Portions of the East Prairie had been burned every spring in late March or early April. These fires were strictly controlled, surrounded by fire lanes with several students carrying wet mops and water tanks standing by to keep the fire in bounds. The wind was an important factor; back fires were started on the down-wind side of the proposed burn to prevent the wind swept flames from jumping the fire lanes and getting into either the forest or parts of the prairie which we did not wish to burn.

Schramm set out Grid I in the spring of 1966 (Fig. 2), at which time the entire East Prairie was burned. Grid I was not burned in subsequent years until this past spring, and thus it was the most heavily littered of the grids. Schramm set out Grid II in the spring of 1968 (Fig. 2), just after this portion of the prairie had been burned. Since then, it too was not burned until this past spring. Grid II therefore, had two years less litter accumulation than Grid I throughout this study.

Two styles of live-traps were used for this study. The English-made Longworth trap has two parts: a 2- x 2- x 5-inch runway, containing the trigger device and the door, and a 2.5- x 3.5- x 5-inch nest box. The floor area of a Longworth trap is 22.5 square inches, and the volume is 43.8 cubic inches. The American-made Sherman trap is a single rectangular prism, measuring 2 x 2.5 x 7 inches. The floor area is 14 square inches, and the volume is 35 cubic inches.

Before each trapping period, except for the spring of 1971, a week of prebaiting was conducted. This consisted of locking all the traps open and, placing within them a large handful of sunflower seeds. Sunflower seeds proved to be an excellent bait, in that they were easily handled, were not messy, and seemed to be as attractive to the animals as rolled oat, or peanut butter.

Traps were placed under nests of dead grass whenever possible to protect them from the direct rays of the sun. On the newly burned Grid II where there was no dead grass, traps were placed in the shade of growing vegetation. An unprotected trap quickly heated up, and would cause the death of an animal caught in it.

During both summers, traps remained set 24 hours a day, five days a week. During the spring, fall and winter periods, traps were set only three nights a week. All traps were adjusted to spring easily. Sherman traps got out of adjustment often, partly because of their temperamental spring door, and partly because seeds would lodge beneath the treadle rendering the trap inoperable.

Because two styles of traps were used, they were placed in a symmetrical pattern throughout each grid to eliminate the possibility of one type being favored over the other, which would have resulted in a disproportionate calculation of population density.

Traps were checked every 12 hours during the summer study periods, just before sunset and just after dawn. They were checked only in the morning during the spring and fall study periods. The traps were checked three hours after sunset and in the morning during the winter trapping period.

In order to examine each animal caught, they were promptly released into a three-foot high wastebasket from which they could be captured with gloved hands. Examination of the animal included sex, reproductive condition, toe-clip number, and weight (Figs. 4-7). If an animal was unmarked, it was immediately toe-clipped. A number sequence was used as often as possible to help remember which identification numbers had already been used and which numbers were still available (Fig. 8).

Clean traps were carried each time the grids were checked in order to replace traps in which shrews had been caught. Shrews will secrete a strong smelling musk which fouls a trap, leaving it repugnant to mice. The odor was sufficiently removed by scrubbing with a stiff bristled brush and water.

The different trapping periods, mentioned hereafter only by season and year, were conducted during the following dates:

Often, the summer data were broken into two portions, early and late, each having the same number of days. Note that data from spring, 1970, were taken later in the season than the spring data of 1971.

A breakdown of trap-nights and total captures per trapping period is shown in Table 1. One of the reasons the number of captures is so low for the summer of 1970 is that only two Microtus pennsylvanicus were caught the entire period. The following fall there were 25.3 captures per 100 trap-nights, primarily because the Microtus population soared in the few weeks between these two trapping sessions.

Most of the captures were Blarina brevicauda, Microtus pennsylvanicus, and Peromyscus leucopus. Several other species were captured, however, in at least one trapping period, including the following: prairie vole (Microtus ochrogaster), deer mouse (Peromyscus maniculatus), western harvest mouse (Reithrodontomys megalotis), jumping mouse (Zapus hudsonius), house mouse (Mus musculus), thirteen-lined ground squirrel (Spermophilus tridecemlineatus), Franklin's ground squirrel (Spermophilus franklinii), eastern chipmunk (Tamias striatus), black-capped chickadee (Parus atricapillus), tufted titmouse (Parus bicolor), fox snake (Elaphe vulpina), and blue racer (Coluber constrictor). Data on these animals were not sufficient to warrant further investigation.

 

Traps were set out on Grid I immediately after this part of the East Prairie had been burned in the spring of 1971 (Fig. 10). Unfortunately, continuous trapping could not be conducted, but it is evident that some of the animals were still on the burn by the fourth day, and that all the animals had gone by the eleventh day. Notice that none of the animals caught on the boundary traps was Microtus, and none of the animals caught on the grid after the burn was Peromyscus.

Fig. 10

On Grid III immediately after the burn, two Microtus from Grid I were caught, one of which was caught in a boundary trap. Beside these, six new (unmarked) adult Microtus were captured, five of which were first caught about 40 feet from the edge of the burn. Only two Microtus were captured that had been on Grid III during the winter trapping period. All five Peromyscus caught on Grid III in the spring of 1971 were former residents. The single Blarina caught was unmarked.

 

Because none of the other trapping periods overlapped, summer was the only season where two similar trapping sessions could be compared. The summers were divided in halves so that density changes throughout the summer could be readily seen. The densities of all these grids increased during both summers. The density of Grid II during the second half of the 1970 summer (24.9 per acre) is not far from the 21.9 per acre found by Schramm (1970) for Grid I in the summer of 1968, two years after it had been burned. His data were collected during the first three weeks in July, a period comparable to my early summer period. The 14.1 shrews per acre found by Schramm for the then newly burned Grid II is slightly higher than the 8.5 shrews per acre I found on the newly burned Grid III.

 

Population densities of the meadow vole, Microtus pennsylvanicus, are shown on Table 3. The greatest density, 58.0 per acre, was recorded on Grid II during the summer of 1970. The sex ratios varied from grid to grid, season to season, and year to year, with neither sex showing a clear majority.

When the data from both summers were compared, we found very little in common. even at a glance it was clear that no trend could be established from these data, and that a yearly cycle was by no means evident.

It is noteworthy that the population decreased on Grid III from a level of 18.2 per acre during the winter of 1971 to a level of 13.0 per acre after the burn, even with the addition of several new individuals to that area.

Population densities of the white-footed mouse, Peromyscus leucopus, are shown for each trapping period in Table 4. The greatest density (19.6 per acre) was recorded on Grid III in the fall of 1970. The lowest density (0.0 per acre) was recorded on Grid I for the spring and summer of 1970.

It is noteworthy that there were no captures made on Grid I during the spring and summer trapping periods of 1970. It is especially noteworthy that during the winter of 1971, the difference in the population density of Grid I (16.3 per acre) and Grid III (17.6 per acre) is so slight. It should also be noted that no Peromyscus was caught on Grid I after the burn in the spring of 1971, but four former residents were caught in the boundary traps near the woods (Fig. 10).

I should mention trap mortality as a part of this study, although it is slightly off the subject. Appendix A goes into the details. Here, let it suffice to say that the highest rate of mortality for Blarina was recorded in the fall of 1970, and for Microtus in the fall of 1970 and the winter of 1971.

There has been no evidence of direct small mammal mortality from prairie fire, neither this past spring nor in the study done in 1966 (Schramm, 1970). Small mammals can easily escape the flames by taking refuge in underground burrows. Riechert (1970) explained that the heat of a prairie fire does not penetrate deeper than 0.5 cm below the soil surface. This was particularly noticeable on Grid I when it was burned this spring. Less than an hour after the fire had swept over it, the ashes could be brushed aside in any portion, and unburned litter could be seen which was still damp. The soil beneath this thin mantle was still cold to the touch; it is obvious that a subsurface burrow would not have been heated up.

The results from Grid III after the 1971 spring burn indicate that several, if not most of the resident animals leave the burned area the first night after the prairie. Two Microtus captured on Grid III had been residents of Grid I during the winter, and it seems likely they were absolutely new to Grid III. Grid III was the only suitable habitat near Grid I to which Microtus could easily go. However, chances are that the resident Microtus of Grid I would travel at random, most going into the forest (a habitat for which Microtus are not suited) and consequently being lost.

Although many animals leave the burrow immediately, the results from Grid I after the burn show at least one Microtus that remained for four days, and four which remained at least over the first night. None of these was captured after the fourth day, and all were presumably lost. All the Peromyscus were gone the day after the burn, and were caught only in boundary traps near the forest. The two Reithrodontomys caught on the eastern boundary were caught in a fire lane (mowed prairie) rather than forest. Reithrodontomys are not suited to a forest habitat, and these two probably remained on that narrow boundary strip until the burned prairie grew tall enough to provide them with sufficient cover. The first capture on the recently burned Grid III in the spring of 1970 was on May 19th. It seemed that very few animals reinvaded between that time and the end of the study two weeks later. Reinvasion probably did not begin to any great extent until early June.

The fact that the population densities of Microtus and Peromyscus were less on Grid III in early spring than during the winter is interesting. Even with the arrival of the six new Microtus and the two from Grid I, the population was lower than in the winter, and only two winter residents were still there. The cause of this decline cannot be solely trap mortality, even though a few Microtus did die during the winter trapping session. One possible explanation for the loss of so many former residents (10 individuals) could be the effects of burning. Since the fire came within 20 feet of Grid III, much of the home range of an individual Microtus that lived near the east side of the grid would have been destroyed. Van Vleck (1968) stated that when one Microtus pennsylvanicus encountered another, the one not within its home range left. Thus, a Microtus with half its home range destroyed would have to encounter another Microtus in the Grid III area if it were to enlarge its home range to its former size. Unfortunately, the former Microtus will submit to the latter, and will be forced to move on to another place less densely populated: in this case right out of the East Prairie. This emigration would have left the eastern portion of the grid vacant long enough for the new immigrants from the burn to establish their home ranges. Van Vleck (1969) stated that female Microtus have smaller home ranges than males. Six of the eight immigrants were females.

Thirdly, there is the inability to capture an animal in all parts of its home range (Van Vleck, 1969). Unless the trapping grid is very large with traps closely spaced, much of an animal's home range is not even touched.

Then there is the problem of catching all the residents. Barbehen (1958) gave evidence that it was primarily the adult Microtus that would be marked and recaptured. Only after these adults were removed by snap-trapping were many pre-adults and juveniles caught.

Finally, there is the problem of how the population density should be calculated. There are many different methods, and often comparisons of population densities among authors is practically impossible. Given the same data, these different authors could arrive at remarkably different densities.

Despite all these shortcomings, I used live-traps, mainly because we wanted to conduct a long-term study, and our only alternative was snap-trapping which would have destroyed the population. Furthermore, live-trapping may be more reliable than snap-trapping (Sealander and James, 1958).

Much was done to control the variables mentioned previously. To prevent trap habituation, traps were alternately set for two days and closed for two days during the summer trapping periods. During the other periods the traps were set only three nights a week. Since it was likely that there were some individuals that did not get trapped, I assumed that the same pressures (be they social or physical) existed in all parts of the East Prairie. If true, then the pressures canceled out. However, the densities shown in the tables must be considered comparative densities and not actual densities.

To alleviate the problem of keeping an animal in a trap all night, traps should have been checked about three hours after sunset as well as in the morning. (This was done only during the winter trapping period.) Pearson (1960) stated that the heights of small mammal activity were from just before sunset to two hours after sunset, and again from two hours before dawn until dawn. To make sure that animals were captured within their entire home ranges, larger grids should have been used. For example, Buckner (1966) stated that the mean home range of Blarina brevicauda was 0.97 + 0.09 acre, an area not covered by a quarter-acre grid.

At any rate, the inequities tend to balance each other out, rendering the data from any one of the trapping sessions comparable to the data of any other trapping session.

Peromyscus leucopus is basically a woodland animal (Hamilton, 1943; Verts, 1957; Gottschang, 1965). It is also well known that Peromyscus are often found in old fields and along fence rows. No doubt it was found in pieces of the original prairie peninsula where the prairie met the bottomland forests. But Peromyscus are not well adapted for prairie life, particularly in regard to locomotion. I have observed all the Peromyscus I handled to run on top of the litter covered surface of the ground. In heavily littered areas of the prairie this type of travel would be a tremendous impediment.

Observe Fig. 11, and note that the mean population density of Peromyscus tended to decrease as the years of litter accumulation increased. The fact that the density did not decrease every year (on the graph) may simply be due to the fact that there were proportionately more Peromyscus in the East Prairie during the summer of 1969 than during the summer of 1970. If data from each summer were considered separately, then the areas with two years of litter accumulation setting them apart show marked decline in densities in the more littered areas. Thus, it is evident that the burning of the prairie (which eliminates the litter) has a density boosting effect on Peromyscus during the summers.

This difference in density does not last the entire year, however. As can be seen from Table 4, the density on Grid III from the summer of 1970 through the winter of 1971 remained nearly constant, while at the same time the density on Grid I increased. By the winter trapping period, the difference in population densities on the two grids was hardly noticeable. The increased use of Grid I was probably due to the snow which covered it most of the winter. About five inches of snow fell on January 4, 1971, and it had a thick crust on it within two days. Therefore, the ease of travel on Grid I was much the same as on Grid III during the winter. The food supply, in terms of grass seed, may have been slightly more abundant on Grid III.

The meadow vole, Microtus pennsylvanicus, is known to inhabit old fields, hay fields, and fence rows. Schramm (1970) found this species in a prairie habitat, and I have found no other references to Microtus pennsylvanicus in prairie. Fleharty (1970) found fairly large numbers of M. ochragaster on a prairie remnant in Kansas. He reported no M. pennsylvanicus.

Microtus ochrogaster is usually found in a somewhat dry habitat (DeCoursey, 1957), and would logically fill the Microtus niche in the drier western prairie. The Illinois prairie receives 30 to 40 inches of precipitation annually. Even though M. ochrogaster is found in Illinois, and even as far east as Ohio, when they are sympatric with M. pennsylvanicus they choose the drier sites of the habitat for their home ranges, and seem to always be subdominant (DeCoursey, 1957; Schramm, 1970).

Schramm found a remarkably high density of M. ochrogaster (100.7 per acre) in the East Prairie in the early spring of 1966, although they were not quite as abundant as M. pennsylvanicus (176.0 per acre). The previous winter was quite wet with standing water much of the time. Why either population was so high at this time is not known. Reproduction in both species does not begin to any large extent until March (DeCoursey, 1957); hence all these Microtus must have lasted through the winter, being the remnant of even larger fall populations. In any case, the prairie fire following Schramm's intensive trapping period presumably caused the elimination of all Microtus from the East Prairie.

Since that time, Microtus have reinvaded (probably from the Woodcock Field, shown in Fig. 1) but the numbers have always been comparatively low. Only one M. ochragaster has been caught in the East Prairie since 1966, although they have been caught fairly often in snap-traps that were set in the West Prairie (Fig. 1).

The immediate effects of burning on Microtus have already been discussed. The long term effects are more difficult to ascertain from my data. Because of the fluctuations found in all microtine populations in any given region, no standard density can stated. Barbehen (1958) noted that Microtus populations fluctuated independently of each other on the three areas he studied. As a rule, microtine cycles are rather regular and therefore predictable. But with the added factor of fire disturbance, the normal cycle could have been disrupted, apparently never reaching a peak. None was observed from the summer of 1968 through the spring of 1971.

Lo Bue and Darnell (1959) stated that their observations substantiated the conclusion that Microtus respond positively to vegetative height and cover. This is undoubtedly true to some extent. The Microtus I released after handling, quickly burrowed beneath the litter. All their runs were beneath the litter. An interesting phenomenon can be noted during the summer of 1969, however. Notice that the density of Microtus on Grid I (Table 3) is substantially lower than the density of Microtus on Grid II, even though Grid I had two years more accumulated litter present. The Microtus of Grids I and II must be considered as members of the same population, being separated by less than 100 yards of continuous prairie habitat. If the amount of litter were the only factor determining which habitat was favorable, then the implication is that Microtus prefer prairie with one year accumulated litter to prairie with three years of litter. There could be other factors involved, however. Barbehen (1958) noted that when live-traps were placed above litter at sites where both Microtus and Blarina had been captured, that only Microtus were caught. This suggests that Microtus do not forage exclusively beneath the litter, in which case too much litter would become an impediment. The entire population of Microtus was so low during the summer of 1970 that it was impossible to tell if this density inequality was present.

From the data it appears that one year of accumulated litter is preferable to three years. From the densities measured on Grids I and III in the fall of 1970, it appears that four years of litter is preferable to no litter. During the winter of 1971, density differences were diminished, possibly because the snow provided sufficient cover on Grid III to either increase immigration, or decrease emigration to another portion of the East Prairie.

The low density of Microtus on the East Prairie during the summer of 1970 is unusual. It was obviously not the result of a crash following a population peak; there was no peak. Also, since there were several individuals present during the spring, and since the height of reproduction is from March to June, it would be logical to expect somewhat higher densities in the summer, not lower densities.

Pearson (1964) observed a population of Microtus with a relatively high density systematically decimated by cats. He noted that Microtus runways are easily recognized by such predators as cats, foxes, raccoons, and skunks, which could simply wait beside a runway for the next Microtus to happen along. Such waiting is not generally worthwhile for a predator unless the population density of Microtus is very high. Cats, as Pearson (1964) observed, will not behave like all other predators which take only excess animals, but rather they will stay until the mice have been virtually exterminated. Predation by cats is one possible explanation of the near absence of Microtus during the summer of 1970. Cats have been seen on the Knox College Field Station, though not in the East Prairie.

Pearson also cited a paper by Brant (1962) in which he reported a population of Microtus being destroyed by carnivores, especially raccoons. Raccoons worked over all three grids the last three trapping nights during the spring period and during the first four nights of trapping during the summer period. The raccoons, adult females with young, would simply follow my trails to each trap, pulling the Longworth traps apart to get the sunflower seeds. They would trip the Sherman traps and the spring-held door prevented them from getting the seeds. Now it seems possible that during the routine of stealing bait, the raccoons would find traps with animals in them. Animals in Sherman traps were safe because the raccoons could not get at them, but animals in Longworth traps would not be difficult to catch. All the traps on Grids I and II were Longworth traps in the spring, and five of every eight traps were Longworth traps during the summer. The odds were greatly in favor of a Microtus getting caught in a Longworth trap and thus having an encounter with a raccoon. I never saw any Microtus remains the morning after raccoons hit the grids, but in the summer of 1968 Schramm noticed a Microtus head near one of his trapping stations, and the head still had an ear-tag in one ear. The grid had been disturbed by raccoons during the night (Schramm, personal communication).

Because of the low number of Blarina brevicauda, the immediate effect of the prairie fire could not be determined. One new individual was captured on Grid III following the burn, but it was caught near the southwestern edge, suggesting it had immigrated from the brushy swale (Fig. 1) rather than from the burned area. It must be assumed that Blarina leave the burned area as the other small mammals do. However, unlike Microtus, Blarina are well suited for life within the forest, and are thus very likely to survive the aftermath of the burn. Barbehen (1958) noted that all preferred Microtus habitat seemed to be favorable Blarina habitat. That was more or less the case here, in that judging from the data in Table 2, Blarina preferred areas with a lot of litter to areas with no litter. They carried this further than did the Microtus (which seemed to favor one year of litter to three years of litter). When the summer data are compared, showing maximum and minimum densities recorded for each years accumulated litter (Fig. 12), it becomes immediately apparent that Blarina prefer areas with at least two years of litter accumulation. There was very little difference among densities in areas with two, three, or four years of accumulated litter. Thus, a density of 26 or 27 shrews per acre seems to be the stable mean density on the restored East Prairie during the summer months. To maintain such a stable density, there should always be portions of the prairie which have not been burned for at least two years.

Peak numbers of shrews were observed during the fall trapping period, not the summer as Buckner (1966) and Christian (1969) reported. Platt (1968) also caught more Blarina in the fall than during any other season, as did Barbehen (1958). There are two possible explanations for this phenomenon. First is that there is late summer reproduction, which does not seem too likely an answer in that few of the captured Blarina appeared to be immature. The second possible answer is that females, which had not been caught during the reproductive period of early summer, were finally in the trapable population. Dapson (1968) reported that female Blarina drop out of the trapable population during pregnancy and birth. Gottschang (1965) reported that the sex ratio of Blarina was very near 1:1 during the fall, as was the case in the East Prairie in the fall (Table 2). At no other time did I find the sex ratio approaching 1:1, except on Grid II in the spring of 1970 when the population density was so low that the data may not have been a true indicator of the facts.

It seems that Blarina rarely feed on Microtus in the wild. Whitaker and Ferraro (1963) found absolutely no small mammal remains in any of the Blarina stomachs they studied. Of the 70 Blarina scats collected by Barbehen (1958), none contained mouse hair. Barbehen concluded that there seemed to be no causal relationship between the fluctuations of Microtus population densities and the population of Blarina in the same area. He noted that there was no ecological interaction between the two species implied by his data. Philips (1956) placed a Microtus in with a caged Blarina. It took 11 minutes for the shrew to kill the mouse, during which time the Microtus frequently ran away from the shrew. Rood (1958) tried a similar experiment using a Peromyscus. The Blarina could only catch the mouse in a very small cage where escape was impossible. He noted also that Blarina were unable to find mealworms at a distance greater than three inches. These data make it appear fairly unlikely that a shrew would prey on mice in the field. Barbehen (1958) suggested that if a Microtus population were high, a great many nests with young in them would exist, thus increasing the chances of a shrew wandering into one. If the mother would not protect the nest, the shrew would dine on the helpless young, which are hairless and would leave no remains in Blarina scats. It is possible, therefore, that Blarina seriously affect the natality rate of Microtus when there is a high population density of voles. During my study there was no high population density of voles, so it seems unlikely that Blarina had much affect on their numbers.

I prefer another answer. These animals have been driven out of the East Prairie, due to the fact that the East Prairie is an isolated island of restored prairie, and the fact that much of it had been burned every year, combined with the fact that the animals mentioned above have home ranges nearly the size of a grid (or larger). Some of them no doubt could have gone to the Woodcock Field (Fig. 1), and some may have been able to survive in the forest. It seems, though, that the great majority were lost to the area, and any new individuals caught were probably reinvaders from either the Woodcock Field or from the West Prairie by way of the Woodcock Field.

Schramm (1970) suggested that Zapus hudsonius were benefited by prairie fires because the lack of litter increased the ease of locomotion. This may be the case. Although my data were too few to substantiate Schramm's suggestion, it is true that I caught more Zapus during the summer of 1970 (after nearly all the East Prairie was burned) than I caught during the summer of 1969 (when only about half the prairie was burned). Zapus seem to have very large home ranges, and they could easily move to the Woodcock Field after a fire, waiting until the growth of the prairie was sufficient for reinvasion.

I would like to thank Nancy Hoover who helped collect data during the spring and summer of 1970, Lou Moreth who worked along with me during the winter and spring of 1971, Dennis Farrell who helped collect data during the spring of 1971, Barb Dunn who typed the original manuscript, and Dr. Peter Schramm who advised and counseled me all along the way.

I would also like to express my appreciation for grants from the Sloan Foundation and the National Science Foundation, which made the two summer trapping sessions possible.

Barbehen, K. R.  1958.  Spatial and populational relationships between Microtus and Blarina.  Ecology, 39:293-304

Blair, W. F. 1940. Notes on home ranges and populations of the short-tailed shrew. Ecology, 21:284-288.

Doremus, H. M. 1964. Live-trapping the short-tailed shrew, Blarina brevicauda. J. Mammal., 45:144-146.

New, J. G. 1959. Causes of mortality in short-tailed shrews. J. Mammal., 40:244-245.

Platt, A. P. 1968. Differential trap mortality as a measure of stress during times of population increase and decrease. J. Mammal., 49:331-335.

Rood, J. P. 1958. Habits of the short-tailed shrew in captivity. J. Mammal., 39:499-507.