Arch. Environ. Contain. Toxicol. 27, 521-526 (1994)

A R C H I V E S

OF

Environmental Contamination a n d Toxicology © 1994Springer.VerlagNew York Inc.

Characteristics of a White-Footed Mouse (Peromyscus leucopus) Population Inhabiting a Polychlorinated Biphenyls Contaminated Site A. V. Linzey 1, D. M. Grant 2 i Department of Biology, Indiana University of Pennsylvania, Indiana, Pennsylvania 15705, USA. 2 Department of Toxicology, North Carolina State

University, Raleigh, North Carolina 27695, USA Received: 24 January 1994/Revised: 13 May 1994

Abstract. Demographic characteristics of a white-footed mouse (Peromyscus leucopus) population inhabiting a woodland containing a tow level of PCBs contamination were compared to an uncontaminated reference site. Although population density was higher on the PCBs-contaminated grid, the population exhibited greater temporal variability among years and had a higher number of transient animals. These population trends are consistent with those observed in structurally suboptimal habitats, indicating that the PCBs-contaminated grid is functioning as a suboptimal habitat. However, the data are not compelling enough to firmly conclude that the differences are due to contamination rather than to undetected environmental variation between sites. If demographic differences due to contamination are slight, they will be difficult to detect against a background of demographic variation resulting from subtle environmental differences. We conclude that animals at this site experienced a level of contamination that was below threshold for unequivocal detection of demographic effects.

Polychlorinated biphenyls (PCBs) are ubiquitous contaminants of global ecosystems (Risebrough et al. 1968; Fleming et al. 1983). The mechanistic effects of these chemicals on physiology and reproduction in mammals are well documented (Kimbrough 1974; Fuller and Hobson 1986). Although there are laboratory studies that can be used as a basis for predicting effects on natural populations, there are few in situ field studies. Laboratory feeding experiments involving wild mammals have resulted in observations that range from a complete cessation of breeding in mink at 2 mg/kg (Platonow and Karstad 1973; Aulerich and Ringer 1977) to no significant effect on breeding performance in cottontail rabbits at 10 mg/kg (Zepp and Kirkpatrick 1976). Studies on white-footed mice (Peromyscus leucopus) have documented impairment of reproductive success, including reduced litter size, poor offspring growth

Correspondence to: A. V. Linzey

and survival, and elimination of breeding by the second generation, as well as a greater impact on individuals exposed at earlier ages (Merson and Kirkpatrick 1976; Linzey 1987; Linzey 1988). The extent to which the results of these studies can be used to predict population-level effects in natural populations is unclear. The study of effects of contaminants in natural situations has lagged far behind the development of laboratory models (McBee and Bickham 1990). There are little data to document that sublethal effects observed in the laboratory alter demographies in the field, with the clearest demonstration of impacts involving highly contaminated areas. Long-term effects on population structure have been noted for three species of small mammals (Peromyscus polionotus, Sigmodon hispidus, and Mus musculus) inhabiting an enclosure treated with the carbamate insecticide sevin (Pomeroy and Barrett 1975). These authors documented reproductive inhibition in cotton rats (Sigmodon) and an increase in numbers of house mice (Mus), with the latter possibly due to reduced species interactions. Studies of meadow voles (Microtus pennsylvanicus) inhabiting the Love Canal waste disposal site revealed reduced population density, lowered survival rates, and a shift in population structure (Rowley et al. 1983). Batty et al. (1990) also documented lower summer populations, increased weights of some internal organs (liver, kidney, spleen, and adrenals), significantly smaller testes, and reduced numbers of juveniles and subadults in a whitefooted mouse (Peromyscus leucopus) population that adjoined a pond heavily contaminated with PCBs and heavy metals. These results are in alignment with predictions based on laboratory studies. Although population effects may be more readily apparent at highly contaminated sites, exposure to low levels of toxicants is probably more common. The development of models for environmental toxicology requires studies at sites with low levels of contamination in order to reveal the existence of possible threshold effects and to determine the level at which effects are likely to be obscured by demographic variability that is unrelated to contamination. The material presented here describes the demographic characteristics of a white-footed mouse population inhabiting a woodland that contains a low level of PCBs contamination.

522

Materials and Methods

Study Areas The study areas were located in Indiana County, Pennsylvania, and consisted of two 90 m by 90 m trapping grids located in woodland habitats. Each grid had 100 trap stations (10 lines of 10 traps each) spaced 10 m apart. The PCBs-contaminated grid adjoined a site where disposal of PCBs-contaminated oils has led to high soil residues and where aerosol releases of these contaminated oils have occurred. The PCBs-contaminated grid was positioned downwind and downhill from the disposal site, so that it would be likely to receive contamination from windbome oils and dust, as well as from runoff during rainfall and snow melt. The reference grid was 1 km west of the contaminated area (prevailing winds are from the west, so that airborne contaminants would not be expected to reach this site) and was also in a different watershed. The two sites were similar in elevation, slope, and aspect, as well as in general habitat characteristics. Ground-level temperatures recorded during the 1989 and 1990 trapping sessions indicate that the sites experienced similar microclimates. Soil samples from both areas were analyzed for PCBs by the Bureau of Laboratories, Pennsylvania Department of Environmental Resources, Harrisburg, PA. Soil samples from 25 regularly distributed locations on the PCBs-contaminated grid contained an average of 0.30 mg/kg (range 0 . ~ . 6 8 mg/kg). The analysis protocol involved testing duplicate samples at a rate of 10% and running standards with every set of 20 samples. Runs in which responses for standards were within 15% of the standard concentration were considered valid. Contamination was not restricted to any specific portion of the grid, but was higher in the area closer to the disposal site. Analysis of soil samples from the reference grid revealed a barely detectable level of PCBs at two stations, both of which adjoined a road.

A.V. Linzey and D. M. Grant

traps (one per station, 100 per grid) were baited with oat seeds, with the addition of apple slices for moisture in hot weather. Monthly or bimonthly trapping sessions lasted for 3 consecutive days and nights, with traps being checked and/or closed shortly after sunrise and checked and/or reopened before sunset. Each animal was marked by toe-clipping for future identification and the following data recorded: weight, sex, age (by pelage color), and reproductive condition (males with enlarged testes; females with perforate vaginae, pregnant, or with enlarged nipples). Mice were considered to be in breeding condition when males had enlarged testes and scrotal sacs and females were pregnant, had perforate vaginae, or had enlarged nipples. Trapping dates were as follows: July-October 1988, May--October 1989, and May-October 1990. The total number of trapnights per grid (one trap set for one night = one trapnight) was 2,100, 2,900, and 1,800 for the 1988, 1989, and 1990 trapping periods, respectively. Demographic data were transformed if necessary, tested for normality and homogeneity of variances, and subjected to the appropriate parametric or non-parametric tests. The acceptable level of statistical significance was p ~< 0.05. Statistical tests were based on data calculated either on a monthly basis (n = 16) or by trapping period (n = 22); the number of animals included in each sample interval varied. For comparability, tests among years used July-October data for 1989 and 1990. Home range indices were calculated by drawing ranges on scaled grid maps, joining the two most distant points in the range (greatest length of range), and drawing the longest possible perpendicular axis to the first line (greatest width of range). The lengths of the two lines (in mm) were added to obtain the home range index.

Results

Habitat Analysis Habitat Analysis Because differences in habitat may lead to differences in small mammal demography that could confound the results of the study, the degree of habitat similarity was determined by quantitative analysis of vegetation structure. The variables quantified provided an estimate of ground cover, shrubhiness, and degree of tree canopy closure. Twenty-seven sampling squares (33% intensity) for habitat analysis were randomly chosen from among the 81 squares on each grid (each 10 m 2 area delineated by four trap stations). Ground cover estimates were taken at the center of the selected sampling squares, so that trampling of vegetation (due to trap checking) did not affect measuremeats. Ground cover (ferns, ground cedar, grasses, and forbs) was estimated at each sample square using the profile board technique as described by Linzey (1984), MacArthur and MacArthur (1961), and Rosenzweig and Winakur (1969). To determine percent canopy cover, the method described by MacArthur and MacArthur (1961) and modified by Van Home (1982) was used. In order to quantify the extent of woody vegetation more precisely, a direct count of trees and shrubs at various heights was made at each sample square. Vegetation variables measured at each station [% cover variables were arcsine transformed as described by Sokal and Rohlf (1981)] were tested for significant differences using the Mann-Whitney U test (MWU) and were considered significant if p ~< 0.05. Means given in the text are followed by standard error estimates.

The vegetation at both sites consisted of a tree overstory, a shrub layer, and leaf litter ground cover. Herbaceous vegetation was locally distributed. The PCBs-contaminated area had a remnant stone wall (about 20 m long) and a blackberry (Rubus sp.) patch, neither of which occurred at the reference site. There were no statistically significant differences between the two study sites with respect to either ground or tree canopy cover estimates. For example, tree canopy cover at the PCBscontaminated and reference grids averaged 85.2% + 1.55 and 88.8% --- 1.16, respectively. Woody vegetation at the two areas differed significantly in regard to two features. The PCBscontaminated area contained a greater number of shrubs taller than 1.0 m (~ = 11.0 + 5.01) than the reference area (£ = 1.5 -+ 0.54; p = 0.03, M W U ) . However, the reference area contained more trees 1-4 m in height (R = 2.81 --- 0.49) than the PCBs-contaminated area (~ = 1.55 + 0.50; p = 0.04, M W U ) . The difference in shrubs taller than 1 m was partly due to the inclusion of two sampling squares located within a patch of blackberry canes on the PCBs-contaminated area.

White-footed Mouse Demography

Density Mammal Trapping Study grids were trapped simultaneously throughout the study to avoid effects of differing weather conditions oa trappability. Sherman live-

When data from all years were pooled, the population density on the PCBs-contaminated grid was significantly greater than that on the reference grid (p = 0.0001; Table 1). Density estimates were calculated by the minimum number alive method

Population Characteristicsof White-FootedMice on a PCBs ContaminatedSite

Table 1. Density of Peromyscus leucopus on reference and PCBscontaminated trapping grids (± standard error). Based on minimum number alive (MNA) calculated on a monthly basis Grid

Reference

Contaminated

P (t-tests)

Years Pooled 1988 1989 1990 Among years (ANOVA) Anova results:

15.7 14.0 20.3 ± 12.2 ±

24.2 20.0 33.8 17.3

0.0001 0.1019 0.0019 0.0294

1.44 1.96 2.33 1.56

+- 2.22 -+ 2.42 --+ 2.24 ± 1.33

p = 0.2328 p = 0.0026 Grids different (p = 0.0037) Years different (p = 0.0014) No interaction (p = 0.5488)

Table 3. Home range indices and overlap for Peromyscus leucopus captured at four or more stations on the reference and PCBs-contamihated trapping grids. Values are means of individual index values --- standard errors. Sample sizes are in parentheses

Home range size Both sexes Males Adults Subadults Females Adults Subadults Home range overlapa a

Table 2. Patterns of residency of Peromyscus leucopus on reference and PCBs-contaminated trapping grids. See text for definitions of immigrants, emigrants, and transients. Values are means of monthly numbers - standard errors

Immigrants Emigrants Transients

Reference

Contaminated

p

3.7 --- 0.65 3.6 - 0.82 3.8 ± 0.71

5.1 - 1.09 5.2 ± 1.01 5.7 ± 0.75

0.1391 0.1096 0.0406

(Jolly and Dickson 1983; Krebs and Boonstra t984). Trappability estimates (Hilborn et al. 1976) were high at both sites (reference grid--86.2%, PCBs-contaminated grid--87.2%), confirming the reliability of density estimates. Density on the PCBs-contaminated grid was consistently higher in each year, with the difference being significant in 1989 and 1990. Comparison among years on each grid indicated that the reference grid population (range 12.2-20.3 animals) was more stable than that of the PCBs-contaminated grid (range 17.3-33.8 animals).

Turnover Rate The number of immigrants (number of new animals each month caught in at least one subsequent month) and emigrants (residents that disappeared between two consecutive months) per month was higher on the PCBs-contaminated grid, but the differences were not significant (Table 2). However, there were significantly more transients (animals trapped only in one month) on the contaminated grid per month (~ = 3.8 vs. = 5.7; p = 0.04, t-test).

Home Range Indices On the reference grid, there was a tendency for males to have larger home ranges than females and for adults of both sexes to have larger home ranges than subadults (Table 3). Similar trends were noted on the PCBs-contaminated grid, except that subadult female ranges were larger than adult female ranges. However, none of these difference were significant. Home range overlap indices were calculated by summing, for each animal, the greatest length plus greatest width of the

523

Reference

Contaminated

144.5 --- 10.67 (33) 159.6 +-- 17.18 (15) 175.0 ± 21.88 (10) 128.8 + 24.33 (5) 133.2 + 13.34 (18) 146.8 + 19.11 (13) 120.2 + 7.11 (5) 165.5 + 28.56 (23)

140.7+ 7.12 (50) 148.7+ 8.17 (33) 157.8- 11.08 (17) 144.7___11.50 (16) 126.7--- 13.10 (17) 122.8--- 15.18 (12) 129.2+ 30.95 (5) 236.8 -+ 24.69 (47)

p = 0.0418; no other significant differences

overlap area with every other animal present at the same time (Table 3). Overlap on the contaminated grid was significantly greater than on the reference grid (p = 0.035, MWU).

Age-Sex Structure of Population When data from all years were pooled, the only significant difference between populations was a greater proportion of adult females on the reference grid (Figure 1). In general, the reference grid had proportionally more adult animals, but the PCBs-contaminated grid had proportionally more young animals (subadults and juveniles). When the data were examined on a yearly basis, only 1988 demonstrated significant differences between grids. In that year, adult females were significantly more abundant on the reference grid, and subadult males were significantly more abundant on the PCBs-contaminated grid. The other noteworthy pattern was that the age-sex structure on the reference grid was relatively stable among years (the only significant difference was that juveniles constituted a greater proportion of the population in 1989 than in 1990), but the relative proportions of adults and subadults in the 1988 PCBs-contaminated grid population differed from the other two years of the study. In 1988, young animals dominated the PCBs-contaminated grid (81% of population), but proportions of young and adults were approximately equal in the other two years.

Reproduction A slightly larger proportion of the population (both males and females) was in reproductive condition on the reference grid than on the PCBs-contaminated grid, but the differences were not statistically significant (Table 4). The proportion of females in breeding condition declined steadily between 1988 and 1990 on the reference grid, but showed no pattern on the contaminated grid. In contrast, proportions of males breeding increased steadily across years at both sites. None of these yearly differences are statistically significant. Analysis of the age structure of the male breeding population indicated that the proportion of adults and subadults that were in reproductive condition did not differ between the two sites or among years at each site (Table 5). However, for females, the

524

Z

A.V. Linzey and D. M. Grant

1.0

[] [] []

C ,-I

0.8

juvenile subadult adult

Males all years 1988 1989 1990 Females all years 1988 1989 1990

0.6-. C

0,

Z

£

[,. ==

0.2'

0 =4

0.0 Ref COn All years

pooled

Ref Con

1988

Ref Con

1989

Z O

F.

1"01 B 0.8

1990

O

0.6'

Z

£

[] []

juvenile subadult adult

04i

==

0.0

Ref Con All years

pooled

Contaminated

0.61 0.55 0.62 0.65 0.67 0.82 0.64 0.61

0.57 0.49 0.56 0.66 0.61 0.66 0.52 0.73

± 0.076 ± 0.189 ± 0.095 ± 0.156 -+ 0.059 ± 0.097 ± 0.093 ± 0.117

- 0.071 ± 0.141 ± 0.105 ± 0.146 ± 0.065 ± 0.153 --4-0.097 ± 0.097

Table 5. Age structure of breeding male and female Peromyscus leucopus on reference and PCBs-contaminated trapping grids. Values are mean proportions of individuals in reproductive condition (± standard errors) that were adults during each trapping period. All other reproductive individuals were subadults

Males all years 1988 1989 1990 Females all years 1988 1989 1990

Reference

Contaminated

0.62 0.44 0.64 0.78 0.99 1.00 0.97 1.00

0.70 0.34 0.86 0.77 0.81 0.20 1.00 1.00

± 0.069 ± 0.113 ± 0.097 ± 0.142 ± 0.015 ± 0.000 a ± 0.030 --- 0.000

-+ 0.010 - 0.145 --- 1.091 ± 0.143 ± 0.088 ± 0.200 ± 0.000 b +- 0.000 c

aReference and contaminated 1988 differ bContaminated 1988 and 1989 differ CContaminated 1988 and 1990 differ

0.2' O

Reference

Ref Con

TIME (YEARS)

O

Table 4. Mean proportion of population of males and females in breeding condition ± standard error during each trapping period

Ref Con

1988

Ref Con

1989

Ref Con

1990

TIME (YEARS)

Fig. 1. Age and sex structure of Peromyscus leucopus on reference (Ref) and PCBs-contaminated (Con) trapping grids. Values are the proportions contributed by each age category to the entire population, but graphed separately for males (A) and females (B). Statistically significant differences are indicated as follows: (a) between the reference and contaminated grids for that parameter in that year; (b) between 1988 and 1989 on the same grid for that parameter; (c) between 1989 and 1990 on the same grid for that parameter; and (d) between 1988 and 1990 on the same grid for that parameter

breeding population on the reference grid consisted exclusively of adults (97-100%), but varied among years on the PCBscontaminated grid (20% in 1988; 100% in 1989 and 1990). Although these differences were statistically significant, some of the monthly values were based on small sample sizes.

Discussion

The PCBs-contaminated grid and reference grid are both located in habitats considered to be primary for this species. Although many of the demographic differences are not statistically significant, they suggest an overall pattern that is reminiscent of populations inhabiting in suboptimal habitats. In gen-

eral, the contaminated grid is higher in density and is characterized by a young population with a high turnover rate. Temporal variation in demographic variables is also characteristic of the contaminated site. The results of this study do not align with predictions from laboratory studies on white-footed mice. For example, predictions made from results of controlled feeding experiments might include reduced population size, fewer juveniles, and elimination of breeding by subadults. However, laboratory experiments may exaggerate potential effects because they employ closed populations that are exposed to uniform (usually high) levels of contaminants. On the other hand, laboratory populations are not subjected to environmental stresses of food shortage, temperature and moisture variation, competition with other individuals, or predation. Because of these multiple stressors, natural populations may exhibit demographic effects at lower contaminant levels than observed in the laboratory. However, population effects may be indirect and/or cryptic because natural populations are not closed, but are constantly altered by colonization from adjoining areas that may be less contaminated. Density of white-footed mice at the PCBs-contaminated site was significantly higher than at the reference site. There are several possible explanations for this observation. First, the habitat may be more productive and hence may support larger numbers of animals. This explanation would lead to a prediction that home range size would be smaller, since animals could presumably meet their requirements in a smaller area. However, although ranges of adults are smaller on the contaminated

Population Characteristics of White-Footed Mice on a PCBs Contaminated Site

grid, ranges of subadults are larger (Table 3). In any event, the differences are not significant. A second possibility is that the higher density figures reflect a higher turnover rate of residents (perhaps due to higher mortality or lower reproduction in situ) and/or there are more transient individuals moving through the area. Both these trends existed at the PCBs-contaminated site (significant only for transients). Finally, mice on the contaminated site may be behaviorally less aggressive, thus tolerating closer spacing of individuals. Aggression levels in rodents appear to be related to levels of circulating testosterone and estradiol (Gaines et al. 1985; Albert et al. 1991). Chlorinated hydrocarbon-induction of hepatic enzyme systems with the ability to metabolize these hormones has been widely reported (Peakall 1970; Welch et al. 1971; Conney and Burns 1972; Sanders and Kirkpatrick 1975, 1977). The observation of greater home range overlap on the contaminated grid would support this explanation. Therefore, the data support a conclusion that higher density on the contaminated site is correlated with tolerance for closer spacing and greater rate of turnover. When the data are sufficient to allow an analysis of temporal changes in population density and structure, two patterns emerge. First, the population in 1988 differed from that in other years. For example, densities at the two sites did not differ in that year (Table 1), but the age-sex structure of the population (Figure 1) and extent of breeding by adult versus subadult females (Table 5) did differ. Secondly, demographic parameters were stable at the reference grid over the three years, but varied at the contaminated site due to differences in 1988 (Table 4). These results may relate to a severe drought (accompanied by unusually high temperatures) experienced by the region during summer 1988. These weather extremes, in combination with stresses relating to PCBs exposure, may have resulted in greater impacts on the contaminated grid population. The use of density data as an indicator of habitat quality has been a common practice, but unusually high densities may be characteristic of suboptimal environments (Van Home 1983). For example, suboptimal habitats may serve as dispersal sinks for surplus animals produced in higher quality habitats, with dispersal being driven by behavioral interactions with socially dominant residents (Linzey 1989). Densities in suboptimal habitats may fluctuate widely, reaching high levels during years of peak productivity in optimal habitat (Lidicker 1975). There are two possible explanations for the results obtained in this study. One interpretation is that the PCBs-contaminated grid serves as a functionally suboptimal habitat; i.e., a lower production and persistence of individuals on the site may result in a habitat "sink" that attracts dispersers (mostly young animals) from a surrounding, less contaminated habitat. The trends observed (higher turnover, larger numbers of young animals, and temporal variability) are consistent with those observed for this species in structurally suboptimal habitats (Adler and Wilson 1987). Furthermore, they align with those exhibited during a 1985-86 population trough in a nearby suboptimal habitat (Linzey and Kesner 1991). However, an equally plausible explanation is that the demographic differences between sites are caused by subtle environmental differences (such as undetected differences in temperature or moisture) that are unrelated to contamination. Therefore, our conclusion must be that the level of PCBs at this site is not high enough to cause demographic effects that can, with certainty, be attributed to contamination. Chronic, low-level habitat contamination is undoubtedly a

525

more common phenomenon than is the "superfund site" syndrome. Ecological toxicologists have proposed a variety of possible "endpoints" to be used in detecting effects of contamination, including the "demographic endpoint" employed in this study (McBee and Bickham 1990). Although this endpoint may be most appropriate for highly contaminated sites, additional studies are needed at various locations along contamination gradients in order to determine the point at which demographic effects can be detected with confidence.

Acknowledgments. This research was supported by grants from the Pennsylvania Academy of Science, Pennsylvania State System of Higher Education, and Indiana University of Pennsylvania Graduate School. We are grateful to James R. Shaw, Bureau of Solid Waste Management, Pennsylvania Department of Environmental Resources, for arranging for analyses of soil samples from our sites. We also thank North Cambria Fuel Company and Garfield Fuel Company for permission to work on their properties.

References Adler GH, Wilson ML (1987) Demography of a habitat generalist, the white-footed mouse, in a heterogeneous environment. Ecology 68:1785-1796 Albert DJ, Jonik RH, Walsh ML (1991) Hormone-dependent aggression in the female rat: Testosterone plus estradiol implants prevent the decline in aggression following ovariectomy. Physiol Behav 49:673-677 Aulerich RJ, Ringer RJ (1977) Current status of PCB toxicity to mink, and effect on their reproduction. Arch Environ Contam Toxicol 6:279-292 Batty J, Leavitt RA, Biondo N, Polin D (1990) An ecotoxicological study of the white footed mouse (Peromyscus leucopus) inhabiting a polychlorinated biphenyls-contaminated area. Arch Environ Contam Toxicol 19:283-290 Conney AH, Burns JJ (1972) Metabolic interactions among environmental chemicals and drugs. Science 178:576--586 Fleming WJ, Clark DR, Henry DJ (1983) Organochlorine pesticides and PCBs: a continuing problem for the 1980's. Trans 48th North Am Wildl Conf: 186-198 Fuller GB, Hobson WC (1986) Effect of PCBs on reproduction in mammals. In: Waid JS (ed) PCBs and the environment. CRC Press, Boca Raton, FL, pp 101-125 Gaines MS, Fugate CL, Johnson ML, Johnson DC, Hisey JR, Quadagno DM (1985) Manipulation of aggressive behavior in male prairie voles (Microtus ochrogaster) implanted with testosterone in silastic tubing. Can J Zoo163:2525-2528 Hilborn R, Redfield JA, Krebs CJ (1976) On the reliability of enumeration for mark and recapture census of voles. Can J Zoo154:10191024 Jolly GM, Dickson JM (1983) The problem of unequal catchability in mark-recapture estimation of small mammal populations. Can J Zool 61:922-927 Kimbrough RD (1974) The toxicity of polychlorinatedpolycycliccompounds and related chemicals. CRC Crit Rev Toxicol:445-498 Krebs CJ, Boonstra R (1984) Trappability estimates for mark-recapture data. Can J Zool 62:2440-2444 Lidicker WZ, Jr (1975) The role of dispersal in the demography of small mammals. In: Golley FB, Petusiewicz K, Ryszkowski L (eds) Small mammals, their productivity and population dynamics. Cambridge Univ Press, NY, pp 103-128 Linzey AV (1984) Patterns of coexistence in Synaptomys cooperi and Microtuspennsylvanicus. Ecology 65:382-393 - - - ( 1 9 8 7 ) Effects of chronic polychlorinated biphenyls exposure

526

A.V. Linzey and D. M. Grant

on reproductive success of white-footed mice (Peromyscus leucopus). Arch Environ Contam Toxicol 16:455--460 --(1988) Effects of chronic polychlorinated biphenyls exposure on growth and reproduction of second generation white-footed mice (Peromyscus leucopus). Arch Environ Contam Toxicol 17:3945 --(1989) Response of the white-footed mouse (Peromyscus leucopus) to the transition between disturbed and undisturbed habitats. Can J Zoo167:505-512 Linzey AV, Kesner MH (1991) Population regulation in white-footed mice (Peromyscus leucopus) in a suboptimal habitat. Can J Zool 69:76-81 MacArthur RH, MacArthur JW (1961) On bird species diversity. Ecology 42:594-598 McBee K, Bickham JW (1990) Mammals as bioindicators of environmental toxicity. In: Genoways HH (ed) Current mammalogy, vol 2. Plenum, NY, pp 37-88 Merson MH, Kirkpatrick RL (1976) Reproductive performance of captive white-footed mice fed a PCB. Bull Environ Contam Toxicol 6:392-398 Peakall DB (1970) p, p' DDT: Effect on calcium metabolism and concentration of estradiol in the blood. Science 168:592-594 Platonow NS, Karstad LH (1973) Dietary effects of polychlorinated biphenyls on mink. Can J Comp Med 37:391-400 Pomeroy SE, Barrett GW (1975) Dynamics of enclosed small mammal populations in relation to an experimental pesticide application. Am Midl Nat 93:91-106 Risebrough RW, Rieche P, Peakall DB, Herman SG, Kirven MN

(1968) Polychlorinated biphenyls in the global ecosystem. Nature 220:1098-1102 Rosenzweig ML, Winakur J (1969) Population ecology of desert rodent communities: habitats and environmental complexity. Ecology 50:558-572 Rowley MH, Christian JJ, Basu DK, Pawlikowski MA, Paigen B (1983) Use of small mammals (voles) to assess a hazardous waste site at Love Canal, Niagara Falls, New York. Arch Environ Contam Toxicol 12:383-397 Sanders OT, Kirkpatrick RL (1975) Effects of a polychlorinated biphenyl (PCB) on sleeping times, plasma corticosteroids, and testicular activity of white-footed mice. Environ Phys Biochem 5:308-313 ,i (1977) Reproductive characteristics and corticoid levels of female white-footed mice fed ad libitum and restricted diets containing a polychlorinated biphenyl. Environ Res 13:358-363 Sokal RR, Rohlf FJ (1981) Biometry. The principles and practice of statistics in biological research. WH Freeman, San Francisco, CA Van Home B (1982) Demography of the longtail vole Microtus long# caudus in seral stages of coastal coniferous forest, southeast Alaska. Can J Zool 60:1690-1709 i(1983) Density as a misleading indicator of habitat quality. J Wildl Manage 47:893-901 Welch RM, Levin W, Kuntzman R, Jacobson M, Conney AH (1971) Effect of halogenated hydrocarbon insecticides on the metabolism and uterotropic action of estrogens in rats and mice. Toxicol Appl Pharmacol 19:234-246 Zepp RL Jr, Kirkpatrick RL (1976) Reproduction in cottontails fed diets containing a PCB. J Wildl Manage 40:491-495