Jannett et al., 2007). Overstories were the following at sites in 1996, which were retrapped in subsequent years: (1) primarily red pine; (2) extremely diverse but with substantial red pine; and (3) primarily paper birch with a small sloped boulder field. In 1997 and 2001 another three and four sites, respectively, were inventoried, including the one with a small talus field. Inventory efforts in 2001 (5111 TN) included 2146 TN using unbaited traps across all four sites and 200 TN using traps baited with sardines at one site. Traps in sites monitored ("M") annually followed straight replicated lines with two traps per station at 6.1 m intervals. Table 2 also includes, for comparison, efforts at the two sites that yielded rock voles previously reported (Jannett et al., 2007), one with a mixed canopy of predominantly red pine and white pine, another a mixed canopy of primarily second-growth jack pine. TN is estimated (~) where exact number of TN is not known because of disturbance by bear (Ursus americanus) and/or pine marten (Martes americana).
We tested the association of the 66 sites with and without rock voles on the Rainy lobe versus on the combined Superior and Des Moines lobes of the Wisconsinan glaciation as delineated by Hobbs and Goebel (1982) using two-tailed [chi square], [chi square] with a Yates correction, and Fisher exact tests (GraphPad, 2015). We also located sites with rock voles in the other inventories with respect to glacial lobes but did not include them in the tests because those sites were not chosen as appropriate for rock voles.
We captured at least one rock vole at each of 53 new sites, including 45 among the 66 inventoried for rock voles and eight among the 25 inventoried for heather voles. The most westerly was only 26 km southeast (21 km south) of the previous westerly-most locality. Localities with rock voles are plotted in Figure 1.
Rock voles were captured at 35 new sites in Cook Co. Sites extend the known range in the county to the southeast and southwest (T. 62N, R. 1W, sec. 21; T. 60N, R. 5W, sec. 8), the furthest being 12 km south of a site previously known to have rock voles. New sites ranged 48 km north to south and 38 km east to west. In Lake Co., we captured rock voles at four localities, extending localities 16 km south of the only previous county record.
In St. Louis Co., new localities include three near the suspected 1921 locality at Burntside Lake. We captured rock voles at several localities near Jeanette Lake on the Echo Trail, the most westerly ca. 4 km west of Jeanette Lake (T. 65N, R. 16W, sec.2). We did not capture them at any of four sites generally suitable for rock voles in the approximately 10 km west of this locality (the vicinity indicated by the large open circle in Fig. 1). Although rocks and crevices were present, the rocks in this area were more angular than glacial rock and boulders just east. We also conducted road-side surveys further west and south of this area (in T. 65N, R. 16W) and found a general absence of appropriate habitat. The most southerly of new sites was approximately 42 km south of the two known sites in VNP.
We did not capture rock voles in VNP farther west than the known rock vole localities in the Park. Each of the 10 additional such sites was approximately 26-43 km west of the most westerly locality with rock voles (Fig. 1). The overall trap-night effort (21,006 TN over 6 y) at the sites where rock voles were not taken was more than six times that at the two sites where they were previously found (Table 2). We also did not capture the species at either site in the Northwest Angle in Lake of the Woods Co., or on Oberg or Carlton Mountains in the southwestern corner of Cook Co.
Among sites associated with the Rainy lobe, each of 44 among the 60 (73%) in the inventory specific for rock voles had one or more rock voles. Among the three sites associated with the Superior lobe to the east of the Rainy lobe, only one, an esker [greater than or equal to] 2 km from the Superior/Rainy lobe margin, had rock voles. There were no rock voles on any of the three sites associated with the Des Moines lobe west of the Rainy lobe, one in western VNP and two in Lake of the Woods Co. The difference in presence of rock voles on the Rainy lobe versus presence on one of six sites on the Superior and Des Moines lobes combined (Table 3) was significant [[chi square] = 8.074 (P = 0.0045), [[chi square].sub.c] = 5.673 (P = 0.0172), and Fisher exact test (P = 0.0107)]. The eight sites with rock voles in the heather vole inventory were all among the 23 sites associated with the Rainy lobe. The remaining two sites, associated with the Superior lobe, were both [less than or equal to] 1 km from the Superior/Rainy lobe margin.
Rock voles were secured in varied habitats. They included six of seven recent clearcuts and all six older clearcuts, with rock voles present in most or all years at each. Among the older clearcuts, two had been barrel scarified, one had a dense cover of aspen saplings, and three had extensive downed timber and slash. Some rock voles were trapped near small patches of rocks but others were apart from such areas. Among the 12 clearcuts with rock voles, each had one or more adult females in one or more years.
Among the 25 dry sites dominated by ericaceous shrubs inventoried for eastern heather voles were 15 sites with bedrock outcrops or boulders. One to 10 rock voles were captured at each of eight of those 15 sites but at none of the other 17 sites. Among the forested sites in St. Louis Co. with rock voles, two were clearcuts not in the aforementioned suites of clearcut sites, each about 25 y old that no longer had downed timber or slash. One site displayed approximately 50% canopy closure primarily by black spruce and alder and the other approximately 75% canopy closure dominated by black spruce, jack pine, eastern larch, alder, and willow, with abundant Labrador-tea and a dense moist Sphagnum substrate.
The 16 other sites in Cook Co. with rock voles out of the 17 surveyed consisted of varying amounts of rock and forbs, but were not open boulder fields. For example one site had a dense but not closed canopy of aspen and white birch, abundant forbs, and abundant but no exposed boulders. The only site in this group without rock voles in 2 y of sampling was a narrow strip of land ~6 m wide between a two-lane unpaved road and a lake. Most of the more westerly sites in St. Louis Co. with rock voles did not have boulder concentrations but rather had few or no exposed boulders and largely closed but not dense canopies; vegetation-covered boulders were present at many of these sites.
Rock voles were obtained in each of the first 6 y (1984-1989 and subsequently through 2011) on three of the four eskers, and one or more adult females were secured at each site in three or four of the first 6 y. Two eskers had slopes averaging about 45[degrees], one of which may be algific. Eskers were topped with mixed deciduous-coniferous forest of varying amounts of canopy closure and groundcover of mosses, forbs, ferns, and grasses. The esker that did not yield rock voles (one year's effort) was even steeper, had been heavily modified as a railroad bed, and had neither exposed rocky areas nor abundant forbs. No rock vole was taken on any of the 25 bogs inventoried for the northern bog lemming.
Running, standing, or obvious subsurface water was not common at or immediately proximate to new sites with rock voles. Two eskers and two other sites were immediately next to streams, while the third esker and each of two other sites was adjacent to a moist black spruce lowland or alder swamp with standing water. One site had pools of water and another was within 100 m of subsurface water. One site had patches of isolated Sphagnum, moss often associated with perched water tables. Water was not noted on or near any other sites.
Other species secured were arctic shrew (Sorex arcticus), masked shrew (S. cinereus), smoky shrew (S. fumeus), pygmy shrew (.S', hoyi), common water shrew (.S', palustris), northern short-tailed shrew (Blarina brevicauda), star-nosed mole (Condylura cristata), least chipmunk (Tamias minimus), eastern chipmunk (T. striatus), northern flying squirrel (Glaucomys sabrinus), deermouse (Peromyscus maniculatus), boreal red-backed vole (Myodes gapperi), eastern heather vole (Phenacomys ungava), meadow vole (Microtus pennsylvanicus), southern bog lemming (Synaptomys cooperi), meadow jumping mouse (Zapus hudsonius), and woodland jumping mouse (Napaeozapus insignis). The red-backed vole was the most widespread and common species secured. The other arvicoline species (M. pennsylvanicus, P. ungava, S. cooperi) were uncommon at sites where M. chrotorrhinus was present.
Our survey findings in northern Minnesota indicate the rock vole is widespread in varied habitats in Cook and Lake Counties, and in northeastern and north-central St. Louis Co. near the westernmost part of its extensive range, and it does not become rarer at the western margin of its range. The range of the rock vole is continuous throughout much of the northern parts of these counties, and rock voles are not confined to scattered fairly restricted habitat islands (Christian and Daniels, 1985).
We found that the broad habitat associations of the rock vole suggested by Kirkland and Jannett (1982) and Christian and Daniels (1985) persist at the western periphery of its range. Rock voles were present in undisturbed forest as well as clearcuts; in various combinations of ground vegetation, tree-species composition, and canopy closure; in areas with and others without water; and with variable presence of rocks and boulders. It is not clear that rock voles persisted over years on the clearcuts where Kirkland (1977) reported them in West Virginia, but we found them, including adult females, in most years on clearcuts; therefore, rock voles are not necessarily transitory on clearcuts. We interpret the presence of lactating and recently lactating females as indicative of sites appropriate for breeding, even though adult female voles are known to disperse from natal nests as an apparent mechanism for weaning (e.g, Jannett, 1978).
The distribution of scattered rocks and forbs in clearcuts and sites toward the western margin of the range reported here appear to be more important for the distribution of the species than the level boulder fields with massive concentrations of glacial till at many other sites (Timm et al., 1977; Christian and Daniels, 1985). Boulder fields themselves do not seem to be necessary for populations of rock voles, but the margins of such boulder fields do provide abundant forbs, the principal food of this species (Kirkland and Jannett, 1982). Boulder fields are also commonly in low-lying areas that are cooler and often damp (Sharp, 1953; DPC and FJJ, pers. obs.). Neither are the talus slides associated with the species in the eastern United States (Kirkland and Jannett, 1982) required as structural elements of the habitats of rock voles in Minnesota. More substantial accumulation of rocks is required by some other mammalian species, e.g., pikas (Ochotona spp.) in talus of western North America (MacDonald and Jones, 1987; Smith and Weston, 1990) and the mountain pygmy possum (Burramys parvus) in alpine periglacial boulder fields of southeastern Australia (Menkhorst et at, 2008). Such geosites are known for associated biodiversity (Hjort et al., 2015), but none of our new sites was near a talus slide and only one was proximate to a large boulder field that could be a source habitat for rock voles (cf., Pulliam, 1988).
The basis for the dependence of rock voles on rocky habitat is unclear. Perhaps scattered rocks are sufficient to provide easy subterranean access and cool moist refuges, or the rocks allow frost heave to create contiguous space for rock voles.
We cannot identify a precise vegetative boundary (Marschner, 1974) that corresponds with areas where we did and did not capture rock voles. Nor are plants commonly associated with rock voles confined to the range of the vole. Bunchberry (Comus canadensis), a common food of rock voles (Whitaker and Martin, 1977; Kirkland and Jannett, 1982) and other plant species associated with rock voles have ranges in Minnesota (Ownbey and Morley, 1991) extending well beyond the known range of rock voles.
THE ROCK VOLE-RAINY LOBE ASSOCIATION
Prior to this work, all known sites for the rock vole in Minnesota were in Cook Co. except one in Lake Co. (Christian and Daniels, 1985) and three in St. Louis Co. (Handley, 1954; Jannett et al., 2007). Now the species has been documented across a broad area of northeast and north-central Minnesota that was covered by three lobes of the late Wisconsinan glaciation (Fig. 1 after Hobbs and Goebel, 1982; Lusardi, 1997; and Patterson, 1997). Only one of the known localities in Minnesota (new herein T. 60N R. 5W, sec. 8, southwestern Cook Co.) is on glacial deposits associated with the Superior lobe and it is an esker. However, this and the sites on Oberg and Carlton Mountains are the only sites we studied that are associated with the Superior lobe. All other localities of the species in the three counties are on ground moraine deposits associated with the Rainy lobe, which is relatively boulder-rich (Hobbs and Goebel, 1982; Matsch, pers. comm; Lusardi, pers. comm.). More specifically, the sites are all on the Vermilion ground moraine of the Rainy lobe, which is even more boulder-rich than the more southerly contiguous Nashwauk and St. Croix ground moraine associations of the Rainy lobe (Hobbs and Goebel, 1982). Therefore, our evidence for a range limit in extreme northwestern St. Louis Co. corresponds with a clear geological marker that relates to habitat requirements of the rock vole (Kirkland and Jannett, 1982). Although our records document the likely western periphery of the range of the species, we predict relictual disjunct populations may be found on isolated eskers and drumlins in the more southerly parts of the Rainy lobe.
The previously reported two sites with rock voles in eastern VNP (Jannett et al., 2007) are also in the area covered by the Rainy lobe, near where the margin of the lobe curves southeasterly and is only about 15 km west of those sites. Nine sites farther west in VNP, where we did not capture rock voles despite seemingly appropriate habitat and extensive trapping effort, are on the extreme margin of the coverage by the Rainy lobe, each less than 15 km from the margin delineated by Hobbs and Goebel (1982). Six of the nine sites are only about 200 m to 2.1 km from the margin. The most westerly site in VNP and the two sites in Lake of the Woods Co. where we did not capture rock voles are associated with a third lobe, the Des Moines lobe, known to be relatively boulder-poor (Hobbs and Goebel, 1982; Matsch, pers. comm.; Lusardi, pers. comm.).
The areas we trapped in north-central St. Louis Co., including those where we caught no rock voles, are also within the Rainy lobe association. The westernmost site with rock voles was ca. 15 km from the margin of the lobe as delineated by Hobbs and Goebel (1982); the four closer sites yielded no rock voles. Presence of rock voles on the Rainy lobe was not sporadic. Except for three scattered sites in Cook Co. (one recent clearcut, one narrow strip of forested habitat, and one esker), presence of rock voles was continuous at sites from east to west until the narrow western margin of the lobe ca. 15 km wide where we trapped no rock voles at any of the 13 sites.
The sites deemed sufficient for rock voles where they were not found at the Rainy lobe margin in VNP or west of Jeannette Lake in St. Louis Co. were determined independently by each of us, and neither suite of sites was chosen on the basis of glacial lobe delineation. Given the already broad concordance of the range and the Rainy lobe, why is there a 15 km hiatus with no rock voles on the extreme western margin of the lobe? Perhaps the following apply: (1) the western margin of the Rainy lobe has not yet been precisely defined; (2) the margin is different from the center of the lobe in terms of rock vole habitat needs; and/or (3) the rock vole is still expanding its range and has not reached the exact western margin.
The hypothesis that the western margin of the Rainy lobe has not yet been precisely defined is suggested by the delineation of the margins of the lobe seeming to be generalized on Hobbs and Goebel (1982) and subsequent maps. The margins are not as intricately defined as lobe boundaries elsewhere, e.g., those of the Superior lobe, St. Louis sublobe, and others in Minnesota and northwestern Wisconsin (Matsch and Schneider, 1986), due to the limited scale of mapping (Lusardi, pers. comm.). Perhaps distribution of sites with and without rock voles is even more closely correlated at the very margins than is presently possible to determine on the basis of geological analyses.
The hypothesis that the extreme western margin is different from the center of the lobe in terms of rock vole habitat needs is suggested by general comments on surficial geology. For example Berthold (2015:ii) commented that "boulder concentrations on the surface of till sheets are ubiquitous features of the Late Wisconsinan Rainy lobe." But in VNP, Kurmis et al. (1986:532) noted that, away from lacustrine deposits, "glacial till deposits and soils are generally shallow, and bedrock outcrops are common." Toward its western margin, the Rainy lobe may not have been as characteristically boulder-laden as in the main body of the lobe, or other processes, e.g., ice streams or glacial Lake Agassiz, may have been important. The hypothesis that the rock vole has not yet reached the western margin of the Rainy lobe can be tested by continued monitoring and inventory in VNP.
OCCURRENCE OF ROCK VOLES
The history of documenting the distribution and occurrence of rock voles in Minnesota is unusual and remains puzzling. In addition to our not finding them in extensive efforts at the western margin of the Rainy lobe, they have not been documented further west or south of the Rainy lobe despite extensive collecting in Minnesota (Herrick, 1892; Surber, 1932; Swanson et al., 1945; Gunderson and Beer, 1953; Hazard, 1982). Our results confirm the presence of the species near its original (1921) locality in the state and extend the finding of Christian and Daniels (1985) by documenting that the species is more widespread throughout the boreal zone in northeastern Minnesota than previously thought (Timm, 1974, 1975; Timm et al., 1977; Buech et al., 1977). In addition, monitoring at localities where rock voles were caught in 1982 (Christian and Daniels, 1985) indicated their continued presence at those sites (Jannett, 1990). However, although the species exhibits remarkably constant population numbers over relatively long periods (e.g., Jannett, 1990), it is occasionally rare for consecutive years or "missing" in all samples (Jannett, pers. obs.). The most extreme instances (1983-2015) were in four recent years, including the last three, when not a single specimen was secured in 18 monitored sites and none in the last 3 y at another site despite a 6 d or 10 d trapping census effort each year. Such periods of extreme rarity may explain the failure of other workers to record their presence between 1921 and the 1970s.
Gaston (2009) suggested "for no single species do we yet have a comprehensive understanding" of the limits of geographic range. The presence of Microtus chrotonhinus on the Rainy lobe and its absence on the extreme margins of the lobe and farther from it are a striking example of the association of the rock vole with a broad geological feature. It is still not a comprehensive understanding inasmuch as we do not know why the species cannot adapt to habitat without the structural component of rocks. However, its general failure to use non-rocky habitat has been suggested to be a competitive exclusion by Microtus pennsylvanicus, a species perhaps excluded in rocky woodlands by both M. chrotorrhinus (Martin, 1971a) and Myodes gapperi (Cameron, 1965). Microtus chrotorrhinus and M. gapperi are characteristically found at the same sites (Martin, 1971b; Kirkland and Jannett, 1982). In Newfoundland with neither M. chrotorrhinus nor M. gapperi, Martin (1971a) reported the presence of M. pennsylvanicus in typical rock vole habitat. The presence of M. chrotorrhinus may be determined by its adaptation to rocky woodland habitats and its ability to exclude the wide-ranging, common, congeneric M. pennsylvanicus in such habitats.
Microtus chrotorrhinus was not secured in Minnesota for more than 50 years after its discovery and has now not been detected at sites where it was recently documented for most of 33 years. Because it is a boreal, arguably relictual, species at its southwestern range limit in Minnesota and shows an enigmatic pattern of disappearances when it is at very low numbers, it may be expected to respond to global warming more quickly or to a greater extent than some other species of small mammals in the state.
Acknowledgments.--We thank the staff of Superior National Forest (SNF), especially W. P. Russ, for maps of clearcut areas and various courtesies that facilitated the work. We gratefully acknowledge financial support from the following: USDA Forest Service (North Central Forest Experiment Station) to DPC; Voyageurs National Park (VNP), Great Lakes Network of the National Park Service, U.S. Fish and Wildlife Service, Science Museum of Minnesota (SMM), and the Minnesota Department of Natural Resources to FJJ; and SNF to each of us. We thank the staffs of VNP and SMM, and students at Northland College and the University of Minnesota-Duluth for assistance. We thank L. H. Grim for sharing his knowledge of the biota and geology of VNP, C. L. Matsch for information on glacial geology of northern Minnesota, and B. A. Lusardi for further discussion of glacial geology and mapping. We thank R. K. Rose for reviewing the manuscript and suggesting changes and additions, K. MacDonald for map changes, and the reviewers for helpful suggestions.
Bailey, V. 1932. Mammals of New Mexico. N. Amer. Fauna, 53:1-412.
Berthold, A. J. 2015. Surface boulder concentrations of the late Wisconsinan Rainy Lobe, Minnesota, USA. Unpubl. M.S. Thesis, Univ. Minnesota, Duluth. 105 p.
Buech, R. R., R. M. Timm, and K. Siderits. 1977. A second population of rock voles, Microtus chrotorrhinus, in Minnesota with comments on habitat. Can. Field-Nat., 91:413-414.
Cameron, A. W. 1965. Competitive exclusion between the rodent genera Microtus and Clethrionomys. Evolution, 18:630-634.
Christian, D. P. and J. M. Daniels. 1985. Distributional records of rock voles, Microtus chrotorrhinus, in northeastern Minnesota. Can. Field-Nat., 99:356-359.
Coffin, B. and L. Pfannmuller, eds. 1988. Minnesota's endangered flora and fauna. University of Minnesota Press, Minneapolis. 473 p.
Connor, P. F. 1959. The bog lemming Synaptomys cooperi in southern New Jersey. Publ. Mus., Michigan State Univ., 1:161-248.
Etnier, D. A. 1989. Small mammals of the Boundary Waters Canoe Area, with a second Minnesota record for the Heather Vole, Phenacomys intermedius. Can. Field-Nat., 103:353-357.
Gaston, K. J. 2009. Geographic range limits: achieving synthesis. Proc. R. Soc. B, 276:1395-1406.
Geber, M. A. 2011. Ecological and evolutionary limits to species geographic ranges. Am. Nat., 178:S1-S5.
GraphPad. 2015. www.graphpad.com.
Gunderson, H. L. and J. R. Beer. 1953. The Mammals of Minnesota. Minnesota Museum of Natural History, Occasional Paper, 6.
Handley, C. O., Jr. 1954. Phenacomys in Minnesota. J. Mammal., 35:260.
Hazard, E. B. 1982. The Mammals of Minnesota. University of Minnesota Press, Minneapolis. 280 p.
Heads, M. 2015. The relationship between biogeography and ecology: envelopes, models, predictions. Biol. J. Linnean Soc., 115:456-468.
Herrick, C. L. 1892. The Mammals of Minnesota. Bull. Geol. Nat. Hist. Surv. Minn., 7:1-299.
Hjort, J., J. E. Gordon, M. Gray, and M. L. Hunter, Jr. 2015. Why geodiversity matters in valuing nature's stage. Cons. Biol., 29:630-639.
Hobbs, H. C. and J. E. Goebel. 1982. Geological map of Minnesota. Quaternary Geology. Minnesota Geological Survey, Quaternary Geology Map S-l, University of Minnesota, St. Paul.
Hoffmann, R. S. and J. W. Koeppl. 1985. Zoogeography, p. 84-115. In: Biology of new world Microtus. Tamarin, R. H. (ed.). Special Publ. 8. American Society of Mammalogists.
Hjort, J., J. E. Gordon, M. Gray, and M. L. Hunter, Jr. 2015. Why geodiversity matters in valuing nature's stage. Cons. Biol., 29:630-639.
Jannett, F. J., Jr. 1978. The density-dependent formation of extended maternal families of the montane vole, Microtus montanus nanus. Behav. Ecol. Sociohiol., 3:245-263.
--. 1990. Population constancy of the rock vole, Microtus chrotorrhinus, in northeastern Minnesota, p. 81-87. In: Social systems and population cycles in voles. Tamarin, R.H., R. S. Ostfeld, S. R. Pugh, and G. Bujalska (eds.). Birkhauser Verlag AG, Basel.
--, M. R. Broschart, L. H. Grim, and J. P. Schaberl. 2007. Northerly range extensions of mammalian species in Minnesota. Am. Midi. Nat., 158:168-176.
Kirkland, G. L., Jr. 1977. The rock vole, Microtus chrotorrhinus (Miller) (Mammalia: Rodentia) in West Virginia. Annals Carnegie Mus. Nat. Hist., 46:45-53.
--and F. J. Jannett, Jr. 1982. Microtus chrotorrhinus. Mammal Spec., 180:1-5.
Kurmis, V., S. L. Webb, and L. C. Merriam, Jr. 1986. Plant communities of Voyageurs National Park, Minnesota, U.S.A. Can. J. Bot., 64:531-540.
Lansing, S. W. 2005. A range extension for the rock vole, Microtus chrotorrhinus, in Labrador. Can. Field-Nat., 119:412-416.
Larson, J. and R. Hallman, compilers. 1980. Equipment for reforestation and timber stand improvement. Equipment Development Center, USDA Forest Service, Missoula, MT. 254 p.
Lusardi, B. A. 1997. Minnesota at a glance: Quaternary glacial geology. Minnesota Geological Survey, University of Minnesota. 4 p.
Macarthur, R. H. 1972. Geographical ecology: Patterns in the distribution of species. Harper and Row, Publishers, Inc., New York. 269 p.
Macdonald, S. O. and C. Jones. 1987. Ochotona collaris. Mammal. Spec.., 281:1-4.
Marschner, F. J. 1974. The original vegetation of Minnesota (map). USDA Forest Service, North Central Forest Experiment Station, St. Paul, Minn. A redraft of the original 1930 edition.
Martin, R. E. 1971a. The natural history and taxonomy of the rock vole, Microtus chrotmrhinus. Unpubl. Ph.D. Dissertation, Univ. Connecticut, Storrs. 123 p.
--. 1971b. Interspecific associations of rock voles. Beta Kappa Chi Bull, 30:5-7.
Matsch, C. L. and A. F. Schneider. 1986. Stratigraphy and correlation of the glacial deposits of the glacial lobe complex in Minnesota and northwestern Wisconsin. Quaternary Sci. Rev., 5:59-64.
Mayr, E. 1963. Animal species and evolution. Harvard University Press, Cambridge, MA. 797 p.
Menkhorst, P., L. Broome, and M. Driessen. 2008. Burramys parvus. The IUCN Red List of Threatened Species. Version 2015.2.
Nordquist, G. E. and E. C. Birney. 1988. Mammals, p. 293-322. In: Coffin, B. and L. Pfannmuller (eds.). Minnesota's endangered flora and fauna. University of Minnesota Press, Minneapolis.
Ownbey, G. B. and T. Morley. 1991. Vascular plants of Minnesota: A checklist and atlas. University of Minnesota Press, Minneapolis. 306 p.
Patterson, C. J. 1997. Southern Laurentide ice lobes were created by ice streams: Des Moines Lobe in Minnesota, USA. Sedimentary GeoL, 111:249-261.
Parmesan, C. 2001. Detection of range shifts: general methodological issues and case studies of butterflies, p. 57-76. In: Walther, G.-R., C. A. Burga, and P. J. Edwards (eds.). "Fingerprints" of climate change--Adapted behaviour and shifting species ranges. Kluwer Academic, New York.
Pearson, D. E. 1999. Small mammals of the Bitterroot National Forest: a literature review and annotated bibliography. Gen. Tech. Rep. RMRS-GTR-25. USDA Forest Service, Rocky Mountain Research Station. 63 p.
Post, E. 2013. Ecology of climate change: the importance of biotic interactions. Monographs in Population Biology, 52. Princeton University Press, Princeton. 373 p.
Pulliam, H. R. 1988. Sources, sinks, and population regulation. Amer. Nat., 132:652-661.
Sharp, R. P. 1953. Glacial features of Cook County, Minnesota. Amer. J. Science, 251:855-883.
Sikes, R. S., W. L. Gannon, and the Animal Care and Use Committee of the American Society of Mammalogists. 2011. Guidelines of the American Society of Mammalogists for the use of wild mammals in research. J. Mammal, 92:235-253.
Smith, A. T. and M. L. Weston. 1990. Ochotona princeps. Mammal Spec., 352:1-8.
Smith, P. W. 1957. An analysis of Post-Wisconsin biogeography of the prairie peninsula region based on distributional phenomena among terrestrial vertebrate populations. Ecology, 38:205-218.
Soule, M. E., ed. 1986. Conservation biology: the science of scarcity and diversity. Sinaeur Associates, Inc., Sunderland, MA. 584 p.
Surber, T. 1932. The Mammals of Minnesota. Minnesota Game and Fish Department, St. Paul. 84 p.
Swanson, G., T. Surber, and T. S. Roberts. 1945. The mammals of Minnesota. Minnesota Department Conservation, Tech. Bull., 2:1-108.
Timm, R. M. 1974. Rediscovery of the rock vole (Microtus chrotmrhinus) in Minnesota. Can. Field-Nat., 88:82.
--. 1975. Distribution, natural history, and parasites of mammals of Cook County, Minnesota. Occasional Papers Bell Mus. Nat. Hist., Univ. Minn., 14:1-56.
--, L. R. Heaney, and D. D. Baird. 1977. Natural history of rock voles (Microtus chrotorrhinus) in Minnesota. Can. Field-Nat., 91:177-181.
Walther, G.-R., E. Post, P. Convey, A. Menzel, C. Parmesan, T. J. C. Beebee, J.-M. Fromentin, O. Hoegh-Guldberg, and F. Bairlein. 2002. Ecological responses to recent climate change. Nature, 416:389-395.
Wetzel, R. M. 1955. Speciation and dispersal of the southern bog lemming, Synaplomys cooperi (Baird). J. Mammal, 36:1-20.
Whitaker, J. O., Jr. and R. L. Martin. 1977. Food habits of Microtus chrotorrhinus from New Hampshire, New York, Labrador, and Quebec. J. Mammal, 58:99-100.
SUBMITTED: 8 OCTOBER 2015
ACCEPTED: 26 SEPTEMBER 2016
FREDERICK J. JANNETT, JR. (1)
Department of Fisheries, Wildlife, and Conservation Biology, University of Minnesota, St. Paul 55108 and The Museum, Texas Tech University, Lubbock 79409
DONALD P. CHRISTIAN (2)
Department of Biology, University of Minnesota-Duluth 55812
(1) Corresponding author: Telephone: (806) 797-0772; e-mail: firstname.lastname@example.org
(2) Present address: President's Office, SUNY New Paltz, New Paltz, NY 12561
Caption: Fig. 1.--Range and localities of the rock vole in Minnesota. Inset map indicates the geographic range of the species (Kirkland and Jannett, 1982; Lansing, 2005). Small triangles represent all peripheral sites (ours and those previously published) at the margin of the range; large triangle posits previously known western-most locality. New localities are represented by small closed circles; when two or more sites were within 1.7 km, they are represented by one circle, and some sites in Cook Co. are not shown because of spatial constraints. The large filled circle represents the original (1921) locality and three new proximate localities. Each small open circle represents one or two localities with habitat appropriate for rock voles where we did not secure any, including seven sites among 10 in VNP near the western margin of the Rainy lobe in extreme northwestern St. Louis Co., and the large open circle represents four such localities. On the Rainy lobe within Cook Co., the three scattered localities without rock voles are not represented among the 35 with rock voles. All three localities on the Superior lobe and the three on the Des Moines lobe are shown. LW = Lake of the Woods Co.; S = St. Louis Co.; L = Lake Co.; C = Cook Co. In bold: D = De Moines lobe; S = Superior lobe; R = Rainy lobe. Small, disjunct segments of the Rainy lobe Vermilion moraine association are not shown. The dashed line represents the border between the Vermilion moraine association to the north and the Nashwauk and St. Croix moraine associations to the southwest and southeast, respectively, within the Rainy lobe. The St. Croix moraine association of the Rainy lobe in disjunct segments to the southwest of the Nashwauk, and other lobes, are not shown
TABLE 1.--Efforts to secure rock voles Sites (n) with rock Habitat type voles Year(s) Recent clearcuts, initially 6 of 7 6 (1984-1989) 11 to 23 months post-cut Older clearcuts, initially 6 of 6 2 or 5 47 to 122 months post-cut (1985-1989) Eskers ** 3 off up to 31 (1984-2015) Mountain slopes 0 of 6 *** 1 or 2 y per site (Oberg + Carlton (1982-1983) Mountains.), scattered rock Forest, largely closed 16 of 17 up to 32 canopy, scattered rock (1983-2015) Forest, largely closed 14 of 18 1 y per site canopy, scattered or no (1987-1989) rock Ericaceous shrubs, 8 of 25 1 (2001) scattered boulders +/or bedrock outcrops at 15 sites Lowland bogs, no surface 0 of 25 1 (2001) rock Forest, largely closed 0 of 10 6 (1996-2001) canopy, scattered rock, some bedrock outcrops Wet forest, sparse rock and 0 of 2 1 (1989) crevices Habitat type County Canopy species Recent clearcuts, initially Cook Mixed, ~no canopy 11 to 23 months post-cut Older clearcuts, initially Cook Mixed, sparse, sapling, 47 to 122 months post-cut pole Eskers ** Cook Mixed, red pine, white pine, paper birch, aspen, black spruce Mountain slopes Cook Mixed, paper birch dominant (Oberg + Carlton Mountains.), scattered rock Forest, largely closed Cook Mixed, paper birch, aspen, canopy, scattered rock balsam fir, black spruce, alder Forest, largely closed St. Louis Mixed, including one of canopy, scattered or no open canopy black spruce; rock most upland Ericaceous shrubs, Cook, Lake, Mixed conifers, jack pine + scattered boulders +/or St. Louis black spruce dominant bedrock outcrops at 15 sites Lowland bogs, no surface Cook, Lake, Mixed conifers, black rock St. Louis spruce dominant, larch common; bog birch Forest, largely closed St. Louis see Methods canopy, scattered rock, some bedrock outcrops Wet forest, sparse rock and Lake of the Mixed, alder dominant; crevices Woods mixed, deciduous-coniferous Habitat type Trap type Trap-nights * Recent clearcuts, initially MS 100 or 200 per site per 11 to 23 months post-cut year Older clearcuts, initially MS 100 per site per year 47 to 122 months post-cut (1 site/2 y; 5 sites/5 y) Eskers ** MS 100 per site per year (1 site/1 y; 1 site/13 y; 2 sites/32 y) Mountain slopes MS+ 120 total/5 sites 1982; (Oberg + Carlton CSSL ****; 400 total/2 sites Mountains.), MS 1983 (including one scattered rock 1982 site) Forest, largely closed MS+CSSL <100 per site per year canopy, scattered rock (10 sites/1 y); 100 per site per year (7 sites/33 y) Forest, largely closed MS <80 per site per year canopy, scattered or no rock Ericaceous shrubs, MS most 300 per site, scattered boulders +/or 7550 total bedrock outcrops at 15 sites Lowland bogs, no surface MS+pit most 300 per site, rock traps 8180 total Forest, largely closed MS 21,006; see Table 2 canopy, scattered rock, some bedrock outcrops Wet forest, sparse rock and MS 297 or 510 per site crevices * Where total trap-night is not given, the effort shown is the minimum per site; at many sites sessions were extended for one or more days because of disturbance (s) or logistic reasons ** Identified by Sharp (1953) *** Because of there being fewer TN per site than in the other suites of sites, each mountain was considered as one site in tests **** CSSL = clean nonfolding small Sherman live-traps baited with peanut butter, rolled oats, and apple TABLE 2.--Inventory and monitoring efforts (in trap-nights, TN) at sites with habitat appropriate for rock voles in Voyageurs National Park Year and month (s) 1998, Sep./ Site code 1996, Aug. 1997, Aug. Oct. 96V1 662 (11) * 1224 (15) -- 96V2 1023 (5) ~477 (0) -- Ml-3 1800 1800 ~900 M8+9 209 1200 800 WP bf 102 236 -- 97V3 -- 500 -- 97V6 -- 828 -- 97V13 -- 321 -- 01V4 -- -- -- 01V5 -- -- -- 01V6 -- -- -- 01V9 -- -- -- Total TN: Sites 96V1 and 96V2 with rock voles (~3386) 1685 ~1701 All 10 other sites without rock voles (~21,006) 2111 4885 ~1700 Year and month (s) 1999, Aug./ Site code Sep. 2000, Aug. 2001, Aug. 96V1 -- -- -- 96V2 -- -- -- Ml-3 1619 1800 600 M8+9 ~900 1200 1000 WP bf 80 -- -- 97V3 -- -- -- 97V6 -- -- -- 97V13 -- -- -- 01V4 -- -- 800 01V5 -- -- 1119 01V6 -- -- 2688 01V9 -- -- 504 Total TN: Sites 96V1 and 96V2 with rock voles (~3386) All 10 other sites without rock voles (~21,006) ~2599 3000 6711 * Numbers of rock voles at the two sites where they were previously reported are in parentheses TABLE 3.--Presence and absence of rock voles on sites associated with the Rainy lobe and on sites associated with the Superior and Des Moines lobes combined Rainy lobe Habitat type Present Absent Recent clearcuts 6 1 Older clearcuts 6 0 Eskers 2 1 Mountain slopes -- -- Forest in Cook Co. 16 1 Forest in St. Louis Co. 14 4 Forest in VNP 0 9 Wet forest -- -- Totals 44 16 Superior and Des Moines lobes combined Habitat type Present Absent Recent clearcuts -- -- Older clearcuts -- -- Eskers 1 0 Mountain slopes 0 2 Forest in Cook Co. -- -- Forest in St. Louis Co. -- -- Forest in VNP 0 1 Wet forest 0 2 Totals 1 5
|Printer friendly Cite/link Email Feedback|
|Author:||Jannett, Frederick J., Jr.; Christian, Donald P.|
|Publication:||The American Midland Naturalist|
|Date:||Jan 1, 2017|
|Previous Article:||Temporal activity levels of mammals in patches of early successional and mature forest habitat in Eastern Connecticut.|
|Next Article:||Freshwater fish inventory of the Aleutian Archipelago, Alaska.|