We captured, marked, and recaptured southern short-tailed shrews (Blarina carolinensis) during a 30-month livetrapping study in a woodlot in Jackson County, Illinois, to compare aspects of their life history with those of the northern short-tailed shrew (B. brevicauda). A total of 106,496 trap checks (15,782 trap nights) resulted in 3,430 captures of 313 B. carolinensis from February 1996 through August 1998. Trapping mortality was only 18 individuals. Sex ratio did not differ from 1:1. Estimated population density peaked at 57 individuals/ha in late summer and autumn then declined during winter. Recruitment, including birth and immigration, peaked in spring and late summer each year. Individuals entering the population in the spring and early summer had higher survival rates than those entering in the autumn. A weak correlation was found between recruitment and precipitation, and between population density and humidity. Shrew activity (timing of captures) showed significant relationships with light condition and season. During summer, shrews were caught more frequently at night. In the winter, they were captured more frequently during the day. Capture rate was negatively related to precipitation and positively related to humidity. Population dynamics and activity patterns were similar to those of B. brevicauda.
Data on life history aspects of Blarina are largely restricted to northern short-tailed shrews (B. brevicauda). Most authors assume that southern short-tailed shrews (B. carolinensis) are similar. However, data for B. carolinensis sometimes are too few to draw meaningful comparisons or are contradictory. For example, Schwartz and Schwartz (1981) reported that sex ratios for B. brevicauda were skewed toward males. Conversely, Gentry et al. (1968) captured more females in a removal study on B. carolinensis.
Most investigators report 2 reproductive peaks for B. brevicauda, 1 in spring and 1 in late summer or autumn (Christian 1969; Gottschang 1981; Hamilton 1929; Hoffmeister 1989; Schwartz and Schwartz 1981). Likewise, Layne (1958) and Hoffmeister (1989) delimited spring and autumn breeding periods for B. carolinensis. Overwinter survival is critical to maintain populations of Blarina and may be influenced by weather, population density, time of birth, or body weight (Pearson 1945). Getz (1994) found that annual peak density of B. brevicauda was influenced by population size at the beginning of the spring breeding season (density of adults that survived the winter), timing of the spring increase in food availability, and the midsummer decrease in reproduction. During a removal study of small mammals in Georgia, numbers of B. carolinensis fluctuated by an order of magnitude (Gentry et al. 1971a; Smith et al. 1974).
The high metabolic demands of shrews necessitate high levels of daily and seasonal activity. In a synthesis of numerous studies, Merritt and Vessey (2000) concluded that shrews exhibit polyphasic activity patterns, consisting of 10–15 shortduration (0.1- to 3-h) bouts a day. B. brevicauda has been reported to be most active on cloudy and rainy days (Getz 1961) and at night (Ingram 1942) during the summer, with a 2nd peak of increased activity during the day in autumn and winter (Pearson 1947; Randolph 1973; Rood 1958). Martinsen (1969) and Merritt (1986) reported that B. brevicauda decreased activity levels in severely cold weather and spent more time inactive at resting metabolic levels. Weather conditions may affect densities or activities of shrews, including precipitation (Pankakoski 1985; Smith et al. 1974) and humidity (Pankakoski 1979; Pruitt 1953). Getz (1994) found only a weak correlation between precipitation and population cycles of B. brevicauda. Similar relationships for B. carolinensis have not been documented.
Our objectives in this study were to test whether population dynamics, and daily and seasonal activity of B. carolinensis were similar to patterns reported for the closely related and better-known B. brevicauda. Specifically, we marked and recaptured a population of B. carolinensis to determine sex ratios, population size, and survival patterns and to quantify relationships between shrew population dynamics and activity patterns relative to light condition, season, and weather variables.
Materials and Methods
Study area.—The study area was located in a woodlot in Jackson County, Illinois (T9S, R1 W, N 1/2, NW 1/4 Sec. 29). It consists of pole-sized 2nd-growth deciduous forest, dominated by shingle oak (Quercus imbricaria), black cherry (Prunus serotina), Osage orange (Madura pomifera), and persimmon (Diospyros virginiand). Understory vegetation is dominated by Japanese honeysuckle (Lonicera japonica) and common pawpaw (Asimina triloba). We established a 12 × 13-station trapping grid (10-m intervals between stations) in January 1996. The 1.32-ha grid was between 2 ridges, included a low area with an intermittent stream, and was bordered on the north by a larger intermittent stream.
Trap types and trapping.—We attempted to conduct 4 trapping sessions each month. A trapping session consisted of 1 trapping period in which traps remained set and were checked, typically over a period of 14–24 h. When possible, 2 trapping sessions were conducted with 1 or 2 days between sessions. Traps were checked every 3 h during a session to minimize mortality of shrews from starvation or stress. Traps were typically set to include at least 1 predusk check and then checked again every 3 h throughout the night until at least after dawn. When possible, traps were checked for 24 h from the time that they were initially set. To reduce mortality, trapping was not conducted during periods of rain or when the animals could not be kept dry.
Two types of live trap were used during each trapping session. During all 100 trapping sessions, we used a live trap (laboratory made) based on a Russian design (Whittaker 2003; Whittaker and Feldhamer 2000). During 13–14 May 1996, standard Sherman traps (7.6 × 7.6 × 25.4 cm, model 3310A, H. B. Sherman Trap Co., Tallahassee, FL) were used. From 21 May 1996 through August 1998, small Sherman traps (5.1 × 6.4 × 16.5 cm, model SNA) were used. During initial trapping sessions, from 9 February through 6 April 1996, we also used pitfall traps. The type of trap set at each station was alternated each trapping session or between paired sessions conducted with only 1 or 2 days between them. Although the Russian-designed traps captured significantly more shrews (Whittaker and Feldhamer 2000), bias from trap type was negated because each trap type was alternated at each station equally throughout the duration of the study. Traps (including pitfalls) were set “aggressively” for shrews, placed near holes or along logs and other structures. Traps were set within 1–2 m of the grid station point.
We baited traps with a mixture of canned (wet) cat food and bread soaked in unrefined peanut oil. Traps were baited at the start of each trap session and as needed after animal captures. Trapping and handling of small mammals were in accordance with guidelines of the American Society of Mammalogists (Animal Care and Use Committee 1998).
Upon capture, small mammals were toe clipped to provide individual marks. Mass and sex (when possible) were determined for captured animals. Because shrews have no external genitalia, several characteristics were used to determine sex. Sex was reliably determined for lactating females, and males may exhibit inguinal swellings during breeding seasons. Also, this species has skin glands midway between the fore and hind legs and on the belly that appear as hairless patches. Although both sexes may have the belly gland, the side or lateral glands are normally restricted to breeding males in B. brevicauda (Gottschang 1981).
Population analyses.—The minimum number known alive (MNKA) method (Krebs 1966) was used as a conservative estimate of shrew density. Additionally, program JOLLY (Pollock et al. 1990) was used to estimate population density, survival, recruitment (which included immigrants as well as births), and probability of capture for each animal known to be alive. Program JOLLY assumes that every individual, regardless of whether marked or not, has an equal probability of capture during a sampling session; every marked animal has an equal probability of survival from one sample to the next; marks are not overlooked or lost; and sampling time is negligible compared to the amount of time between trapping sessions (Jolly 1965; Pollock et al. 1990). For both estimates, shrew captures were pooled by month for analyses.
Survival was the probability a shrew within the study population survived (persisted) for a given period of time. We considered survival synonymous with persistence because we could not determine whether the animals died or emigrated from the study grid. Data from MNKA estimates were analyzed by using the Kaplan-Meier procedure (Kaplan and Meier 1985; Pollock et al. 1989a, 1989b) by using StatView 4.5 (SAS Institute, Inc. 1999). The date of 1st capture for an individual was considered the event time, and survival was calculated based on subsequent captures. We used MNKA because it is a more conservative measure than JOLLY. Additionally, it is more likely that assumptions of program JOLLY, particularly equal probability of capture, could lead to biased estimates. We tested for differences in survival between the sexes, and for individuals entering the population in spring versus late summer and autumn. Exact times of “death or emigration” were not known, but intervals between trap checks were brief enough to provide a reasonable estimate (at most several weeks) of actual persistence.
We excluded individuals captured during the 1st spring of trapping (1996, before the 1st recruitment peak) based on the assumption they would have overwintered and probably biased survival rates because they were nearing the end of their lives. Likewise, shrews caught during the final spring sessions (1998) were excluded because our study ended early in their lives and would have lowered overall survival estimates. The remaining shrews lived during the duration of the study. Observations of these individuals were considered uncensored (as complete observations) because we assumed complete persistence data for them. To determine whether including only shrews caught during the middle of the study impacted survival estimates, Kaplan-Meier procedures also were conducted on the entire data set, but data for individuals previously excluded were considered censored (as incomplete observations—Muenchow 1986; Pollock et al. 1989a).
Precipitation and humidity data (measured 2 m above ground) in Carbondale, Illinois, were acquired from the Water and Atmospheric Research Monitoring Program, Illinois Climate Network (1999). Simple linear regressions were used to determine relationships between levels of precipitation and humidity and estimates of population size, as well as entry of new, unmarked individuals to the population. We used current levels of precipitation and humidity as well as levels during the previous 1 and 2 months to compensate for potential time lag between weather and population-level effects.
Activity.—We estimated shrew activity by grouping capture times into light condition categories. Captures occurring during daylight trap checks were grouped as “light,” captures occurring at night were grouped as “dark,” and captures occurring when an animal may have entered the trap during dusk or dawn were grouped as “transition.” Trapping sessions also were categorized seasonally. A climograph (Nicholson et al. 1997) was prepared for the study period and used to assign months to seasonal categories based on mean monthly temperatures and precipitation during years sampling was conducted (Water and Atmospheric Research Monitoring Program, Illinois Climate Network 1999). November through March grouped tightly as “winter.” April and May grouped as “spring.” September and October grouped as “autumn.” June, July, and August grouped together as “summer” (Whittaker 2003). We used a 2-way analysis of variance to test for effects of light condition and season on shrew captures, including initial captures and recaptures. Initial captures and recaptures were analyzed separately to determine if shrews potentially new to the study site (either through birth or immigration) might be behaving differently than “resident” shrews already established within, or familiar with, the study area. Bonferoni-Dunn post hoc tests were used to test where significant variation occurred.
We used simple linear regressions to test temperatures within each season for a relationship with shrew captures. Mean shrew captures and mean air temperature (measured on site while checking traps) were used for each trap-checking period (e.g., midmorning, predawn, postdawn, and so on) during a particular season. Trap-checking periods with <3 observations were deleted from analyses. A mean of 19.1 observations was recorded for each trap checking period (range 3–41 observations). We used simple regressions to examine relationships between levels of precipitation and activity. To compensate for potential time lag between precipitation and its effect on shrew activity, precipitation was regressed concurrently with mean shrew captures per trap-checking period and staggered by 1 and 2 months.
Population dynamics.—We conducted 100 trapping sessions from February 1996 through August 1998. A total of 106,496 trap checks (15,782 trap nights) resulted in 3,430 captures of 313 southern short-tailed shrews. Trapping mortality was low; 18 individual short-tailed shrews died (0.5% of captures; 5.8% of individuals) in traps during the 30-month study. Of the 313 individual B. carolinensis captured, 61 were identified as females, 52 as males, and sex could not be determined for 200 individuals. For individuals whose sex could be determined, the ratio did not differ from 1:1 (χ2 = 0.717, P > 0.05).
Density fluctuated seasonally (Fig. 1). Population peaks occurred in October 1996 at 38 individuals known alive (28.8 individuals/ha), August and September 1997 with 75 individuals known alive (57.0 individuals/ha), and in June 1998 with 56 individuals known alive (42.4 individuals/ha). Unmarked or “new” individuals entering the population exhibited 2 yearly peaks (Fig. 1), 1 in the spring and a 2nd in early autumn. Captures of new individuals peaked just before peaks of MNKA. Survival rate as determined from MNKA data averaged 0.52 (SE = 0.07). Estimates of population size with JOLLY (model A, see Whittaker  for details of model choice) peaked in summer 1997 and 1998 (Fig. 1). Overall mean population as calculated by program JOLLY was 55.6 individuals, with an estimated average survival rate of 0.79 (SE = 0.02). Survival rates decreased below 0.70 only during 3 months: June 1996 (to 0.53), November 1997 (to 0.69), and February 1998 (to 0.57). Survival was ≥0.90 during 6 months: July and October 1996, and March, June, July, and September 1997. Probability of capturing an individual known to be alive in any month averaged 0.65. Capture probability was lowest in July 1997 (0.38) and highest in November 1996 (0.92) and 1997 (0.74).
For 199 individuals captured during the middle of the study, 50 were captured only once, whereas 1 individual persisted 516 days. Males (n = 45), females (n = 55), and individuals of unknown sex (n = 99) differed significantly in their respective survival, or persistence, rates. Individuals of unknown sex had the lowest rate (χ2 = 61.95, d.f. = 2, P < 0.01). Considering survivorship, or persistence, only in known males and females, the sexes did not differ significantly. Shrews 1st captured in spring (n = 111) had significantly higher survival, or persistence, rates (χ2 = 6.13, d.f. = 1, P = 0.01) than those initially captured in late summer and autumn (n = 87).
Minimum number known alive was not significantly related to concurrent precipitation, previous month's precipitation, or previous 2 month's precipitation. Number of new individuals was positively related to concurrent precipitation (R2 = 0.223, F = 7.47, d.f. = 1,26, P = 0.01), but was not related to previous 1 or 2 month's precipitation.
Minimum number know alive was positively related to current humidity (R2 = 0.261, F = 9.52, d.f. = 1,27, P< 0.01; Fig. 2), as well as the previous month's humidity (R2 = 0.196, F = 6.60, d.f. = 1,27, P = 0.02), but not previous 2 month's humidity. New individuals were not significantly related to concurrent humidity, or previous 1 or 2 month's humidity.
Activity.—A significant interaction (F = 14.90, d.f. = 6, 663, P < 0.0001) was found between total number of shrew captures by light conditions (n = 229 trap-checking periods in dark, 244 light, and 202 transition) and season (n = 96 trap-checking periods in autumn, 117 spring, 168 summer, 294 winter). During autumn, mean numbers of shrew captures were highest during transitional periods and least during light periods. During summer, more shrews were captured during the dark periods. During winter this pattern reversed, with mean shrew captures highest during light periods (Fig. 3). Light condition also was a significant factor when considering only number of initial captures (F = 3.58, d.f. = 2, 663, P < 0.01) or recaptures (F = 3.89, d.f. = 2,663, P = 0.02). Initial shrew captures were highest during dark periods for all seasons except winter. Mean recaptures of shrews were highest during the light period in winter, during transition in autumn, and in the dark period only during summer. No relationship was found between number of recaptures and light conditions during spring.
Season also was a significant factor in average total shrew captures (F = 20.78, d.f. = 3, 663, P < 0.0001). Captures were significantly greater in autumn than in any other season (P < 0.0001), and greater in winter than in spring (P < 0.0001) or summer (P = 0.0003). Captures in summer and spring were not significantly different from each other. Season remained a significant factor when considering only initial shrew captures (F = 43.70, d.f. = 3, 633, P < 0.01) or recaptures (F = 36.77, d.f. = 3, 633, P < 0.01). Initial captures were highest in spring and summer, whereas recaptures were significantly greater in autumn and winter.
Mean shrew captures per trap-checking period were not related to concurrent mean air temperature during spring or autumn. Captures were positively related to air temperatures during winter (R2 = 0.713, F = 19.91, d.f. = 1, 8, P < 0.01) and were negatively related during summer (R2 = 0.552, F = 7.41, d.f = 1, 6, P = 0.04). For all capture data combined regardless of season, a polynomial trend with ambient air temperature was evident (Fig. 4), although the relationship was not statistically significant (R2 = 0.084, F = 3.04, d.f. = 1, 33, P = 0.09).
The number of captures per trap-checking period during a session and concurrent precipitation were not significantly related. However, captures per trap-checking period were negatively related to precipitation in the preceding month (R2 = 0.164, F = 5.11, d.f. = 1,26, P = 0.03), and 2 preceding months (R2 = 0.168, F = 5.25, d.f. = 1,26, P = 0.03). Likewise, captures per trap-checking period were not significantly related to current humidity, but were directly related to humidity during the previous 1 month (R2 = 0.202, F = 6.82, d.f. = 1, 27, P = 0.02) and 2 months (R2 = 0.380, F = 16.58, d.f. = 1, 27, P < 0.01).
Our population estimate (MNKA) peaked at 57 individuals/ha, similar to the 60 shrews/ha reported for B. brevicauda by Piatt (1976). Both were higher than the estimate of 18 shrews/ha reported for B. brevicauda by Getz (1994), and 14.8 shrews/ha found for B. carolinensis by Smith et al. (1974). We found population density peaked in late summer and autumn before declining during winter. A similar pattern was described for B. carolinensis (Briese and Smith 1974) and B. brevicauda (Getz 1989,1994), except that our population tended to peak later. The higher mortality rate of 42.2% reported by Getz (1994), compared to 5.8% in our study, may have been a contributing factor to the later peaks and higher densities in our study. Had more of the animals in his study survived capture, they could have influenced population dynamics.
The low estimated population density in 1996 was likely an artifact of the MNKA method at the start of our study. Density fluctuations in 1997 and 1998 were similar except for the final month, again a likely artifact of the MNKA method. Lima et al. (2002) and Merritt et al. (2001) showed distinct multiannual fluctuations for B. brevicauda in Pennsylvania. They suggested that fluctuations were driven primarily by intraspecific competition for resources during the winter when food was most limiting.
Two peaks in new individuals entering the population occurred during each year of our study. Recruitment pulses in spring and late summer correspond to the pattern described by Gottschang (1981), Hamilton (1929), Hoffmeister (1989), and Schwartz and Schwartz (1981). Recruits entering the population during other times may represent shrews immigrating to the study site. New individuals appear to be relatively constant other than peaks associated with breeding pulses. This suggests a relatively consistent number of shrews adjusting territory size or attempting to establish new territories throughout the year.
Survival estimates varied little during our study and were consistent with those of Lima et al. (2002). As with population density, our 2 survival minima (53% and 57%) may be artifacts. The 1st occurred in summer 1996, near the beginning of the study, and the 2nd occurred in winter 1998, near the end. No such dramatic decrease occurred during 1997, the middle portion of the study. Decreases in April and August 1997 probably indicate the loss of overwintering individuals, and new individuals that were unable to maintain territories on the study site and died or dispersed. Churchfield et al. (1995) suggested that levels of mortality would be higher as young shrews leave the nest and attempt to establish territories. Because there are indications that members of the genus Blarina are highly territorial (Merritt 1987; Platt 1976), dispersal, whether as a juvenile leaving the nest or an adult seeking a new territory, would certainly increase the likelihood of death.
Pearson (1945) suggested that tooth wear is a useful measure of age in shrews (Blarina). We captured 1 female repeatedly from 14 August 1996 to 14 March 1997. During the winter of 1996–1997 we noted that her teeth were worn to the gumline. We think she was born in 1995 and was living through her 2nd winter. Shortly after she was 1st caught in 1996, she had prominent nipples and was probably lactating. In March 1997, during the last session in which she was caught, she also had prominent nipples. If our estimate of her age is correct, it supports conclusions of Carraway and Verts (1999) that older shrews do not suffer reproductive senescence and can produce litters.
Known male or female shrews had equal survival rates on our study site, but individuals whose sex could not be determined— 200 of 313 individuals (64%)—had significantly lower survival. Most shrews of unknown sex were captured only once and quickly disappeared from the population. This contributed to the difficulty of determining their sex. These individuals may have been juveniles with no recognizable secondary sexual characteristics. As noted by Churchfield et al. (1995), juveniles probably experienced the highest levels of mortality. Recognition of sex often required multiple captures of an individual, by which time it was sexually mature. Additionally, individuals of unknown sex may have been transients, with higher mortality rates than resident shrews (Platt 1976). Even though some of the unknown-sex individuals persisted on the area, none approached the maximum persistence of the known males and females. Our ability to differentiate sex of individuals likely related to whether shrews were residents or transients. We speculate that transient shrews may not be able to put as much energy into reproduction, may be less successful reproducing, and may have fewer obvious secondary sexual characteristics. Also, because a transient shrew moves through established territories, it may be adaptive for it to produce less scent to draw less attention from residents.
Season also was a significant factor in survival. Individual B. carolinensis that 1st appeared during the spring had significantly higher survival rates. Decreased survival of the autumn cohort is likely because of increased intraspecific competition with survivors of the spring cohort. Comparatively less competition in the spring may result from overwinter mortality of adults as well as increased abundance of invertebrate prey. Members of the 2nd cohort would be less experienced and likely forced to establish territories in lower-quality habitat. Thus, both increased competition and less available habitat would increase chances of an individual becoming a transient and decrease its survival rate.
Little consensus exists on potential effects of weather on population dynamics of shrews, but density and activity frequently are related to precipitation (Getz 1994; Merritt and Vessey 2000). However, Pankakoski (1985) and Smith et al. (1974) suggest that precipitation is an indirect factor that increases prey densities to which shrew populations then respond. We found a negative relationship between captures per trap-checking period and precipitation during the previous 1 and 2 months. This supports the hypothesis that shrews respond to prey density rather than the precipitation itself, if invertebrate density increases a month or two after increased precipitation. Once prey levels increase, shrews may have an easier time finding food, and bait in traps may become less attractive.
Pruitt (1959) suggested that humidity was a critical factor for maintenance of populations of B. brevicauda in Michigan. We found humidity during the previous 1 and 2 months accounted for significant variation in captures of B. carolinensis per trap-checking period. Again, the population may not respond directly to humidity. Humidity is highest during summer, which is when prey densities are highest, and coincides with reproductive peaks. As with precipitation, no significant relationship was found between concurrent captures and humidity. Nonetheless, shrews tend to be active in burrows with high humidity relative to aboveground levels (Pruitt 1959), which helps shrews maintain respiratory water balance.
A shrew must be active at ground level to be trapped. Fewer captures may indicate reduced seasonal activity of shrews or a change in habitat use. Two opposing factors may affect capture frequency during winter. When ambient temperatures are below a shrew's thermoneutral zone, it may remain below ground, where temperature is moderated; thus, these shrews are not susceptible to traps. Although invertebrate prey is less available in winter than summer, the resting metabolic rate of B. brevicauda increases 38–43% during winter (Lima et al. 2002; Merritt 1986). With a higher metabolic rate and decreased prey available, the bait in traps may become an attractive alternative to cached food, thus increasing shrew capture rate and decreasing their reliance on cached food. Thus, captures per trap-checking period might be a biased representation of activity during seasons when food is scarce. Additionally, older shrews entering their 2nd winter with more worn teeth might be attracted to the soft bait and be more likely to enter traps.
Despite the short diurnal period in winter, shrews were captured more frequently during the day. Alternatively, despite a shorter nocturnal period in summer, shrews were captured more frequently at night (Fig. 3). Schmid-Holmes and Drick-amer (2001) reported increased captures of B. carolinensis in habitats where temperatures were moderated. Pearson (1947), Platt (1971), Randolph (1973), and Rood (1958) observed similar seasonal shifts in activity for B. brevicauda. Reduced activity likely reflects behavioral changes consistent with minimizing energetic costs of activity when environmental temperatures are outside thermoneutrality. Body temperature of B. carolinensis in Florida was 36.8°C ± 0.07 SE (McNab 1991); the zone of thermoneutrality was 30–34°C. Thus, ambient temperatures below or above this range would require shrews to increase energy expended. Observed activity shifts may not be entirely dependent on temperature, but also may be affected by levels of invertebrate prey (Merritt and Vessey 2000). Temperature changes may affect behavior and distribution of soil invertebrates. Shrews might balance their activity patterns relative to ambient temperature and prey availability.
Initial captures and recaptures shared a similar pattern relative to season and light conditions. Initial captures during spring and summer occurred more frequently during dark periods. As noted, new (primarily young) individuals entered the population in highest numbers in spring. Recaptures during spring showed no relationship to light conditions. Temperature during spring may be moderate enough to have little effect on activity of resident shrews. Total captures and recaptures of shrews were most frequent in the autumn and least frequent in spring and summer. Fewer shrew captures during summer were consistent with the results of Feldhamer et al. (1993), in which they noted decreased trapping susceptibility of B. brevicauda and Sorex spp. on sites in Kentucky and Tennessee during summer.
To efficiently operate live traps to maximize capture frequency and minimize mortality of shrews, investigators should consider the effect (direct or indirect) of ambient temperature on activity. It would be prudent to open traps at dusk in summer, check them at regular intervals throughout the night, and close them again shortly after dawn. If trapping in winter, it would be best to check traps during daylight and close them at dusk. The low trapping mortality for shrews in our study (0.5% of captures and 5.8% of individuals), compared to 66% of Gentry et al. (1971b), 42% of Getz (1994), and 25–30% of Blair (1940) demonstrates the importance of frequent trap checks. Although we recognize that any field study will have impacts on animals, studies with low mortality rates may produce less biased estimates of demographic trends.
Funding for this project was provided by an Austin Peay State University Senior Research Fellowship and the support of the Zoology Department at Southern Illinois University at Carbondale (SIUC). Greg Moroz, Central Research Shop, SIUC, provided assistance with live-trap construction. We thank many SIUC volunteers for help with the field work, including C. W. Whittaker, K. Taylor, D. Varble, J. Uherka, B. Lawrence, C. Marsan, F. Dahlkamp, J. Jenkot, S. Sudkamp, S. Wolff, S. Phillippe, and A. Morzillo. We also thank G. Waring, J. Whitaker, E. Schauber, C. Krajewski, D. Leitner, J. Kie, M. Nicholson, J. Newman, R. Gates, and K. Dugger for useful suggestions throughout this study. We also thank 2 anonymous reviewers for suggestions and comments that improved this manuscript.