Invasion of the meadow vole (Microtus pennsylvanicus) in southeastern Kentucky and its possible impact on the southern bog lemming (Synaptomys cooperi).
From the 1930s to the 1980s, mild climatic cooling and increased precipitation have caused species of small mammals belonging to the boreal faunal element to expand their distributions southward in the midwestern United States (Frey, 1992). Such a spread could have drastic impact on existing mammalian communities as new species invade and form new sympatric relationships. This is because interspecific competition is considered one of the major factors influencing structural organization of small mammal communities (Linzey, 1984). If, for example, two species with similar diets and habitat preferences come in contact with one another, one of the species may decline in abundance or shift to other habitats as a result of competitive exclusion.
The meadow vole (Microtus pennsylvanicus) is one member of the boreal faunal element spreading S in Kansas, Missouri and Illinois (Frey, 1992; Hoffmeister, 1989). In the past, M. pennsylvanicus was confined to the Bluegrass Plateau of N-central Kentucky, whereas the southern bog lemming (Synaptomys cooperi) was the dominant microtine on the Cumberland Plateau of southeastern Kentucky where it was isolated from M. pennsylvanicus (Barbour and Davis, 1974).
Barbour and Hardjasasmita (1966) provided a checklist of the mammals of Robinson Forest in Kentucky based on a 10-wk survey conducted during 1960 and 1961 in addition to Barbour's 12 yr of working in the forest. These authors considered Synaptomys cooperi to be common in forest clearings, collecting 20 specimens during the 10 wk. Many of these animals were caught in clearings near Camp Robinson (a research camp near the forest's western boundary). The pine vole, Microtus pinetorum, was the only other microtine rodent trapped during their study. Barbour never reported M. pennsylvanicus from Robinson Forest during 20 yr of collecting from the 1950s to the 1970s (Wayne Davis, pers. comm.).
In 1980, a single Microtus ochrogaster was collected from the clearing adjacent to Camp Robinson (Wayne Davis, pers. comm.). By 1983, Bill McComb (pers. comm.) was collecting both M. pennsylvanicus and M. ochrogaster on reclaimed surface mines adjacent to the forest. He also reported that both M. pennsylvanicus and Synaptomys cooperi were present in the forest clearings near Camp Robinson. Specific data are not available, although McComb reported that both species were trapped together, with M. pennsylvanicus being more abundant. From a 1987-1988 study on bobcats (Lynx rufus) in Robinson Forest, Frederick et al. (1989) and J. Whitaker (pers. comm.) found M. pennsylvanicus in 50% of all bobcat scats during the autumn and winter months and 67% of the scats in the spring and summer months. The frequency of S. cooperi in scat was 4% in autumn and winter and 11% in spring and summer. No additional information is available for Robinson Forest until the beginning of our study in 1992.
The purpose of this study was to determine the distribution of Microtus pennsylvanicus in Robinson Forest, which microhabitats it occupies, and its relative abundance compared to other small mammals. In addition, these results were compared to the microhabitat use and relative abundance of Synaptomys cooperi.
This study was conducted on the main body of the University of Kentucky's Robinson Forest, a 4450-ha, second-growth forest located in Breathitt, Knott and Perry counties in southeastern Kentucky. This is the largest remaining forest stand of any kind on the Cumberland Plateau. Elevations range from 240 to 490 m above sea level. It is completely surrounded by 1 to 20-year-old reclaimed surface mines and some active mining. The forest is a maze of deep, narrow valleys, steep slopes and narrow winding ridges (Overstreet, 1989). Clearings occur along streams flowing through the valleys. Most clearings are man-made with some over 70 yr old which are maintained by periodic mowing or periodic flooding. Robinson Forest includes four general habitats:
Mixed mesophytic forest. - Robinson Forest is near the center of the mixed mesophytic region (Braun, 1950). Such deciduous forest is the predominant habitat at our study site. Approximately 50 woody species dominate the area (Overstreet, 1989). Tree species vary from the most mesic beech-dominated areas characterized by beech-poplar (Fagus-Liriodendron) and hemlock (Tsuga) zones, to intermediate moisture levels characterized by oak-hickory (Quercus-Carya) and oak-red maple (Quercus-Acer) zones to the most xeric oak-dominated areas characterized by oaks and pines (Pinus; Kalisz et al., 1987).
Low-elevation wildlife clearings. - Approximately 26 clearings ranging in area from 0.1-3.0 ha existed adjacent to the three major streams. These included narrow; mowed roadsides and other, larger mowed areas. Elevations ranged from 250-300 m. Many of these clearings were extremely wet. They varied greatly with respect to dominant vegetation, but generally were either grass-dominated or herb-dominated. Any one grass-dominated site had from 3-5 species of grass including broomsedge (Andropogon virginicus), smooth crabgrass (Digitaria ischaemum), tall rescue (Festuca elatior), camus (Microstegium uminium), redtop panic-grass (Panicum rigidulum), switch grass (P. virgatum), bushy panic grass (P. dichotomum), and tall redtop (Triodia flava). No exposed earth was present at these sites. In contrast, herb-dominated sites were sparsely vegetated and dominated by smartweeds (Polygonum sp.), tall ironweed (Vernonia gigantea), goldenrods (Solidago sp.) and horse-nettle (Solanum carolinense). Only two grasses were present in these clearings (small crabgrass, Digitaria ischaemum, and spreading witchgrass, Panicum dichotomiflorum) and both were subdominants.
High-elevation wildlife clearings. - Periodically, ridgetops have had trees removed and successional plant communities have become established. During the current study, a single 2.0-ha high-elevation clearing existed. It was formed in 1989 and was 3 yr old at the beginning of our study. It was located on a ridgetop and the adjacent steep slopes; the elevation ranged from 320-430 m. This clearing was isolated from the nearest low-elevation clearings by 6 km of dense forest. The entire area was covered with a dense stand of broomsedge.
Reclaimed surface mines. - Reclaimed areas ranged in age from 1-20 yr and completely surround the forest. These areas were generally sparsely vegetated and the characteristic vegetation was quite variable. They range from being damp with small ponds to well-drained and dry. The soil is very hard and compact (resembling a desert hard-pan surface) and comprised of shale, sandstone and coal ranging in size from fine to coarse gravel. Some areas were littered with small stones. Virtually no leaf litter or organic matter was present. These areas were sparsely vegetated with bunch grasses, herbaceous species, and a variety of introduced plants used for their nitrogen-fixing abilities.
Four distinct subhabitat types were present on the reclaimed surface mines. First, low-lying areas typically had small, shallow ponds dominated by cattails (Typha sp.), bulrush (Scirpus cyperinus) and rushes (Juncus sp.). The second subhabitat was higher, drier habitat surrounding the cattails and characterized by a mixture of grasses that included broomsedge, redtop (Argostis gigantea), spreading witchgrass and purpletop (Tridens flavus). Other characteristic plants included goldenrod, horseweed (Conyza canadensis), and three species of lespedeza (Chinese lespedeza, Lespedeza cuneata; Korean clover, L. stipulacea; Japanese clover, L. striata). Small (up to 5 m tall) black locust (Robinia pseudoacacia) were typically scattered though this habitat. A third subhabitat type was dominated by crowned vetch (Coronilla varia) and up to five species of lespedeza (the three species listed above plus violet lespedeza, L. violacea, and the bush, L. bicolar). Redtop (A. gigantea) was the only grass present and was subdominant. The fourth subhabitat type was on the steep rubble slopes along the forest boundary. These slopes were sparsely vegetated and dominated by tall fescue (Festuca elatior) that was planted to stabilize the soil. The slopes sampled were N- or NE-facing.
We trapped at 45 sites in Robinson Forest and the surrounding reclaimed surface mines. During our study, active mining began at the northern boundary of Robinson Forest destroying two of our 45 trap sites before the study was completed. Distances between sites varied from 200 m to 6 km. Trapping was conducted on a total of 22 nights during March, April, September and October of 1992; during January, February, March, September and December of 1993; and in April of 1994. Several sites were trapped repeatedly; however, we waited 4-6 mo between trapping periods. A total of 23 low-elevation wildlife clearings were trapped along the three major streams in the forest. Eighty percent of all high-elevation trapping was conducted on the single 2.0-ha ridgetop site. In addition, we trapped three small roadside clearings near this clearing. Six sites were trapped on reclaimed surface mines, and 12 sites in the mixed-mesophitic forest. Forest sites were within 500 m of wildlife clearings. We trapped these forest sites to determine if they were used as marginal habitat by either Microtus pennsylvanicus or Synaptomys cooperi.
We used a combination of snap traps (museum specials and rat traps) and Sherman live traps in trap lines. Snap trap bait was a mixture of bacon grease, peanut butter and rolled oats. For live traps, this bait mixture was combined with additional rolled oats and bird seed mix. Traps were used in approximately equal proportions (50% live traps, 50% snap traps) in the largest wildlife clearings, reclaimed surface mines, and in the forest habitat. To avoid impacting small populations, we relied more on Sherman live traps (80-100% live traps; 0-20% snap traps) in the small wildlife clearings and in larger clearings where we trapped repeatedly. Voucher specimens were prepared from each site and are currently deposited in the University of Kentucky vertebrate collection. Specimens were kept whenever their identification was in question.
In the eastern part of its range, Synaptomys cooperi is noted for being elusive and difficult to trap (Linzey, 1983). To compensate for this, we tried to improve capture success by searching clearings for active runs wherein we placed traps. Synaptomys cooperi uses other voles' runs and tends to be caught when traps are placed in their paths, as opposed to being attracted to bait (Linzey, 1983). We also searched for their characteristic shiny, green droppings; and if any were found, traps were placed in all nearby runs. Most of this trapping was done during the winter, when Synaptomys cooperi in Kentucky tends to be more easily captured (Robinson, 1981).
We used a nonparametric ANOVA (SAS, 1985) to compare trap success of Microtus pennsylvanicus and Synaptomys cooperi in habitats where the two species were found. The following within-species comparisons were made: (1) for capture rates in grass-dominated vs. herb-dominated low-elevation clearings; (2) for capture rates in all low-elevation clearings vs. reclaimed mines; and (3) for capture rates in grass-dominated clearings vs. reclaimed mine sites. Only data collected from December through March were used. These were the months of highest trap success and when all habitats were trapped during a given collecting trip.
Based on 2817 trap nights over 2 yr, Microtus pennsylvanicus was the most abundant rodent in low-elevation clearings and second most abundant in all other open habitat, whereas Synaptomys cooperi was the least commonly collected rodent at all sites (Table 1). One M. pennsylvanicus was caught in the forest habitat. Microtus pennsylvanicus were caught 18 times more often than S. cooperi in the low-elevation wildlife clearings (47.0% vs. 2.6% of all small mammals caught, respectively) and 4.3 times more often on reclaimed surface mines (26.5% vs. 6.1%, respectively). Furthermore, M. pennsylvanicus accounted for 31% of all small mammals caught on the high-elevation clearing where no S. cooperi were captured.
Microtus pennsylvanicus were most often trapped in grass-dominated habitats. This included low- and high-elevation wildlife clearings as well as some areas on the reclaimed surface mines where grass was one of the dominants. Cattail habitats were the only exception. Microtus pennsylvanicus were trapped at the edge of this habitat and active runs were in abundance through stands of sedge in the cattail habitat. Some grasses were present along these edges; however, active mining destroyed this habitat before mature grasses could be sampled. No small mammals were caught in areas where only crowned vetch and lespedeza dominated and grasses were virtually lacking.
Of the six small mammal species trapped in forest clearings, only Microtus pennsylvanicus was caught during each of the 7 mo. The other five species were missing from 2-5 of the months sampled [ILLUSTRATION FOR FIGURE 1A OMITTED]. Synaptomys cooperi was caught less frequently than any other species, and only one individual was caught during any given month from October to February. Drastic seasonal fluctuations in trap success were experienced over the 2 yr of the study [ILLUSTRATION FOR FIGURE 1B OMITTED]. Trap success in forest clearings was highest in January and December (20 mammals per 100 trap nights), intermediate from February to April (11 mammals per 100 trap nights), and lowest in September and October (1.5 mammals per 100 trap nights). [TABULAR DATA FOR TABLE 1 OMITTED] This sharp drop in trap success was most extreme in September, when a total of 476 trap nights during September 1992 and 1993 yielded a single M. pennsylvanicus. This low yield occurred despite our efforts to rebait traps during the night to offset the high rate at which invertebrates removed the bait. With very few exceptions, we failed to find signs of active runs.
Microtus pennsylvanicus and Synaptomys cooperi were both caught during October, December, January and February [ILLUSTRATION FOR FIGURE 1A OMITTED]. Thus comparing trap success for the different habitat types during these months, M. pennsylvanicus was significantly more likely (15.2 times more likely) to be trapped in grass-dominated clearings than in herb-dominated clearings (F = 34.11, P = 0.0001; Table 2), whereas S. cooperi was only 3.5 times more likely to be caught in grass-dominated clearings than herb-dominated clearings (F = 0.719, P = 0.41; Table 2). Synaptomys cooperi was 7.2 times more likely to be trapped on reclaimed surface mines (F = 4.16, P = 0.053), whereas M. pennsylvanicus was as likely to be caught on reclaimed mines as in forest clearings (5.2% vs. 5.3%, respectively; F = 0.00, P = 0.99; Table 2). Microtus pennsylvanicus was 2.6 times more likely to be caught in grass-dominated forest clearings than on reclaimed mines (F = 3.4, P = 0.09; Table 2), whereas S. cooperi was 3.9 times more likely to be caught on reclaimed mines (F = 1.06, P = 0.33; Table 2).
Documented distributions in the midwestern United States and available information for Kentucky suggest that the presence of Microtus pennsylvanicus in Robinson Forest is recent. This range expansion has probably been the combined result of climatic cooling (as suggested by Frey, 1992) and deforestation (as suggested by Linzey, 1984). This species is extending its range southward in the midwestern states (Frey, 1992). Microtus pennsylvanicus was not known in Kansas until 1963 but is now well-established (Frey, 1992). Meadow voles were first reported in Missouri in 1973 and have become increasingly more abundant in [TABULAR DATA FOR TABLE 2 OMITTED] southeastern Nebraska (Frey, 1992). Hoffmeister (1989) noted that M. pennsylvanicus has been extending its range southward in Illinois since the late 1950s. The apparent southeastern expansion in Kentucky is concordant with the fact that other members of the boreal faunal element have recently become more widespread in Kentucky. This is the case for Zapus hudsonicus (Houtcooper, 1982), and is most striking for the least weasel (Mustela nivalis). The least weasel was first reported from Kentucky in 1976 (Davis and Barbour, 1979) and since then, abundant records have become known for N-central Kentucky (Meade, 1992). Frey (1992) considered the mild decrease in temperatures and increase in precipitation in the midwest from 1935 to 1985 to be a likely explanation for the southward expansion of the northern boreal faunal element in the midwest. Similar cooling trends in the southeastern United States (Trapasso and Al Kolibi, 1994) may suggest a similar scenario in Kentucky.
Generally, Synaptomys cooperi is rare and elusive in the eastern part of its range (Linzey, 1983). This characteristic coupled with the lack of quantitative data, makes it difficult to determine how its population at Robinson Forest may have changed over the years or how it has been affected by the arrival of Microtus pennsylvanicus. We do know that S. cooperi is now rarely caught in forest clearings, whereas M. pennsylvanicus is the most commonly caught small mammal in forest clearings. One S. cooperi was caught for every 20 M. pennsylvanicus. During our study, we failed to collect S. cooperi or find signs of it in clearings (especially those near Camp Robinson) where Barbour and Hardjasasmita (1966) found them to be common. It is the same spot where, W. McComb (pers. comm.) found them in the early 1980s (when M. pennsylvanicus probably first arrived).
Linzey (1983) stated that isolated, man-made clearings in eastern forests are not readily colonized by Microtus pennsylvanicus and are favored by Synaptomys cooperi, although the clearing of eastern woodlands and replacement of native grasses by introduced species favors M. pennsylvanicus. These conditions are similar to those seen in Robinson Forest. First, forest clearings in southeastern Kentucky were more isolated in the 1950s and 1960s before extensive deforestation from surface mining. Second, an increase in the number of paved highways, coupled with the planting of fescue grass along the roadsides, occurred during the 1960s and 1970s (W. Davis, pers. comm.). Both the deforestation and grassy roadsides reduced the isolation of Robinson Forest and may have favored the spread of M. pennsylvanicus.
The current rarity of Synaptomys cooperi in Robinson Forest is more difficult to explain. Linzey (1983, 1984) emphasized that S. cooperi is difficult to trap; nonetheless, Barbour and Hardjasasmita (1966), W. Davis (pers. comm.) and W. McComb (pers. comm.) caught it readily in Robinson Forest and in near-by Cumberland Gap National Historical Park (Barbour et al., 1979) from the 1950s to the early 1980s. Two possible explanations for the rarity of S. cooperi include: (1) the population may be at a low point in its population cycle in Robinson Forest, or (2) Microtus pennsylvanicus may be competitively excluding S. cooperi from the forest clearings.
Linzey (1984) demonstrated in a study 170 km E of Robinson Forest that Microtus pennsylvanicus competitively excludes Synaptomys cooperi from high-quality habitat (damp areas with dense grass), forcing them to move into lower quality habitat (areas with less grasses and more herbaceous species). A similar situation may be occurring at Robinson Forest. Microtus pennsylvanicus is now present in grassy clearings once occupied by S. cooperi, and the few S. cooperi caught were in herb-dominated clearings. Barbour and Davis (1974) considered it a general phenomenon in Kentucky for S. cooperi to avoid competing with M. pennsylvanicus by occupying habitat that cannot support M. pennsylvanicus.
Establishing the relationship between Microtus pennsylvanicus and Synaptomys cooperi in Robinson Forest will require long-term trapping and/or experimental manipulations. Competitive exclusion may be operating if either S. cooperi abundance increases when M. pennsylvanicus population cycles reach low densities or if experimental manipulations of M. pennsylvanicus numbers in clearings influence S. cooperi abundance. Without such studies, we can only speculate how species that are expanding their ranges affect the small mammal communities they invade.
Acknowledgments. - We thank John Overstreet, Will Marshall and Bob Muller for providing access to Robinson Forest as well as housing and vehicles. We also thank Paul VanBooven for obtaining clearance from the Arch Mining Company to trap on reclaimed surface mines. We are grateful to Mike Martin, Deb Walker, Kim Tarter, Martina Hines, Mike Alfieri, Chris Carroll and the 1992 mammalogy class for field assistance. Will Marshall, John Overstreet, Bill McComb, Wayne Davis and Bob Frederick provided helpful information. Jeff Walck identified the plants; his efforts were crucial to this study. Finally, Wayne Davis and Ken Geluso provided helpful comments and suggestions on this manuscript.
BARBOUR, R. W. AND S. HARDJASASMITA. 1966. A preliminary list of the mammals of Robinson Forest, Breathitt County, Kentucky. Trans. Ky. Acad. Sci., 27:85-89.
----- AND W. H. DAVIS. 1974. Mammals of Kentucky. University of Kentucky Press, Lexington. 321 p.
-----, ----- AND R. A. KUEHNE. 1979. The vertebrate fauna of Cumberland Gap National Historical Park. National Parks Service Final Report. Contract C X 5000 71 232.
BRAUN, E. L. 1950. Deciduous forests of North America. Hafner, New York. 596 p.
FREDERICK, R. B., T. L. EDWARDS, D. J. PAINTER AND J. WHITAKER. 1989. Bobcat densities and population dynamics in Kentucky. Kentucky Department of Fish and Wildlife Resources FINAL REPORT: P-R Project W-45-19, Study B-R-I.
FREY, J. K. 1992. Response of mammalian faunal element to climate changes. J. Mammal., 73:43-50.
HOFFMEISTER, D. F. 1989. Mammals of Illinois. Univ. Illinois Press, Urbana. 348 p.
HOUTCOOPER, W. C. 1982. Current distribution and status of jumping mice (Zapodidae) in Kentucky. Trans. Ky. Acad. Sci., 43:97-102.
KALISZ, P. J., R. W. ZIMMERMAN AND R. N. MULLER. 1987. Root density, abundance, and distribution in the mixed mesophytic forest of eastern Kentucky. Soil Sci. Soc. Am. J., 51:220-225.
LINZEY, A. V. 1983. Synaptomys cooperi. Mamm. Species, 210:1-5.
-----. 1984. Patterns of coexistence in Synaptomys cooperi and Microtus pennsylvanicus. Ecology, 65:382-393.
MEADE, L. 1992. New distributional records for selected species of Kentucky mammals. Trans. Ky. Acad. Sci., 53:127-132.
OVERSTREET, J. C. 1989. Second-growth forest communities on the Cumberland Plateau of southeastern Kentucky. M.S. Thesis, Univ. of Kentucky, Lexington. 160 p.
ROBINSON, T. S. 1981. A contribution to the biology of the southern bog lemming in Kentucky. Trans. Ky. Acad. Sci., 42:90-94.
TRAPASSO, L. M. AND F. M. AL KOLIBI. 1994. Regional temperature trends and variations in the contiguous United States from 1935 to 1986. Trans. Ky. Acad. Sci., 55:131-138.
DAVIS, W. H. AND R. W. BARBOUR. 1979. Distributional records of some Kentucky mammals. Trans. Ky. Acad. Sci., 40:111.
S.A.S. INSTITUTE INC. 1985. SAS user's guide: statistics. Cary, North Carolina. 956 p.
|Printer friendly Cite/link Email Feedback|
|Author:||Krupa, James J.; Haskins, Kristin E.|
|Publication:||The American Midland Naturalist|
|Date:||Jan 1, 1996|
|Previous Article:||Factors affecting nest box use by Southern flying squirrels (Glaucomys volans) and gray squirrels (Sciurus carolinensis).|
|Next Article:||Sex-biased response of raccoons (Procyon lotor) to live traps.|