Weasel population response, home range and predation on rodents in a deciduous forest in Poland.
Linked fluctuations in numbers of small rodents and their specialist predators, the small mustelids, have been observed in the Holarctic and New Zealand (Nasimovich 1949, Tapper 1979, Krivosheev 1981, Delattre 1983, Gorbunov 1983, King 1983, Oksanen and Oksanen 1992). It is unclear, however, if mustelids simply follow the changes in rodent abundance with little consequences for prey demography, or if they drive the cycles or fluctuations (MacLean et al. 1974, Fitzgerald 1977, Delattre 1984, Henttonen et al. 1987).
Hansson (1987) and Hanski et al. (1991) hypothesized that small mustelids drive microtine cycles in northern Europe or at least increase the amplitude and length of the cycle and deepen and prolong the crash. This hypothesis assumes that small mustelids, as rodent specialists, can only make a small dietary shift to alternative prey, and the changes in their density lag behind those in voles. Korpimaki et al. (1991) tested this hypothesis by studying the response of stoats Mustela erminea and weasels M. nivalis to the weakly cyclic bank vole Clethrionomys glareolus, field vole Microtus agrestis, and common vole M. epiroticus populations in Alajoki, Finland. They found that density of weasels (but not stoats) lagged 0.5-1 yr behind the vole abundance, and thus, weasels may have increased the amplitude and length of the vole cycle.
Hansson (1987) further predicted that in temperate Europe, where the predator community is dominated by generalists (relying heavily on other prey, alternative to rodents), small mustelids would play a minor role in total predation and rodents should be noncyclic. In two localities where rodents maintain fairly stable seasonal fluctuations over several years (Revinge, southern Sweden, Erlinge et al. 1983; and Bialowieza, eastern Poland, Jedrzejewski and Jedrzejewska 1993) weasels and stoats contributed only 7-9% and 15-18% to the total predation, respectively. Rodents in Revinge and Bialowieza, as well as noncyclic forest rodents in Turew (western Poland, Goszczynski 1977) were predominantly preyed upon by generalist predators.
Estimates of the magnitude of weasel predation on cyclic rodent populations are not yet available. Small mustelids are not necessary for rodent populations to cycle. Lemmings (Lemmus sibiricus and Dicrostonyx torquatus) on Wrangel Island (Chukchi Sea) show 4-5 yr cycles (Chernyavskii 1979, Chernyavskii and Dorogoi 1981) despite the absence of weasels and stoats and the fact that no other predator (Stercorarius pomarinus, S. longicaudus, Larus hyperboreus, Nyctea scandiaca, and Alopex lagopus) is a resident rodent specialist.
This paper analyses the pattern and magnitude of, and constraints on, weasel predation on forest rodents (the bank vole, Clethrionomys glareolus, and the yellow-necked mouse, Apodemus flavicollis) during 7 yr in the pristine deciduous forests of Bialowieza National Park, eastern Poland. The population dynamics of rodents was characterized by 4-6 yr at moderate densities followed by 2 yr of outbreak and crash. The outbreak was triggered by a superabundant, synchronous seed crop of oak Quercus robur, hornbeam Carpinus betulus, and maple Acer platanoides. The aims of our study were to investigate: (1) numerical response of weasels to the changes in forest rodent abundance; (2) size of home ranges of male weasels during the rodent outbreak and crash; and (3) the role of weasel predation in regulating the numbers of forest rodents.
The waves of outbreak and crash of rodent populations in the deciduous forests of Bialowieza shared some features with the northern "true" cycles of microtines: (1) they begin with winter breeding of rodents, (2) the ratio of peak to crash numbers may be up to 100-fold, and (3) the crash lasts 1 yr (Pucek et al. 1993). There is, however, an essential difference: populations of north European rodents have consistent 3-4 yr cycles and no noncyclic periods (Hansson 1987).
In Bialowieza, weasels coexist with [greater than]20 species of carnivores and birds of prey (Jedrzejewski and Jedrzejewska 1993). Stoats are uncommon (winter density 0.2 individual/[km.sup.2]). Detailed data on the autumn-winter predation on rodents exerted by eight species of predators were presented by Jedrzejewski and Jedrzejewska (1993).
MATERIALS AND METHODS
Bialowieza National Park (=BNP, 52 [degrees] 43[minutes] N, 23 [degrees] 54[minutes] E), covering 47.5 [km.sup.2], forms part of the extensive woodlands (1450 [km.sup.2]) located on the Polish-Belorussian border. BNP (UNESCO's Man and Biosphere Reserve and World Heritage Site) protects the last remnants of the pristine forests of European lowland, still unaltered by human activity. The forest is dominated by mature stands of oak, hornbeam, lime (Tilia cordata), and maple, with admixtures of spruce (Picea abies). Alderwoods (with black alder Alnus glutinosa) cover wet localities, and floodplain forest with ash (Fraxinus excelsior), black alder, and elm (Ulmus scabra) is found in the vicinity of small rivers. Mixed coniferous forests of spruce and pine (Pinus sylvestris) grow on poor, sandy soils. Detailed information on the vegetation of BNP is given by Falinski (1986). The western and northern borders of BNP are small rivers (Narewka and Hwozna) with narrow belts of treeless marsh. From the south, BNP adjoins the open meadows and fields of the Bialowieza Glade. The climate of BNP is of transitional character but continental features prevail (Olszewski 1986).
Estimating rodent abundance
Rodents were sampled by three methods: (1) removal trapping conducted once yearly on seven sites in various forest types in BNP; (2) removal trapping conducted three times a year in oak-lime-hornbeam forest; and (3) capture-mark-recapture trapping conducted nine times during 2 yr of outbreak and crash on a 3.24-ha plot in oak-lime-hornbeam forest.
In method 1, rodents were trapped in October of 1986 through 1991. Each site was operated for 6 nights using 4 snaptraps, 4 livetraps (all baited with oats, oil, and parsley), and 2 cones.
Method (2) was part (1985-1992) of a long-term trapping program of small mammals in BNP (see Pucek et al. 1993 for details). The trapping area was in the southern part of a 11.2-[km.sup.2] area in the southwest portion of BNP where weasels were studied intensively. Rodents were sampled by 50 cones, 50 livetraps and 50 snaptraps. The removal trapping was conducted three times a year: in spring (43-47 d from 15 April), summer (28-31 d from 1 July), and autumn (30-47 d from 15 September). The first 21 d of each trapping series were used to make the results of all series comparable (Appendix 1).
Indices of combined autumn numbers of bank voles and yellow-necked mice obtained by methods 1 and 2 in 1986 through 1991 were correlated (Kendall's [Tau] = 0.87, P = 0.05). Both methods, however, underestimated the abundance of rodents in autumn 1989, when the rodents were not attracted to traps due to the heavy seed crop. (Overwintered adults were more numerous in the spring 1990 trapping series than all rodents captured in autumn 1989.) Thus, the value interpolated between summer 1989 and spring 1990 was used as a more reliable estimate of rodent numbers in autumn 1989.
In method 3, rodents were livetrapped using the capture-mark-recapture method on the 3.24-ha plot (southeast part of the 11.2-[km.sup.2] study area) in May, July, September, and November 1990, and April, May, July, September, and November 1991 (i.e., during rodent outbreak and crash). Each trapping series lasted 6 d. Four traps baited with parched oats were set at each station on a 12 x 12 grid spaced at 15 m. Traps were checked twice daily. The density of rodents was estimated by the minimum number alive method for the entire trapping plot and for the inner square (1.44 ha, two outer trap belts removed) to check for an edge effect (Hansson 1969). Since the two estimates differed, on average, by only [+ or -] 10%, we used that for the entire 3.24-ha plot.
For estimating rodent densities in spring, summer, and autumn during 1985/1986-1992, we correlated the results of studies done by various researchers within our 11.2-[km.sup.2] study area on densities of mice (19 seasons) and bank voles (9 seasons) with our indices of rodent numbers obtained concurrently by method 2 [ILLUSTRATION FOR FIGURE 1 OMITTED]. Then, we converted the indices (N/100 trap-nights; data in Appendix 1) into density (N/ha) by reading the appropriate values from the regression lines.
Estimating abundance and home range size of weasels
Livetrapping of weasels was conducted in the summers (July-August) of 1986, 1987, and 1989. More frequent trapping (8 series during 19 mo) was done in 1990 and 1991 to follow weasel numbers during the outbreak and crash of rodents. Wooden flip-door box traps with a separate bait compartment were set at ground level near fallen logs or in dense vegetation at 50-m intervals. The trapping transect, located in the southwest part of the main study area and going through the oak-lime-hornbeam forests most typical of BNP, ranged from 1000 m in 1987 to 3050 m in 1991, but the core section was always the same. Traps were operated for 9-25 d (426 to 999 trap-nights per 1 trapping series; mean [+ or -] 1 SD, 695 [+ or -] 173 trap-nights, n = 11).
TABLE 1. Home range sizes (ha) of male weasels radiotracked in Bialowieza National Park in 1990 and 1991. N = total number of localizations (at 15-min intervals) used for home range calculation and accepted as 100% data in both methods. Mean HR = mean home range size; D mean HR = diameter of the mean home range (m) calculated on the assumption of a circular home range.
Minimum convex polygon Harmonic mean
Weasel N 100% 95% 90% 95% 90% 75%
1 256 37.4 24.1 18.7 27.0 17.9 9.7 2 515 30.9 16.1 12.0 19.2 13.6 7.3 4 109 12.6 10.5 9.1 7.6 6.7 3.5 8 132 10.6 6.5 4.9 7.0 4.6 2.5 10 74 29.5 25.6 23.4 10.3 6.2 2.6 Mean HR 24.2 16.6 13.6 14.2 9.8 5.1 (SD) (11.9) (8.3) (7.4) (8.7) (5.7) (3.2) D mean HR 555 460 416 425 353 255
12 503 215.8 109.2 83.8 139.0 100.5 42.6 16 471 117.4 88.9 75.6 91.8 73.2 42.2 Mean HR 166.6 99.1 79.7 115.4 86.9 42.4 (SD) (69.6) (14.4) (5.8) (33.4) (19.3) (0.3) D mean HR 1457 1124 1007 1212 1052 735
In each trap, two live laboratory mice were used as bait and given food and bedding. Traps were checked twice daily. When a weasel was caught, it was anaesthetised with ether, its sex and mass were determined and its ears marked by punching. In total, 36 weasels were caught (24 in the summer trapping series and 12 in other seasons). N weasels/100 trap-nights was used as an index of weasel numbers.
Weasels were radiotracked in 1990 and 1991 (see list in Appendix 2). A weasel to be radiocollared was taken to the laboratory, anaesthetised with ether, had its sex determined, and mass and neck circumference measured. On the following day, the weasel was sedated with Ketalar and a radio collar with a loop antenna (3.5-4.5 g, AVM Instrument Company, Livermore, California) was fitted around its neck. The weasel was then released into an outdoor enclosure (4.5 x 12 m) and provided with live rodents and a nest box with meat and water. During the next 1-3 d (after the first rodents killed by a weasel had appeared in its nest box), we released the weasel back to the forest in the place of capture. Such "ambulatory" treatment was necessary because weasels are extremely sensitive to handling and collaring (see Delattre et al. 1985). All collared weasels remained in their territories around the place of initial capture. In summer 1990, both males and females were caught, but transmitters small enough for females were not available. In autumn of 1990 through 1991, no females were trapped.
The range of the signal emitted by radio collars was 100-400 m, usually 200 m. The weasel was followed at a distance of [less than]50 m and its position estimated every 15 min in relation to the grid of transects for snow-tracking. The error of such positioning was [less than]50 m. Weasels were not disturbed by radiotracking and were often seen at 5-15 m (Jedrzejewski et al. 1992).
Out of 12 radio-collared weasels, 7 yielded enough data to calculate home ranges (Table 1). Weasels were located at 15-min intervals during activity (i.e., when not in a den) and once during inactivity (when in a den for [greater than]15 min). Home range size was estimated with the program McPAAL (Stuwe 1988). Two estimators were used (White and Garrott 1990): (1) minimum convex polygon (MCP) with 100, 95, and 90% of data points; (2) Dixon-Chapman harmonic mean (HM) with 95, 90, and 75% of data points (reliable calculation with 100% of data was not possible).
During the winters of 1985/1986 through 1991/1992 (except for the mild winter of 1989/1990; see Appendix 1) snowtracking was conducted on a grid of transects (total length 59.25 km) covering the 11.2-[km.sup.2] study area. During the 2nd and 3rd d after a new snowfall, the transects were walked and weasel tracks crossing each section were counted. Most sections were 533 m long. Each section was tracked 3-11 times during a winter. Total length of snowtracking in consecutive winters varied from 74.7 km in 1990/1991 to 331.9 km in 1987/1988 (mean 177.2 km). For those days when snowtracking was done on 20 or more sections (i.e., [greater than]10 km of transects) the number of recorded tracks was converted into N tracks per kilometre per day. This index was used to assess the effect of snow depth on detectability of weasels and relative changes in weasel numbers in the course of one winter and between winters.
To estimate the winter numbers of weasels, all fresh (1-2 d old) tracks recorded during both the regular snowtracking and any field work were mapped. Then, the results of radiotracking done in 1990 (rodent densities high in summer, moderate in winter) were applied to the snowtracking data from winters 1985/1986 through 1990/1991 (all winters with moderate rodent densities), and the results of radiotracking in 1991 to snowtracking data from winter 1991/1992 (very low rodent numbers). We assumed that: (1) there was little territory overlap in winter (Lockie 1966, King 1975; radiotracking data for two weasels in BNP); (2) the diameter of an average winter home range was [less than]555 m in 1985/1986-1990/1991 and [less than]1457 m in 1991/ 1992, i.e., the mean values for weasels radiotracked in 1990 and 1991, respectively; and (3) the tracks of the same day that were located [greater than]400 m apart (in winters 1985/1986 through 1990/1991) or 600 m (in winter 1991/1992) were tracks of two different weasels. These values were derived from the frequency distribution of the maximum straight line distances of daily movements of weasels. For those days when a weasel was radiotracked continuously for 24 h, we calculated the maximum distance between any two locations during that day. In 1990 (n = 27 d), 80% of the maximum daily distances were [less than]400 m (280 [+ or -] 156 m; [mean] [+ or -] SD, range 20-605 m), whereas in 1991 (n = 30 d), 73% were [less than]600 m (457 [+ or -] 241 m, range 50-930 m). Moreover, the mean diameter of usable home ranges (95% HM estimates or 90% MCP estimates) was slightly above 400 m in 1990, and 1000 m in 1991.
After mapping of weasel tracks for each winter, the schematic round "territories" were drawn that included the tracks estimated to belong to one weasel. To calculate density, the number of such "territories" was divided by 13.9 [km.sup.2] (a boundary belt 200 m wide, equal to half of the assumed weasel "territory" diameter, was added to the 11.2-[km.sup.2] study area in the winters of 1985/1986 through 1990/1991) or by 15.25 [km.sup.2] (300-m belt added in winter 1991/1992). With the above assumptions, the numbers of weasels in winters may have been somewhat overestimated, but the maximum error (derived front the data on daily movement distances of radiotracked weasels) should not have exceeded 20% in 1985-1991 and 27% in 1991/1992. However, an underestimation was also possible due to low detectability of weasels during deep snow cover and possible territory overlap. The overlap could have been proportional to the frequency of occurrence of [greater than]1 track laid in a given "territory" on the same day at least once during winter. If so, the average underestimate of weasel numbers in all winters does not exceed 37%.
Weasel densities in summers, and in other seasons when indices of densities only were available from live-trapping, were estimated from a correlation between indices and density estimates for the same seasons [ILLUSTRATION FOR FIGURE 2 OMITTED]. Three of the density estimates were from snow-tracking in winter and two came from radiotracking, visual observations and livetrapping of weasels along our 2122 m of trap line. The area covered by the trap line was assumed to be the mean diameter of home ranges of radiotracked weasels (555 x 2122 m = 1.18 [km.sup.2]; see Appendix 1). For five seasons for which we had density indices (N per 100 trap-nights), we converted the indices to density estimates (N per square kilometre) from the correlation [ILLUSTRATION FOR FIGURE 2 OMITTED]. Although the two summer density points are the minimum estimates (some weasels might not have been noticed), the regression is reliable, as its origin lies slightly above the point (0, 0), which is a prerequisite for relating trapping indices to absolute numbers.
Analysis of weasel-rodent relationships
The seasonal and between-year changes in the ratio of prey to predator densities and the predation of weasels on voles and mice in autumn-winter (1 October-15 April 1985/1986 through 1991/1992) were calculated. The numbers of rodents removed by weasels from an average hectare during 197 d of autumn-winter seasons were estimated in two ways from three sets of data: (1) early winter densities of weasels obtained by mapping of "territories" from snowtracking; (2) minimum estimates of weasel food consumption (Jedrzejewska and Jedrzejewski 1989); and (3) maximum estimates of kill rates of weasels (Jedrzejewska and Jedrzejewski 1989). Calculations were made for bank voles and yellow-necked mice, separately, and then combined to estimate the total predation on both species. Method 1 was designed to provide a minimum estimate of numbers of rodents killed. Method 2 provided a maximum.
Radiotracking of weasels revealed that 99% of weasels' attacks were on rodents (Jedrzejewski et al. 1992). We assumed that in all years they preyed exclusively on rodents and killed the two species proportionally to their abundance in the rodent community (after Erlinge 1975, King 1980a).
In method 1, the minimum reliable number of bank voles ([N.sub.bv]) removed by weasels from 1 ha was calculated as:
[N.sub.bv] = D [multiplied by] (M [multiplied by] [B.sub.bv])/17,
where D = density of weasels (N/ha); M = total biomass (in grams) eaten by one weasel during 197 d of autumn and winter, calculated from data on food consumption by captive weasels at various ambient temperatures (Jedrzejewska and Jedrzejewski 1989). Daily food consumption averaged 30.0 g at ambient temperature [greater than]15.1 [degrees] C, 32.0 g at 5.1 [degrees] -15 [degrees], 33.2 g at -5 [degrees] to 5 [degrees], and 35.0 g below -5 [degrees]; [B.sub.bv] = fraction of bank voles in the number of rodents captured during autumn trapping (by method 1 of rodent trapping); average body mass of bank voles captured during autumn trappings was 17 g.
The minimum reliable number of yellow-necked mice removed by weasels from 1 ha was calculated in the same way, with [B.sub.ym] = fraction of yellow-necked mice in autumn trapping; and 31 g = mouse body mass.
In method 2, the kill rate by weasels was used instead of the consumption rate, to take into account the seasonal surplus killing governed by ambient temperature in autumn and winter (Jedrzejewska and Jedrzejewski 1989).
Numbers of bank voles ([N.sub.bv]) and yellow-necked mice ([N.sub.ym]) removed from 1 ha were calculated as:
[N.sub.bv] = D [multiplied by] (K [multiplied by] [B.sub.bv])/17 and [N.sub.ym] = D [multiplied by] (K [multiplied by] [B.sub.ym])/31.
K = total biomass of rodents killed by a weasel during 197 d, calculated from data on temperature and mean kill rate at given ambient temperatures (Jedrzejewska and Jedrzejewski 1989). The kill rates (in grams of rodents per day) were: 36.0 at temperatures [greater than]5.1 [degrees] C, 46.8 at -5 [degrees] to 5 [degrees], and 11.1 at [less than]-5 [degrees].
Data on weasel predation during the winters of 1986/1987 through 1988/1989 were briefly presented by Jedrzejewski and Jedrzejewska (1993) as a part of the impact of the predator community on forest rodents.
Home range size in male weasels
Home ranges (minimum convex polygons using all locations) of 5 male weasels radiotracked during 1990 (rodent outbreak) averaged 24.2 [+ or -] 11.9 ha (mean [+ or -] 1 SD, range 10.6-37.4 ha) (Table 1, [ILLUSTRATION FOR FIGURE 3 OMITTED]). The home ranges of two males (numbers 1 and 2) radiotracked during the mating season were larger than ranges held by males numbers 4 and 8 in autumn-early winter (i.e., after the reproductive season), although the numbers of rodents declined in the cold season. The two males tracked during 1991 (crash of rodents) had, on average, home ranges seven times larger than weasels tracked in 1990 (Table 1). The difference was statistically significant for all six variants of home range estimation (Student t test, t = 5.17-15.49, df = 5, P [less than] 0.001 to [less than] 0.01).
In continuous forests of BNP, weasel home ranges were approximately circular. The intensively used parts of weasels' home ranges (25% of outer localizations excluded) covered, on average, only 24 [+ or -] 8.2% of the maximum area utilized by weasels (Table 1). Although the radiotracking could not show the degree of home range overlap, other sources of information (livetrapping in the same area, casual trapping of weasels in rodent traps, and visual observations) showed the presence of 5 other weasels (2 females and 3 males) within the home range of male number 1, and 3 other weasels (1 female, 1 male, and 1 undetermined) within the range of male number 2 in summer 1990 [ILLUSTRATION FOR FIGURE 3 OMITTED]. Towards autumn, the number of other weasels recorded in the ranges of radiotracked specimens decreased. In early winter 1990 and in 1991, no other weasels were recorded in the home ranges of radiotracked males [ILLUSTRATION FOR FIGURE 3 OMITTED].
Weasel abundance in winter and summer seasons
Two factors influenced numbers of weasel tracks recorded during snowtracking censuses on transects. The detectability of weasels decreased with increasing snow depth as weasels used subnivean spaces, and the numbers of recorded tracks decreased through the course of winter [ILLUSTRATION FOR FIGURE 4 OMITTED]. Snow cover (SC) and consecutive day of winter (D) explained nearly 40% of variation in the mean number of tracks per kilometre per day during six winters (multiple regression, Y = 1.306 - 0.0011(SC) - 0.005(D), [R.sup.2] = 0.393, P [less than] 0.0005). Contributions of these two factors to the relationship were similar (squared semipartial correlations, [[sr.sup.2].sub.D] = 0.10, and [[sr.sup.2].sub.sc] = 0.12). (In standard multiple regression, squared semipartial correlation ([[sr.sub.j].sup.2]) for an independent variable is the amount by which [R.sup.2] would be reduced if that independent variable were not included in the regression equation [Tabachnick and Fidell 1983:107-108]).
In four winters, when tracking covered most of the cold season, fewer tracks were observed in late winter (late February-early April) than in early winter (December) [ILLUSTRATION FOR FIGURE 4 OMITTED]. During this period, the percent decrease of the averaged track indices (indicating winter mortality of weasels) was 81% in 1985/1986, 67% in 1986/1987, 57% in 1987/1988, and 100% in 1991/1992, for an overall average of 76 [+ or -] 18.6%.
Mapping of all tracks in the entire study area [ILLUSTRATION FOR FIGURE 5 OMITTED] showed that weasel density (representative of early winter) varied from 5.2 individuals/10 [km.sup.2] in 1991/1992 (rodent crash year) to 27.3 individuals/10 [km.sup.2] in 1987/1988 and averaged 18.4 [+ or -] 7.4 individuals [ILLUSTRATION FOR FIGURE 6 OMITTED]. If only the "territories" that persisted after the 151st d of winter (i.e., after 28 February) were counted, winter mortality was 75% in 1985/1986, 42% in 1986/1987, 63% in 1987/1988, 40% in 1990/1991, and 100% in 1991/1992 and averaged 64 [+ or -] 24.9%.
The index of weasel abundance in July/August in 1986-1991 varied from 0.2 individual/100 trap-nights (corresponding to 19.1 weasels/10 [km.sup.2]) in 1991 to 1.6 (101.7 weasels/10 [km.sup.2]) in 1990 [ILLUSTRATION FOR FIGURE 6 OMITTED]. In the years of moderate rodent numbers it was 0.6 or 0.7 weasel/100 trap-nights (41.9 and 47.6/10 [km.sup.2], respectively). The average index of summer numbers of weasels was 0.76 [+ or -] 0.51 individual/100 trap-nights.
Rodents and weasels during the outbreak-crash years
After the superabundant seed crop of oak, hornbeam, and maple in 1989, both mice and bank voles bred in winter 1989/1990 (Pucek et al. 1993). Capture-mark-recapture trapping of rodents showed that their numbers continued to increase until July-September 1990 [ILLUSTRATION FOR FIGURE 7 OMITTED]. In September 1990, 269 rodents/ha were recorded, and breeding nearly ceased: only 13% of female mice and 7% of female voles were pregnant. Winter decline was rapid. In May 1991, only 1.8 rodents/ ha were found. Very low numbers were recorded throughout 1991 [ILLUSTRATION FOR FIGURE 7 OMITTED].
In response to a rodent outbreak in 1990, weasel numbers increased from May till July/August, when they attained the highest abundance [ILLUSTRATION FOR FIGURE 7 OMITTED]. Weasels began to decline in late summer. In September, their abundance was only slightly higher than those in summers of 1986-1989 with moderate numbers of rodents. The whole year of rodent crash was characterized by scarcity of weasels; their abundance in July/August 1991 was one-third of that in years of moderate rodent numbers [ILLUSTRATION FOR FIGURES 6 AND 7 OMITTED]. The same dynamics of weasel numbers emerged from counting the radiotracked, observed, and captured weasels in the area. In July/August 1990, 12 weasels were recorded, in late autumn 5, and in 1991 one weasel at any time [ILLUSTRATION FOR FIGURE 3 OMITTED]. Twelve data points of weasel abundance in 1990 and 1991 (obtained by the two methods and shown in Fig. 7) were expressed as a percentage of the maximum values recorded by these methods. These standardized indices were significantly related to rodent density with no time lag (Y = 1.19 + 0.265X, df = 10, [R.sup.2] = 0.78, P [less than] 0.0005).
Long-term dynamics of weasel and rodent populations
Conversion of indices of rodent and weasel numbers into densities allowed analysis of the dynamics of predator and prey populations and the predator/prey ratios. The weasel population was characterized by regular, though variable, increase from spring (April) to midsummer (July/August) [ILLUSTRATION FOR FIGURE 8 OMITTED]. The rate of weasel increase in spring-summer (W[I.sub.sp-su], in N weasels per day per square kilometre) depended on the spring density of rodents ([R.sub.sp] in N rodents per square kilometre): W[I.sub.sp-su] = 0.208 + 0.00004[R.sub.sp]; [R.sup.2] = 0.84, n = 5, P = 0.03. From midsummer to winter (December), weasels always decreased in numbers and the rate of this decrease was correlated with their spring-summer increase (r = 0.924, n = 5, P [less than] 0.025). The rate of weasel population decline in winter (December-late March) was related neither to the rate of rodent population decline nor to the mean temperature of winter. This decline was the most stable part of the multiannual dynamics of weasel population (CV = 36%).
Numerical response of weasels to changes of rodent numbers was always log-shaped. Weasels attained higher densities in summer than in other seasons at the same densities of rodents [ILLUSTRATION FOR FIGURE 9 OMITTED].
In years of moderate rodent density (1985/1986-1989), the prey/predator ratio did not exceed 3400 rodents/weasel [ILLUSTRATION FOR FIGURE 8 OMITTED]. After the fall of mast in autumn 1989, rodents reproduced throughout the winter (Pucek et al. 1993). Weasels were never recorded to reproduce in winter in Poland (Jedrzejewska 1987). Their low numbers in early spring 1990 testify that they did not breed in winter 1989/1990. Thus, from early spring 1990 until late winter 1990/1991, the rodent/weasel ratio was high (up to 12 000 rodents/weasel). In the crash year, it was low until April 1992, when the weasels disappeared from the area [ILLUSTRATION FOR FIGURE 8 OMITTED].
The predator/prey ratio (N weasels/1000 rodents) expresses the relative impact of weasels on rodents [ILLUSTRATION FOR FIGURE 8 OMITTED]. Rodent numbers varied [greater than]50-fold, and weasels followed these changes within the fairly narrow range of 0 to 2.5 weasels/1000 rodents, on average 0.80 [+ or -] 0.62 weasel/1000 rodents (mean [+ or -] 1 SD). At moderate rodent densities, this ratio was on average 0.82 [+ or -] 0.49 weasel/1000 rodents. During the rodent outbreak, weasels exerted the smallest impact on rodents. The mean ratio was then 0.25 [+ or -] 0.25 weasel/1000 rodents, i.e., only 30% of that in the years of moderate densities of rodents. During the rodent crash the weasel/rodent ratio was highest, on average 1.46 [+ or -] 0.77 weasels/1000 rodents, i.e., 178% of the value recorded in moderate years.
The weasel/rodent ratio declined rapidly with increasing density of rodents [ILLUSTRATION FOR FIGURE 10 OMITTED]. Additionally, a clear seasonal pattern emerged. At the same density of rodents, this ratio was higher in summer and autumn (when weasels reproduced) than in winter and spring [ILLUSTRATION FOR FIGURE 10 OMITTED].
Weasel predation on rodents in autumn-winter seasons
Weasel predation was quantified in autumn-winter seasons when neither rodents (except for 1989/1990) nor weasels reproduced and the densities of weasels were estimated directly on the 11.2-[km.sup.2] study area. From 1 October through 15 April weasels removed from 1.6 (in 1991/1992) to 9.5 (in 1987/1988) forest rodents from an average hectare (Table 2). Predation amounted to 2-28% of the combined autumn densities of bank voles and yellow-necked mice (Table 2). The percent predation function (resulting from log-shaped numerical response of weasels), fitted to the empirical points, explained 96% of variation in weasel predation on rodents [ILLUSTRATION FOR FIGURE 11 OMITTED].
Weasel response to changes in rodent numbers
This study analyzed the role of weasels in a three-stage linked system of seed crop-rodents-predators in the primeval deciduous forests typical of the temperate zone of Europe. The seed-rodent part of this system is described by Pucek et al. (1993), who have presented data on rodent dynamics since the late 1950s. Here, the mechanisms promoting and inhibiting weasel numerical response in this system are shown.
Weasels responded to spring numbers of rodents. Population growth of weasels was, however, constrained by seasonality of climate, their own spacing behavior, and predation on weasels. While rodents started the 1990 outbreak with winter breeding, the weasels did not begin to respond numerically until spring. Weasels from Bialowieza belong to the vulgaris subspecies of Mustela nivalis (see Schmidt 1992), and they do not breed in winter (Jedrzejewska 1987, King 1989). Thus, weasels' numerical response occurred from spring to summer. During that time, depending on prey availability, weasels are biologically capable of breeding once or twice. When resources are superabundant, the litter size may increase to up to eight young (Jedrzejewska 1987) and weasels of both sexes can breed in the year of their birth. King (1989) estimated that given excellent food conditions (such as 1990 in BNP) weasels can increase their spring numbers by 30-fold by the end of summer. We observed a fivefold increase (from 19 to 102 weasels/10 [km.sup.2]) during 2.5 mo, which was possible given the reproduction parameters of weasels (see Jedrzejewska 1987).
Weasel dependence on rodents' spring density explains the negligible growth of weasel numbers during the rodent crash in spring-summer 1991. Erlinge (1974) estimated that the threshold spring density of voles at which weasels would breed was 5 reproducing female voles/ha. In BNP, only 1.8 female rodents/ha were recorded in April 1991, none in May, and 5 rodents/ha in June. Thus, the extinction of weasels by early spring 1992 was a consequence of a paucity of 1-yr animals. Only [approximately equal]2-5% of weasels can survive until their 3rd yr of life (King 1980b).
Every autumn, irrespective of rodent abundance, weasels declined proportionally to their increase in numbers from spring to summer. This decline can be explained by: (1) dispersal of young weasels from their natal territories (see King 1989), and (2) predation on weasels. The high midsummer density of weasels and the presence of young inexperienced weasels probably expose them to a higher risk of predation. Out of 8 collared weasels radiotracked for a total of 271 d in 1990, 1 was killed by a red fox Vulpes vulpes, 1 by another weasel, and 2 disappeared suddenly, supposedly due to predation (see Appendix 2). During the study, weasels were recorded in pellets of Common Buzzards Buteo buteo and Tawny Owls Strix aluco, and in scats of red foxes and pine martens Martes martes (0.2-0.8% frequency in scats or pellets, 0.5 on average; Jedrzejewski and Jedrzejewska 1992, Jedrzejewski et al. 1993a, 1994; W. and B. Jedrzejewski, unpublished data). Korpimaki and Norrdahl (1989) showed that such occurrence of weasels in the diets of predators can indicate fairly high predation on the weasel population. Powell (1973) showed mathematically that raptors can keep weasel numbers below the level set by abundance of Microtus voles.
The significant feature of weasel population dynamics in the pristine deciduous forests of Bialowieza National Park was the high correlation, with no time lag, of weasel and rodent numbers. A similar coincidence (no time lag) of changes in weasel and rodent numbers was reported from the Alaskan tundra by MacLean et al. (1974), the French farmland by Delattre (1983), and the Turkmen desert by Gorbunov (1983).
In contrast, Tapper (1979) and Korpimaki et al. (1991) reported a time lag in weasel numerical response to rodent numbers. Tapper (1979) found that, on English farmland, the relative changes in weasel numbers lagged behind those in Microtus agrestis by 9 mo. However, both the lag and its time span might have been a product of weasel carcasses being collected from January until June, i.e., before the birth of young cohorts and the seasonal peak of weasel numbers. The sample represented the survival of the previous year weasels rather than the current year numbers. In contrast to the indices of weasel numbers, the reproduction parameters given by Tapper (1979) (percent of females [TABULAR DATA FOR TABLE 2 OMITTED] pregnant and mean number of embryos per pregnancy) correlated well with the rodent numbers in the same year (Tapper 1979: Table 2). Korpimaki et al. (1991) snowtracked weasels twice a year (November/December and February/March) in Finnish farmland, on a 0.5-2.5 km transect. The effect of snow depth on tracking indices was not evaluated. Since the late autumn and late winter censuses missed the yearly midsummer peak in numbers of weasels, the multiannual variation might be "flattened." Positive correlations between weasel and rodent indices of abundance were found for no time lag, 0.5-yr, or 1-yr lag data, though more lag-time correlations were significant (Korpimaki et al. 1991).
Evidently, time lag in weasel response to changes in rodent numbers may occur, but especially careful methodology and year-round censusing is needed to document it. This study, and that by Tapper (1979), indicated that weasels adjust their breeding effort to spring numbers of rodents. In the decline phase of cyclic microtines, high spring numbers of voles are often followed by summer decline. In such years, a belated outbreak of weasels may occur and strongly deepen the [TABULAR DATA FOR TABLE 3 OMITTED] decline of voles (see, e.g., Goszczynski's 1977 study on cyclic Microtus arvalis and Kucheruk and Dunaeva's 1948 observations on cyclic Microtus brandti in Choybalsan, Mongolia).
Magnitude of weasel predation on small rodents
"All known techniques for counting small mammals are inaccurate, and errors of estimation will increase at compound interest through the series of calculations needed to estimate predation rates." Only with this caution by King (1989: 89) in mind can we proceed to interpret the results of weasel impact on forest rodents in this, and in other studies.
Six studies (this included) have attempted to quantify the impact of weasel predation on rodents (Table 3). Four of them estimated winter predation by weasels and yielded similar results: at their best, weasels removed up to 35% of available prey. Goszczynski (1977) and King (1980a) documented that the role of weasel predation decreased at higher rodent densities.
In Bialowieza National Park, the weasel/rodent ratio (as an index of predation impact) increased with decreasing numbers of rodents until some critical point, below which weasel numbers were cut by starvation and inability to reproduce. This critical point was [approximately equal]400 rodents/weasel. When rodents became scarce, weasels tried to cope with famine by extending their home ranges. The largest home range, at an extremely low density of rodents, was 225 ha and "included" [approximately equal]400 rodents. There is, however, a bioenergetic limit to expansion of a weasel's home range, which is also a threshold of territorial behavior and triggers the extinction of the local weasel population. Oksanen et al. (1992) reported that, in north Fennoscandia, weasels became nomadic during the crash of cyclic rodents in winter. Temporary extinctions of weasels during and after crashes of rodents were reported by Nasimovich (1949) in the taiga on Kola Peninsula (Russia), by Kadochnikov (1953) on the steppes of Azerbaijan, by MacLean et al. (1974) on Point Barrow tundra (Alaska), and by Gorbunov (1983) in the Turkmen desert. In the latter study conducted in the Zauzboi region, in marginal habitat for weasels, this predator was not trapped for 6 yr after two consecutive years of very low densities of gerbils.
The impact of weasels on rodents (expressed as weasel/rodent ratio) in the temperate forest has two rhythms: (1) it varies with changing density of rodents, and (2) at the same density of rodents, the impact of weasels is heavier in the reproductive season of both prey and predator (summer-autumn) than in winter and early spring, when weasels do not reproduce. Thus, the numerical response of weasels and their heaviest impact on rodents take place in summer-autumn, when rodents reproduce at a rate that more than compensates for losses from predation.
The pattern of weasel predation on rodents in Bialowieza Forest is not typical of temperate forests only. The same pattern emerged from data recalculated from Goszczynski's (1977) work on cyclic Microtus arvalis in western Poland and Gorbunov's (1983) 20-23 yr trapping of weasels in the colonies of noncyclic Meriones lybicus and Rhombomys opimus in the Turkmen desert [[ILLUSTRATION FOR FIGURE 12 OMITTED]. Both in Poland and Turkmenistan, weasels breed seasonally (from March through June in Turkmenistan, Gorbunov 1983). Perhaps this is a general model of the weasel-rodent relationship. It was formulated verbally by Pearson (1966), Goszczynski (1977), and Korpimaki et al. (1991), who hypothesized that resident predators would deepen the decline and prolong the low phase of the rodent cycle. It is essential to note that no time lag in weasel numerical response to variations in rodent numbers, which was believed to be necessary to produce such a pattern, is actually needed.
However, it must be borne in mind that in deciduous forests of Europe, the weasel is one of [[approximately equal to]10 species of predators that prey on rodents intensively (e.g., Jedrzejewska and Jedrzejewski 1993, Jedrzejewski and Jedrzejewska 1993, Jedrzejewski et al. 1993a, b). Thus, the role of predation in regulating rodent populations cannot be inferred from this paper without knowing the pattern of predation by other important predators such as Tawny Owls, pine martens, and red foxes.
We are grateful to Earthwatch Research Corps, E. McNeish, K. Zub, and W. Nowakowski for their help in livetrapping and radiotracking of weasels. The MRI technical staff (especially S. Bogdanska, E. Bujko, M. Szlachciuk, M. Swic, and J. Siemieniuk) greatly helped in snowtracking of weasels and trapping of rodents. Mr. L. Siemieniuk collected and elaborated the climatic data. K. Zub prepared the figures in Corel-Draw program. We thank Mrs. Barbara Kermeen (AVM Instrument Co.) for her cooperation and unfailing support. The project was financed by the Mammal Research Institute budget, the Earthwatch grant, and grant KBN 4 4416 91 02. We express our gratitude to Drs. J. Goszczynski, L. Hansson, C. M. King, and Z. Pucek for their comments on the earlier draft, and to Drs. R. A. Powell and L. Oksanen for their thorough reviews.
Chernyavskii, F. B. 1979. Population dynamics and population indices of Lemmus sibiricus and Dicrostonyx torquatus on the Vrangel Island. (In Russian with English summary.) Zoologicheskii Zhurnal 58:553-563.
Chernyavskii, F. B., and V. I. Dorogoi. 1981. O roli khishchnikov v dinamike chislennosti lemmingov (na primere ostrova Vrangelya.) [On the role of predators in the dynamics of lemming numbers (on the example of the Wrangel Island.)] Pages 32-50 in Ekologiya mlekopitayushchikh severo-vostochnoi Sibiri. [Ecology of mammals of NorthEast Siberia.] (In Russian.) Izdatelstvo Nauka, Moskva, USSR.
Delattre, P. 1983. Density of weasel (Mustela nivalis L.) and stoat (Mustela erminea L.) in relation to water vole abundance. Acta Zoologica Fennica 174:221-222.
-----. 1984. Influence de la pression de predation exercee par une population de belettes (Mustela nivalis L.) sur un peuplement de Microtidae. Acta Oecologica, Oecologia Generalis 5:285-300.
Delattre, P., J.-M. Angibault, and S. Poser. 1985. Influences du port d'un collier emetteur sur les deplacements et le rythme d'activite de la belette (Mustela nivalis). Gibier Faune Sauvage 2:5-13.
Erlinge, S. 1974. Distribution, territoriality and numbers of the weasel Mustela nivalis in relation to prey abundance. Oikos 25:308-314.
-----. 1975. Feeding habits of the weasel Mustela nivalis in relation to prey abundance. Oikos 26:378-384.
Erlinge, S., G. Goransson, L. Hansson, G. Hogstedt, O. Liberg, I. N. Nilsson, T. Nilsson, T. von Schantz, and M. Sylven. 1983. Predation as a regulating factor on small rodent populations in southern Sweden. Oikos 40:36-52.
Falinski, J. B. 1986. Vegetation dynamics in temperate lowland primeval forests. Dr. W. Junk, Dordrecht, The Netherlands.
Fitzgerald, B. M. 1977. Weasel predation on a cyclic population of the montane vole (Microtus montanus) in California. Journal of Animal Ecology 46:367-397.
Gorbunov, A. V. 1983. Kolebanie chislennosti kunikh v severo-zapadnoi Turkmenii i yuzhnom Ustyurte [Changes in numbers of mustelids in the north-west Turkmenistan and south Ustyurt]. (In Russian with Turkmen summary.) Proceedings of the Academy of Sciences of the Turkmen SSR, Series of Biological Sciences 4/1983:51-56.
Goszczynski, J. 1977. Connections between predatory birds and mammals and their prey. Acta Theriologica 22:399-430.
Hanski, I., L. Hansson, and H. Henttonen. 1991. Specialist predators, generalist predators, and the microtine rodent cycle. Journal of Animal Ecology 60:353-367.
Hansson, L. 1969. Home range, population structure and density estimates at removal catches with edge effect. Acta Theriologica 14:153-160.
-----. 1987. An interpretation of rodent dynamics as due to trophic interactions. Oikos 50:308-318.
Henttonen, H., T. Oksanen, A. Jortikka, and V. Haukisalmi. 1987. How much do weasels shape microtine cycles in the northern Fennoscandian taiga? Oikos 50:353-365.
Jedrzejewska, B. 1987. Reproduction of weasels Mustela nivalis in Poland. Acta Theriologica 32:493-496.
Jedrzejewska, B., and W. Jedrzejewski. 1989. Seasonal surplus killing as hunting strategy of the weasel Mustela nivalis - test of a hypothesis. Acta Theriologica 34:347-359.
Jedrzejewska, B., and W. Jedrzejewski. 1993. Summer food of the Pygmy owl Glaucidium passerinum in Bialowieza National Park, Poland. Ornis Fennica 70:196-201.
Jedrzejewski, W., and B. Jedrzejewska. 1992. Foraging and diet of the red fox Vulpes vulpes in relation to variable food resources in Bialowieza National Park, Poland. Ecography 15:213-221.
Jedrzejewski, W., and B. Jedrzejewska. 1993. Predation on rodents in Bialowieza primeval forest, Poland. Ecography 16:47-64.
Jedrzejewski, W., B. Jedrzejewska, and M. Brzezinski. 1993b. Winter habitat selection and feeding habits of polecats (Mustela putorius) in the Bialowieza National Park, Poland. Zeitschrift fur Saugetierkunde 58:75-83.
Jedrzejewski, W., B. Jedrzejewska, and E. McNeish. 1992. Hunting success of the weasel Mustela nivalis and escape tactics of forest rodents in Bialowieza National Park. Acta Theriologica 37:319-328.
Jedrzejewski, W., A. Szymura, and B. Jedrzejewska. 1994. Reproduction and food of the Buzzard Buteo buteo in relation to the abundance of rodents and birds in Bialowieza National Park, Poland. Ethology, Ecology and Evolution 6:179-190.
Jedrzejewski, W., A. Zalewski, and B. Jedrzejewska. 1993a. Foraging by pine marten Martes martes in relation to food resources in Bialowieza National Park, Poland. Acta Theriologica 38:405-426.
Kadochnikov, N. P. 1953. O vzaimootnoshenii khishchnykh ptic i obshchestvennoi polevki v stepnom Azerbaidzhane [On the relationships between birds of prey and the social vole in the steppe of Azerbaijan]. (In Russian.) Zoologicheskii Zhurnal 32:1222-1233.
King, C. M. 1975. The home range of the weasel (Mustela nivalis) in an English woodland. Journal of Animal Ecology 44:639-668.
-----. 1980a. The weasel Mustela nivalis and its prey in an English woodland. Journal of Animal Ecology 49:127-159.
-----. 1980b. Population biology of the weasel Mustela nivalis on British game estates. Holarctic Ecology 3:160-168.
-----. 1983. The relationships between beech (Nothofagus sp.) seedfall and populations of mice (Mus musculus), and the demographic and dietary responses of stouts (Mustela erminea), in three New Zealand forests. Journal of Animal Ecology 52:141-166.
-----. 1989. The natural history of weasels and stouts. Christopher Helm, London, England.
Korpimaki, E., and K. Norrdahl. 1989. Avian predation on mustelids in Europe. 1: Occurrence and effects on body size variation and life traits. Oikos 55:205-215.
Korpimaki, E., K. Norrdahl, and T. Rinta-Jaskari. 1991. Responses of stoats and least weasels to fluctuating food abundances: is the low phase of the vole cycle due to mustelid predation? Oecologia (Berlin) 88:552-561.
Krivosheev, V. G. 1981. Faktory regulyacii chislennosti myshevidnykh gryzunov i khishchnykh mlekopitayushchikh taigi Kolymskoi nizmennosti [Factors regulating the numbers of small rodents and predatory mammals in the boreal forests of the Kolyma Lowland]. (In Russian.) Pages 61-82 in Ekologiya mlekopitayushchikh severo-vostochnoi Sibiri [Ecology of mammals of North-East Siberia]. Izdatelstvo Nauka, Moskva, USSR.
Kucheruk, V. V., and T. N. Dunaeva. 1948. Materialy po dinamike chislennosti polevki Brandta (K voprosu o vozdeistvii epizootii na populyaciyu) [Data on dynamics of numbers of the Brandt's vole (On the influence of epizooty on population)]. (In Russian.) Fauna i ekologiya gryzunov 3:111-178.
Lockie, J. D. 1966. Territory in small carnivores. Symposia of the Zoological Society of London 18:143-165.
MacLean, S. F., Jr., B. M. Fitzgerald, and F. A. Pitelka. 1974. Population cycles in arctic lemmings: winter reproduction and predation by weasels. Arctic and Alpine Research 6:1-12.
Meshcheryakova, I. S., and T. N. Dunaeva. 1976. Discovery of tularemia in tundra rodents in West Yakutia. (In Russian.) Zoologicheskii Zhurnal 55:788-789.
Nasimovich, A. A. 1949. Biologiya laski na Kol'skim poluo-strove v svyazi s konkurentnymi otnosheniyami s gornostaem [Biology of the weasel on the Kola peninsula in relation to its competitive interactions with the stout]. (In Russian.) Zoologicheskii Zhurnal 2:177-182.
Oksanen, L., and T. Oksanen. 1992. Long-term microtine dynamics in north Fennoscandian tundra: the vole cycle and the lemming chaos. Ecography 15:226-236.
Oksanen, T., L. Oksanen, and M. Norberg. 1992. Habitat use by small mustelids in north Fennoscandian tundra: a test of the hypothesis of patchy exploitation ecosystem. Ecography 15:237-244.
Olszewski, J. L. 1986. The role of forest ecosystems in modifying local climate of the Bialowieza Primeval Forest, as revealed by air temperature characteristics. (In Polish with English summary.) Wydawnictwo Polskiej Akademii Nauk (Prace Habilitacyjne). Ossolineum, Wroclaw, Poland.
Pearson, O. P. 1966. The prey of carnivores during one cycle of mouse abundance. Journal of Animal Ecology 35:217-233.
Powell, R. A. 1973. A model for raptor predation on weasels. Journal of Mammalogy 54:259-263.
Pncek, Z., W. Jedrzejewski, B. Jedrzejewska, and M. Pucek. 1993. Rodent population dynamics in a primeval deciduous forest (Bialowieza National Park) in relation to weather, seed crop, and predation. Acta Theriologica 38:199-232.
Schmidt, K. 1992. Skull variability of Mustela nivalis Linnaeus, 1766 in Poland. Acta Theriologica 37:141-162.
Sikorski, M. D. 1989. System kojarzenia sie i charakterystyka sukcesu reprodukcyjnego w naturalnej populacji nornicy rudej [Mating system and reproductive success in a free-living population of bank vole]. Dissertation. (In Polish.) Mammal Research Institute, Bialowieza, Poland.
Stuwe, M. 1988. McPAAL - microcomputer programs for the analysis of animal locations. Version 2.0. Conservation and Research Center, National Zoological Park, Smithsonian Institution, Washington, USA.
Tabachnick, B. G., and L. S. Fidell. 1983. Using multivariate statistics. Harper & Row, New York, New York, USA.
Tapper, S. 1979. The effect of fluctuating vole numbers (Microtus agrestis) on a population of weasels (Mustela nivalis) on farmland. Journal of Animal Ecology 48:603-617.
White, G. C., and R. A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic Press, New York, New York, USA.
Wolk, E., and J. Kozlowski. 1989. Changes of body weight and hematological parameters in a fluctuating population of Apodemus flavicollis. Acta Theriologica 34:439-464.
Wojcik, A. M. 1993. Genetic variation in a fluctuating population of the yellow-necked mouse Apodemus fiavicollis. Acta Theriologica 38:273-290.
|Printer friendly Cite/link Email Feedback|
|Author:||Jedrzejewski, Wlodzimierz; Jedrzejewska, Bogumila; Szymura, Lucyna|
|Date:||Jan 1, 1995|
|Previous Article:||Interactions between predation risk and competition: a field study of kangaroo rats and snakes.|
|Next Article:||Resource consumption variance within and among individuals: on coloniality in spiders.|