Effects of Population Density on Stress and Maternal Care in a
Wild Rodent (Peromyscus maniculatus)
by
Gillian M. Merritt
A Thesis
presented to
The University of Guelph
In partial fulfillment of requirements
for the degree of
Masters of Science
in
Integrative Biology
© Gillian M. Merritt, August 2014
ABSTRACT
EFFECTS OF POPULATION DENSITY ON STRESS AND MATERNAL CARE IN A WILD RODENT
(PEROMYSCUS MANICULATUS)
Gillian Marie Merritt
University of Guelph, 2014
Advisors:
Professor Amy E.M. Newman
Professor Andrew G. McAdam
Laboratory studies have demonstrated that variation in maternal care is linked to differences of offspring
stress axis dynamics. Offspring who receive a low level nursing face increased exposure to elevated
CORT compared to high nursing offspring which, according to the CORT-fitness hypothesis, leads to a
reduced fitness. There is a growing body of empirical data which suggests that this relationship between
CORT and fitness may depend on the environmental conditions. In the wild, if a female has access to
environmental cues, it may be advantageous for a female to program her offspring’s stress axis to match
the environment they will experience upon independence. Here, I tested the effects of population density
on maternal behaviour by bringing wild-caught Peromyscus maniculatus into captivity to measure their
frequency of nursing while experimentally manipulating perceived density cues through the presence of
soiled bedding. I predicted the mechanism for this behavioural response to environmental cues to be
related to a change in maternal glucocorticoids. As population density increased over the breeding
season the frequency of nursing decreased by approximately 27% however, olfactory cues alone do not
experimentally alter nursing. We report a U-shaped relationship between population density and maternal
stress levels that can be explained by a significant interaction between population density and time of
breeding. This suggests that CORT is not the mechanistic link between maternal behaviour and
population density but rather may be related to the level of maternal care the female received as a pup.
Wild mice face this mismatch of environments where the density in which they were born can be very
different from the density in which they breed. These complex relationships between natal and breeding
environments might require complex systems for appropriately programming offspring for the current
environmental conditions. Matching behavioural phenotypes to a changing environment could be the
driver responsible for maintaining variation in maternal behaviour in a wild environment.
iii
iii
ACKNOWLEDGEMENTS
First and foremost, I would like to thank my advisors Amy Newman and Andrew McAdam for their
continuous support and mentorship. Thank you for empowering me to strive for excellence and never
settling for good enough. I am truly fortunate to have had the opportunity to work with both of you. I would
like to thank my committee member Rob McLaughlin for your valuable insight and encouraging me to
“stretch the elastic”.
I am grateful to all members of the McAdam and Newman labs. I would like to thank J. Pakkala for
sticking with me through all of my RIA analyses and E. Siracusa for being there to help problem solve no
matter the dilemma. I would also like to acknowledge my office mates L. Sicoly and A. McLean for
spending countless moments considering “Does this make sense?”.I would especially like to thank my
partner in crime, K.Deasley for having my back and keeping me grounded over the last two years.
I could not have completed my two successful field seasons without the love and support of the Wildlife
Research Station family. Thank you R. Eckenswiller and D. Miller for dealing with all my, at times
ridiculous, requests and my cabin mates V.Mather and K.W ozniak for our late night talks and your
acquired ability to ignore my 5:00am alarm. Most of all I would like to thank my field technicians; E.
Schmidt, T. Koeth, M.Freeman and K.Smith. You guys got up every morning at the crack of dawn, despite
the snow, rain, heat or mosquitoes, ready to get the job done with a couple of laughs and a lot of ice
cream along the way.
Finally thank you to my family for their endless love, support and patience. I could not have done this
without you.
This research was supported in part by funding provided by the Ontario Graduate Scholarship and the
National Science and Engineering Research Council of Canada.
iv
iv
TABLE OF CONTENTS
LIST OF TABLES .........................................................................................................................................V
LIST OF FIGURES ......................................................................................................................................VI
1.0 INTRODUCTION..................................................................................................................................... 1
2.0 METHODS .............................................................................................................................................. 6
STUDY LOCATION AND TRAPPING.......................................................................................................................................6
LABORATORY PROCEDURES...............................................................................................................................................6
Animal care ............................................................................................................................................................6
Experimental density manipulation .......................................................................................................................7
Maternal observation ............................................................................................................................................8
Measuring stress response ....................................................................................................................................9
Fecal sample extraction .......................................................................................................................................10
STATISTICAL ANALYSIS ...................................................................................................................................................11
Measuring population density of wild Peromyscus .............................................................................................11
Natural density ....................................................................................................................................................11
Perceived density experiment ..............................................................................................................................12
3.0 RESULTS ............................................................................................................................................. 13
Effects of natural population density...................................................................................................................13
Effects of experimentally manipulated density....................................................................................................14
4.0 DISCUSSION ........................................................................................................................................ 14
Relationship between maternal behaviour and population density in wild mice ................................................14
Maternal behaviour, fecal corticosterone and population density......................................................................17
5.0 CONCLUSION ...................................................................................................................................... 20
6.0 LITERATURE CITED............................................................................................................................ 21
7.0 TABLES ................................................................................................................................................ 27
8.0 FIGURES .............................................................................................................................................. 28
9.0 APPENDIX ............................................................................................................................................ 35
Effect of food supplementation on maternal behaviour......................................................................................35
v
v
LIST OF TABLES
Table 1: Description of the behaviours and locations used to classify a female’s maternal care. Location
and behaviour were categorized every three minutes during four one-hour focals for the first three nights
postnatal. Behaviours were not mutually exclusive as a female could be performing multiple behaviours at
a time........................................................................................................................................................... 27
vi
vi
LIST OF FIGURES
Figure 1: The number of wild P. maniculatus captures significantly increased over the course of the 2013
breeding season (Julian Date). Mouse population density was measured as the total number of deer
mouse captures per 100 trap nights on nearby 100m transects within three habitat types at the time of
initial capture. .............................................................................................................................................. 28
Figure 2: The proportion of time spent nursing by P.maniculatus females significantly declined with
increasing population density (measured as captures/100 trap nights; n=30). The population density
measurement for each female represents the local density the female was experiencing at the time she
was brought into the lab. ............................................................................................................................. 29
Figure 3: There was a significant positive quadratic relationship between corrected-CORT on PND3 and
population density at the time of the female’s capture. There are fewer females with corrected-CORT
values as not all females produced a fecal sample on PND3 (n=25). ........................................................ 30
Figure 4: Maternal CORT levels show an inconsistent response to population density. Early breeders
show no effect of population density on maternal CORT (closed circles; n=12) while late breeders show
the predicted positive relationship between population density and maternal CORT (open circles; n=13).
.................................................................................................................................................................... 31
Figure 5: There was no significant difference in the proportion of time spent nursing between control
females (n=14) and treatment females (n=16) who were exposed to the soiled bedding of four unknown
individuals. Females were exposed to the bedding during their pregnancy up until PND3........................ 32
Figure 6: There was not a significant difference in corrected-CORT values on PND3 between control
females (n=13) and treatment females (n=12) who were exposed to the soiled bedding of four unknown
individuals....................................................................................................................................................
33
Figure 7: The proportion of time spent nursing decreased over the course of the season in both 2013
(solid line) and 2011 (dotted line; A). In 2011 (dotted line) the proportion of time spent nursing declined
almost 50% from the beginning to the end of the breeding season compared to just over 27% in 2013
(solid line). In contrast, in both 2011 and 2013 there were remarkably consistent declines in nursing in
response to a seasonal increase in population density, despite overall densities being lower in 2013 than
2011 (B)....................................................................................................................................................... 34
1
1.0 Introduction
Maternal care plays a critical role in an offspring’s growth and development. This is especially true in
mammals as the extended period of mother-infant contact is not only essential for survival, but the
behaviour of the mother can be highly influential on the long term wellbeing of her offspring (Myers et al.,
1989; Meaney, 2001; Champagne et al., 2003). Elegant laboratory studies have identified individual
variation in the maternal care of laboratory rodents, such as the variation in the frequency of nursing over
the first seven days post partum (Myers et al., 1989; Champagne et al., 2003; Champagne et al., 2007).
These studies have traditionally categorized females as being either high- or low-nursing mothers (Liu et
al., 1997; Meaney, 2001; Champagne et al., 2003; Champagne et al., 2007), but this classification is
based on an artificially constructed dichotomy as the frequency of nursing is a normally distributed
variable (Champagne et al., 2003). In addition to providing essential nutrition to the offspring during the
early life environment, the frequency of nursing has been shown to affect long-term aspects of the
offspring’s physiological and psychological health including cardiovascular functions, cognitive abilities,
immune response and the development of the stress axis (Liu et al., 1997; Meaney, 2001; Champagne et
al., 2003; Champagne et al., 2007).
The stress response of an individual is controlled by the Hypothalamic-Pituitary-Adrenal (HPA) axis where
glucocorticoids (GCs) are maintained at basal levels through negative feedback (Smith and Vale, 2006).
When an individual is presented with a stressor, levels of circulating corticosterone (CORT), the primary
GC in rodents, increase (Meaney, 2001; Smith and Vale, 2006). Elevated CORT mobilizes glucose and
lipid reserves thereby diverting energy towards physiological functions essential for survival (i.e.
increased cardiac function, attentional processes, learning and memory) and activating the ‘fight or flight
response’ (Meaney, 2001). A large body of literature on laboratory rodents has shown that variation in
the frequency of nursing is linked to differences in offspring stress axis dynamics. Offspring who receive a
low level of nursing have a reduced number of glucocorticoid receptors (GRs) in the brain compared to
high nursing offspring. During a stress response, a fewer number of GRs ultimately leads to prolonged
CORT elevation as the negative feedback cannot respond as quickly (Liu et al., 1997; W eaver et al.,
2
2004; Hellstrom et al., 2012). Cross-fostering studies provide evidence for a nongenomic transmission of
individual differences in stress reactivity and maternal behaviour (Meaney, 2001). Female offspring
birthed by low-nursing mothers but reared by high-nursing mothers will experience reduced exposure to
elevated CORT as a result of an increased number of GRs and will also subsequently exhibit a high level
of nursing towards their offspring (Meaney, 2001). These laboratory studies have been fundamental in
our understanding of the consequences of maternal care for physiological and psychological
development, but the ultimate question of why variation in maternal behaviour persists remains
unanswered.
It is commonly assumed, especially in the biomedical literature, that an individual or population with a
slower HPA negative feedback has reduced fitness (Bonier et al., 2009; Boonstra, 2012). This has been
termed the CORT-Fitness hypothesis and is based on the observation that individuals exposed to
elevated CORT for a prolonged period of time are at increased risk of heart disease, diabetes, anxiety
disorders and depression (Meaney, 2001; Bonier et al., 2009). However, if offspring who receive a low
level of nursing always have lower fitness when compared to offspring that received more nursing, the
low maternal behaviour phenotype should be eliminated from the population through natural selection,
thereby reducing and/or eliminating any variation in the frequency of nursing. As this is not the case, there
must be a mechanism present in the natural world that maintains this variation in nursing frequency. In
contrast to the ‘stress as pathology’ paradigm, there has been a recently growing body of empirical data
that suggest the relationship between GCs and fitness is not static, but might vary with changing
environmental conditions (Bonier et al., 2009; Boonstra, 2012). In stressful environments, an earlier and
longer activation of the fight or flight response may outweigh the cost of increased exposure to elevated
CORT (Meaney, 2001; Harris et al., 2012). Laboratory studies with rodents have demonstrated that
maternal behaviour is altered by stressors experienced during gestation; females who experience high
CORT during the prenatal environment exhibit reduced maternal behaviour towards their pups compared
to pregnant females in non stressful environments (Moore and Power, 1986; Patin et al., 2002; Smith et
al., 2004; Champagne and Meaney, 2006; Baker et al., 2008; Brummelte and Galea, 2010). Furthermore,
3
experimental manipulations of maternal stress during gestation have been found to affect the CORT
levels of the pups indicating that maternal stress hormones can affect the stress levels of the offspring
(Weinstock 2001, 2008; Smith et al., 2004; Champagne and Meaney, 2006; Baker et al., 2008). In
stressful environments, low maternal care would result in offspring with increased exposure to elevated
CORT that would display behaviours essential for survival (i.e. increased cardiac function, attentional
processes, learning and memory) faster and for a longer period of time than offspring who experienced
high levels of maternal care with reduced exposure to elevated CORT (Meaney, 2001; Harris et al.,
2012). In a wild environment, where gestational stressors vary in space and time, if females have access
to cues of the environment that offspring will experience then it might be advantageous for a mother to
program her offspring’s stress axis to match the environment they will experience upon independence
(Uller, 2008; Dantzer et al., 2013). Specifically, a female might program an HPA axis with a reduced
number of GRs in her offspring when offspring are likely to experience consistent or repeatedly stressful
environments.
One biotic factor that fluctuates and could have important implications for the appropriateness of a hyperor hypo-responsive stress axis is population density. High population density has been hypothesized to
be a stressful environment for mouse populations as a result of the positive associations between
population density and food competition, negative social interactions and predation (Kaiser and Sachser,
2005; Creel et al., 2012). This hypothesis is based on laboratory and field studies that have shown a
decrease in reproductive output at high population density (laboratory studies: Christian 1950, 1956,
1961; Terman 1965, 1968, 1969; Rogers and Beauchamp, 1976; Lidicker, 1976; Lombardi and Whitsett,
1980; Creigh and Terman, 1988; field studies: Morris, 1989; Lusk and Millar, 1989; Montogermy, 1989;
Drost and Fellers, 1991; Terman, 1993). This reduction in reproductive output has been attributed to the
reduced size of reproductive organs, decreases in fertility and the inhibition of juvenile maturation
(Christian 1950, 1956, 1961; Terman, 1969; Rogers and Beauchamp, 1976; Lidicker, 1976; Lombardi and
Whitsett, 1980; Creigh and Terman, 1988). Laboratory studies working with Mus, Acromys and
Peromyscus show a positive relationship between the number of individuals housed in each cage and
4
CORT levels suggesting that stress may be the possible mechanism responsible for the reduction in
reproductive output observed at high population densities in the wild (Hunt and Hambly, 2006; Nováková
et al., 2008; Peng et al., 2009; Thomason et al., 2013). To date only one field study has investigated the
relationship between CORT and population density in wild mice; Harper and Austad (2004) documented
that CORT increased with rising population density in a deer mouse (Peromyscus maniculatus)
population.
Many small mammal species, including deer mice (Peromyscus maniculatus), exhibit large seasonal and
annual fluctuations in population density (Fryxell et al., 1998; Falls et al., 2007). In addition to
reproductive output, population density has been previously shown to affect immune function (Pedersen
and Greives, 2008; Thomason et al., 2013) behaviour (Wolff, 1985) and home range size (Taitt, 1981;
Wolff, 1985; Boutin et al., 1990) in wild deer mice. It is unclear what social cues deer mice use to assess
population density, but these could be tactile, auditory, olfactory or visual (Terman, 1979, 1980). Scent
appears to be the most likely cue for assessing population density as reproductive inhibition and the delay
of sexual maturation in juveniles, two phenomena observed under high population density, can be
mediated through olfactory signals in the absence of other social cues (Champlin, 1970; Drickamer, 1974;
Ma et al., 1998). These findings are derived from laboratory experiments in which the effect of multiple
individuals in a cage was simulated by exposing singly housed females to the urine or soiled bedding of
conspecifics (Mus musculus: Champlin, 1970; Drickamer, 1974; Ma et al., 1998; Peromyscus maniculatus
bairdii: Lawton and W hitsett, 1979; Lombardi and Whitsett, 1980; Creigh and Terman, 1988; Peromyscus
leucopus: Rogers and Beauchamp, 1976; Terman, 1984). In the wild, as population density increases
there is greater home range overlap between individuals (Wolff, 1985; Boutin et al., 1990), which
presumably results in an increase in the number of unique deer mouse scents or scent marking within an
individual’s home range. Unfortunately, it is virtually unknown how scent marking is used by wild mice
(Wolff, 2003).
5
While numerous laboratory studies have examined variation in maternal care in mice (Myers et al., 1989;
Liu et al., 1997; Meaney, 2001; Champagne et al., 2003 Champagne et al., 2007), only one previous
study has applied these laboratory measures of maternal behaviour to a wild rodent (Stewart and
McAdam, 2014). Stewart and McAdam (2014) used standard laboratory rodent protocols (Myers et al.,
1989; Champagne et al., 2003; Champagne et al., 2007) to measure the maternal behaviour of wildcaught female deer mice. The maternal behaviour of deer mice was found to be repeatable but changed
plastically over the summer as population density increased (Stewart and McAdam, 2014). Females at
the beginning of the breeding season spent twice as much time nursing as females that gave birth at the
end of the breeding season when population density was much higher (Stewart and McAdam, 2014). This
plasticity differs from previous studies of laboratory rodents, which have found maternal behaviour to be
very consistent within an individual (Meaney, 2001), and suggests that mother’s might adjust their
behaviour toward offspring depending on cues of population density. This one field study, however, was
observational and could, therefore, not definitively attribute the changes in maternal behaviour to changes
in population density since population density was confounded with seasonality (Stewart and McAdam,
2014). Experiments are, therefore, needed to test the effects of population density on maternal behaviour,
and identify the modality by which density is assessed. Furthermore, although laboratory studies have
shown CORT to affect maternal behaviour and field studies have documented increases in CORT levels
with increasing population density the role of maternal stress in mediating the linkage between seasonal
changes in density and maternal behaviour has not yet been explored.
Here, I tested the effects of population density on maternal stress and maternal behaviour, by bringing
pregnant, wild-caught Peromyscus maniculatus gracilis into captivity to measure their maternal
behaviours following standard protocols under controlled conditions (Myers et al., 1989; Champagne et
al., 2003; Champagne et al., 2007). In the laboratory, I experimentally manipulated cues of perceived
density and collected fecal samples from the female to measure CORT levels. I hypothesized that female
deer mice would adjust their maternal behaviour in response to population density and that this would be
mediated by maternal stress levels. I therefore, predicted that as deer mouse population density
6
increased as a result of natural seasonal fluctuations in the wild, the frequency of nursing would decrease
and maternal CORT levels would increase. I further predicted that female deer mice exposed to the scent
of unknown deer mice in the lab would have a lower frequency of nursing and higher maternal CORT
levels than control females not exposed to conspecific scent.
2.0 Methods
Study Location and Trapping
Research was conducted at the Wildlife Research Station (WRS) in Algonquin Provincial Park, Ontario,
Canada (45°37'N, 78°21'W) from April to August 2013. Deer mice were trapped in Longworth live traps
(Roger’s Manufacturing, British Columbia Canada) on 11 transects that were each 100m long with two
traps placed every 10m along each transect. Traps were baited with water soaked sunflower seeds
(Helianthus spp) and freeze-killed mealworms (Tenebrio molitor) were provided to reduce accidental
shrew mortality (Soricomorpha: Soricidae; Do et al., 2013). Cotton bedding was provided in the traps for
warmth. Traps were set three consecutive nights per week at dusk and checked at dawn. Upon capture,
deer mice were removed from the trap, placed in a large handling bag and weighed with a Pesola spring
balance (Pesola® Switzerland) to the nearest 0.5g. Deer mice were tagged in both ears with metal ear
tags (National Brand and Tag Co.) for identification. Age of the individual was determined based on
dorsal pelage where juveniles are grey, sub-adults are a mixture of grey and brown and adults are brown
(Schug et al., 1991). All females who were not perforate were assumed to be pregnant and were then
placed in a clean Sherman trap (H. B. Sherman Inc., Tallahassee Florida USA) and transported to the
animal care facility at the WRS.
Laboratory Procedures
Animal care
Females were housed in an animal proof laboratory with natural diel light hours. Windows in the
laboratory were left open so that temperature and humidity conditions were as close to the current
7
weather conditions as possible. When external temperatures dropped below 5°C, the heat was turned on
and maintained at 15°C to prevent dam or pup mortality. Individuals were housed in Plexiglas cages (35
cm x 15 cm x 17 cm) with ad lib supply of food (Rodent Chow 14% protein) and water. The cages
contained Corn Cob bedding (Harlan Corn Cob Bedding), paper (Andersons CN Crink-l’nest™) and
additional cotton bedding (Ancare Nestlets) as well as a cardboard tube for environmental enrichment. All
bedding was replaced every three days when the female’s cage was cleaned except when pups were
present. During the cage cleaning, the female was placed in a clean handling bag and weighed to the
nearest 0.5g with a Pesola spring balance. The female usually produced a fecal sample while inside the
handling bag. If the female did not defecate the mouse was held until a fecal sample was produced, so all
feces collected were known to be fresh. Fecal samples were collected in a clean, labelled 1.5mL
microcentrifuge tube and frozen until extraction. Cages were thoroughly cleaned with a 10% bleach
solution between occupants.
Experimental density manipulation
Upon arrival in captivity, females were randomly assigned to either control (n = 28) or treatment (n = 26)
groups. Treatment females were exposed to four cotton ball sized pieces of frozen soiled bedding from
four different adult deer mice; two males and two females. Control females were exposed to four similar
sized pieces of clean, frozen cotton bedding. Soiled bedding was collected from live traps that contained
adult deer mice on live trapping transects used to assess the density of various small mammal species in
Algonquin Park (Fryxell et al., 1998; Falls et al., 2007). The transects were located well outside the home
range of the females studied in the laboratory, so individuals from which the soiled bedding was collected
were presumably unfamiliar. Soiled bedding was collected directly from the trap upon release of the
captured mouse and placed in a clean plastic bag that was labelled and sealed. Upon return to the
laboratory, the cotton bedding was cut into cotton ball sized pieces and immediately frozen. The bedding
pieces were replaced every three days when the females’ cage was cleaned unless pups were present.
By cutting the soiled bedding into pieces at the time of collection, I was able to expose each treatment
female to the scent of the same four unknown individuals over the course of her time in the laboratory.
8
Maternal observation
Females were checked at least every 12 hours in order to determine the day of birth, which was
designated as postnatal day (PND) 0. When females gave birth, pups were weighed to the nearest 0.1g
using an electronic balance (Ohaus Corporation, New Jersey USA) and females were weighed to the
nearest 0.5g with a Pesola spring balance. A fecal sample was also collected from the female at this time.
rd
Females who gave birth prior to July 3 were defined as early breeders while females who gave birth on
rd
or after July 3 were defined as late breeders. Early breeders were likely over-wintered (OW) animals
born in the previous fall since the young-of-the-year (YY) in this study and a concurrent long-term study of
th
small mammals in the area (Fryxell et al., 1998; Falls et al., 2007) were first captured on May 29 , 2013.
Based on an age of sexual maturation of one month and a gestation period between 22.4 and 25.5 days
rd
(Millar, 1982; Millar, 1985; Kirkland and Layne, 1989) known YY could not have given birth before July 3 ,
2013. Females defined as late-breeders could have been either OW or YY mice. The female and pups
were then left undisturbed until PND3. During this three-day period maternal behaviour was observed by
videotaping the female with her pups under infrared conditions from dusk to dawn, when deer mice are
most active (Champagne et al., 2007).
Maternal behaviour was quantified following the methods of Myers et al. (1989) and Champagne et al.
(2003) during four predetermined 1-hour focal periods within each night; 1 hour after sunset, 23:30, 1:30
and 2 hours before sunrise. All recordings were made using a SONY HANDYCAM (MODEL #DCR-SR68)
video camera. Maternal behaviour was scored every 3 minutes within each 1-hour focal period (21
observations per 1-hour focal × 4 focals per night × 3 post natal nights = 252 observations per female)
with the observer blind to conditions. At each 3-minute interval maternal behaviour was first scored based
on the location of the female 1) “in nest” when the female was in physical contact with the nesting
material or 2) “out of nest” when the female was not in contact with nesting maternal. The female’s
behaviour was then classified from the following list in which behaviours were not mutually exclusive 1)
nursing, 2) licking/grooming pups, 3) manipulating nest shavings, 4) manipulating non nest shavings 5)
9
self grooming, 6) eating, 7) drinking, 8) rearing/climbing, 9) resting, 10) carrying pup, 11) out of sight
(Table 1). Nursing postures are often concealed by bedding material rendering them difficult to classify
(Stewart and McAdam, 2014). Therefore, for the purpose of this study all nursing postures were
combined and simply described as nursing.
On PND3, pups were sexed and weighed with an electronic balance to the nearest 0.1g. Ano-genital
distance (AGD; distance from the pups genital papilla to their anus in mm; Ryan and Vandenbergh,
2002), and body length (distance from the tip of the nose to the base of the tail in mm) measurements
were also taken. Adult females were weighed to the nearest 0.5g with a Pesola spring balance and a
fecal sample was collected. Females and their offspring were then placed in a nest box (12.5 cm x 12.5
cm x 15 cm) with the bedding from their nest, an additional nestlet and approximately 50 sunflower seeds.
Nest boxes were hung on a tree at the site of their original capture location at approximately chest height
and entry holes were exposed so that the female could depart the nest box with her pups.
Measuring stress response
Fecal samples were used to assess stress levels as they are a non-invasive, integrated measure of an
individual’s basal GC profile. Fecal samples are also not as sensitive to handling disturbances compared
to blood sampling as they represent GC levels over a number of hours instead of a specific instance in
time (Harper and Austad, 2000; Palme, 2005). GCs in the blood are broken down by the liver into
species-specific GC metabolites and then excreted as bile into the small intestine (Palme, 2005). In
Peromyscus, the GC metabolites can then be measured in the fecal samples to determine the stress level
of the individual 4-6 hours prior to the sample being produced (Harper and Austad, 2000; Harper and
Austad, 2004; Palme, 2005).
10
10
Fecal sample extraction
Prior to extraction, all fecal samples were freeze dried to remove variation in water content (Wasser et al.,
1993). For the extraction, all samples were weighed on an electronic balance to the microgram and then
transferred to individually labelled 16x100mm glass test tubes. A small amount of 100% methanol was
then added to the test tube and samples were homogenized with a glass rod. Samples were then topped
up to 2mL with 100% methanol. Samples were vortexed for 20 seconds, placed in a fume hood, covered
and shaken for 2 hours in a room temperature water bath. After 2 hours, samples were centrifuged for 10
minutes at 2500xg and the supernatant was poured into a clean 12x75mm test tube. Samples were then
dried in a fume hood and frozen at -80°C until analysis.
To quantify CORT, we used a double-antibody l-125 radioimmunoassay (MP Biomedicals LLC,
Orangeburg NY; catalog no. 07120103) that was modified following (Newman et al., 2008). Briefly,
manufacturer’s directions were followed except all reagents were halved and extracts were diluted to
between 0.004-0.006g per 50µL of assay buffer (0.004-0.006 g per tube). CORT concentrations are
expressed as nanogram of hormone per gram of fecal sample. Further dilution of a sample was
completed if the initial results were beyond the extrapolation potential of the standard curve. W e tested for
parallelism between a serially diluted pool of fecal samples and the standard curve using an analysis of
covariance in order to determine the biochemical validity of the assay (ANCOVA: Main effect of Sample
Type (standard or feces), F1, 11 = 717.0, p < 0.001; Main effect of Sample Dose (nanograms of CORT or
milligrams of feces), F1, 11 = 634.27, p < 0.001; Interaction between Sample Type and Dose, F1,11 = 0.25,
p = 0.63). The lack of a significant interaction between serial dilution and standard curve indicates that the
line slopes did not differ.
ETHICAL NOTE
All laboratory and field procedures were approved by the University of Guelph Animal Care Committee
(AUP #1418).
11
11
Statistical Analysis
Measuring population density of wild Peromyscus
Nearby 100m transects, located less than 300m away from the 100m transects used in this study, were
used to assess the local density the female was experiencing at the time she was brought into the lab.
These transects are part of a long-term small mammal study in Algonquin Park (Fryxell et al., 1998; Falls
et al., 2007). These density lines were trapped biweekly throughout the summer, with two traps placed at
10m intervals and there were 2-3 density transects in each of the three habitat types. This created eight
discrete density time periods for each habitat type. The number of captures for each two week period
within one habitat type was combined and density was measured as the number of captures of deer mice
per 100 trap nights. Females were brought into the laboratory every week, so the density measurement
for each female represented the population density on the nearby density lines in the same habitat type
either the week she was brought into the lab or the week before. In order to test for a seasonal change in
population density, I regressed total density and the density in each habitat type for each two-week period
against date (measured in Julian dates).
Natural density
The numbers of nursing observations were summed over the three days maternal behaviour was
measured and divided by the total number of observations scored for that female in order to calculate the
proportion of time spent nursing by that female for that litter. This resulted in a single nursing value for
each female. Two females were brought into captivity more than once for successive litters, but only the
first litter was used in the analysis. I tested whether natural seasonal fluctuations in density was a
significant predictor of maternal behaviour using linear mixed-effect models where the proportion of time
spent nursing was modeled by the population density at the time of capture (captures/100 trap nights).
Despite being a proportion, the proportion of time spent nursing was normally distributed so general linear
models of nursing were used. Time period was included as a random effect but a likelihood ratio test
indicated that the inclusion of this random effect of time period did not significantly improve the fit of the
-9
model (likelihood ratio test: χ² = 3.17x10 , df = 1, p = 1), so it was excluded.
12
12
Only PND3 fecal samples were used in the analysis in order to remove any variability in CORT resulting
from the female’s stage of gestation when she entered the laboratory (Pawluski et al., 2009). All of the
fecal pellets collected from the female on PND3 were used in the analysis. CORT concentrations were
log-transformed to obtain a normal distribution. There was a significant negative relationship between
CORT concentrations and the mass of the fecal sample collected (ß = -438.25 ± 84.42, t23 = - 5.19, p <
0.0001), so the residuals of this model were used in subsequent analyses. These residuals reflect the
CORT concentration after correcting for the mass of the original fecal sample collected from the female
and are referred to as corrected-CORT measures. I used a general linear model to test for an effect of
natural population density at the time of capture on corrected-CORT three days after parturition in
captivity with time of breeding (i.e. early or late) included in the model as a covariate.
Perceived density experiment
I compared the proportion of time spent nursing and corrected-CORT between treatment and control
groups using separate general linear models to test whether manipulation of perceived density affected
maternal behaviour and maternal CORT. There was no significant interaction between the treatment and
natural density (ß = 0.0018 ± 0.005, t26 = 0.37, p = 0.71), so the interaction was excluded from
subsequent models however, natural density remained as a covariate. Finally, I tested whether there was
an effect of corrected-CORT on maternal behaviour using a general linear model.
All statistical analyses were performed in R version 2.15.1 (R Development Core Team, 2012). Linear
mixed-effect models used the nlme package (Pinhero et al., 2011) with a Gaussian error distribution. All
values are presented as means ± s.e.m. unless otherwise stated.
13
13
3.0 Results
During the 2013 breeding season in Algonquin Provincial Park, natural population density measurements
of wild deer mice ranged from 0 to 26.7 captures/100 trap nights. The number of deer mice increased
-1
over the course of the breeding season overall (ß = 1.92x10 ± 0.33, t 6 = 5.78, p = 0.001; Figure 1) and
-1
within each of the three habitat types (cut over mixed woods: ß = 2.19x10 ± 0.053, t 6 = 4.12, p = 0.006;
-1
dense mixed woods: ß = 2.21x10 ± 0.053, t 6 = 4.18, p = 0.006; and white pine and white spruce: ß =
-1
1.21x10 ± 0.02, t 6 = 5.94, p = 0.001).
Effects of natural population density
Thirty mothers spent an average of 48.3 ± 25.7% of their time nursing their litters. Females who
experienced greater population densities in the field at the time of capture spent significantly less time
nursing their pups as predicted (ß = - 0.005 ± 0.002, t22 = - 2.18, p = 0.038; Figure 2).
The twenty-five mothers that produced a fecal sample on PND3 had an average CORT level of 58.2 ±
221.78ng/g. There was a significant nonlinear effect of the natural density of deer mice at the time of initial
capture on corrected-CORT values on PND3. Overall, there was a negative effect of density on correctedCORT (linear: ß = - 0.1 ± 0.036, t22 = - 2.75, p = 0.012) but there was also a significant positive quadratic
relationship (quadratic: ß = 0.004 ± 0.001, t22 = 2.91, p = 0.008; Cook’s Distance <0.05). Females that
experienced very low and very high population densities at the time of capture had the highest correctedCORT levels, whereas those that experienced moderate densities had the lowest corrected-CORT levels
(Figure 3). This quadratic relationship is consistent with a significant interaction between timing of
breeding and population density (ß = 0.087 ± 0.041, t21 = 2.13, p = 0.045; Figure 4). There was a positive
effect of population density on corrected-CORT for late breeders (ß = 0.077 ± 0.023, t11 = 3.40, p = 0.006)
but no effect of density on corrected-CORT for early breeders (ß = -0.102 ± 0.035,
t10 = - 0.29, p = 0.778). Finally, nursing was not negatively affected by CORT as we had predicted.
Instead, there was a significant positive relationship between maternal stress hormones and maternal
14
14
behaviour as females with elevated corrected-CORT levels spent significantly more time nursing their
pups than those with lower corrected-CORT levels (ß = 0.085 ± 0.039, t23 = 2.20, p = 0.038).
Effects of experimentally manipulated density
There was no effect of the experimental manipulation of perceived density on the proportion of time spent
nursing (ß = 0.066 ± 0.034, t27 = 1.77, p = 0.094; Figure 5) or corrected-CORT levels (ß = - 0.281 ± 0.186,
t21 = - 1.51, p = 0.146; Figure 6). Control females spent on average 44.6 ± 22% of their time nursing with
an average CORT measurement of 59.5 ± 220.6 ng/g. Conversely, treatment females spent on average
51.6 ± 16.3% with an average CORT measurement of 56.9 ± 198.9 ng/g. Furthermore, the presence of
the soiled bedding did not result in any significant differences between control and treatment females for
any other variable measured (date into lab, natural density, lab density, length of gestation, female
weight, litter size, pup weight, percent male, AGD, body size: t28 < 1.98, p > 0.05).
4.0 Discussion
Relationship between maternal behaviour and population density in wild mice
In contrast to previous laboratory studies that report maternal behaviour in captive mice to be repeatable
within an individual, I found that maternal behaviour in wild-caught mice was plastic; specifically, the time
spent nursing drastically decreased over the course of the breeding season as population density
increased. This result supports the only other study on maternal behaviour in mice from the natural
environment (Stewart and McAdam, 2014), and highlights the importance of studying and understanding
behaviour in a natural context. Interestingly, Stewart and McAdam (2014) studied this same deer mouse
population in 2011, a year in which densities increased from 5 to 45 deer mouse captures per 100 trap
nights while nursing behaviour declined by almost 50%. In comparison, density peaked in the current
study (in 2013) at 27 captures per 100 trap nights and nursing behaviour declined by 27% (Figure 7B).
While it is possible that another seasonal variable is responsible for the change in nursing behaviour, I
would then expect a consistent seasonal decline in nursing behaviour across years, independent of peak
15
15
population density (Figure 7A). The similarity in the quantitative response of nursing behaviour to
population density between years of high versus low density suggests a causal relationship. Nonetheless,
experimental density manipulations in the wild are needed to conclusively demonstrate the effect of
population density on maternal behaviour.
To disentangle the effects of population density and season, I attempted to experimentally increase the
perceived local density in my wild-caught female mice using olfactory cues. I predicted that exposure to
the scents of unknown individuals would increase perceived density and cause a reduction in nursing.
Furthermore, I predicted that the mechanism for this behavioural response to environmental cues would
be an increase in maternal CORT levels, similar to Dantzer et al. (2013). Surprisingly, despite notable
odour differences between the treatment and control bedding, there was no significant difference in
maternal behaviour or fecal CORT levels between the two groups. There are two possible explanations
for why the soiled bedding did not elicit the predicted response. First, it is possible that olfaction is not the
correct social cue modality. Despite the considerable evidence of conspecific scent decreasing
reproductive output and retarding sexual maturation in laboratory mice, including Peromyscus (Mus
musculus: Champlin, 1970; Drickamer, 1974; Ma et al., 1998; Peromyscus maniculatus bairdii: Lawton
and Whitsett, 1979; Lombardi and W hitsett, 1980; Creigh and Terman, 1988; Peromyscus leucopus:
Rogers and Beauchamp, 1976; Terman, 1984), the results from this study clearly demonstrate that scent
alone does not cause wild female deer mice to respond to local density conditions. Some studies suggest
that tactile communication may also be required to elicit the physiological responses demonstrated at
high population density in Peromyscus maniculatus bairdii (Terman, 1979, 1980; Lombardo and Terman,
1980). These laboratory studies show that individuals who remain in a colony without any tactile cues (i.e.
kept in a “no contact” box) do not show a decline in reproductive output at high population density
(Terman, 1979, 1980; Lombardo and Terman, 1980). Future studies are needed to isolate individual
and/or combinations of social cues i.e. olfactory, tactile, visual and/or auditory in order to identify what
cues are required to elicit a response in wild female deer mice as a result of fluctuations in population
density.
16
16
Alternatively, it is possible that despite my evidence for a causal relationship between population density
and maternal behaviour, maternal behaviour might also be related to another seasonally changing
variable. Previous studies have identified that population density is often confounded with food
availability in wild mice (Taitt et al., 1981; W olff, 1985). To address this possibility, I conducted a food
supplementation experiment during the 2013 breeding season in a subset of our population where one
section of our study site received additional food over the course of 12 weeks and another section did not
(Appendix 1). The food supplementation had no effect on density, thus females in the supplemented
section had access to more food resources than control females. Despite an increase in food resources,
there was no effect of this additional food on the proportion of time spent nursing (Appendix 1),
suggesting that food resources, which might otherwise be confounded with population density, do not
affect maternal behaviour in deer mice.
Another variable that changed seasonally was female mass at parturition (PND 0) and three days post
parturition (PND 3) where female mass decreased significantly over the breeding season (female weight
PND0: ß = - 6.51 ± 2.59, t28 = - 2.52, p = 0.018; PND3: ß = - 6.36 ± 3.04, t28 = - 2.09, p = 0.045). Female
mass could also explain a significant proportion of the variation in the frequency of nursing (female weight
PND0: ß = 0.024 ± 0.008, t28 = 2.87, p = 0.008; PND3: ß = 0.034 ± 0.009, t28 = 3.87, p = 0.0006) notably
however, maternal weight is not considered an accurate predictor of maternal condition in Peromyscus
(Millar and Teferi, 1992), but it is possible that body mass is related to time spent nursing. Finally,
laboratory density might also be a seasonally changing variable responsible for the decline in maternal
behaviour. As pregnant females were opportunistically brought into captivity, the number of females
2
-5
housed in the laboratory was also correlated with natural density (PND0: R = 0.44, p = 3.58x10 ; PND3:
2
R = 0.20, p = 0.008), and as such increased over the course of the summer (laboratory density PND0: ß
-5
= 4.4 ± 0.95, t 28= 4.61, p= 8.04x10 ; PND3: ß = 3.17 ± 1.07, t28 = 2.95, p = 0.006). The number of females
housed in the lab was a weak predictor of the proportion of time spent nursing (PND0: ß = - 0.009 ±
0.004, t28 = - 2.34, p = 0.027; PND3: ß = - 0.008 ± 0.004, t28 = - 1.94, p = 0.062). Future studies should
17
17
control for or manipulate housing density (number of individually housed mice in a single room) to test for
the possible effects on maternal behaviour.
Maternal behaviour, fecal corticosterone and population density
A positive relationship between population density and stress hormone levels has been documented in
the majority of vertebrate studies in both the laboratory and field (Newman et al., submitted) including the
one other field study to investigate the relationship between population density and CORT in wild
Peromyscus (Harper and Austad, 2004). This relationship has been attributed to the positive association
between population density and the number of predators, resource competition and negative social
interactions (Creel et al., 2012). Here, I report a U-shaped relationship between post-natal maternal
CORT in the laboratory and natural population density at the time the female was brought into the
laboratory; females captured under high and low density conditions experienced higher stress levels than
females under moderate density conditions. When population densities were greater than 12 deer mouse
captures per 100 trap nights there was a positive relationship between CORT and population density;
stress levels increased with increasing population density. Interestingly however, when population
densities were low, there was the negative relationship between population density and CORT. It is
known that many social interactions induce stress in wild female deer mice (Harper and Austad, 2004;
Creel et al., 2012) but it is possible that too few interactions also increase CORT. Laboratory studies
reveal that laboratory mice that are removed from a colonial environment and placed into isolation
experience higher stress levels than individuals who remain in the colony (Guidotti et al., 2001;
Matsumoto et al., 2007), however, the data are inconsistent and currently unresolved (Brian, 1975; Hunt
and Hambly, 2006; Arndt et al., 2009).
CORT was measured in fecal samples collected from females three days after parturition, but because
females were brought into the laboratory at various stages of pregnancy, pregnant mice were housed in
the laboratory anywhere between 1 to 20 days before giving birth (mean: 9.13 days). The CORT levels
measured in the fecal samples of wild deer mice are an integrated measure of the individual’s GC profile
18
18
4-6 hours prior to the sample being produced (Harper and Austad, 2000), therefore, our CORT
measurements are not likely to reflect a direct response to natural population density at the time of
capture. A more plausible explanation is that the U-shaped relationship between CORT and population
density may be explained by the early-life environment. Early breeders in 2013 are those females that
were born at the end of the 2012 breeding season, under high-density conditions, and over-wintered
(OW) to reproduce the following spring (Millar, 1979). Conversely, late breeders are probably young of
the year (YY), but it is also possible that some could have been long-lived OW females that were not
captured until late summer (Millar, 1979; Havelka and Millar, 2004). YY would have been born in early
spring 2013 under low-density conditions. Early-born Peromyscus mature and breed during the summer
of their birth and typically do not survive over the winter (Millar, 1979). In contrast to OW females, YY are
born at low density but tend to breed at high density. It is possible that natal density interacts with current
density resulting in distinct behavioural and physiological phenotypes and that the high CORT phenotype
of late summer carries over in the OW female’s physiology, but not their nursing behaviour.
There is considerable evidence for the long term consequences of early life programming on the HPA
axis; maternal behaviour experienced during early life is significantly correlated with the magnitude of a
stress induced CORT response as an adult (Liu et al., 1997; Caldji et al., 1998; Weaver et al., 2004;
Meaney and Szyf, 2005; Hellstrom et al., 2012). W ithin Rodentia, laboratory studies show that the
frequency of tactile stimulation through licking/grooming and nursing alters the methylation pattern in the
promoter region of the glucocorticoid receptor (GR) gene within the offspring (Liu et al., 1997; W eaver et
al., 2004; Hellstrom et al., 2012). The consequence of these epigenetic changes is that individuals who
received high maternal care, have greater GR expression, and thus have a rapidly responding HPA axis,
ultimately resulting in lower overall amounts of GC release (Liu et al., 1997; Caldji et al., 1998; W eaver et
al., 2004; Meaney and Szyf, 2005; Hellstrom et al., 2012). However, in laboratory studies, the
environment is intentionally kept constant, thus interpreting the link between hormone-behaviour
relationships in a natural context is difficult. I propose that variation in maternal CORT is indeed related to
the level of maternal care females received when they were pups, but CORT is not the mechanistic link
19
19
driving future maternal behaviour; rather maternal behaviour is related to the current and anticipated
environmental conditions rather than early-life programming of the HPA axis. My data support this
hypothesis as OW females born under high density conditions at the end of the previous season, who
received less maternal care and were programmed with a “high CORT” phenotype, exhibited high levels
of nursing to their own offspring born under low density conditions the following spring. Further, early
born YY, born under low density conditions, would have received higher maternal care and had a “low
CORT” phenotype, but as density increased over the breeding season, later born YY would have
received correspondingly less maternal care, accounting for the positive relationship between population
density and CORT later in the season. Interestingly this would also explain the positive relationship
reported in the one previous field study to investigate the relationship between CORT and population
density in Peromyscus as the study only took place over a portion of the breeding season (late July – mid
October) when the majority of the samples were most likely from YY.
The adaptive nature of these physiological and behavioural phenotypes is evident when I consider the
variation in environmental conditions experienced within and among individuals’ life-history. In the
beginning of the breeding season, under low-density conditions, early-born offspring may have greater
fitness with a “low CORT” phenotype, as there is little advantage of the more reactive “high CORT”
phenotype (i.e. increased cardiac function, attentional processes, consumptive behaviour; Bonier et al.,
2012; Boonstra et al., 2012). Conversely, offspring born at high population density, where they face an
increased number of both predators and competitors, may have increased fitness with a “high CORT”
reactive behavioural phenotype (Meaney, 2001; Bonier et al., 2012; Boonstra, 2012; Creel et al., 2012).
Matching physiological and behavioural phenotypes to a changing environment could be the driver
responsible for maintaining variation in maternal behaviour in a wild population. Future field experiments
should investigate how maternal behaviour shapes offspring HPA axis in wild mice and quantify the
fitness consequences of variation in the HPA axis.
20
20
5.0 Conclusion
In conclusion, in wild deer mice maternal behaviour is a plastic phenotype as the frequency of nursing
drastically decreased over the course of the breeding season as population density increased. Despite
the substantial laboratory evidence demonstrating mice responding to the scent of conspecifics, I have
shown that olfactory cues alone do not cause wild female deer mice to respond to local density
conditions. Future studies are needed to disentangle the effects of population density and seasonality on
maternal behaviour and to investigate what environmental cues deer mice use to assess local density
conditions.
Maternal CORT is not the mechanistic link between population density and maternal behaviour but rather
the stress phenotype of the female may be related to the level of maternal care the female received in the
early life environment as a pup. Alternatively, a female’s maternal behaviour is related to the current or
anticipated environmental conditions her offspring will experience. The relationships reported in this study
exemplify the ways in which wild mice are different from laboratory mice. W ild mice face this mismatch of
environments where the density in which they were born can be very different from the density in which
they breed. These complex relationships between natal and breeding environments might require
complex systems for appropriately programming offspring for the current environmental conditions.
This project provided a rare opportunity to incorporate proven laboratory techniques with a wild
population to demonstrate how females alter in maternal behaviour in response to a changing
environment, a phenomena not observed in a laboratory environment. Laboratory studies have identified
significant consequences of a reduced number of GRs, but little work has been done in the natural
environment. The results from this study suggest that prolonged exposure to elevated CORT might not
always result in a reduced fitness, as often assumed by the biomedical literature, but that programming
hyper-responsive offspring might be adaptive under high-density conditions. Future studies are needed to
determine how maternal behaviour affects HPA axis development in wild rodents and the fitness trade-off
between HPA axis sensitivity and behavioural response under various population density conditions.
21
21
6.0 Literature Cited
Arndt, S. S., Laarakker, M. C., van Lith, H. A., van der Staay, F. J., Gieling, E., Salomons, A. R., van't
Klooster, J., Ohl, F., 2009. Individual housing of mice—impact on behaviour and stress responses.
Physiology & Behavior, 97(3), 385–393.
Baker, S., Chebli, M., Rees, S., LeMarec, N., Godbout, R., Bielajew, C., 2008. Effects of gestational
stress: 1. Evaluation of maternal and juvenile offspring behavior. Brain Research, 1213, 98–110.
Bonier, F., Martin, P. R., Moore, I. T., Wingfield, J. C., 2009. Do baseline glucocorticoids predict fitness?
Trends in Ecology & Evolution, 24(11), 634–42.
Boonstra, R., 2012. Reality as the leading cause of stress: rethinking the impact of chronic stress in
nature. Functional Ecology, doi: 10.1111/1365-2435.12008.
Boutin, S., 1990. Food supplementation experiments with terrestrial vertebrates: patterns, problems and
the future. Canadian Journal of Zoology, 68(2), 203-220.
Brain, P., 1975. W hat does individual housing mean to a mouse? Life Sciences, 16(2), 187–200.
Brummelte, S., Galea, L. A., 2010. Chronic corticosterone during pregnancy and postpartum affects
maternal care, cell proliferation and depressive-like behavior in the dam. Hormones and Behavior, 58(5),
769–779.
Caldji, C., Tannenbaum, B., Sharma, S., Francis, D., Plotsky, P. M., Meaney, M. J., 1998. Maternal care
during infancy regulates the development of neural systems mediating the expression of fearfulness in
the rat. Proceedings of the National Academy of Sciences of the United States of America, 95(9), 5335–
40.
Champagne, F. A., Francis, D. D., Mar, A., Meaney, M. J., 2003. Variations in maternal care in the rat as
a mediating influence for the effects of environment on development. Physiology & Behavior, 79(3), 359–
71.
Champagne, F. A., Meaney, M. J., 2006. Stress during gestation alters postpartum maternal care and the
development of the offspring in a rodent model. Biological Psychiatry, 59(12), 1227–35.
Champagne, F. A., Curley, J. P., Keverne, E. B., Bateson, P. P. G., 2007. Natural variations in
postpartum maternal care in inbred and outbred mice. Physiology & Behavior, 91(2-3), 325–34.
Champlin, A., 1971. Suppression of oestrus in grouped mice: the effects of various densities and the
possible nature of the stimulus. Journal of Reproduction and Fertility, 27(2), 233–241.
Christian, J. J., 1950. The adreno-pituitary system and population cycles in mammals. Journal of
Mammalogy, 247–259.
22
22
Christian, J. J., 1956. Adrenal and reproductive responses to population size in mice from freely growing
populations. Ecology, 258–273.
Christian, J. J., 1961. Phenomena associated with population density. Proceedings of the National
Academy of Sciences of the United States of America, 47(4), 428.
Creel, S., Dantzer, B., Goymann, W., Rubenstein, D.R., 2012. The ecology of stress: effects of the social
environment. Functional Ecology, doi: 10.1111/j.1365-2435.2012.02029.x
Creigh, S. L., Terman, C. R., 1988. Reproductive Recovery of Inhibited Male Prairie Deermice
(Peromyscus maniculatus bairdii) from Laboratory Populations by Contact with Females or Their Urine.
American Society of Mammalogists, 69(3), 603–607.
Dantzer, B., Newman, A. E., Boonstra, R., Palme, R., Boutin, S., Humphries, M. M., McAdam, A. G.,
2013. Density triggers maternal hormones that increase adaptive offspring growth in a wild mammal.
Science, 340(6137), 1215–1217.
Do, R., Shonfield, J., McAdam, A. G., 2013. Reducing accidental shrew mortality associated with smallmammal livetrapping II: a field experiment with bait supplementation. Journal of Mammalogy, 94(4), 754–
760.
Drickamer, L. C., 1974. Sexual maturation of female house mice: social inhibition. Developmental
Psychobiology, 7(3), 257–265.
Drost, C. A., Fellers, G. M., 1991. Density Cycles in an Island Population of Deer Mice, Peromyscus
maniculatus. Nordic Society Oikos, 60(3), 351–364.
Falls, J. B., Falls, E. A., Fryxell, J. M., 2007. Fluctuations of deer mice in Ontario in relation to seed crops.
Ecological Monographs, 77(1), 19–32.
Fryxell, J. M., Falls, J. B., Falls, E. A., Brooks, R. J., 1998. Long-term dynamics of small-mammal
populations in Ontario. Ecology, 79(1), 213–225.
Guidotti, A., Dong, E., Matsumoto, K., Pinna, G., Rasmusson, A., Costa, E., 2001. The socially-isolated
mouse: a model to study the putative role of allopregnanolone and 5α-dihydroprogesterone in psychiatric
disorders. Brain Research Reviews, 37(1), 110–115.
Harper, J. M., Austad, S. N., 2000. Fecal glucocorticoids: a noninvasive method of measuring adrenal
activity in wild and captive rodents. Physiological and Biochemical Zoology, 73(1), 12–22.
Harper, J. M., Austad, S. N., 2001. Effect of Capture and Season on Fecal Glucocorticoid Levels in Deer
Mice (Peromyscus maniculatus) and Red-Backed Voles (Clethrionomys gapperi). General and
Comparative Endocrinology, 123(3), 337–344.
Harper, J.M., Austad, S.N. 2004. Fecal Corticosteroid Levels in Free-living Populations of Deer Mice
(Peromyscus maniculatus) and SouthernRed-backed Voles (Clethrionomys gapperi)., 2004. American
Midland Naturalist, 152,400-409.
23
23
Harris, B. N., de Jong, T. R., Yang, V., Saltzman, W., 2013. Chronic variable stress in fathers alters
paternal and social behavior but not pup development in the biparental California mouse (Peromyscus
californicus). Hormones and Behavior, 64(5), 799–811.
Havelka, M., Millar, J., 2004. Maternal age drives seasonal variation in litter size of Peromyscus leucopus.
Journal of Mammalogy, 85(5), 940–947.
Hellstrom, I. C., Dhir, S. K., Diorio, J. C., Meaney, M. J., 2012. Maternal licking regulates hippocampal
glucocorticoid receptor transcription through a thyroid hormone-serotonin-NGFI-A signalling cascade.
Philosophical Transactions of the Royal Society B: Biological Sciences, 367(1601), 2495–510.
Hunt, C., Hambly, C., 2006. Faecal corticosterone concentrations indicate that separately housed male
mice are not more stressed than group housed males. Physiology & Behavior, 87(3), 519–26.
Kaiser, S., Sachser, N., 2005. The effects of prenatal social stress on behaviour: mechanisms and
function. Neuroscience & Biobehavioral Reviews, 29(2), 283–294.
Kalcounis-Rueppell, M. C., Metheny, J. D., Vonhof, M. J., 2006. Production of ultrasonic vocalizations by
Peromyscus mice in the wild. Frontiers in Zoology, 3(3), 1–12.
Kirkland, G. L., Layne, J.N., 1989. Advances in the study of Peromyscus (Rodentia). Kirkland, G. L., and
J. N. Layne, eds. Texas Tech University Press, Texas.
Lawton, A. D., W hitsett, J. M., 1979. Inhibition of sexual maturation by a urinary pheromone in male
prairie deer mice. Hormones & Behavior, 13(2), 128–38.
Lidicker, W. Z., 1976. Social behaviour and density regulation in house mice living in large enclosures.
The Journal of Animal Ecology, 677–697.
Liu, D., Diorio, J., Tannenbaum, B., Caldji, C., Francis, D., Freedman, A., Sharma, S., Pearson, D.,
Plotsky. P.M., Meaney, M.J., 1997. Maternal care, hippocampal glucocorticoid receptors, and
hypothalamic-pituitary-adrenal responses to stress. Science, 277(5332), 1659–1662.
Lombardi, J. R., W hitsett, J. M., 1980. Effects of urine from conspecifics on sexual maturation in female
prairie deermice, Peromyscus maniculatus bairdii. Journal of Mammalogy, 61(4), 766–8.
Lombardo, D. L., Terman, C. R., 1980. The influence of the social environment on sexual maturation of
female deermice (Peromyscus maniculatus bairdi). Researches on Population Ecology, 22(1), 93–100.
Lusk, S. J., Millar, J. S., 1989. Reproductive inhibition in a short-season population of Peromyscus
maniculatus. The Journal of Animal Ecology, 329–341.
Ma, W., Miao, Z., Novotny, M. V., 1998. Role of the adrenal gland and adrenal-mediated chemosignals in
suppression of estrus in the house mouse: the lee-boot effect revisited. Biology of Reproduction, 59(6),
1317–1320.
24
24
Matsumoto, K., Puia, G., Dong, E., Pinna, G., 2007. GABAA receptor neurotransmission dysfunction in a
mouse model of social isolation-induced stress: possible insights into a non-serotonergic mechanism of
action of SSRIs in mood and anxiety disorders. The International Journal on the Biology of Stress, 10(1),
3–12.
Meaney, M. J., 2001. Maternal care, gene expression, and the transmission of individual differences in
stress reactivity across generations. Annual Review of Neuroscience, 24, 1161–92.
Meaney, M. J., Szyf, M., 2005. Environmental programming of stress responses through DNA
methylation: life at the interface between a dynamic environment and a fixed genome. Dialogues in
Clinical Neuroscience, 7(2), 103.
Millar, J. S., 1979. Energetics of lactation in Peromyscus maniculatus. Canadian Journal of Zoology,
57(5), 1015–1019.
Millar, J. S., 1982. Life cycle characteristics of northern Peromyscus maniculatus borealis. Canadian
Journal of Zoology, 60(4), 510–515.
Millar, J. S., 1985. Life cycle characteristics of Peromyscus maniculatus nebrascensis. Canadian Journal
of Zoology, 63(6), 1280–1284.
Millar, J. S., Teferi, T., 1993. W inter survival in northern Peromyscus maniculatus. Canadian Journal of
Zoology, 71(1), 125–129.
Modi, W .S., 1984. Reproductive tactics among deer mice of the genus Peromyscus. Canadian Journal of
Zoology, 62, 2576-81.
Montgomery, W., 1989., Population regulation in the wood mouse, Apodemus sylvaticus. II. Density
dependence in spatial distribution and reproduction. The Journal of Animal Ecology, 477–494.
Moore, C. L., Power, K. L., 1986. Prenatal stress affects mother-infant interaction in Norway rats.
Developmental Psychobiology, 19(3), 235–245.
Morris, D. W., 1989. Density-dependent habitat selection: testing the theory with fitness data.
Evolutionary Ecology, 3(1), 80–94.
Myers, M.M., Brunelli, S.A., Squire, J.M., Shindeldecker, R.D., Hofer, M.A., 1989. Maternal Behavior of
SHR Rats and Its Relationship to Offspring Blood Pressures. Developmental Psychobiology, 22(1), 29–
53.
Newman, A.M., Chin, E.H., Schmidt, K.L., Bond, L., W ynne-Edwards, K.E., Soma, K.K., (2008). Analysis
of steroids in songbird plasma and brain by coupling solid phase extraction to radioimmunoassay.
General and Comparative Endocrinology, 155(3), 503–510.
Nováková, M., Palme, R., Kutalová, H., Janský, L., Frynta, D., 2008. The effects of sex, age and
commensal way of life on levels of fecal glucocorticoid metabolites in spiny mice (Acomys cahirinus).
25
25
Physiology & Behavior, 95(1-2), 187–93.
Palme, R., 2005. Measuring fecal steroids: guidelines for practical application. Annals of the New York
Academy of Sciences, 1046(1), 75–80.
Patin, V., Lordi, B., Vincent, A., Thoumas, J., Vaudry, H., Caston, J., 2002. Effects of prenatal stress on
maternal behavior in the rat. Developmental Brain Research, 139(1), 1–8.
Pawluski, J. L., Brummelte, S., Barha, C. K., Crozier, T. M., Galea, L. A., 2009. Effects of steroid
hormones on neurogenesis in the hippocampus of the adult female rodent during the estrous cycle,
pregnancy, lactation and aging. Frontiers in Neuroendocrinology, 30(3), 343–357.
Pedersen, A. B., Greives, T. J., 2008. The interaction of parasites and resources cause crashes in a wild
mouse population. Journal of Animal Ecology, 77(2), 370–377.
Peng, X., Lang, C., Drozdowicz, C., Ohlsson-W ilhelm, B., 1989. Effect of cage population density on
plasma corticosterone and peripheral lymphocyte populations of laboratory mice. Laboratory Animals,
23(4), 302–306.
Pinhero, J., Bates, D., DeRoy, S., Sarkar, D.,. 2011. nlme: Linear and Nonlinear Mixed Effects Models. R
package version 3.1-102.
R Core Team., 2012. R: A language and environment for statistical computing. R Foundation for
Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org/.
Rogers Jr, J. G., Beauchamp, G. K., 1976. Influence of stimuli from populations of Peromyscus leucopus
on maturation of young. Journal of Mammalogy, 320–330.
Ryan, B. C., Vandenbergh, J. G., 2002. Intrauterine position effects. Neuroscience & Biobehavioral
Reviews, 26(6), 665–678.
Schug, M.D., Vessey, S.H., Korytko, A. I., 1991. Longevity and Survival in a Population of W hite-Footed
Mice (Peromyscus leucopus). Journal of Mammalogy, 72(2), 360–366.
Smith, J., Seckl, J., Evans, A., Costall, B., Smythe, J., 2004. Gestational stress induces post-partum
depression-like behaviour and alters maternal care in rats. Psychoneuroendocrinology, 29(2), 227–244.
Smith, S. M., Vale, W. W., 2006. The role of the hypothalamic-pituitary-adrenal axis in neuroendocrine
responses to stress. Dialogues in Clinical Neuroscience, 8(4), 383.
Stewart, F. E., McAdam, A. G., 2014. Seasonal plasticity of maternal behaviour in Peromyscus
maniculatus. Behaviour, doi: 10.1163/1568539X-00003211.
Taitt, M. J., 1981. The effect of extra food on small rodent populations: I. Deermice (Peromyscus
maniculatus). The Journal of Animal Ecology, 111–124.
26
26
Terman, C. R., 1965. A study of population growth and control exhibited in the laboratory by prairie
deermice. Ecology, 890–895.
Terman, C. R., 1968. Inhibition of reproductive maturation and function in laboratory populations of prairie
deermice: a test of pheromone influence. Ecology, 1169–1172.
Terman, C. R., 1969. Weights of selected organs of deer mice (Peromyscus maniculatus bairdii) from
asymptotic laboratory populations. Journal of Mammalogy, 50(2), 311.
Terman, C. R., 1979. Inhibition of reproductive development in laboratory populations of prairie deermice
(Peromyscus maniculatus bairdii): influence of tactile cues. Ecology, 455–458.
Terman, C. R., 1980. Social factors influencing delayed reproductive maturation in prairie deermice
(Peromyscus maniculatus bairdii) in laboratory populations. Journal of Mammalogy, 61(2), 219–223.
Terman, C. R., 1984. Sexual maturation of male and female white-footed mice (Peromyscus leucopus
noveboracensis): influence of physical or urine contact with adults. Journal of Mammalogy, 97–102.
Terman, C. R., 1993. Studies of natural populations of white-footed mice: reduction of reproduction at
varying densities. Journal of Mammalogy, 678–687.
Thomason, C., Hedrick-Hopper, T., Derting, T., 2013. Social and nutritional stressors: agents for altered
immune function in white-footed mice (Peromyscus leucopus). Canadian Journal of Zoology, 91(5), 313–
320.
Uller, T., 2008. Developmental plasticity and the evolution of parental effects. Trends in Ecology &
Evolution, 23(8), 432–438.
Wasser, S., Thomas, R., Nair, P., Guidry, C., Southers, J., Lucas, J., Wildt, D.E., Monfort, S., 1993.
Effects of dietary fibre on faecal steroid measurements in baboons (Papio cynocephalus cynocephalus).
Journal of Reproduction and Fertility, 97(2), 569–574
Weaver,I.C.G., Cervoni, N., Champagne, F.A., D'Alessio, A.C., Sharma, S., Seckl, N.A., Dymov,S., Szyf,
M., Meaney, M.J., 2004. Epigenetic programming by maternal behavior. Nature Neuroscience, 7(8), 847–
854.
Weinstock, M., 2001. Alterations induced by gestational stress in brain morphology and behaviour of the
offspring. Progress in Neurobiology, 65(5), 427–451.
Weinstock, M., 2008. The long-term behavioural consequences of prenatal stress. Neuroscience &
Biobehavioral Reviews, 32(6), 1073–1086.
Wolff, J. O., 1985. The effects of density, food, and interspecific interference on home range size in
Peromyscus leucopus and Peromyscus maniculatus. Canadian Journal of Zoology, 63(11), 2657–2662.
27
27
Wolff, J. O., 2003. Laboratory studies with rodents: facts or artifacts? Bioscience, 53(4), 421–427.
7.0 Tables
Table 1: Description of the behaviours and locations used to classify a female’s maternal care. Location
and behaviour were categorized every three minutes during four one-hour focals for the first three nights
postnatal. Behaviours were not mutually exclusive as a female could be performing multiple behaviours at
a time.
Location
Description
In Nest
Female was in physical contact with nesting material
Out of Nest
Female not in physical contact with nesting material
Behaviour
Description
Nursing
It is difficult to visualize the female nursing, so anytime the female was in contact with her pups
the behaviour was classified as nursing. Previous studies have found mother-pup contact and
nursing to be highly correlated behaviours (Champagne et al., 2007). In addition to this,
laboratory studies have determined that females are always active when in contact with their
pups (Myers et al., 1989; Champagne et al., 2003; Champagne et al., 2007).
Licking/Grooming
Manipulating Nest
Shavings
Female was licking/grooming pups.
The female was either retrieving nesting material or moving around existing nesting maternal
with her front paws or snout.
Manipulating Non
Nest shavings
The female was moving around material in the cage but not retrieving the material to the nest.
Self Grooming
Female was licking herself.
Eating
Female was eating either the rodent chow or the sunflower seeds that were provided in the cage.
Drinking
Female was drinking from the water bottle provided.
Smelling
Female was smelling the air or an object in the cage.
Rearing/Climbing
Female was on her hind legs or climbing and biting at metal bars on the lid of the cage.
Resting
Female did not appear to be doing anything. Only occurred when the female was outside of the
28
28
nest.
Carrying Pup
Female was carrying a pup either in her paws or mouth.
Out of Sight
Female was not visible.
8.0 Figures
Figure 1: The number of wild P. maniculatus captures significantly increased over the course of the 2013
breeding season (Julian Date). Mouse population density was measured as the total number of deer
29
29
mouse captures per 100 trap nights on nearby 100m transects within three habitat types at the time of
initial capture.
Figure 2: The proportion of time spent nursing by P.maniculatus females significantly declined with
increasing population density (measured as captures/100 trap nights; n=30). The population density
measurement for each female represents the local density the female was experiencing at the time she
was brought into the lab.
30
30
Figure 3: There was a significant positive quadratic relationship between corrected-CORT on PND3 and
population density at the time of the female’s capture. There are fewer females with corrected-CORT
values as not all females produced a fecal sample on PND3 (n=25).
31
31
Figure 4: Maternal CORT levels show an inconsistent response to population density. Early breeders
show no effect of population density on maternal CORT (closed circles; n=12) while late breeders show
the predicted positive relationship between population density and maternal CORT (open circles; n=13).
32
32
Figure 5: There was no significant difference in the proportion of time spent nursing between control
females (n=14) and treatment females (n=16) who were exposed to the soiled bedding of four unknown
individuals. Females were exposed to the bedding during their pregnancy up until PND3.
33
33
Figure 6: There was not a significant difference in corrected-CORT values on PND3 between control
females (n=13) and treatment females (n=12) who were exposed to the soiled bedding of four unknown
individuals.
34
34
Figure 7: The proportion of time spent nursing decreased over the course of the season in both 2013
(solid line) and 2011 (dotted line; A). In 2011 (dotted line) the proportion of time spent nursing declined
almost 50% from the beginning to the end of the breeding season compared to just over 27% in 2013
(solid line). In contrast, in both 2011 and 2013 there were remarkably consistent declines in nursing in
response to a seasonal increase in population density, despite overall densities being lower in 2013 than
2011 (B).
35
35
9.0 Appendix
Effect of food supplementation on maternal behaviour
Population density is often confounded with the availability of food resources (Taitt, 1981; Wolff, 1985), so
studies disentangling these two factors experimentally can identify the independent effects of each of
these important variables separately (Dantzer et al., 2013). I conducted a food supplementation
experiment during the 2013 breeding season as a second way to attempt to experimentally increase
population density in order to disentangle the effects of density and seasonality on maternal behaviour.
Starting the first week of June, 2013 I broadcasted a total of 26kg of sunflower seeds over 7.4 ha of a
20ha study grid weekly for a total of 12 weeks. A 7.8 ha control area was located adjacent to the foodst
supplemented grid with a buffer zone in between. Beginning May 1 , the grid was trapped biweekly with
Longworth live traps placed at 20m intervals on transects that were 20m apart. Traps were baited with
water soaked sunflower seeds, freeze-killed mealworms and cotton bedding for warmth. Traps were set
at dusk and check the next morning at dawn. Upon capture, deer mice were removed from the trap,
placed in a large handling bag and weighed with a Pesola spring balance to the nearest 0.5g. Deer mice
were tagged in both ears with metal ear tags for identification. Age of the individual was determined
based on dorsal pelage where juveniles are grey, sub-adults are a mixture of grey and brown and adults
are brown (Schug et al., 1991). Pregnant females (>22g or obvious pear shape) from control and food
supplemented populations were brought into the lab to measure their maternal behaviour. Prior to the
beginning of the food supplementation, females trapped on either side of the grid prior were designated
as control. After this, females were categorized based on their trapping location i.e. food supplementation
or control.
Previous field studies working with wild Peromyscus have demonstrated that population density will
increase with increased food availability (Taitt, 1981; Wolff, 1985), but I found no effect of the food
supplementation on the local population density around each female’s home range (ß= - 0.32 ± 1.32, t32 =
- 0.243, p= 0.81). As a result, females on the food-supplemented population had more per capita access
to food resources than females on the control population. The home range of a female was defined as the
36
36
eight traps surrounding the individual’s home trap (trap most frequently visited) three weeks prior to
coming into the lab. Density was measured as the number of deer mouse captures per 100 trap nights,
within those nine traps, over the three-week period. A density measurement was only calculated for
females who were captured on two or more occasions on the grid in order to correctly identify a home
trap.
Animals were cared for and maternal behaviour was measured using the same methods as those already
described above with the exceptions that females were not exposed to any unknown scent and fecal
samples were not collected.
There was no effect of this supplemental food on the frequency of nursing so maternal behaviour does
not appear to depend on food resources experienced in the field during gestation prior to coming into
captivity (F32 = 0.009, p = 0.93). There was also no significant difference between food supplementation
and control females for any of the variables measured (days in lab, lab density, number of pups, pup
weight, AGD, body length, female enter laboratory weight, parturition weight, PND3 weight: t35 = <1.66, p
= > 0.05)