THz-zgm w MWWilllfii‘l'flfilliflllmiflififill a 3 1293 10641 0545 This is to certify that the thesis entitled HABITATS OF THREATENED SMALL MAMMALS ON THE ROSE LAKE WILDLIFE RESEARCH AREA, CLINTON COUNTY, MICHIGAN--EARLY FORTIES AND LATE SEVENIIES presented by Jacquelyn L. Shier has been accepted towards fulfillment of the requirements for M. S. degree in Zoology @M Major professor 0-7 639 g. lJBRAnY Michigan State University H Em3f§7§i*““ fl OVERDUE FINES: 25¢ per day per ite- RETURNI'G LIBRARY MATERIALS: Place in book return to name charge from ctrculetion records {EM 3 WE? .‘; i l 2 , a r _, ,X 5' Rpk *64399 that '-' I: i; .21 . neg I @5124. cg‘l I LL“ . l ' hr ”W878 J 3 :Jneiiiz DECOéfibflis 02 6-70 51 6 HABITATS OF THREATENED SMALL MAMMALS ON THE ROSE LAKE WILDLIFE RESEARCH AREA, CLINTON COUNTY, MICHIGAN--EARLY FORTIES AND LATE SEVENTIES By Jacquelyn L. Shier A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Zoology-T l981 ABSTRACT HABITATS 0F THREATENED SMALL MAMMALS ON THE ROSE LAKE WILDLIFE RESEARCH AREA, CLINTON COUNTY, MICHIGAN--EARLY FORTIES AND LATE SEVENTIES By Jacquelyn L. Shier In view of succession on the agricultural and forested lands of a game area in southern Michigan a follow-up investigation was conducted to determine the differences in composition and distribution within habitats between current small mammal populations and those obtained in the 19405, particularly for three threatened species. Transition from farm-game to wildlife management removed land from agricultural use. Habitat management continues with shrub strip planting, food patches, and brushpiles. Farmed areas have reverted to persistent brome grass (Bromus inermis), Oldfield situations. Introduced grasses, low plant diversity, and inappropriate vegetative structure within the oldfields are cited as responsible for the apparent absence of least shrews (Cryptotis parva) and avoidance by southern bog lemmings (Snyaptomys cooperi). Lemmings persist due to their versatile association with transition habitats. Canopy closure within woodlands reduced the abundance of ground flora such that wood- land voles (Microtus pinetorum) apparently no longer inhabit the study area . ACKNOWLEDGMENTS I wish to thank Sylvia Taylor, director of the Michigan Endangered Species Program, for making this study possible; and to John Lerg who helped in many ways with administration of the grant that financed this investigation. Other personnel of the Wildlife Division and particularly members of the Rose Lake Staff were helpful in many ways. Assistance in trapping was given by Kevin Jay in 1978, Pete Verbanac and Brent Danielson in 1979, and Kimberly Rizzo in 1980. I am grateful to have had free access to the results of other small mammal studies conducted on the Rose Lake Wildlife Research Area by Christopher Carmichael in 1978 and by Brent Danielson in 1979. John Johnson of Michigan State University generously loaned traps for predator control during the study. Special thanks to Rollin H. Baker fbr conceiving of this project and fbr encouragement and guidance throughout. Chris Faulkner generously contributed to all facets of this study. This project was supported by federal aid, the United States Fish and Wildlife Service and the Michigan Department of Natural Resources cooperating through the Federal Endangered Species Act of 1973 and the Michigan Endangered Species Act of 1974. ii LIST OF TABLES LIST OF FIGURES INTRODUCTION Background Objectives . METHODS OF STUDY Small Mammal Survey . Questionnaire . Vegetation Surveys STUDY AREA . RESULTS AND DISCUSSION . Croplands TABLE OF CONTENTS Annual Grass and Weed Associations . Perennial Grass and Weed Associations . Perennial Grasslands Mixed Perennials and Shrub Association Intolerant Trees, Shrubs, and Grass Patches : Mid—Tolerant Woodlands . Wetlands SPECIES ACCOUNTS Cryptotis parva Microtus pinetorum Synaptomys cooperi SUMMARY . RECOMMENDATIONS APPENDIX BIBLIOGRAPHY Page iv vi Table 10. ll. 12. LIST OF TABLES Monthly and annual mean temperatures recorded at East Lansing, Michigan (°F) Monthly and annual precipitation recorded at East Lansing, Michigan (in inches) . Comparison of percent land use in the central 325- hectare unit under intensive study . . Overall small mammal captures on the Rose Lake Wildlife Research Area from 1940 to 1946. Totals include all methods . . . . . . . . Overall small mammal captures on the Rose Lake Wildlife Research Area from 1978 to 1980. Totals include all methods . . . . . . . . . . . . Importance values of plants in perennial grass and weed associations in the central study area, 1979 . . Prey items identified from owl pellets collected on the Rose Lake Wildlife Research Area from February 24 to April 1, 1980 . . . Dates of collection by hand of Cryptotis parva. Entries are taken from the literature Dates of trap capture of Cryptotis parva. Entries are taken from the literature . Dates of collection of pellets containing Cryptotis. parva . . Synaptomys cooperi captured on the Rose Lake Wildlife Research Area, Clinton County, Michigan, for the years 1979 to l980--P1anted Conifer Stand-A . . . Synaptomys cooperi captured on the Rose Lake Wildlife Research Area, Clinton County, Michigan, for the years 1979 to l980--Planted Conifer Stand-B . . . . iv Page 15 I6 20 22 25 34 65 66 68 69 77 78 Table Page 13. Synaptomys cooperi captured on the Rose Lake Wildlife Research Area, Clinton County, Michigan, for the years 1979 to l980--P1anted Conifer Stand-C . . . . . . 79 l4. Synaptomys cooperi captured on the Rose Lake Wildlife Research Area, Clinton County, Michigan, for the years 1979 to l980--Conifer Hedgerow A . . . . . . . . 80 15. Synaptomys cooperi captured on the Rose Lake Wildlife Research Area, Clinton County, Michigan, for the years 1979 to 1980—-Conifer Hedgerow B . . . . . . 81 16. Population changes for resident microtines over three years . . . . . . . . . . . . . . . . . 85 17. Population levels in Synaptomys cooperi in two Michigan Counties . . . . . . . . . . . . . . . . 86 LIST OF FIGURES Figure Page 1. Vegetative cover map for the Rose Lake Wildlife Research Area in 1970 . . . . . . . . . . . . . . . ll 2. Vegetative cover map for the central 325 hectares . . 12 vi INTRODUCTION Background A major study of small mammals and their habitats in the mixed agricultural and forested lands at Rose Lake Wildlife Research Area (hereafter called the Rose Lake Area) was conducted in the early 19405 by joint investigators from the Michigan Department of Conser- vation (DNR) and Michigan State College (MSU), with published results by Joseph Linduska (1950) appearing in bulletin form. These threat- ened small mammals, the least shrew (Cryptotis parva), the southern bog lemming (Synaptomyscooperi), and woodland vole (Microtus pjne- torum),(Jones,_gt_al., 1979), were among the small insectivores and rodents found as members of the local mammalian communities. Periodic observations of small mammals on the Rose Lake Area in the past fifteen years have shown no captures or other evidence of the pres- ence of the least shrew or the woodland vole, and only an occasional record of the bog lemming. There has been no concerted and system- atic attempt to appraise or monitor small mammals at Rose Lake since the mid-19405, especially in view of successional changes in habitats. This project was, therefore, of great importance to the long-term research objectives of the DNR. Evidence concerning the reasons why these small mammals formerly seemingly common and/or widespread are now rare and/or restricted, might be revealed. Objectives It was proposed that a survey of small mammals on the Rose Lake Area be conducted systematically to determine (a) the kinds, distributions, habitat preferences, and relative abundance of the small mammals found there currently; (b) the differences in species composition between current mammalian populations and those obtained by the Linduska study in the early 19405; (c) reasons for the apparent suppressed state or likely disappearance of species on the threatened list; and (d) a plan to monitor small mammals in selected sites at Rose Lake in the future so as to maintain basic ecologic and demo- graphic data on the threatened species. METHODS OF STUDY Small Mammal Survey One objective of this study was to measure the comparative numbers ofsmall mammals in the several habitats at various times of the year, and compare them to species diversity and population levels observed in the 19405 study. True censuses of small mammals usually require intensive live-trapping on quadrats or assessment lines several acres in size (Dice, 1938; Blair, 1941; Stickel, 1946a; and Smith, et_al,, 1971). For various reasons, these types of procedures were not adopted in the earlier study by Linduska (1950). Instead, he found it desirable to sample concurrently in more situations than could be trapped con- veniently by intensive means. The approach he adopted was one of live-trapping along lines of measured length as suggested by Dice (1938) and used by Blair (1939). For the most meaningful comparison, the same methodology was used in this present follow-up study. In the earlier fieldwork, trapline placement was systematic and record keeping thorough, making it possible for the subsequent investigators to relocate every trapline. Unless otherwise indicated, all population indices were obtained by using a series of 25 traps spaced at 7-meter (22-foot) intervals along a straight line 170 meters (8 chains) long. Traplines extended from the field border toward the center and were located midway between the field edges 3 paralleling the line of traps. Fields larger than 10 hectares (25 acres) received two lines of traps, each was placed 40 meters (two chains) in from the field border which the lines paralleled. A few very long hedgerows received 50 traps when it was important to sample along their entire lengths. When sampling situations demanded the use of more, or less, than 25 traps in a line, the capture results have been standardized to a 25-trap equivalent for their inclusion in Tables 4 and 5. It was desirable to use the same treadle-operated pressed fiberboard live traps (4.5 x 4.5 x 16 cm) employed by Linduska in the 19405 study. Unfortunately, they had been discarded years ago when they became water damaged. Instead, a treadle-operated galvanized- steel live trap (6 x 9 x 27 cm) made by Leathers and Sons of Athens, Georgia, was used. Treadles could be adjusted to where traps were sufficiently sensitive to capture 8- to lO-gram animals. A few traplines were established in each habitat type (crop- land, perennial grasslands, hedgerow, and woodlot) examined by Linduska in the 19405. These were used to determine seasonal abun- dance and habitat preferences of all members of the small mammal community in light of successional changes and present land-use practices. These lines were operated for single three-night sessions in the fall of 1978, spring and fall of 1979, and in the spring of 1980. Traps were baited in late afternoon and checked and closed the following morning, for each of three successive nights of trap- ping. Those on shaded lines were also operated during the day to increase the capture rate for Synaptomys and Microtus. A dry bait combination of peanut butter and rolled oats was used in most situa- tions. The data recorded upon capturing an animal included its toe- clipped number, species, sex, reproductive condition, weight, location on the trapline, and the date. Reproductive condition was assessed in the manner described by Krebs (1966). For males, the position of the testes, whether scrotal or abdominal, was noted. For females, a combination of measures indicated the onset of reproductive activity in subadults and the stage of pregnancy in adults. These include the condition of the vagina, whether perforate or not, the size of the nipples and the surrounding subcutaneous lactation tissue, as well as degree of protrusion of the abdominal area. The degree of separation in the pubic symphysis gave evidence of recent par- turition (Hall and Newton, 1946). Although effective for microtines, deermice, and jumping mice, these measures were useful for shrews only in the most advanced stages of pregnancy or lactation. Since the major objective of this study was to determine the reasons for the apparent suppressed state or likely disappearance of the three "threatened" species of small mammals, the least shrew, the woodland vole, and the southern bog lemming, decidedly different procedures were employed in the search for these species in areas other than those surveyed by Linduska. Additional study sites having properties similar to those reported in the literature for these three species were selected for trapping using less conventional methods. Removal-trapping with Museum Special snap traps were employed for the purpose of catching "rarer" species not especially evident until the removal of the more "dominant" mammals which may initially monopolize trap sites. Pit- fall can-traps set alone and along drift fences were used to search for all three of the species. These sets were put into operation when low population levels the first year gave the impression that animals might be avoiding the conventional live and snap traps. Both Microtus and Synaptomys are reported to frequently avoid the latter (Bailey, 1929; Jackson, 1961; and Master, 1978). When the trapping success for Synaptomys with both snap and live traps improved in 1979 the can-traps were abandoned, since the yield from these proved to be lower in upland situations than with the standard traps. Examination of the 19405 trapping records and past litera- ture on Synaptomys suggested that dispersing individuals of this species may be captured almost anywhere with intensive trapping effort. However, concentrations were more often reported from bogs, riparian habitats, and upland in moist-mid-successional stages having monocotyledonous vegetation. In these situations Synaptomys was searched for, initially with snap traps. If an individual was cap- tured within the three- to six-day search period, the snap traps were replaced with live traps for additional investigation. In one pine plantation where a colony of 16 to 20 animals was located, an exhaustive analysis of their use of this habitat was performed to aid in predicting the presence of the lemming at other sites on the Research Area. Runway systems were excavated to locate food caches, nesting sites, and latrines, and to determine what locations within the grove were most heavily used. Traps for Synaptomys were vari- ously baited with apple and/or grass (after Burt, 1928) and with the peanut butter and rolled oats combination as used for the general trapping. The woodland vole was searched for by trapping in underground runways located in the humus layer of area woodlots, orchard remnants, and hedgerows of ornamental shrubs. Four woodlots were selected on the basis of herbaceous ground-layer vegetation and humus proper- ties. Those with deep, mgr humus (Jameson, 1949) and an abundance of herbs with tuberous root systems that would provide food and cover for the woodland vole (Paul, 1970) were more likely to support breed- ing populations of the species. Runways were then located and a portion large enough to accommodate a Museum Special snap trap was excavated and set with a trap, as suggested by Paul (1970). Paul also advised that these sets be covered over to block out light. Eadie (1936) and Miller (1964) found that it made little difference in the capture success. Here, initial trapping with the set covered over had little effect on what species were captured, or whether soil excavation by animals during the night covered the traps, so they were left uncovered in subsequent settings. A special search had to be made for Cryptotis since 3- to 4-gram shrews are too light to consistently trip the treadles of the large live traps and too short to be caught by Museum Special snap traps from the front. The least shrew is frequently reported to have been found beneath logs, boards, and other debris in fields of herbaceous plants or in pastures (Moseley, 1930; McCarley, 1959). Whenever such items were found on the Rose Lake Area, they were examined carefully for these animals or any leaf nests or runways of pencil diameter that would suggest their presence in the area. Fifty pieces of aluminum sheeting measuring 1 x 2 meters were also laid out in several fields having properties similar to those reported by Howell (1954) and Getz (1962) in the hope of attracting these shrews. Examination of these sheets for nests and runways was made periodically from August 1979 to June 1980. Mouse traps were set around the edge of those sheets where sign suggested the possible presence of the least shrew. Questionnaire A questionnaire was mailed to field mammalogists within the range of the southern bog lemming to collect recently gathered information on the species' local habitat use and small mammal associates (see Appendix). A comparison between pepulations in Michi- gan and those in other regions where the lemming appears to be thriv- ing will aid in determining why the species is apparently threatened in Michigan. Vegetation Surveys Vegetation types on the Rose Lake Area for the years 1941 to 1946 were accurately determined by combining references from several sources. Brief descriptions on the trap cards by Linduska provided a species list of plants present on the site and a vague idea of habitat fine structure for each of the areas he trapped. A more complete picture of habitat structure for these areas was obtained from camera point photographs periodically taken of the area from 1938 to 1960 by Rose Lake area personnel. Photographs also helped to characterize the earlier vegetative cover present on areas sampled in this study which had not been trapped in the 19405. Although the latter information was not crucial to this follow-up investigation, it helped to place the results of this recent study in better perspective. In addition, the Rose Lake Station files con- tained photographs of numerous management activities occurring on the area throughout the years, and of their subsequent effects on the habitat. Finally, personnel employed by the Station over the years were consulted for details not available from the above sources. Thorough description of the habitat was made for each of the traplines run in this recent survey. Plant species were identified according to Gleason and Cronquist (1963). Photographs were taken of noteworthy situations. In addition, a photographic record was systematically made of those portions of the area where trapping occurred, most of which correspond to regions in which the original study was conducted. Systematically recording the remainder of the Rose Lake Area on film would have been a major undertaking of limited value, as most of it has grown up to a thick tangle of shrub and understory trees. These photographs have been accessioned into the film file located at the Rose Lake Wildlife Research Station, 8562 E. Stoll Road, Bath Township, Clinton County. 10 Vegetative cover maps for the Rose Lake Wildlife Research Area in 1970 and the central 325 hectares (800 acres) where most of the trapping occurred are shown in Figure l and Figure 2, respectively. .onF cw moc< cosmmmmm ww_Fu_w3 oxmu mmom mza Low nae cm>oo o>wpnpmmo>nu.P mesmpu 1.7. v ,. a0u > m¥<._ mmOm Sulaal 11:} I Sun—ell 5.3;: 88. 1.1- ..B BEJA< mug; 9.56:, a mug n! In. '11:] um: m 1 111 L liltoulllzllr i_ u... 1.: [IS-rel " 1* (~_ I v—(., , . “traffic! ‘_‘ " >- t‘ffi-Cfl ~ ' 11mm I 1 was: COVER LEGEND Upland Woods - Cultivated Area and Food Patch Upland Brush [:1 Follow Fields and Grass Areas Lowland Woods °—°—°- Brush and Shrub Planting- Narrow Lowland Brush Evergreen Planiinqs m Marsh 5 a a Slump Piles or Rows Figure 2.--Vegetative cover map for the central 325 hectares. STUDY AREA The Rose Lake Wildlife Research Area comprises 1350 ha (3,334 ac) (rf state-owned mixed agricultural land, forests, and wetlands in Bath Township, Clinton County and Woodhull Township, Shiawassee County, Michigan (see Figure l). The study area was included in the initial purchase of land for the Research Area by the DNR in 1938. Topography varies between comparatively level and rolling countryside. The steepest slopes are wooded in most cases, but in the late 18005 early settlers cleared much of the rolling terrain for agriculture; subsequently, it became badly eroded. Early efforts (in the late 19305) by personnel at the Rose Lake Wildlife Research Station to stabilize those areas prone to erosion with shrub plant- ings have not only stopped the erosion, but accelerated the return of many areas to a wooded condition. The sandy, less productive soils were removed from production while farming was continued on the better soils (Rose Lake Wildlife Experiment Station, 1941). The amount of farming activity on the area has continued to decrease. Less than 40 ha (99 ac) are currently farmed by share- croppers and Rose Lake personnel. Before the land was cleared, the uplands were covered by beech (Fagys grandifolia), maple (Acer spp.), hickory (Carya 5pp.), and oak (Quercus 5pp.) forests with a few scattered white pines 13 l4 (Pinus strobus). Lowland forests were dominated by maple, ash (Fraxinus spp.), and tamarack (Larix laricina). The soil pattern is complex as is characteristic of heavily glaciated areas where there is a wide variation in drainage condi- tions. Twenty-seven soil types were recorded for the Area in the 19405. In the well-drained uplands were Newaygo loam, Bellefontaine sandy loam, and Coloma loamy sand. The poorly drained soils of kettle-holes, marshes, and lake and stream bottoms were predominantly Rifle and Greenwood peats, and Kerston and Arlisle mucks. A slightly different soil profile was reported for the 19805. Well-drained soils include Boyer sandy loam, Marlette loam, and Spinks loamy sand. The poorly drained soils include Adrian muck, Carlisle muck, Gilford sandy loam, Houghton muck, Tawas muck, and Wasepi sandy loam (Threlkeld and Feenstra, 1974; and Pregitzer, 1978). Mean winter (October-March) temperature is 0.4°C (32.7°F) while mean summer (April-September) temperature is 16.7°C (62.1°F). Periods of extremely hot or cold weather are usually of short dura- tion. Precipitation averages 87.5 cm (34.4 in) per year. The pre- vailing winds are from the west with an average speed of 16.6 km/hr (10.3 mi/hr) (U.S. Weather Bureau, pers. comm.). Details of meteorological conditions during both periods of investigation are given in Tables 1 and 2. During the 19405 study period, both snow and rainfall were above normal. For the recent study period, annual precipitation was 4 to 5 inches below the normal 31.6 inches, despite heavy snowstorms in .cmucmu :ogmmmmm wczppzowuco: :mz soc mmmngcoz .cmmwzowz .mpmo Pmuwmopoumewpu Eocm ommpuwnmp 15 .ommp .mxmsvzw4 Eocc o¢m_-oem_ "mocsom w.c¢ N.mm m.mm m.om ¢.Fo m.mo m.o~ ¢.mm m.om o.m¢ N.Nm m.- ¢.N~ peacoz o.~m _.mm ¢.¢e m.m~ w.m_ m.mm owmp o.o¢ m.m~ ¢.wm m.m¢ N.No m.mm o.mm m.um m.om e.e¢ m.~m F.ep m.mp mmmp o.m¢ m.om ~.mm m.m¢ m.¢m ¢.mo o.mo P.oo m.mm _.¢¢ m.m~ m.ep P.m_ mump m.m¢ N.om o.Pe “.mm m.mm m.mm m.o~ o.mo m.em m.oe m.e¢ o.e~ ¢.m~ mvmp e.o¢ m.- m.mm m.~¢ m.om n.5m c.5c o.~o ~.om e.me e.o¢ e.m~ N.~P memp m.~¢ m.F~ o.mm ~.w¢ P.No o.Pn v.0“ e.wm m.oo m.P¢ w.mm m.m~ N.m~ camp o.me ¢.mm N.mm ¢.w¢ m.~m o.mo ¢.ou o.mo w.¢m ~.p¢ w.om m.mm «.mp memp m.n¢ o.m~ m.wm o.om «.mm m.~o N.o~ o.no o.mm ¢.Fm m.mm m.o~ «.mm Nemp P.mv e.¢m N.pe m.~m e.em «.mm m.FN F.mm o.om m.Fm m.m~ ¢.m~ m.¢~ Pemp m.me N.om o.om m.om ¢.om m.mo N.~N «.mo w.¢m F.~e F.n~ o.¢m o.mp oemp szcc< .omo .>oz .uoo .pamm amzma< x—za mesa Am: pwca< :ocmz .aod .cmw ALOV cmmcgucz .mcwmcm; pmom um cmucoomc mmcaumcwasou cams Fmaccm ucm apcpcozuu.p m4moz .uoo .uamm pm=m=< apza mczw he: FPLQ< socmz .nmm .cmn Ammgocw cwv campsowz .mcwmcmA pmmm pm emucoowc cowpmuwnwomca pmzccm use ngucoz--.~ ugm S a; I; 3 «a Land- '— 3 xesugmra o 3.5- ; ”5e teases“. s 2%.»; .9... a: 5‘” m8 .2“ a .2.= ele Uplands Crapland No. of (106 lines) Individuals 0 0 7 0 11 301 0 40 19.2 No. per Trapline O 0 0.07 0 0.10 2.8 0 0.37 Annual grasses No. of and weeds Individuals 0 0 22 0 0 77 0 17 9.4 (24 lines) No. of Traplines 0 0 0.90 0 0 3.36 0 0.70 Perennial grasses No. of and weeds Individuals 0 2 5 1 0 7 0 1 20.0 (8 lines) No. of Trapline 0 0.25 0.63 0.13 0 0.88 0 0.13 Perennial grasses No. of (6 lines) Individuals 0 1 5 0 0 18 O 2 3.9 No. of Traplines O 0.16 0.85 0 0 3.0 0 0.3 Mixed perennials No. of and shrubs Individuals 3 17 27 20 7 32 0 6 18.3 (22 lines) No. of Trapline 0.13 0.76 1.21 0.89 0.31 1.43 0 0.27 Intolerant No. of Trees/shrubs Individuals 0 3 O O 5 0 0 0 2.9 rass patches 7 lines) No. of Trapline 0 0.43 0 0 0.75 0 0 0 Mid-tolerant No. of trees/shrubs Individuals 0 52 0 2 142 0 O 2 10.2 (44 lines) No. of Traplines 0 1.18 0 0.05 3.22 O 0 1.0 Wetlands Grass-sedge No. of marshes Individuals 3 0 4 0 0 0 0 0 6.5 (3 lines) No. per Trapline 1.0 0 1.3 0 0 0 0 0 23 greatly outnumbered the deermouse. In contrast, fields where corn was stored as shocks during the winter were favored as'winter refuge by most of the local small mammals, including the three species that are presently considered threatened. Linduska (1950) actually found as many individuals of these latter three species in one winter by censusing under corn shocks as he did in all his years of trapping. Seventies Of the land previously used for the growing of grains, 70 percent was retired from cultivation by the late 19605. Since then, these lands have advanced to the perennial grass and weed stage. The 30 percent remaining is planted only with corn on a share crop basis for a program which provides for 25 percent of the crop to be left standing for wildlife food and cover. Annuals and biennials thrive between the rows of corn. When these fields are left fallow, even for one season, the foxtails (Setaria 5pp.), pokeweed (Phytolacca americana), and Lamb's quarters (Chenopodium album) invade the field en masse. Today, corn shocking is no longer practiced. Instead, the whole stalk is cut and removed from the field in the fall for silage. In addition, only a small portion of each field of corn is left standing over the winter for wildlife. Such a practice may still allow mice to feed on the fruit in winter, however, its value as cover when standing, compared to when it is in shocks, is greatly reduced. This fact alone probably diminishes the value of cornfields to small mammals in winter compared to what it was in the forties. 24 In this recent investigation only prairie deermice were cap- tured in croplands during the winter. This evidence, although gathered from a limited amount of winter trapping in January 1979, indicates that cornfields harvested by modern methods may not provide sufficient cover for meadow voles or the rest of the local small mammals as they did in Linduska's day. This also means that winter censusing for any of the threatened small mammals should now be conducted in other suitable cover types, rather than in cornfields. The principal species in croplands today are still the prairie deermouse and the house mouse (see Table 5). An average catch of 3.8 deermice and 0.77 house mice per trapline was obtained. Since the prairie deermouse is a species which often shows distinct fluctuations in population levels over a period of years (Blair, 1940; Linduska, 1950; and Terman, 1966), this figure in no way indi- cates the presence of a higher mean population level than existed in the forties. In fact, Linduska observed a marked increase in live trap catches in each succeeding year of study. This trend started with a low of 0.6 individuals per trapline in July 1940 and ended with a high of 4.3 per line in March 1946. Although no data were reported for 1943 to 1945, Station records show that eight lines set in craplands in 1944 caught an average of 3.9 individuals per line. Since this value lies between those found for 1942 and 1946, it gives credibility to the uniform increase inferred by his graph, rather than a crash, during the intervening years. Trapping during the fall of 1978 in cornfields yielded a catch of 5.6 prairie deermice per line. By the following April a 25 TABLE 5.--Overall small mammal captures on the Rose Lake Wildlife Research Area from 1978 to 1980. Totals include all methods is U a E m Percent s .‘23 s: s 35 3 s m" a; so u a: TE .- Use all; 1:; 8F: 8 §8 §E 3 LC ‘0) U: 0 3 mm 8'6 3* s as 8‘” 5‘4 3|: Uplands Cropland No. of (21 lines) Individuals 0 1 l 0 0 58 0 10 5.6% No. per Trapline O 0.04 0.04 0 0 3.8 0 0.77 Annual grasses No. of and weeds Individuals -- -- -- -- -- * * * 0.18% (sightings‘) No. per Trapline Perennial grasses No. of and weeds Individuals 2 64 391 0 16 26 52 3 14.5% (36 lines) No. per Trapline 0.05 4.8 22.4 0 0.44 0.85 3.94 0.07 Perennial rasses No. of (389 lines? Individuals 3 273 553 0 101 13 91 1 28.5% No. per Trapline 0.05 1.10 6.85 0 0.42 0.37 0.77 0.08 Mixed perennials No. of and shrubs Individuals 0 5 5 0 9 4 1 0 1.0% (4 lines) No. per Trapline 0 0.87 0.4 0 2.25 0.6 0.25 0 Intolerant No. of Trees/shrubs Individuals 11 71 102 6 73 3 14 0 11.6% rass patches I15 lines) No. per Trapline 0.42 1.58 2.23 0.21 1.83 0.09 0.59 0 Mid-tolerant No. of trees/shrubs Individuals 6 45 7 19 72 5 8 0 26-0% (50 lines) No. per Wetlands Trapline 0.12 0.90 0.14 0.38 1.44 0.10 0.16 0 Grass-sedge No. of marshes Individuals 24 0 129 0 3 0 10 0 12.2% (28 lines) N0. per Trapline 1.81 0 4.82 0 0.04 0 0.93 0 26 catch of 0.82 per line, compared to the 3.6 obtained by Linduska in March of 1944, gave every indication that a crash in their population level had occurred. However, the population in cr0p1ands had increased to 6.3 individuals per trapline by June. Apparently, the cornfields are somewhat inhospitable to deermice during the wet spring when the fields are saturated with meltwater. It is probable that the summer population consisted of individuals which had dis- persed into the cornfields after they had dried out a bit. There was no sharecropping on the Rose Lake Area in 1980. Only a few small corn patches were planted by station personnel that spring. Should this trend continue, it will probably have little direct effect on the three threatened species of small mammals under consideration here, since cornfields were of fairly limited value to these species by 1980. Their movements into corn shocks in the winter of 1941 is viewed as having been a preferential act engaged in by these animals, rather than a necessary one stemming from any lack of sufficient cover in their summer range or in alternative ranges. Annual Grass and Weed Associations First-year hayfields of alfalfa and red clover, and grain fields in the late fall period (cornfields more than other grains), are similar in many respects to the annual grass and weed stage in succession from plowed ground which Allen (1938) describes. What is bare ground in spring closes in quickly by summer with the matura- tion of cultivated plants and weed species. Consequently, classifi- cation of cover type relative to what is significant to the small 27 mammal community is complicated by drastic seasonal changes in the habitat. For the purpose of tabulating data in Tables 4 and 5, percent land-use was based exclusively on spring cover maps. However, the assignment of individual traplines to the various seral stages was determined by the condition of existing cover each time the trapline was operated. Forties Newly planted hayfields, cornfields in the fall, and retired croplands contributed to the annual grass and weed cover type in the 19405. At that time, hay production and crop rotation placed an average of 30 hectares per year into this early stage of succession. Most of these autumn cornfields were occupied by common ragweed (Ambrosia artemisifolia), yellow and green foxtails (Setaria glauca and S. viridis), lamb's quarters, knotweed (Polygonum aviculare), and in lesser numbers by other annuals. By the following spring the dominant vegetation in these same cornfields consisted of early—stage perennial grasses and weeds such as quack grass (Agropyron repens), broom sedge (Andropogon virginicus), and wormwood (Artemisia campestris), red sorrel (Rumex Acetosella) and pokeweed. Seventies Because of limited tilling or use of herbicides during the present investigation, annual grass and weed associations developed within the cornfields. The plant species diversity closely matched that described by Linduska (1950) for the forties. Sampling for small mammals occurred here in the fall of each year. 28 Also supporting annual grasses and weeds were strips of land 10 meters wide by 100 to 1000 meters long that were plowed for the planting of shrub hedgerows along the roadways in 1978. Subsequent colonization by annual grass and weed species,gradually increased the cover available. By 1980 these strips were being taken over by perennial grasses and weeds. Now, as in the 19405, these recently disturbed upland habitats appear to support only the prairie deermouse and the house mouse, and infrequently, the meadow jumping mouse (Zapus hudsonius). In late fall after harvest and in spring before maturation of the planted species only the prairie deermouse inhabitats open farmland. Accom- panying the summer transition in abundance of cover there is an influx of the house mouse into croplands when adequate shelter becomes available. Both Linduska (1950) and Whitaker (1967) docu- mented these moves by house mice, and they were observed again in this follow-up survey. In fact, 22 of the 57 house mice captured by Linduska in cr0plands and annual grass and weed associations were taken after August 1 when the ground cover had matured, and all were captured after July 8. Likewise, the ten house mice captured in the seventies from cornfields were taken after July 24. Although these annual grass and weed associations support a modest growth of monocotyledonous vegetation and some species of arthropods, they are unlikely fall and winter refugia for either the least shrew, the southern bog lemming, or the woodland vole. The woodland vole has only been reported from such situations when tuberous root crops are present (Linduska, 1950; Paul, 1970). 29 There is not enough litter accumulated or high enough stem density for either the bog lemming or the more common meadow vole to estab- lish the runways they usually construct to move, under cover, between foraging sites and nests. Whether they forage here to any extent in the winter was not determined in this present investigation. However, Linduska (1942a) did observe a significant movement of voles into cornfields once appropriate cover existed in the form of field- shocked corn and a moderate accumulation of snow. During the spring, summer, and early fall, he caught only 7 meadow voles from 106 trap- lines (an average of 0.07 voles per line), and no bog lemmings or woodland voles, in unaltered cornfields. By January a seven-day session of removing rodents from 28 acres of field-shocked corn yielded 166 voles (the equivalent of 0.8 voles per line), several woodland voles, and a few Synaptomys. And there was not enough litter or debris accumulated in these recently disturbed fields for the least shrew to build nests in such "stable" sites as seem to be preferred by this species (Whitaker, 1974). Perennial Grass and Weed Associations The transient annual-biennial stage on both cultivated and small grain lands is followed by a longer period in which perennial grasses and weeds predominate. This usually begins about three years after abandonment. Canada bluegrass (Poa compressa) and Kentucky bluegrass (Poa pratensis) are of particular importance in southern Ml°¢higan. Either one or both of these species may completely cover the .surface of the ground over large areas; Canada bluegrass is 30 likely to occur on the heavier soil types, whereas Kentucky bluegrass tends to be more common on the lighter upland soils. Many of the weedy associates have long tap roots that allow them to co-exist with the grasses which use mainly the upper soil levels. Forties In 1940, 20 percent of the central area was maintained as a perennial grass and weed association for the pasturing of a dairy herd. The dominant grass species were Canada bluegrass and Kentucky bluegrass, the latter was referred to as June grass in notes by Linduska. Considerable variation existed between fields with respect to species diversity and abundance of the herbaceous types mixed with these grasses. Among them were goldenrods (Solidago spp.), milk- weeds (Asclepias spp.), yarrow (Achillea Millefolium), common mullein (Verbascum Thapsus), sour dock (Rumex crispus), fleabane (Erigeron ppppg), flowering spurge (Euforbia corollata), and wormwood. Second- and third-year hayfields are included in this cate- gory because the associated weeds were perennial and the litter layer had developed considerably by this time. Oldfields of this early seral stage contained a slightly higher diversity of small mammal life than did either the croplands or annual grasslands (see Table 4). However, the overall density of small mammals was substantially lower. Linduska captured only 0.88 prairie deermice and 0.63 meadow voles per trapline. The short- tailed shrew first appears in this seral stage where there is enough ground litter to provide both cover and abundant food of the types 31 usually selected by the shrew. The diet of the short-tailed shrew in Indiana consists of 36 percent earthworms, 9 percent slugs and snails, 8 percent lepidopterous larvae, crickets, centipedes and beetles (Whitaker, 1972); all of these are more abundant where the ground litter is substantial, or where a moderate diversity of flower- ing herbaceous plants occur. Both the meadow vole and the short-tailed shrew were strictly fall and winter residents. Linduska (1950) observed an annual migration in September by the meadow vole and the southern bog lemming (both from wetlands) and the woodland vole (from the woodlots) into these upland grass- lands. Slightly fewer had settled here than had invaded the crop- lands. Summer trapping failed to take any bog lemmings or woodland voles, and relatively few meadow voles were captured in these areas in summer. Yet, from December to February of 1942, both bog lemmings and meadow voles were common in grasslands. Linduska believed that microtines avoided these areas in summer because most of the vegeta- tion was either severely dehydrated or grazed short by the livestock. When these rodents finally moved upland in the fall, cooler weather had allowed rehydration of the grasses and there was a considerable amount of new fall sprouts for food. The winter of 1942 was the beginning of an upward population trend for the local bog lemmings which appeared to peak in 1946, although sparse data from 1943 to 1946 during the war makes this conclusion tentative. In December of 1942 lemmings were especially 32 numerous in a 1.5 hectare opening beside a stream bottom grown up to St. John's wort (Hypericum perforatum), dewberry (Rubus hispidus), wormwood, milkweed,nullein,ragweed, and bluegrass. In this area during mid-day these animals could be seen running along the snow surface as they moved between tunnels. Since winter trapping under snow cover was conducted only within pine hedgerows, cornfields, and along the edges of vernal ponds, it is unknown whether current peren- nial grass and weed associations serve as winter refugia for either the lemming or the woodland vole. The only locality, other than field-shocked corn in winter, from which Linduska (1950) reported capturing a least shrew was an old field-bog margin, on 2 October 1941. Although the site bordered on an extensive acreage of old field and pastureland, no additional captures were made during regular trapping operations. Several were taken later in February from beneath corn shocks only 400 meters from the bog margin. Seventies The abundance of habitat in the perennial grass and weed association has decreased only slightly, from 20 to 14.5 percent, over the intervening years. However, major differences in success- ional stage and plant species composition exist, and are probably responsible for the dramatic differences seen in the accompanying small mammal community. Prior to their abandonment in 1967 when the dairy herd was sold, hay and pasturelands were subject to a major change in principle 33 grass species. Native bluegrasses were replaced by smooth brome grass imported from Europe for hay production. The lands presently in the perennial grass and weed stage derive from lands abandoned from grain cultivation which were subsequently invaded by brome grass and herbaceous weed species from the pasturelands. Free from cultivation for a considerable period of time, these tracts have reached a more advanced seral stage than had the perennial grass and weed associations that Linduska had available to sample. While the weeds were dominant on his frequently cultivated hay and pasturelands, long-term abandonment of today's grass and weed fields has allowed a thickening of the sod and the grasses dominate. Consequently, the weedy associates are different and fewer in number. The relative importance values of the grass and weed species found in two abandoned fields in 1979 are shown in Table 6. When one looks at which cfldfields are inhabited by the brome grass, it becomes apparent that some prior clearing of the land was necessary for its successful invasion. Hence, it dominates on the cultivated portion of the Rose Lake Area. Here it established ini- tially by direct seeding onto tilled lands, and later it spread to nearly all lands left fallow. The more mature perennial grass and weed associations of the seventies appear to support a more stable small mammal community. Shrews and microtines are year-round residents and occur in much greater numbers than in the younger, and drier, grasslands of the forties (see Table 5). The short-tailed shrew has increased from 0.25 34 TABLE 6.--Importance values of plants in perennial grass and weed associations in the central study area, 1979 SPECIES Importance Value Grid A Bromus inermis 168.27 Lychnis alba 5.96 Daucus carota 5.88 Berteroa incana 5.82 Solidago canadensis 4.96 Xanthium strumarium 3.18 Agropyron repens 2.92 Solidago graminifolia 1.55 Trifolium pratense 1.46 Grid 8 Bromus inermis 128.92 Daucus Carota 22.54 Medicago sativa 11.57 Taraxacum officianale 6.36 Solidago canadensis 5.92 Solidago graminifolia 5.64 Cirsium arvense 5.63 Bertaroa incana 4.52 Agropyron repens 2.69 Lychnis alba 2.64 Xanthium strumarium 1.84 Convolvulus arvensis 1.74 35 to 4.8 individuals per trapline and the meadow vole has increased from 0.63 to a phenomenal 22.4 mice per line of 25 traps. The meadow vole still exhibits a multi-annual fluctuation in population density. Since the value of 22.4 was obtained primarily from trapping in the fall of 1980 during a p0pulation high, it probably falls well on the high side of the actual average for the entire study period. More realistic as an average value is that of 6.8 per line given for the perennial grasslands. There has been no significant change in either prairie deer- moust or house mouse populations in this habitat type. In addition, the more mature fields of today support a modest 0.44 white-footed mice. The apparent increase in meadow jumping mice from none to 3.9 per trapline may be, in part, an artifact of the difference in the trap type used in the two studies. Linduska (1950) mentions that, even though only one individual was taken by trapping during the course of his work, at least a dozen were removed from post holes that were dug near a marshy area. His livetraps were much smaller than mine; measuring only 4.5 x 4.5 x 16 cm. It is plausible that a trap this small might be avoided by the jumping mouse, for they readily entered the livetraps used here. They also were caught frequently in snap traps in the present study. However, a comparison of snap-trapping results for the jumping mouse is not possible since Linduska's use of these traps was almost exclusively confined to grain fields where no jumping mice were captured, and to winter grasslands when these mice would be hibernating. 36 These perennial grass and weed associations are the first stage in the seral development from cropland to forest where condi- tions supposedly would be suitable for the least shrew. There is a great diversity of broad-leaved herbaceous plants and enough stability of the plant community to support a varied insect diet for the least shrew. Whitaker and Mumford (1972) examined stomach contents which primarily included lepidopterous larvae, earthworms, spiders, orthopterans, and aphids. There also appears to be sufficient sta- bility in the vegetative structure to provide suitable nest sites, and there is lodged grass for winter refuge (Davis and Joeris, 1945). Whether these factors are sufficient to attract the least shrew if it were presently on the area is unknown. The inferences drawn here about potential habitat selection, foraging sites, and trappability mostly derive from publications of minor observations made while the investigators were engaged in research on other species, and some come from incidental encounters with the shrew. This wide range of observations is summarized by Whitaker (1974), but even this large body of literature does not give us a very clear picture of the ecology of this species. Consequently, the previous statements about likelihood of current residency or habitat use by the least shrew on the Rose Lake Area are a "best guess" based on an overall considera- tion of this body of data. 37 Perennial Grasslands Forties In 1940 only 4 percent of the central area possessed a thick enough sod ground cover to have excluded those weeds that were pres- ent in the younger grasslands. These tracts are considered here in the perennial grassland stage as a midpoint in the seral development of plowed ground to woodland. These areas were also used for the pasturing of the dairy herd. Canada bluegrass and Kentucky bluegrass were the dominant grass species, forming dense uniform stands with timothy (Phleum pratense). The principal species of small mammal captured by Linduska in bluegrass sod were the prairie deermouse, house mouse, and the meadow vole. In all seasons he took a number equal to that collected from adjacent croplands (see Table 4) of both the deermouse (3.0 per line) and the house mouse (0.3 per line). Subsequent to the movement of microtines to upland habitats in the fall, he captured an average of 0.8 meadow voles per trapline. Pasturelands and other sod areas were attractive to bog lemmings as winter refuge, just as some perennial grass and weed associations sustained winter colonies. In February 1942, 50 snap traps were operated a total of four nights at the edge of a bluegrass pasture and 18 lemmings were taken from less than half an acre. Seventies All the grasslands retired from pasturing the dairy herd since the introduction of brome grass onto the Rose Lake Area in 1950 38 eventually became occupied by brome. In comparison to those grass- lands that were abandoned while under cultivation for hay, these pastures have had a longer period of time without disturbance to the sod layer. Hence, today these fields are essentially monotypic, and very stable in this state of succession. Their lack of plant divers- ity discourages seed-eating birds and mammals from inhabiting these fields, and consequently the introduction of propagules from other plant species is slow. The few associates found with brome grass in these mature fields were goldenrods and soapwort (Saponaria officinalis). The transition of perennial grasslands to a woody condition is further curtailed by the apparent resistence the growth form of brome grass has to the germination of other species. Colonization by woody species occurred in one field only, where the grade was sufficiently steep to have allowed a mild erosion of the topsoil and a thinning of the grasses there. Although the sod areas have been undisturbed by man for at least 15 years, they can be expected to remain this open for another 15 years, and probably longer (Bump, 1950). In a study similar to this one Bump found that, even though apen fields were located a short distance from existing seed sources, colonization was much less rapid on sites that supported a particu- larly lush growth of grass and other herbs. In his study area, several such sites remained quite open even though untouched by man for 30 years or more. The perennial grass association now occupies 28.5 percent of the central 325 hectare study area (see Table 5) and was the 39 predominant habitat type during the present investigation. Even more importantly, the overwhelming success of smooth brome grass has sig- nificantly altered the patterns of seasonal distribution and abun- dance in the grassland small-mammal assemblage. The widespread establishment of monotypic brome grass cover has substantially reduced the amount of grassland habitat that will support the prairie deermouse. In the forties, Linduska caught an average of 3.0 deermice per trapline in bluegrass sod, whereas in the seventies only 0.37 per line were taken. Closer examination of the seventies data revealsthat captures of deermice within brome grass- lands occurred only in early seral stage grasslands directly adjacent to croplands. Here food was readily available from both the mature crop plants and from weeds that grew among the rows, and abundant cover could be had within the grasslands. The discovery of several occupied deermouse nests within the adjacent brome field verified the use of grassland for cover in these situations. While reduced in value to the prairie deermouse, mature perennial grasslands were more attractive to the meadow vole and the short-tailed shrew. Linduska obtained an average of only 0.85 meadow voles in the few sod fields that there were to examine, whereas extensive trapping in the seventies netted an average of 6.8 voles per trapline. With the dramatic population fluctuations exhibited by this species, trapping results varied from an average of 1-2 per trapline during a papulation low to a high of 22 or more per line. More significant is the fact that mature brome grass 40 fields support populations of voles even in the driest years. Granted, populations also became higher in wetlands when grasslands were very dry in 1978, but the upland fields always had viable breeding p0pula- tions, numbering about 20 per hectare. There was no noticeable evi- dence of an annual upland migration of microtines in the fall like those that Linduska observed. Or if they did occur, then the grass- lands are now extensive enough to absorb these individuals into wide- spread resident population with little effect on local numbers. Whatever the case may be, the grasslands were not completely aban- doned during the hot summer. That moisture retention by brome grass fields is probably greater than in the bluegrass fields in summer, and therefore, has less affect on annual movements by meadow voles, was discussed in the section on perennial grass and weed associations. The greater moisture retention of brome grass cover seems to have had a different, but equally significant, affect on the short- tailed shrew. In both brome and weed fields, and in brome alone, Blarina reached densities far greater (4.8 and 1.1 shrews per trap- line, respectively) than what had been observed by Linduska for bluegrass fields (0.25 and 0.16). It is likely that the greater moisture content of brome sod has enhanced the soil environment of earthworms and other arthropods on which the short-tailed shrew feeds. All of the other local small mammal species were secured from brome habitat type except for the southern bog lemming. Although mature grasslands are not a habitat frequented by the house mouse, it occasionally appears in these situations. This 41 may depend more on the nature of the surrounding habitat than on characteristics of the grassland through which it is probably passing to reach more favorable areas. Whitaker (1967) suggests that the combination of its foraging behavior and its desire for cover restrict it to mature cultivated locations. Since these locations usually prevail for only a short period of time, the house mouse leads a nomadic existence in search of suitable habitat. Given this, the decline from 0.3 individuals per trapline in the forties to 0.08 in the seventies probably reflects the reduction in cultivated and annual grass and weed associations that these animals would be moving between. Undoubtedly the same situation pertains for the woody hedgerows and forest margins from which house mice were captured in the forties, but not in the recent surveys (see Tables 4 and 5). The seed-eating meadow jumping mouse was much less abundant in the monotypic grass associations than in those having both grass and weeds. It was captured at the rates of 0.77 and 3.94 per trap- line, respectively. This difference may, in part, be due to reduced seed abundance. However, for the other seed-eaters, the insignifi- cant difference in captures of the white-footed mouse and house mouse, and the small reduction in capture rate for prairie deermice suggests that another habitat requirement of the jumping mouse may not have been met. A higher stem density in the perennial grasslands was an obvious situation that may make travel in these fields difficult for the jumping mouse. 42 Probable reason for why the meadow jumping mouse was not caught in the 19405 investigation was discussed in the section on perennial grasses and weeds. The white-footed mouse is a seed-eater of late Oldfield succession and of woody habitats. Yet trapping in both perennial grass and weed associations and perennial grasslands yielded a modest catch of 0.4 mice per trapline. The distribution of captures with respect to trapline placement revealed that these captures were con- fined to traps in close proximity to bordering hedgerows. Grasslands should, therefore, be considered marginal habitat for the white- footed mouse, and the extent of their utilization of these habitats be considered dependent on the management of directly adjacent tracts. Smooth brome grass fields in various stages of succession now cover 43 percent of the 325-hectare study area. Such dominance attests to the success this grass species has had in establishing in the more fertile lands abandoned following farming operations. The brome grass has not, however, spread into the bluegrass-weed fields that developed on the infertile sandy soils before the introduction of brome grass. There is apparently little threat of this grass spreading into the native grasslands that developed as a seral stage following the lumbering in southern Lower Michigan. However, its dominance in abandoned farmland and on highway rights-of—way (where it has spread following construction or been sodded in) seems to have been of great consequence to the grassland small mammal community in Lower Michigan. 43 Mixed Perennials and Shrub Association Perennial forbs are replaced by a number of woody plants that gradually become of increasingly greater importance in the vegeta- tion. Shrubs seldom completely occupy an area to the exclusion of all other types of plants. Usually they are present in aggregations, with perennial grasses and forbs in the open spaces between and a number of trees scattered over the area. However, group plantings for wildlife management result in more discrete, monotypic stands or rows of shrubs. Forties In his study of field borders in relation to farming prac- tices, Linduska (1950) examined eight sod fencerows for comparison with brush-tree borders. These sod borders were grown up principally to stands of bluegrass (Poa pratensis and P. compressa) with a variety of vines and shrubs providing a secondary vegetative structure absent from the open grasslands. Numerous shrub plantings for wild- life food cover and for erosion control placed scattered patches directly into this advanced oldfield seral stage. Small mammal populations in these transition type communities had lower densities than did those in adjacent croplands during the same period. These two habitat types showed parallel seasonal vari- ation in the relative abundance of the resident species. During early August of 1942, four of these sod borders yielded 11 individuals including 6 prairie deermice, 2 meadow voles, 2 bag lemmings and l short-tailed shrew. In late October and early November of that year 44 four sod borders, including two examined before and two new ones off the area, were trapped again. Of 77 small mammals captured in this period, 50 were meadow voles, 19 were short-tailed shrews, 3 were bog lemmings, 3 were prairie deermice and 2 were house mice. Concurrent October and November trapping in craplands resulted in a higher catch of prairie deermice but a far lower catch of meadow voles. During April 1946, a few border habitats were sampled once more. In this trapping ten lines took a total of 58 individuals, which included 21 meadow voles, l7 bog lemmings, 8 short-tailed shrews, 6 house mice, 3 prairie deermice, and 3 masked shrews. Several of these lines also appear to have been operated off the area in nearby farmland. Considering only results from traplines Operated on the central research area, these semi-open oldfields and border strips supported the most diverse small mammal assemblage of all cover types examined. The observed small mammal species also were present in more nearly equal numbers than in other habitats (see Table 4). In general, transition areas like grass-shrub fields support a mix of vegetation types and structural profiles. By this very fact one would expect that some individuals of species normally found within areas supporting the various components of the mix might venture into this border habitat, with some taking up permanent residence. In addition, lands with border vegetation frequently act as dispersal routes between major habitats. Both of the above circumstances would promote greater species diversity than that found in the surrounding major cover types, as was found in this study. However, more frequent and long term trapping would have been necessary to determine the 45 real extent to which any such dispersing non-residents may have con- tributed to the net catch here. The predominance of grass cover over shrubs in these units still favors the prairie deermouse. The capture rate for the prairie deermouse was 1.43 individuals per trapline and for the white-footed mouse was 0.31 per line. This di5parity existed whether in semi-open oldfields, (1.41 and 0.08 mice, respectively) or in sod strips border- ing croplands (1.75 and 0.43). The mixed perennial and shrub association was the only upland cover type from which the masked shrew was secured in the forties. Only three individuals were captured in regular trapping. However, six other of these shrews were trapped in November and December and two additional ones in March during an eight-month study on the homing ability of prairie deermice in oldfields over the winter of 1942-43. The short-tailed shrew was also more abundant in bluegrass sod borders than in the more open grasslands or croplands. The over- all distribution of the short-tailed shrew was found by both Linduska (1950) and Dambach (1944) to parallel the location of thick ground litter and loose soil. Populations were much higher in ungrazed woodlands and in field border situation than they were in similar unites where grazing occurred. In explanation of these population differences Dambach stated that invertebrates in an ungrazed tract were two and one-halftimesas plentiful as in a grazed area and that freezing of the soil is much less severe in ungrazed woodlands. With reference to this point, a comparison of short-tailed shrew 46 abundance in perennial grass and weed associations, in perennial grasslands, and in intolerant tree, shrub and grass patch regions upholds this conclusion. Perennial grass and weed fields of the forties were young hayfields where the soil had been tilled in the last three to four years. Trapping here averaged 0.25 shrews per line compared to 4.8 in grasslands that had been fallow for up to ten years. Perennial grasslands in the forties were grazed or recently grazed pasturelands. Only 0.16 shrews were captured per trapline compared to 1.10 obtained from grasslands retired from grazing for nearly 15 years. In the brushy roadside and fencerow borders supporting vege- tation in the intolerant tree and shrub stage, the vegetation had closed in to such a degree that little or no herbaceous undergrowth was present. However, most of the border vegetation from this cate— gory examined in the seventies supported a light growth of grasses and herbs amidst a thick accumulation of litter. Trapping in these two situations yielded 0.43 shrews per line in the forties compared with 1.58 in the present investigation. In contrast, although fields supporting mixed perennials and shrubs were of somewhat different seral stages in the two time per- iods, both were undisturbed areas with a substantial accumulation of litter. Here the relative catch was more nearly equal, with 0.76 shrews per trapline in the forties and 0.87 in the seventies. The differences in number per trapline may not seem that great compared to differences seen for other species. However, 47 considering that captures of shrews are relatively rare events and that individuals are rarely recaptured in subsequent efforts due to high mortality when confined in a trap, the numbers are rather convincing. In Linduska's survey, the southern bog lemming was most often captured in early-stage transition habitats. He obtained an average of 0.89 lemmings per trapline. As was the case with other grassland habitats, there were seasonal differences in residency. Lemmings and voles avoided these areas in the hot summer months, preferring the wetland habitats instead. Bluegrass sod hedgerows and border strips were the first to contain bog lemmings at the onset of the fall upland migration. From there they appeared to move out into shrubby oldfields (0.25 per line, 24 lines), stream borders (observed 18 on a winter day at noon), and isolated grass and weed patches having shrubby borders (7.5 per line, 0.4 lines). However, the hedgerows still maintained relatively high, (0.78 per line, 16 lines) and widespread (12 out of 16 trap- lines caught lemmings) populations. Linduska's survey of transition habitat concentrated on areas with young shrub and conifer plantings from two to four years old that were recently planted for wildlife, and on borderlands with young shrubs or vines. The recent survey was not to replicate the earlier study, but to examine how the maturation of these same lands has influenced small mammal species diversity and relative abundance. Consequently, trapping efforts on lands supporting mixed perennials 48 and shrubs were kept to a minimum, both by design and by scarcity of qualifying habitats (less than 1 percent). Seventies The hedgerows surveyed in the recent investigation were not very comparable to those just discussed above. They contained rela- tively fewer seed- and fruit-bearing species and the shrubs were more mature (8 to 10 years old). Red cedar (Juniperus virginiana), white cedar (Thuja occidentalis), apple (Pyrus Malus), staghorn sumac (Rhustyphina) and numerous oak saplings (Quercus 5pp.) were among the common woody plants and either brome grass or bluegrass was the principle ground cover. Hence, the white-footed mouse was more abundant than the prairie deermouse, with relative densities per trapline of 2.25 and 0.6, respectively. When this trapping occurred in the Spring of1980, the prairie deermouse was at a low in its population fluctuation, as well. The southern bog lemming population was also at a low in the spring of 1980 (see Table 8). There was abundant sign in the form of characteristic droppings and runways from the previous fall. However, prebaiting for one week and trapping for another in two separate locations where they occurred before failed to catch any lemmings. Even so, it appears that these hedgerows do sustain lemmings in years when their populations are high enough to fill most appropriate habi- tats. 49 Intolerant Trees,,Shrubs, and Grass Patches This stage is composed of trees that are relatively intoler- ant, that is, they require considerable amounts of light, they are able to tolerate only moderate root competition, and their seedlings are unable to grow under their own shade. Forties Vegetative cover composed of tolerant trees, shrubs, and grass patches occurred in two different forms, isolated shrub or tree patches (wasteland or plantings) and hedgerows, and comprised only 2.9 percent of all cover types. Uncommon were dry upland areas with scatterings of small intolerant trees (average dbh 5 cm) including American elm (Ulppp americana), black cherry (Prunus serotina), and black locust (Robinia Psuedoacacia). The ground cover consisted of a thick leaf litter, sparse grasses, and several species of bramble (Rpppg 5pp.) including dewberry, blackberry, and raspberry. The only small mammal species associated with these pre— dominantly woody cover types was the short-tailed shrew (0.87 per line). The significance of thick leaf litter to the presence of short-tailed shrews was discussed in the previous section on mixed perennials and shrubs. In a comparative study Linduska (1950) examined fencerows to determine the potential of hedgerow vegetation to form reservoirs of crapland pests that could invade cultivated tracts at opportune times. The wide variety of species found in the common bluegrass sod 50 borders was discussed in the previous section. Less common, and only occasionally continuous with the bluegrass-shrub fencerows, were the brush roadside fencerow representing the other extreme. In some of these, woody vegetation had closed in to such a degree that little or no herbaceous undergrowth was present. The principle trees and shrubs in these situations included white and black oak (Quercus alba and Q. velutina), shagbark and pignut hickory (Carya ovata and C, glabra), choke cherry (Prunus virginiana) and black cherry, red maple (Acer rubrum), hawthorn (Crataegus), staghorn sumac (Sambucus canadensis), wild grape (Vitis), and others. Hedgerows with primarily woody vegetation attract relatively few small mammals. In early August of 1942, trapping in four wooded borders secured only five individuals; four were prairie deermice and one a white-footed mouse (Linduska, 1950; not included in Table 4). In April 1946, two traplines in wooded borders caught only one white- footed mouse. An analysis of Linduska's raw data (see Table 4) shows that the only two small mammal species associated with these woody fencerows were the white-footed mouse (0.75 individuals per trapping, 4 lines) and the short-tailed shrew (0.23 per line). At this stage of woody plant succession the white-footed mouse is now obviously the predominant resident deermouse. The scarcity of leaf litter in these hedgerows, in conjunction with a low shrew population may corroborate earlier statements about what are acceptable burrowing conditions. 51 Seventies ‘ From what can be determined, these woody hedgerows have not . changed much from photos taken of the area in 1940. Sumac and the other woody plants have increased in diameter and height. The larger trees have also grown and spread into areas previously occupied by the sumac. Thickness of the leaf litter has increased to the extent that most of the grasses have been smothered out. Oak and hickory seedlings are sprouting through the leaf litter in places where sig- nificant seed pools exist. In addition to these border strips, sev- eral of the wildlife plantings may now be categorized with the intolerant trees and shrubs. Transition to a game area program called for extensive wild- life plantings in the form of food patches and cover plants. Woody plants were placed in hedgerows and plantations crisscrossing the study area, connecting many of the previously separated habitat types with a common transition zone providing a network of dispersal routes for small mammals. The many species planted include red and scotch pine (Pinus resinosa and P. sylvestris), black locust, silver buffalo berry (Sheperdia argentea), autumn olive (Elaeagnus umbellata), honey- suckles (Lonicera spp.), multiflora rose (Rosa multiflora), hawthorn, apple, and others. Woody plants associated with these plantings, but not restricted to them, were staghorn sumac, grey-stemmed dogwood (Cornus racemosa), the blackberry and the raspberry. Maturation of these plantings by the late seventies has placed a considerable por- tion of the land into the intolerant tree and shrub stage with brome grass sod frequently beneath in patches. 52 Since a variety of different aged plantings are included in this category, their combined capture results reflect a much more diverse small mammal assemblage than what occurred in the more mature woody hedgerows that Linduska examined (see Tables 4 and 5). His objective was to obtain a clear contrast between sod fencerows and those with woody vegetation and little ground cover. However, my objective was to find the southern bog lemming. Since some of the most productive cover types for lemmings in the forties were fence- rows with grass sod and shrubs, this recent survey concentrated on these trasition-stage border strips from the available range in maturity of the wildlife plantings. Only 0.09 prairie deermice and no house mice were captured in these moderately-woody hedgerows, which are probably too woody for these species (see earlier sections of more thorough discussions of habitat use by these species). On the other hand, most of the other local small mammal species were moderately abundant relative to their other more preferred habitats. Mark and recapture analysis indicated that three of the species had individuals that were predominantly residents in their habitat. These were the white-footed mouse, short-tailed shrew, and masked shrew, whereas individuals of the other three species, the meadow vole, bog lemming, and meadow jumping mouse, were largely transients passing along these borders. 53 Mid-Tolerant Woodlands In this stage an assortment of mid-tolerant trees, princi- pally oaks and hickories, predominate. It is generally believed that a mixture of beech, basswood, and sugar maple is the most toler- ant, and therefore climax, vegetative type in Clinton County under present conditions. The beech-maple-basswood stage will thus slowly replace the intermediately tolerant oak-hickory type, although the rate of succession will be extremely slow. Forties 0f the central 225-hectare study area, 22 hectares were in the woodland condition in the 19405. The great diversity in soil, litter, vegetation structure, and plant species diversity among the area woodlots (see Table 16) allowed for a broad examination of how wood- land type might affect small mammal abundance and species diversity. Seven woodlots, 4 to 10 hectares in size, were present in this unit. Six were of the typical upland oak-hickory type, the principle species being black oak, jack-oak (Quercus ellipsoidalis), white oak, pignut and shagbark hickory, with an average dbh of 15 to 25 cm. Most had little or no middle story. The sparse under- stories contained seedlings of the principle tree species along with black cherry, sassafras (Sassafras albidum), ground juniper (Juni- peris communis depressa), various brambles (Rubus spp.) and occa- sionally other woody plants. One woodlot was a lowland type in which maple dominated. The principle species here was red maple, mixed with white, red, and 54 black oaks, pignut hickory, large-toothed aspen (Populus grandi- dentata) and American elm. This woodlot has a very dense understory of red maple seedlings and a tangle of grape (Vitis 5pp.), dogwood (Cornus spp.), hazel-nut (Corylus sp.), witch hazel (Hamamelis virginiana) and service-berry (Amelanchier sp.). No sod areas were present in this woodlot. Their past history showed that all were cut over for large timber sometime within the 50 years prior to 1940. In the recently grazed upland woodlots the soil tended to be compacted and bluegrass sod was a predominant ground cover. This provided a contrast to the more typical woodland conditions present in those lots which had nearly recovered from light grazing. Even then, most of the woodlots were young enough in 1940 that breaks in the canopy allowed substantial patches of sad to develop. The diversity of small mammal species captured in the mid- tolerant tree and shrub stage for woodlands probably reflects the great abundance of ground-layer vegetation and sad patches. Not only were the white-footed mouse and short-tailed shrew captured, but also the southern bog lemming, house mouse, and the woodland vole (see Table 4). Previous investigations have stressed how important soil and ground cover properties appear to be for the short-tailed shrew (Linduska, 1950) and the woodland vole (Hamilton, 1938; Jameson, 1949; Benton, 1955). The latter usually requires loose, sandy soils and deep humus suitable for burrowing, as well as an abundance of root plants (grasses, shrubs, etc.) and tubers (woodland herbs) for food. Young, open woodlots of the uplands provided these conditions. 55 The more mature lowland woodlots supporting tolerant tree species still had open canopies and lush understories, but there were no sod patches and few woodland herbs. Consequently, even these woodlots should be considered in the mid-tolerant stage with respect to small mammal habitat. The heavy accumulation of litter over the entirety of the lowland woodlots still attracted short-tailed shrews, but there were no bog lemmings or woodland voles captured in these woodlands during Linduska's survey. Seventies Since the early l940s, additional cutting has occurred in only one of the central-area woodlots and all have been protected from grazing by domestic animals. with the cessation of these activities the woodlots have matured following the pattern of suc- cession typical of most upland deciduous woodlots. The canOpy has closed in above with the aid of maturing second-story trees. This closure has reduced the amount of available light for understory herbs and shrubs, allowing only spring woodland perennials and shade-tolerant shrubs and saplings to persist. Sod patches no longer prevail in the ground story except at the margins of the woodlands where light penetrates from the side. Woodlots with these characteristics were placed in the mid-tolerant tree category in the present investigation because the ground story still supports more vegetation than one would expect in a tolerant woodland. 56 The small mammal communities for woodlands now in the mid- tolerant seral stage are more diverse than those present in the forties (see Table 5). Although more diverse, the substantial reduction in the density of understory shrubs and saplings may have caused a two- fold reduction in the population of white-footed mice, from 3.22 to 1.44 individuals per trapline. Considering the relative abundance of white-footed mice in all the sampled habitats in the 19405 versus the 19705, the above conclusion is preferred over the alternatives of (l) an overall low in p0pulation levels on the area; or (2) it having been an artifact of more limited sampling in this seral stage. For example, ample increases occurred in cropland (0 to 0.1), grass and weed associations (0 to 0.4), grasslands (0 to 0.4), grass and shrub areas (0.3 to 2.3), and in intolerant tree and shrub stands (0.8 to 1.8). The amount of land supporting woodland vegetation has more than doubled since the 19405 and now totals 81 hectares (200 acres). The majority of this land is in the mid-tolerant tree and shrub category. Some of these areas result from the maturation of oldfields and very young woodlands; the rest is from tree and shrub plantings. Several younger stands are open enough to support brome grass sod. The greater diversity of small mammal species associated with this seral stage in the seventies can mostly be attributed to the fact that, in addition to trapping Linduska's lines, some trapping was conducted nearer the woodland borders in search for the woodland vole and the bog lemming. 57 Included here are the pine plantations that have reached a stage of maturity where the litter layer, sod patches, and sparse shrub associates resemble that present in the mid-tolerant deciduous woodlands. Small, relatively stable bog lemming colonies were found in these upland pine plantations throughout the summer of 1979. Their only small mammal associates within these groves were white-footed mice and short-tailed shrews. In the shrub hedgerows bordering one such grove, these three species were joined by meadow voles, jumping mice, and masked shrews. In another, more Open, grove individuals of all five species were captured in neighboring traps. Wetlands Linduska Operated a few traplines in wetland habitats in con- junction with his farmland study. Since no vegetation descriptions were made of these sites, photographs taken during the study period were relied upon for some basic information. Areas on or bordering wet lowlands were primarily covered with sedges. Reed canary grass (Philaria arundinacea) occupied stream banks and formed patches around shallow ponds and lakes. The few photographs available to characterize the willow- maple-dogwood association showed shrubs approximately 1 to 2 meters tall with a ground cover of sedges and mixed herbaceous plants. Shrubs were fairly thick within a few feet of the water's edge and thinned out farther from the edge. The principle shrub species were red osier and grey—stemmed dogwood (Cornus stolonifera and Q, racemosa), black willow (Salix nigra) and a variety of shrubby willows. 58 Also present were posion ivy (Rhus radicans), poison-sumac (Rhus Vernix), water-hemlock (Cicuta maculata), marsh ash (Fraxinus pennsylvanica), and some small red maples. In the very wet areas, most sedges and reeds grew on hummocks Of soil and organic matter emergent from the surface of the water. Linduska captured only three masked shrews and four meadow voles on three lines operated each for one three-day period. In such moist situations the masked shrew predominates over the short-tailed shrew; the latter rarely ever is captured in wetlands. Although the only bog lemmings Linduska ever actually captured in the vicinity of wetlands were taken from an adjacent lowland wood- let, he makes several references to their association with wetlands. Linduska (1950) observed an upland migration from the bogs and marshes by Synaptomys in the fall. These movements coincided with the invasion Of meadow voles into upland sod areas that neither species of mouse occupied during the summer months. He speculated that both microtines were moving into the grasslands, which in summer were dry and inhos- pitable, but by fall were moist and suitable refuge from the season- ally colder and eventually frozen wetlands. He also observed several lemmings in a grass-weed Opening beside a stream bank in winter. By the 19705, lowland areas have acquired a new growth Of mid- tolerant trees. Box elder (Acer Negundo), black willow and black locust have grown to an average dbh of 5 cm. multiflora rose, honey- suckle, willow and dogwoods have matured, with stems approximately 3 cm in diameter. Vegetation is very thick around water edges and 59 new sprouts, 10 to 15 cm tall, appear further out from the shore in the sedge, canary grass, and brome patches. Much more extensive trapping was conducted in wetlands for the present survey. This was primarily to assess whether upland populations Of the bog lemming were higher or lower than those that occurred simultaneously in the wetlands, or whether there were sea- sonal migrations between the two regions. 0n the solid boggy borders Of lakes an equal mix of short- tailed shrews, meadow voles, and meadow jumping mice was Obtained. At both Burke Lake and Moon Lake these species each averaged 3 indi- viduals per trapline. Tamarack stands yielded no captures Of small mammals on any of three occasions. 0f greatest significance to this study was the total absence of bog lemming captures in wetland habitats even though a variety of trap types were used, including Museum Special snaps, pitfall-can traps, and livetraps. Microtine sign was abundant in many of these locations. It had qualities similar to that described for bog lem- mings by numerous authors, namely a light green color and blunt-ended fecal pellets. But these qualities may be just as dependent on the nature of the vegetation eaten as on the microtine's intestinal tract. Evidence supporting this possibility was Observed in the upland pine plantations. Here, fecal pellets that could unquestionably be attribu- ted to the bog lemming appeared very dark green and blunt when formed from brome grass and yet were small and tapered, much like that of 60 deermice, when taken from a trap from which the animal had eaten oats. Hence, it is conceivable that a blunt, light green pellet could just as likely be formed by meadow voles feeding on light green sedges as by bog lemmings. If this is the case, then much of this sign may be attributed to meadow voles. This is indeed a ripe area for further investigation. SPECIES ACCOUNTS Cryptotis parva In the northeastern United States, the least shrew is found predominantly in Open grasslands (Blossum, 1931; Burt, 1946; and Barbour and Davis, 1974) and abandoned Oldfields (Getz, 1962; Gotts- chang, 1965; and Choate and Fleharty, 1973). Although this species has occasionally been found in forest- or marsh-edge situations in the northeast (Hoffmeister and Mohr, 1957), it is more often reported from grassy or weedy fields in close proximity to wetlands. What these weedy fields may provide is a diversity of broad- leaved herbaceous plant species necessary to support a varied enough insect diet for the least shrew. In a survey by Whitaker and Mumford (1972) the five commonest foods of Cryptotis parva were found to be lepodopterous larvae (18 percent), earthworms (11 percent), spiders (7 percent), orthopteran internal organs (7 percent), and aphids (4 percent). They also derived significant portions Of their diet from centipedes, crickets, and mast. Such a variety of prey items might only be supported by a large variety Of leafy herbaceous plants. Howell (1954) saw evidence of habitat selection highly correlated with cover vegetation within Old-fields duringa.live-trapping study where he captured Cryptotis in only three Of seven cover types exam- ined. He caught four individuals in Open growths Of bush—clover (Lespedeza), 17 in goldenrod (Solidago) and sassafras, and three in 61 62 honeysuckle (Lonicera). None were caught in dense bluegrass stands (Poa pratensis) with sparse goldenrod, in Johnson grass (Sorgygi halepense), left fallow or in dewberry or blackberry stands. Oldfield habitat supposedly suitable for the shrew was avail- able in great abundance on the Rose Lake Area as both annual and perennial grass and weed associations in the forties. This cover type constituted 34 percent Of all major habitat types present on the area. Even though 15 percent of the land supported this same cover type in the seventies, trapping efforts in excess of 2000 trap nights for each study period captured so few animals as to sug- gest that this species was, and still is, very rare on the Research Area. Many of the oldfields on land more recently acquired by the State were never planted to hay, and were retired from cultivation before the introduction Of brome grass. They, therefore, support a greater diversity of weed species, and they support the native blue- grasses. It is more likely that the least shrew will be found in these Oldfields, since there is a great diversity of braOd-leaved herbaceous plants and enough stability Of the plant community to support a varied insect diet for the least shrew. There would also be a greater amount of debris in these Older fields that might provide suitable nest sites for the shrew (Whitaker, 1974). Whether these factors are sufficient to attract the least shrew if it were presently on the area is unknown. The inferences drawn here about potential habitat selection, foraging sites, and trappability mostly derive from publications of minor observations 63 made while the investigators were engaged in research on other spe- cies, and some come from incidental encounters with the shrew. This wide range of Observations is summarized by Whitaker (1974), but even this large body Of literature does not give us a very clear picture Of the ecology Of this species. Consequently, to say that the least shrew is probably absent from the central study area, but potentially present on the out-lying Oldfields more typical of southern Michigan, is a "best guess" based on an overall considera- tion Of literature published on the shrew. Linduska (1950) reports of snap-trapping only two individuals on the study area during the six-year period. A third individual was found on 2 October 1941, apparently left by a predator beneath an oak tree at an Oldfield-bog margin. In addition, one was col- lected during the examination of field-shocked corn in January, 1941 (Linduska, 1942a; and Aldrich, 1978). Four additional specimens were obtained from activities on the Area over the years since Linduska's work. One was reported killed by'a fox in 1946 (Gordinier, 1946), one was trapped in 1957 and two in 1960 (Aldrich, 1978). In my own trapping in all seasons and with various trap types (pitfall, snap, live) not a single least shrew was taken. There is independent verification of this scarcity Of least shrews if one looks at the owl pellets collected on the Area during both periods. In 681 items taken in all seasons from five species of hawks and three species of owls, Linduska (1950) found no least shrews. I also found no least shrews in the 90 pellets I identified 64 from three species Of owls (see Table 7). These items were all from winter collections. Should the species be extinct from the Rose Lake Area at the present time (it probably is not), there is abun- dant habitat to support viable populations should a re-invasion or ‘ population increase occur. Despite the fact that this species readily builds nests just below ground in Open fields (Hamilton, 1944; and Davis and Joeris, 1945), it has been most frequently collected during incidental dis- coveries from above-ground nests beneath logs, rocks, sheet metal, and boards (see Table 8). Davis and Joeris (1945) state that this shrew is difficult to catch from March to November, which happens to coincide with its reproductive season in the northern parts of its range (Hamilton, 1944). Consequently, it is the information gleaned from incidental discoveries, along with several captive studies (Hamilton, 1944; Conaway, 1958; and Pheiffer and Gass, 1963), which comprises the bulk of our knowledge about the least shrew. The examination Of debris occasionally used for nesting by the shrew may be a useful way to locate and capture these animals (Barbour and Davis, 1974). However, one should not consider these Objects to be a necessary component Of the least shrew's habitat. Careful examination of the literature published on this species suggests that there is a strong seasonal aspect to their activity. It is interesting to note that in 90 percent of the reported instances where this animal has been taken by trapping, the captures have been made during the winter months from November 65 TABLE 7.--Prey items identified from owl pellets collected on the Rose Lake Wildlife Research Area from February 24 to April 1, 1980 a s c: .0 g "a? E = (U f6 0‘6 x 1: > W '0 LIJ 3 r— 3 'F" I6 >1 0 U S- m m o m o. -v- 0 4.! 3 0:- : Q- '- cu m > s: to 8f '4- ' .— s. Q) m z 31 r— d) S- Q. m (n a) C .D 3 O 3 U) «A Q. Ol- U, U C m U U 06 :3 m I'd m S— M- c 44 >1 m '- 3 "- O X °I- O E 3 'l" L .0 d.) S- S- O +1 > 3 U, . s. to u s. u v— -r- C O O I— ‘I- 0) '0 >1 0 0 Z V) a: 2 O. a: (I) V) U) Screech Owl 6 (Otus 6510) 3 Long-eared Owl 12 2 4 5 1 -- -- -- -- (Asio otus) Great Horned Owl 23 __ __ 7 3 3 1 1 9 (Bubo virginianus) 66 pomp .cOmxuaa m~o_ _.La< m mopcmuca mN ..=.me_> mma. .»a_tauux mam. scaacaa a mop u_o mu c_o_cu_o maxop pmop .ucax xgaaunmu mm wowzm c_u Love: a» v .uo N Lw>vc com: xcoa Ammucu moxo» consoouo o ocaogmcovca ~— momp .m_cooa a mv>mo cansounmm ow ucaogmgouca .o» m assumoa mmmsu mmxw» Nmm— .mxoocnuaocm Aguagnum or pages uomsm Love: a» m .vo ~ gouge an «cannon moxw» mwm. .Lou»=m ama. ».=a-o==a coozuc.to a guaon no u=.p ua_to oregano emwp .xmpomoz onmp ..uuo a, mo. Love: N «cabana u_o aegc nwo. .ampe_cm .uaom .a.: .30“ m a:..oc.u .z omm. .comgauua .naa n - .>oz e “sopca> e »u_=_o_> we. mm=.u_,=m see» :62 «ea. .coup_saz «mo_ .soz N a» m .ua . one: cauootaa Goa .Lo> so: mam. _.ca< m acaocazovca _ u_o_cu_o ”mm, sue»: a. veaocaamuca m u.~_cu_o «no. .eo»,.2a: Luneo>az on xuoc Luvs: m ego» zaz «No, .coS__soz awe. Logsosoz p _ u_m_c eucnupu-c.a= xtO» :uz mmo. .Loccou .mop »_=a --m mucaoa Lace: 4 axouaue _av.h xomcua so: amm. .Lmotan_.om a emu—o: mma. ._La< c. coo; rayon a» o .ua . mtsumoa--u.o_cupo xxuaucux neg. .mmoa a Luce.uca .om. coco: a» e .u. . m.o=,ppu «ea. .coa..sa: NNa. um=a=< om F ass)» uoxoc_coxo awmzouu eem_ .eou__eaz Dem. gut»: m~ N new»: atom «uptopa “no. .Lmocpcam mma. .uao .m «Fen emaca Laue: _ o__a saute .ooaeao xco outcopa Louum_umo>c_ mama cowuwmmu .oz uuu_nm: wuoum mczuogmupp mg» soc» cmxmu ago mm_cucw .o>com mwuoumxcu mo new: An cowuump—ou mo moans..-.m ubmm umcsuamu mcoz Azopv ummp mumpwnm; campa: ages at Aacaucznav 8mm, ommp cw cmwzm omm— .mxmzucwg memp .gwmpm camp .aszm me> xuczou copcwpu zucaou copmmcw>m4 mOFucaou cmmwzomz ozu cw wcmaoou wxsoummcaw c? mpm>mp cowumpzaoa.irmpm4m