Effect of Rock Cover on Small Mammal Abundance in a Montana Grassland

Scientific Disciplines
Biological Sciences - Terrestrial
Keywords
Montana
University of Montana
deer mice
peromyscus maniculatus
hantavirus
Sin Nombre virus
SNV
abundance
Years
Authors
Volumes
Volume 17, No. 1-4

Effect of rock cover on small mammal 
Abundance in a montana grassland
 
Kyle Richardson, Department of Biology, Montana Tech of the University of Montana,    
Butte, Montana 59701
Scott Carver1, Department of Biology, Montana Tech of the University of Montana,  
 
Butte, Montana 59701
Richard Douglass, Department of Biology, Montana Tech of the University of Montana,    
Butte, Montana 59701
Amy Kuenzi, Department of Biology, Montana Tech of the University of Montana,  
 
Butte, Montana 59701
Abstract
We examined the influence of rock cover, as an indicator of presumable retreat site availability 
on the abundance of deer mice (Peromyscus maniculatus) and prevalence of Sin Nombre virus 
(SNV) using long-term live trapping and habitat data from three live trapping grids and a short-
term (three month), spatially replicated study across three slopes in Cascade County, Montana. 
In our long-term study, we found that deer mice were more abundant at a live-trapping grid with 
greater rock cover, than two grids with less rock cover. There was a non-significant trend (P = 
0.053) for deer mice to be more abundant in rocky sites in the short term study. In the long-term 
study, average SNV antibody prevalence among deer mice was slightly greater (5.0 vs. 3.5 % 
on average) at the live trapping grid with more rock cover, than the grid with less rock cover. 
We were unable to demonstrate differences in SNV antibody prevalence among treatments in 
the short-term study. Further studies are needed to elucidate the multiple determinants of deer 
mouse abundance and SNV prevalence in grassland ecosystem and other habitat types.
Key Words: deer mouse, Peromyscus maniculatus, Sin Nombre virus, retreat sites
Introduction
The abundance of small mammals 
have on pathogen prevalence, are needed to 
varies temporally and spatially among and 
help reduce human exposure.
within habitat types (e.g., Krebs, 1996). 
Deer mice (Peromyscus maniculatus
In many cases, however, the underlying 
are widespread omnivorous rodents, which 
determinants of variability in small 
occur in a variety of habitat types across 
mammal abundance are unknown. Variation 
North America (Kirkland and Layne 1989, 
in abundance can have consequences 
Douglass et al. 2001). These rodents have 
for pathogen transmission because host 
been ideal organisms for studies of habitat 
abundance may influence contact and 
relationships for many decades (e.g., 
transmission rates among individuals 
Smith 1940, Douglass 1989, Matlack et al. 
in a population (Keeling and Rohani 
2001, Johnston and Anthony 2008) and in 
2008). For example, human incidence of 
Montana their abundance varies among and 
nephropathia epidemica (caused by Puumala 
within many habitat types (e.g., Douglass 
hantavirus) was related to abundance of the 
et al. 2001). Deer mouse abundance has 
host Clethrionomys glareolus in Sweden 
been found to be positively related to shrub 
(Niklasson et al., 1995). Studies that 
cover in a short grass prairie environment 
examine determinants of host abundance 
(Stapp and Van Horne 1997) throughout 
in nature, and the effects these factors may 
central and western Montana (Douglass 
1989a, Douglass et al. 2001) and Western 
1  Current  address:  Department  of  Microbiology, 
Colorado (Douglass 1989b). Furthermore, 
  Immunology and Pathology, Colorado State  
University, 1619 Campus Delivery, CO 80523
deer mice are reservoirs for Sin Nombre 


virus (SNV, Bunyaviridae:Hantavirus
open environments with high shrub cover 
(Nichol et al., 1993), a directly transmitted 
(Douglass et al. 2001, Douglass 1989a, 
zoonotic pathogen that is transmissible to 
Douglass 1989b, Douglass and Frisina 
humans causing Hantavirus Pulmonary 
1993), in this environment rock cover may 
Syndrome (HPS), which has a high 
be important for avoiding large predators, 
mortality rate (CDC, 2009). The abundance 
caching resources, or nesting. Other habitat 
of deer mice is an important component 
characteristics differed among our long-term 
of SNV transmission (e.g., Madhav et al. 
live trapping grids (see results) and likely 
2007) and human exposure risk (Childs 
also contributed to variation in deer mouse 
et al. 1995). Both abundance of deer mice 
abundance within this grassland habitat type 
and SNV prevalence among mice can 
(Wecker 1963, Douglass 1989a, Douglass 
vary significantly over small spatial scales 
1989b, Morris, 1997). The multivariate 
(Glass et al. 2000, Douglass et al. 2001). 
effect of habitat characteristics on deer 
Therefore, studies that investigate variability 
mouse abundance will be the subject of 
in deer mouse abundance and how this is 
further investigations.
affected by habitat characteristics within 
We hypothesized that in a specific type 
an expansive habitat type are of value to 
of Northern Great Plain grassland, deer 
understanding SNV transmission and human 
mice are more abundant, with a greater 
exposure risk.
number of individuals and prevalence of 
In Montana deer mouse abundance 
SNV antibodies in the population, in sites 
and prevalence of antibodies to SNV are 
with similar vegetative characteristics but 
on average highest in sagebrush (Artemesia 
greater rock cover (potential retreat site 
tridentata) dominated environments 
availability) than sites with less rock cover. 
(Douglass et al. 2001), where sagebrush 
We investigated the relationship between 
itself may provide retreat sites. However, 
population abundance and prevalence of 
deer mice can also be abundant, with 
SNV in relation to rock cover in two ways. 
high SNV antibody prevalence in other 
We used a long-term (1994-2010) study on 
habitat types, such as grassland and forest 
population dynamics of deer mice on three 
environments (Kuenzi et al. 2001, Douglass 
live trapping grids and prevalence of SNV 
et al., 2001). Here, we investigated how 
on two live trapping grids (Douglass et al. 
abundance of deer mice and prevalence of 
1996, Douglass et al. 2001) from which our 
SNV antibodies varied within a grassland 
observations led to the hypothesis above. 
ecosystem located in Cascade County, 
Because the long-term study lacked spatial 
Montana. We focused on rock cover, which 
replication in rock cover, we also undertook 
may be an indicator of potential retreat site 
a short-term (three month) spatially-
availability, as a determinant of variation 
replicated study to evaluate the effect of 
in deer mouse abundance in this habitat 
rock cover on small mammal abundance. 
type. Within this grassland several habitat 
characteristics were different among 
Materials and methods
sampling locations and some, e.g., such as 
shrubs or patches of tall grass, may provide 
Long-term Data Collection
retreat sites for deer mice. However, we 
Data on deer mice and habitat 
chose to focus on rock cover in this study 
characteristics were collected on three one 
because the abundance of deer mice at 
hectare grids (grid numbers 10, 11 and 12) 
three long-term grassland live trapping 
located near Cascade (46° 59.3  N, 111° 
grids differed (Douglass et al. 2001) as 
35.3 W, 1408 m AMS), Montana. Grids 
did rock cover. In addition, we frequently 
were situated in grassland habitat supporting 
observed deer mice using rocks as retreat 
an active cattle ranch (Douglass et al., 
sites upon release from live traps. Although 
1996). We live-trapped deer mice for three 
rock cover is clearly not critical to deer 
consecutive nights/month on all three 1-ha 
mouse survival in all habitats especially in 
grids for 174 consecutive months between 
Effect of Rock Cover on Small Mammal Abundance in a Montana Grassland           21


June 1994 and November 2008. Trapping 
present at all sites, (2) small burrows were 
grids consisted of 100 equally-spaced 
common around rocks at our sites, (3) mice 
Sherman live traps (H. B. Sherman Traps. 
were frequently observed retreating to 
Tallahassee, Florida), baited with rolled 
rocks for cover when released from traps, 
oats and peanut butter and provisioned with 
and (4) mouse tracks and burrows in snow 
polyester Fiberfil bedding. Upon capture, 
commonly originated from rocks.
each rodent was given a uniquely numbered 
Short-term Data Collection
ear-tag (model #1005-1, National Band and 
Tag Co., Newport, KY) and species, gender, 
To determine if abundance of deer 
body mass, reproductive condition, and 
mice differed among sites with more or less 
presence of scars or wounds were recorded. 
rock cover (retreat sites), we established 
We routinely collected blood samples, which 
six 0.25-ha grids on three slopes (named 
were later tested for antibodies to SNV, 
Hill One, Hill Two and Hill Three for this 
from grids 11 and 12 only. We followed 
study) of approximately the same elevation 
animal handling, blood collection and 
and aspect. Within each slope, one grid was 
safety precautions described by Mills et 
located in a rocky area and one in a non-
al. (1995) and approved by the University 
rocky area. Grids were live-trapped for small 
of Montana IACUC. Serological testing 
mammals monthly from August-October 
was performed at the Montana State Public 
2008 in the same manner as the long-term 
Health Laboratory and at Special Pathogens 
live trapping grids, but with 25 equally 
Branch, U.S. Centers for Disease Control 
spaced rather than 100 Sherman live-capture 
and Prevention, using methods described by 
traps. Small mammal handling, testing for 
Feldmann et al. (1993).
SNV antibodies, calculation of MNA, MNI 
We used the enumeration technique 
rock cover, and SNV prevalence among 
of Chitty and Phipps (1966) to provide a 
deer mice was carried out in the same 
minimum number of individuals known to 
manner as the long-term trapping grids. 
be alive (MNA) during a 3-day trapping 
In September, we also measured habitat 
session as an index of population abundance 
characteristics at each of the six grids in the 
for each month. The minimum number of 
same way as the long-term grids, but with 
deer mice antibody-positive to SNV (MNI) 
habitat characteristics determined from six 
during each trap session was calculated in 
randomly-assigned locations at each grid (a 
the same way for grids 11 and 12. Estimated 
similar proportion/unit area as the long-term 
antibody prevalence (EAP) of deer mice for 
grids).
a given month was calculated by dividing 
Analyses
MNI by MNA (Mills et al. 1999a). 
We employed a Friedman analysis 
Data on habitat characteristics of the 
(Zar, 1996) to determine if abundance of 
three long-term grids were collected twice 
deer mice, number of deer mice antibody 
annually (Jun and Sep) at 25 randomly-
positive for SNV, and antibody prevalence 
assigned locations/grid using a 10-pin point 
among deer mice differed among long-
frame. Percent cover of lichens mosses, 
term trapping grids. A Friedman analysis 
grass, forbs, shrubs, rock, leaf litter and bare 
was also used to determine if habitat 
ground was determined at each location. 
characteristics differed among the long-term 
A contact with rock and a point frame rod 
grids over time. Friedman analysis enabled 
was counted as rock cover.  Although, 
examination of direction of differences 
direct use of rocks as retreat sites by deer 
among variables for individual sampling 
mice was not quantified in this study, our 
occasions collectively and, accordingly, 
assumption of rocks offering retreat sites 
was independent of the effects of seasonal 
seemed reasonable. Although our analysis 
and climatic forcing on dynamics, which 
did not directly account for use of rocks 
are certainly important but beyond the 
as retreat sites we found (1) large rocks 
scope of this investigation. For the short-
that mice were able to move beneath were 
term study, habitat characteristics among 
22          Richardson et al.


grids were compared using a Generalized 
mice (Luis et al. 2010). Among the three 
Linear Model (GZLM), with hillside as a 
long-term live trapping grids we detected 
random effect. Optimal distribution and 
significant differences in each of the habitat 
transformation of data (Poisson or negative 
variables except bare ground (Table 2). Of 
binomial probability distribution and 
particular focus in this study, rock cover was 
identity or log-link function respectively) for 
greatest at grid 11 (0.556 %) and least at grid 
the GZLM were specified using an Omnibus 
12 (0.147 %; Table 2).    
test. A linear mixed model (LMM) was 
We initially confirmed that the 
used to determine if deer mouse abundance 
proportion of cover that consisted of rock 
and antibodies to SNV among deer mice 
was greater among sites we had selected 
(number antibody positive and antibody 
as rocky sites, than non-rocky sites, in 
prevalence) differed between rocky and 
the short-term study (Table 3). Cover of 
non-rocky sites, with month as a repeated 
most other habitat variables did not differ, 
measure and hillside as a random effect. All 
except cover of lichens, which were found 
analyses were performed using SPSS 15.0 
exclusively on rocks (Table 3). Over the 
(SPSS Inc. Chicago, U.S.).
3-mo short-term study, 172 individual deer 
mice were captured 381 times in 1350 trap-
Results
nights.  Deer mouse abundance was greatest 
Over the 15-yr long-term study, 
at Hills One and Two and least abundant at 
we captured 4458 individual deer mice 
Hill Three (Table 4). Average deer mouse 
12,265 times in 155,700 trap-nights of 
abundance was higher at rocky sites than 
effort. Abundance of deer mice fluctuated 
non-rocky sites but this difference was not 
among months and years. Average monthly 
statistically significant (Tables 4 and 5). The 
abundance of deer mice was greatest at grid 
number of deer mice antibody positive for 
11 (MNA = 25.8), least at grid 12 (MNA 
SNV and deer mouse antibody prevalence 
= 15.3), and intermediate at grid 10 (MNA 
was unrelated to rocky or non-rocky habitat 
= 22.9) (Table 1). The average monthly 
(Tables 4 and 5).
number of antibody positive deer mice 
(MNI) and estimated antibody prevalence 
Discussion
(EAP) was greater at grid 11 (MNI = 1.3, 
Determining factors that influence 
EAP = 0.05) than grid 12 (MNI = 0.3, 
zoonotic host abundance is important to 
EAP 0.035) (Table 1). However, on several 
understanding pathogen transmission, and 
occasions during 15 yrs of sampling, both 
subsequently human exposure risk. Here, 
grids 11 and 12 experienced many months 
we investigated if abundance of deer mice 
with no presence of antibody-positive deer 
(a host reservoir for SNV) in a grassland 
Table 1. Mean (± SE) abundance of deer mice, , number of antibody positive deer mice to 
SNV (MNI) and antibody prevalence of deer mice (EAP) for the three long-term (Jun 1994- 
Nov 2008) live trapping grids. SNV data was collected for grids 11 and 12 only. Friedman 
analysis comparing rodent abundance and SNV infection data among grids. Significant results 
in bold.
 
 
 
Live Trapping Grid 
Friedman Analysis
  
10  11 12  x2 df  p
Deer mice mean  22.9  25.8  15.3  155.780 

< 0.001
  
SE 
1.651 
1.388  1.211 
 
 
Mni 
mean   
1.3 0.3  40.238  1  < 0.001
  
SE 
 
0.157  0.044 
 
 
Eap 
mean   
0.05 0.035  21.248  1 
< 0.001
    
SE 
  
0.005 
0.007
Effect of Rock Cover on Small Mammal Abundance in a Montana Grassland           23


Table 2. Mean (± SE) percent cover of habitat types for the three long-term (Jun 1994-Nov 
2008) live trapping grids and Friedman analysis of how vegetation cover differed among 
grids. Significant results in bold.
  
  
 
 
 
Leaf  
Grass
 
Bareground rock  moss  lichens litter  grass  height  forbs  shrubs
  10  Mean 0.2  0.4  0.1  0.1  8.6  7.7  18.4  2.1 
0.5
  
SE 
0.057 
0.106 
0.071 
0.018  0.455  0.415 
7.0 
0.372 
0.100
 11  Mean 0.2  0.6  0.9  0.2  7.4  6.6  14.3  1.5  0.2
  
SE 
0.059 
0.090  0.2 
0.068  0.523  0.458  5.868  0.243 
0.091
 12  Mean 0.2  0.1  0.000 0.02  8.9  8.5  22.7  1.1 
1.3
  
SE 
0.07 
0.042 
0.000 
0.017  0.427  0.350  8.213  0.238 
0.163
  
X2 
0.03  17.643  32.411  17.930 22.769  35.766  10.308 13.470  31.735
 
 df 2  2  2  2 2 2  2 2  2
   
P 0.983 
< 0.001  < 0.001  < 0.001  < 0.001  < 0.001 
0.006 
0.001 
< 0.001
ecosystem was related to availability of rock 
in this grassland ecosystem. In previous 
cover, i.e., as an indicator of potential refuge 
work (Douglass 1989b) found a negative 
site availability, and if prevalence of SNV 
correlation between deer mouse abundance 
antibodies was positively related to deer 
and grass cover. Douglass et al. (2001) also 
mouse abundance between rocky and less 
reported deer mice to be more abundant in 
rocky areas. Using long-term data from three 
sagebrush habitats in Montana. Shrubs in the 
live trapping grids, we found deer mice more 
current study area were mostly snowberry 
abundant at a site with more rock cover, than 
(Symphoricarpus spp.) with a growth form 
sites with less rock cover over many years. 
quite different from (particularly density of 
Deer mice were on average more abundant 
stems over the ground) sage. Collectively, 
at rocky sites in the short-term study, but 
our study, Douglass (1989b), and Douglass 
this was a non-significant trend (P = 0.053). 
et al. (2001) indicated that deer mice are 
Based on the long-term study, we also found 
captured in less complex habitat matrices. 
the number and prevalence of SNV antibody 
We focused on rock cover as a source of 
positive deer mice was greater at sites with 
variation in deer mouse abundance because 
more rock cover, than sites with less rock 
of rock cover’s potential use as retreat 
cover, which supported our hypothesis. Our 
sites by deer mice. While releasing deer 
short-term study did not detect a significant 
mice after capture we often observed them 
difference in SNV between sites with more 
seeking refuge under rocks. In this grassland 
or less rock cover, which could have been 
environment other types of retreat sites 
due to the short duration and relatively low 
(such as logs and thick shrub cover) are 
number of deer mice trapped. 
absent or rare. 
Habitat characteristics varied among 
The habitat composition in our short-
the three long-term grids in this study. 
term study only differed by rock cover and 
This variation in habitat composition, in 
lichens (which were observed on rocks only) 
addition to rock cover, included moss, 
among rocky and non-rocky grids. However, 
lichens, grasses, forbs, shrubs, and leaf 
deer mice were generally but not statistically 
litter. Among the three long-term grids, 
significantly so, more abundant at rocky 
more deer mice were captured where cover 
grids. Our long-term study demonstrated 
of rocks, moss and lichens were high, 
that the number of deer mice with SNV 
and fewer were captured where cover of 
antibodies and SNV antibody prevalence 
leaf litter, grasses and shrubs were high, 
was higher on the grid with greater deer 
suggesting these variables could all be 
mouse abundance. This lends support to 
determinants of deer mouse abundance 
our hypothesis linking rock cover, host 
24          Richardson et al.


Table 4. Abundance of deer mice, number of 
deer mice antibody positive to SNV (MNI) 
and antibody prevalence of deer mice (EAP) 
P
141
043
003
281
0.
0.
0.
0.945
0.598
0.681
0.
0.322
between rocky and non-rocky sites among 
hillsides for each trapping occasion, for the 
. Generalized 
short-term (Aug-Oct 2009) study. 
Zlm
  
 Hill 
Grid  
Mice  mni  eap
G
 
  Aug  One Non-rock  6  0  0
5
2 1,x
8.293 
11.456 
17.857 
1.203 
3.671 
3.124 
6.265 
4.675 
  
 
Rock 
10 

0
  
Two  Non-rock 


0
  
 
Rock 
22 
3  13.6
   
Three Non-rock 


0

     
 
Rock 


0
 
2) 
79
667
619) 
Ock
1.927) 
 sept  One Non-rock  17  0  0
R
  
 
Rock 
17 

0
1.2(0.
1.3(0.
1.0(0.516) 
8.0(0.775) 
31.5(5.005) 
7.5(1.057) 
1.5(0.
11.4(
  
Two  Non-rock 

2  28.6
Hree 
  
 
Rock 
23 

4.3
Ill T
 
   
Three Non-rock 

2  25.0
H

167
5) 
401) 
674) 
     
 
Rock 
15 

0
(2.
On-rock
2(0.
5(0.
  Oct  One Non-rock  30  0  0
N
0.
0.
0.3(0.333) 
8.0(0.856) 
31.2(5.134) 
9.0(0.516) 
1.2(0.
13.0
  
 
Rock 
26 

0
  
Two  Non-rock  14 
2  14.3
  
 
Rock 
27 
4  14.8
   
Three Non-rock  10 

0
 
333) 
543) 
     
 
Rock 
19 

0
Ock
1.658) 
R
7(0.
8(0.
0.
2.8(1.641) 
2.5(1.455) 
6.3(0.989) 
34.0(3.812) 
6.5(1.522) 
0.
9.4(

w

Table 5. LMM, with sampling occasion as 
Ill T
 
H
46) 
a repeated measure and site as a random 
543) 
833) 
563) 
(5.0
effect, of how abundance of deer mice 
On-rock
8(0.
8(0.
N
0.
0.
0.2(0.167) 
9.0(0.516) 
30.3(3.201) 
8.3(0.843) 
1.5(0.
15.0
(Mice), number of deer mice antibody 
positive for SNV(MNI), and antibody 
prevalence (EAP) differed between non-
rocky and rocky sites, for the short-term 

 
2) 
(August – October 2009) study.
fer between rocky and non-rocky areas, with variables nested within hillside. Significant results in bold.
79
667
619) 
Ock
1.488) 
R
  f  p
1.2(0.
1.3(0.
1.0(0.516) 
8.0(0.775) 
31.5(5.005) 
7.5(1.057) 
1.5(0.
8.4(
1,4
 Mice 
7.403 
0.053
Ne 
 MNI 
0.300 
0.613
Ill O
 EAP 
1.128 
0.432
H
 
333) 
On-rock
82(.307) 
abundance and pathogen abundance in 
N
2.2(1.014) 
0.
0.82(.307) 
5.2(0.91) 
35.7(9.017) 
7.0(0.816) 
2.7(0.
7.8(1.350) 
this grassland ecosystem. This relationship 
suggests that SNV prevalence is positively 
related to deer mouse abundance, which is 
round 
eight 
tter 
consistent with some studies (e.g., Calisher 
. Mean (± SE) percent cover of vegetation types among hillsides and sites within hillsides for the short-term (Aug-Oct 2009) study
ass 
et al. 1999, Yates et al. 2002, Carver et al. 
ock 
chens 
rass 
rass h
Bare g
R
Li
G
G
Leaf li
Forbs 
Biom
2011a, and Madhav et al. 2007 and Carver 
Table 3 Linear Model of how vegetation types dif
  
  
 
 
 
 
 
 
 
 
et al. 2011b in a delayed density fashion). 
Effect of Rock Cover on Small Mammal Abundance in a Montana Grassland           25


Transmission of SNV is horizontal due to 
Our informal observations indicate 
intraspecific interactions among deer mice, 
that deer mice use rocks as retreat sites 
such as aggressive encounters (Mills et al. 
throughout the year. However, variation in 
1997, Root et al. 1999, Botten et al. 2002). 
deer mouse abundance across this landscape 
Intraspecific interactions among deer mice 
is likely related to a complex combination 
may increase with increasing abundance, 
of habitat, climatic, and density dependant 
resulting in increased transmission events.
variables (i.e., Luis et al. 2010). It is also 
Our short-term study did not detect 
possible that at certain times of the year 
a relationship of deer mouse abundance 
predators, such as weasels and snakes 
with the number of SNV antibody positive 
during summer months, may influence 
individuals. Our short-term grids were only 
use of rocks by deer mice. Future spatially 
0.25-ha in size (consisting of 25 Sherman 
replicated analyses of other significant 
live traps each), which limited the number 
relationships among habitat characteristics 
of deer mice that could be trapped compared 
and deer mouse abundance, particularly 
to the numbers captured on the larger long-
using multiple long-term studies, would 
term grids. Given the size of trapping grids, 
be valuable to delineate all the underlying 
the lack of detectable differences in SNV at 
habitat determinants in this grassland 
rocky and non-rocky grids may be due to 
ecosystem landscape. Studies evaluating 
study design. Conducting this short-term, 
interactions among retreat site use, other 
spatially replicated study for a longer period 
fauna and climatic factors would also be 
of time and with one hectare trapping grids 
valuable. Telemetry studies (Douglass 
may enable better detection of relationships 
1989a) and food habit studies (Van Horne 
between 1-ha rock cover, deer mouse 
1982) would be valuable approaches to 
abundance and SNV among deer mice 
accurately determine how deer mice use this 
across this grassland landscape.
grassland habitat.
Other habitat characteristics can also 
function as sources of retreat sites to deer 
Acknowledgments
mice, particularly in other habitat types, 
We thank the private ranch owner 
leading to variation in abundance and 
at Cascade for allowing us access to his 
infection. Douglass et al. (2001) found deer 
property. Numerous individuals provided 
mouse abundance and SNV prevalence 
valuable assistance in the field including K. 
was greater in sagebrush habitat (where 
Coffin, R. Van Horn, C. Rognli, T. Wilson, 
deer mice may use sagebrush as retreat 
W. Semmens, K. Hughes, A. Skypala, D. 
sites) than grassland and forest habitats in 
Waltee, B. Lonner, J. Wilson, A. Leary, 
western Montana. Root et al. (1999) found 
J. Bertoglio, A. Alvarado, J. Trueax,C. 
abundance of deer mice and the number of 
Richardson and F. Arneson. K, Wagner 
SNV positive individuals to be greater at a 
provided database support, encouragement 
site dominated with sagebrush/juniper/pine 
and general advice. S. Zantos provided 
than a site with oak/mixed grass /forbs in 
valuable laboratory assistance. Financial 
western Colorado. Lehmer et al. (2008), in a 
support was provided by NIH grant P20 
brief study, found deer mice were on average 
RR16455-06-07,08 from the INBRE-
more abundant and with higher SNV 
BRIN program of the National Center for 
prevalence at sites with less mechanical (off 
Research Resources and the U.S. Centers 
road vehicle) anthropogenic disturbance 
for Disease Control and Prevention, Atlanta, 
in the Great Basin Desert, Utah.  However, 
GA, through cooperative agreement. K. 
in general, deer mice have been found to 
Richardson was additionally supported by a 
increase in numbers when habitats were 
grant from the Montana Tech Undergraduate 
opened by grazing (Smith 1940, Douglass and 
Research Program. This work followed all 
Frisina 1993, Matlack et al. 2001, Johnson and 
relevant environmental and institutional 
Anthony 2008) and forest treatments (Sullivan 
regulations in the collection of data 
1979, Douglass et al. 1999).  
presented here. The findings and conclusions 
26          Richardson et al.


presented here are those of the authors and 
States. American Journal of Tropical 
do not necessarily represent the views of the 
Medicine and Hygiene 52: 393-397.
funding agencies.
Chitty, D, and E. Phipps. Seasonal changes 
L
in survival in mixed populations of 
Iterature cited
two species of vole. Journal of Animal 
Botten, J., K. Mirowsky, C. Y. Ye, K. 
Ecology 35:313-331.
Gottlieb, M. Saavedra, L. Ponce, and 
Douglass, R. J., 1989a. An Evaluation of 
B. Hjelle. 2002. Shedding and intracage 
Trap-revealed microhabitat selection; 
transmission of Sin Nombre hantavirus 
Using Radio-telemetry to test critical 
in the deer mouse (Peromyscus 
assumptions. Journal of Mammalogy 
maniculatus) model. Journal of Virology 
70:648-652
76:7587-7594.
Douglass, R.J. 1989b.  Assessment of the 
Calisher, C. H., W. Sweeney, J. N. Mills, 
use of selected rodents in ecological 
and B. J. Beaty. 1999. Natural history of 
monitoring. Environmental. Management 
Sin Nombre virus in western Colorado. 
13:155-363. 
Emerging Infectious Diseases 5:126-134.
Douglass, R. J. and M. R. Frisina. 1993. 
Carver, S., J. N. Mills, A. Kuenzi, T. 
Mice and management on the Mount 
Flieststra,  and R. Douglass. 2010. 
Haggin Wildlife Management Area.  
Sampling frequency differentially 
Rangelands 15:8-12.
influences interpretation of zoonotic 
pathogen and host dynamics: Sin Nombre 
Douglass, R.J., J. Quinn, K.W. Coffin, 
virus and deer mice. Vector-Borne and 
and G. Mariani. 1999. Initial effects 
Zoonotic Diseases 10:575-583. 
of a landscape ecology treatment of a 
coniferous forest on small mammals. 
Carver, S., A. Kuenzi, K. H. Bagamian, J. 
Intermountain Journal of Sciences 5:12-22.
N. Mills, P. E. Rollin, S. N. Zanto, and 
R. Douglass. 2011a. A temporal dilution 
Douglass, R. J., R. VanHorn, K. W. Coffin, 
effect: hantavirus infection in deer mice 
and S. N. Zanto. 1996. Hantavirus 
and the intermittent presence of voles 
in Montana deer mouse populations: 
in Montana. Oecologia. Ahead of Print
preliminary results. Journal of Wildlife 
DOI 10.1007/s00442-00010-01882-z.
Diseases 32:527-530.
Carver, S., T. J. Trueax, R. Douglass, and 
Douglass, R. J., T. Wilson, W. J. Semmens, 
A. Kuenzi. 2011b. Delayed density-
S. N. Zanto, C. W. Bond, R. C. Van 
dependent prevalence of Sin Nombre 
Horn, and J. N. Mills. 2001. Longitudinal 
virus infection in deer mice (Peromyscus 
studies of Sin Nombre virus in deer 
maniculatus) in central and western 
mouse-dominated ecosystems of 
Montana. Journal of Wildlife Diseases 
Montana. American Journal of Tropical 
47:56-63
Medicine and Hygiene 65:33-41.
CDC. 2009. All about hantaviruses. 
 Feldmann  H., A. Sanchez, S.  Morzunov, 
Available: http://www.cdc.gov/ncidod/
C. F.  Spiropoulou, P. E. Rollin, T. G. 
diseases/hanta/hps/noframes/epislides/
Ksiazek, C. J. Peters and S. T. Nichol. 
episl5.htm. [Accessed Aug 18, 2009].
1993. Utilization of autopsy RNA for the 
synthesis of the nucleocapsid antigen of 
Childs, J. E., J. W. Krebs, T. G. Ksiazek, G. 
a newly recognized virus associated with 
O. Maupin, K. L. Gage, P. E. Rollin, P. 
hantavirus pulmonary syndrome. Virus 
S. Zeitz, J. Sarisky, R. E. Enscore, J. C. 
Research 30:351-367.
Butler, J. E. Cheek, G. E. Glass, and C. 
J. Peters. 1995. A household-based, case-
Glass, G. E., J. E. Cheek, J. A. Patz, T. M. 
control study of environmental-factors 
Shields, T. J. Doyle, D. A. Thoroughman, 
associated with hantavirus pulmonary 
D. K. Hunt, R. E. Enscore, K. L. Gage, 
syndrome in the southwestern United 
C. Irland, C. J. Peters, and R. Bryan. 
Effect of Rock Cover on Small Mammal Abundance in a Montana Grassland           27


2000. Using remotely sensed data to 
T. Davis, D. T. Tanda, J. W. Frampton, 
identify areas at risk for hantavirus 
C. R. Nichols, C. J. Peters, and J. E. 
pulmonary syndrome. Emerging 
Childs. 1997. Patterns of association with 
Infectious Diseases 6:238-247.
host and habitat: Antibody reactive with 
Johnston, A.N. and R.G. Anthony. 2008.  
Sin Nombre virus in small mammals 
Small-mammal microhabitat associations 
in the major biotic communities of the 
and response to grazing in Oregon. 
southwestern United States. American 
Journal of Wildlife Management 
Journal of Tropical Medicine and 
72:1736-1746. 
Hygiene 56: 273-284.
Keeling, M. J. and P. Rohani. 2008. 
Morris, D. W. 1997. Optimally foraging deer 
Modeling infectious diseases in humans 
mice in prairie mosaics: a test of habitat 
and animals. Princeton University Press, 
theory and absence of landscape effects. 
Princeton.
Oikos 80: 31-42.
Kirkland, G. L. and J. N. Layne. 1989. 
Nichol, S. T., C. F. Spiropoulou, S. 
Advances in the study of Peromyscus 
Morzunov, P. E. Rollin, T. G. Ksiazek, 
(Rodentia). Texas Tech University Press, 
H. Feldmann, A. Sanchez, J. Childs, S. 
Lubbock.
Zaki, and C. J. Peters. 1993. Genetic 
identification of a hantavirus associated 
Krebs, C. J. 1996. Population cycles revisited. 
with an outbreak of acute respiratory 
Journal of Mammalogy 77:8-24.
illness. Science 262: 914-917.
Lehmer, E. M., C. A. Clay, J. Pearce-Duvet, 
Niklasson, B., B. Hornfeldt, A. Lundkvist, S. 
S. S. Jeor, and M. D. Dearing. 2008. 
Bjorsten, and J. Leduc. 1995. Temporal 
Differential regulation of pathogens: the 
dynamics of Puumala virus antibody 
role of habitat disturbance in predicting 
prevalence in voles and of nephropathia 
prevalence of Sin Nombre virus. 
epidemica incidence in humans. 
Oecologia 155:429-439.
American Journal of Tropical Medicine 
Luis, A. D., R. J. Douglass, J. N. Mills and 
and Hygiene 53:134-140.
O. N. Bjørnstad. 2010. The effect of 
Root, J. J., C. H. Calisher, and B. J. Beaty. 
seasonality, density and climate on the 
1999. Relationships of deer mouse 
population dynamics of Montana deer 
movement, vegetative structure, and 
mice, important reservoir hosts for Sin 
prevalence of infection with Sin Nombre 
Nombre hantavirus. Journal of Animal 
virus. Journal of Wildlife Diseases 
Ecology 79:462-470.
35:311-318.
Madhav, N. K., K. D. Wagoner, R.  J. 
Smith, C.C. 1940. The effect of overgrazing 
Douglass, and J. N. Mills. 2007. Delayed 
and erosion upon the biota of the mixed-
density-dependent prevalence of Sin 
grass prairie of Oklahoma. Ecology 
Nombre virus antibody in Montana deer 
21:381-397.
mice (Peromyscus maniculatus) and 
implications for human disease risk. 
Stapp, P. and B. Van Horne. 1997. Response 
Vector-Borne and Zoonotic Diseases 
of deer mice (Peromyscus maniculatus
7:353-364.
to shrubs in shortgrass prairie: linking 
small-scale movements and the spatial 
Matlack, R. S., D. W. Kaufman, and G. A. 
distribution of individuals. Functional 
Kaufman. 2001. Influence of grazing by 
Ecology 11:644-651.
bison and cattle on deer mice in burned 
tall grass prairie. American Midland 
Sullivan, T.P. 1979.  Clear-cut habitat and 
Naturalist 146:361-368.
conifer seed predation by deer mice. Journal 
of Wildlife Management 43:861-871.
Mills, J. N., T. G. Ksiazek, B. A. Ellis, P. E. 
Rollin, S. T. Nichol, T. L. Yates, W. L. 
Van Horne, B. 1982. Niches of Adult 
Gannon, C. E. Levy, D. M. Engelthaler, 
and juvenile deer mice (Peromyscus 
28          Richardson et al.


maniculatus) in seral stages of coniferous 
Castle, C. H. Calisher, S. T. Nichol, K. 
forest. Ecology 63:992-1003.
D. Abbott, J. C. Young, M. L. Morrison, 
Wecker, S. C. 1963. The role of early 
B. J. Beaty, J. L. Dunnum, R. J. Baker, J. 
experience in habitat selection by 
Salazar-Bravo, and C. J. Peters. 2002 The 
the prairie deer mouse, Peromyscus 
ecology and evolutionary history of an 
maniculatus bairdi. Ecological 
emergent disease: Hantavirus pulmonary 
Monographs 33:307-325.
syndrome. Bioscience 52:989-998.
Yates, T. L., J. N. Mills, C. A. Parmenter, T. 
Zar, H. Z. 1996. Biostatistical Analysis.  
G. Ksiazek, R. R. Parmenter, J. R. Vande 
Prentice Hall
Received  16 November 2010
Accepted  29 June 2011
Effect of Rock Cover on Small Mammal Abundance in a Montana Grassland           29