SITE SELECTION FOR PARTURITION BY GRAVID FEMALE RATTLESNAKES
(CROTALUS OREGANUS) IN OSOYOOS, BRITISH COLUMBIA

by

Anna Skurikhina

A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE
OF
BACHELOR OF NATURAL RESOURCE SCIENCE (HONS.)

In the Department
Of
Natural Resource Sciences
At
Thompson Rivers University
Thesis Examining Committee
Dr. Karl Larsen, Thesis Supervisor and Professor, Dept. Natural Resource Science
Sheri Watson, Lecturer, Dept. Natural Resource Science
Dr. Lauchlan Fraser, Professor, Dept. Natural Resource Science
Dated the 18th day of April 2018, in Kamloops, British Columbia, Canada

Abstract
Thesis Supervisor: Professor Karl Larsen
Knowledge of habitat use and resource selection behaviours are critical to understanding
the population ecology of any animal species. In some species, genders and age classes
exhibit different preferences for foraging, reproduction, and shelter. The Western
Rattlesnake (Crotalus oreganus) is a good example of a species that shows variation in
habitat use between genders, mainly due to different reproductive life histories. My
research focussed on identifying and describing gestation sites referred to as rookery sites
used by gravid female rattlesnakes for parturition. Radiotelemetry was used to monitor
twelve gravid females from April to September 2017 in Osoyoos, British Columbia. I
assessed vegetation cover and other habitat features at the rookery sites across three
different spatial scales (1 m, 3 m, 10 m radius plots). To compare rookeries to random
habitat plots, I used a matched case-control study design. Conditional logistic regressions
revealed that at the 1 m scale, females selected sites with higher surface temperature,
whereas at the 3 m scale, they selected for higher air temperature and the sufficient cover
of rocks. At the largest scale, 10 m radius plot, average air temperature once again was
selected for by gravid females. Overall, rookery sites had 80% rock cover and were
found on south-facing slopes (on average 26% slope incline). Information on the location,
habitat structure and use of rookeries provides important insight into understanding the
ecology of Western Rattlesnakes and facilitating their conservation.

Keywords: rattlesnakes, rookery, telemetry, site selection, habitat analysis, conditional
logistic regression, Crotalus oreganus

ii

ACKNOWLEDGEMENTS
I would like to thank my primary supervisor Dr. Karl Larsen for his profound support,
guidance and mentorship throughout this project. Without his help, I would have never
had a chance to complete my first scientific investigation in a sound manner. I especially
thank him for finding time to meet with me and discuss my paper, educate me about
scientific method, providing me with connections and network opportunities and being
patient throughout the entire process. I would like to acknowledge members of my
examining committee, Sheri Watson and Dr. Lauchlan Fraser, for taking the time to
evaluate my work and provide valuable comments on my paper.
This project would not be possible without contributions and support of Dr. Christine
Bishop, a research scientist in the wildlife research division of the Science and
Technology Branch, Environment and Climate Change Canada. I thank the Osoyoos
Indian Band and the Nk’Mip Desert Cultural Centre for allowing me access their
protected sites to research the Western Rattlesnake population. Working at the research
lab in Nk’Mip Desert Cultural Centre made the research process much more valuable
because it gave me an opportunity to meet other snake enthusiasts and participate in
public outreach programs. Conversing with John Herbert helped me gain insight into the
location of little-known spots and activity areas for Western Rattlesnakes living within
Nk’Mip area.
I am grateful for Dana Eye who helped me with data collection, taught me snake handling
techniques and dealt with me and my accent throughout the summer. I also thank Jared
Maida and Stephanie Winton who provided me with support and background knowledge
on the ecology of Western Rattlesnakes and helped me with statistical analysis, data
representation and report writing. A special thank you to volunteers who assisted me with
habitat plots and temperature sampling in the field: H. Wasstrom, J. Barden, M. AubertinYoung, K. Mancuso, M. Bieber, V. Robinson, S. K. Pereira, and K. Legutky.
Field work was conducted under TRU Animal Ethics protocol #101546 , and research
permits from the federal (SAR-PYR-2015-0306) and provincial (MRPE15-171661 )
governments.

iii

Table of Contents

Introduction .................................................................................................................................... 7
Site Selection ................................................................................................................................ 7
Western Rattlesnakes Reproductive Ecology ............................................................................... 9
Population Status ......................................................................................................................... 9
Objectives................................................................................................................................... 10
Materials and Methods ................................................................................................................ 10
Study Site Description ................................................................................................................ 10
Snake Capture and Assessment .................................................................................................. 10
Telemetry ................................................................................................................................... 11
Rookery Identification & Habitat Measurements ...................................................................... 12
Statistical analysis ..................................................................................................................... 14
Results ........................................................................................................................................... 16
Snake Capture and Assessment .................................................................................................. 16
Location and Rookery Description ............................................................................................ 16
Temperature ............................................................................................................................... 20
Site Selection .............................................................................................................................. 20
Discussion ..................................................................................................................................... 24
Literature Cited ........................................................................................................................... 27
Appendix A ..................................................................................................................................... A
Appendix B ..................................................................................................................................... B

iv

List of Figures
Figure 1. Changes in body weight shown for gravid female Crotalus oreganus in
Osoyoos, 2017. Each female is depicted separately from one another (1-9). ....... 17
Figure 2. Location of nine identified rookery (⌾) and three den (△) sites within
bunchgrass zone in Osoyoos, British Columbia. Colour coding was used to show
location of rookery sites used by female rattlesnakes relative to the den sites from
which they emerged. Average distance traveled by a gravid female rattlesnake
from her den to her rookery site was 142.0 m (SE  41.4 m). ............................. 18
Figure 3. Four rookery sites () used by gravid Western Rattlesnakes for parturition in
August, Osoyoos, 2017. ........................................................................................ 19
Figure 4. Comparison between average air temperature in Osoyoos retrieved from the
Osoyoos Weather Station and the air temperature at the rookery sites and the
reference sites recorded with air loggers in 2017. The period of snake parturition
is outlined in red.................................................................................................... 21
Figure 5. Fluctuations in 24 hr maximum air, surface and cavity temperatures between
the rookery sites and reference sites in August, 2017. .......................................... 22
Figure 6. Estimated rock cover on rookery sites was significantly higher (80%) than on
reference sites (50%) at 3 m scale (depicting SE for each site). ........................... 23

Figure A 1. A prototype of a double cone used for air temperature loggers. Distance
between top and bottom cone edges is 3-5cm. An air logger was placed inside the
top cone secured in a plastic Ziploc bag. Cone design was adopted from literature
(Tarara and Hoheisel 2007). .................................................................................. A

v

List of Tables
Table 1. Summary of explanatory environmental variables assessed at three scales
(column 2) and identified significant variables (column 4) for individual scales.
See text for explanation of variables. Modelling of habitat variables was
conducted in COXREG (IBM SPSS Statistics). ................................................... 13

Table B 1. Summarized correlations for air, surface and cavity temperatures collected at
rookery sites and reference sites (Pearson’s correlation, n=27, Minitab). .............. B
Table B 2.Pairs of strongly correlated explanatory habitat variables measured at rookery
sites and associated controls (Pearson’s correlation, n=27, Minitab). .................... B

vi

Introduction
Site Selection
Patterns of habitat use as shown by animals is a key factor in understanding the basic
ecology of species. Heterogeneity of resource distribution, competition, environmental
pressures and disturbances on the land can influence aspects of niche selection
demonstrated by individuals in a population (Cornell and Lawton 2016). Site selection is
evident when animals choose a certain habitat type more often than expected by overall
availability (Koper and Manseau 2012). Available habitat includes all habitat types found
within an area that are used or not used by the species (Jones 2001). Selected habitat can
be characterized by specific vegetation cover, sun exposure, prey availability, and other
landscape features within available habitat (McLoughlin et al. 2002). Resource selection
is assessed within the existing home range of an animal, whereas site selection can be
interpreted as the selection of a home range in an area defined by the entire population
(Aebischer et al. 1993). Since habitat variables, structure, and resource availability differ
depending on the size of the assessed area, resource selection could vary based on spatial
scale (Jones 2001). For example, large spatial scale could refer to a patch of forest with a
heterogenous stand structure, whereas a small scale could incorporate an area within that
forest patch where individual tree features are assessed. Investigating site selection at
different spatial scales is important because it may demonstrate varying criteria used by
animals at different levels, such as broader habitat range or within a microsite (Jones
2001).
Selectivity of resources or sites is a tactic expressed by individuals to increase overall
fitness and survivability (McLoughlin et al. 2006). Differences in habitat use and
resource selection within one population are often observed due to the biological
characteristics of individuals such as age class, reproductive status or gender. For
example, studies show that habitat use can differ between juvenile and adult fish of the
same population due to the ontogenetic differences in energy needs for growth,
competition and predation pressures (Bay et al. 2001, Hofmann and Fischer 2002, King
2004). Habitat needs also may vary between reproductive females and males, for
7

example, in site selection for breeding and nesting in avian populations. Male birds tend
to select breeding sites that provide resources and opportunity for acquiring food, mates,
and appropriate acoustic environment for communication (Barg et al. 2006). On the other
hand, female birds often utilize habitat in close proximity to males to increase chances of
copulation (Mennill et al. 2004). In addition, sexual segregation can arise from higher
energetic costs for females associated with reproduction where energy demands increase
during lactation, gestation, and replenishment of body mass after giving birth. Depending
on reproductive mode, energy costs to embryo development and parturition would be
different. For example, egg-laying (oviparous) animals might have to allocate less energy
for reproduction due to external embryonic development of their young (Shine 2003).
However, this strategy poses higher risks to the development of the eggs, since they are
fully dependent on the environmental conditions at the point of oviposition (Shine 2003).
On the other hand, energetic costs of live-bearing (viviparous) animals that rely on the
internal embryonic development are much higher, and, therefore, may restrict activity of
gravid females and greatly exploit their energy reserves (Macartney and Gregory 1988,
Shine 2003). Viviparous strategy makes reproductive females more susceptible to
predation and extreme environmental conditions compared to males (Barten et al. 2001,
McLoughlin et al. 2002). These costs are counterbalanced in viviparous animals by
improved protection of embryos from unpredictable environment, predation, microbial
attacks and pathogens (Blackburn 1999).
Gravid, viviparous females require habitat that provides protection (refuge) from
predators, while at the same time affording opportunities to replenish energetic
expenditures (Barten et al. 2001, McLoughlin et al. 2002). Locations where animals give
birth are often referred to as ‘rookery’ sites (nest sites), particularly for birds and reptiles
(Sage and Vernon 1978, Graves and Duvall 1993a, Kasprzykowski 2003, Anderson et al.
2005). However, characteristics of rookery sites can vary between endothermic and
ectothermic animals due to their metabolic needs. Since ectotherms (e.g. reptiles) rely on
external sources of heat, rookery habitat intuitively should provide conditions for
efficient thermoregulation during embryonic development in reproductive females
(Johnson et al. 2016). Temperature, along with sufficient cover that provides shade and
refuge for snakes, are therefore likely two critical characteristics of snake rookeries
8

(Graves and Duvall 1995, Johnson et al. 2016). Environmental temperature is of
particular interest because gravid female snakes often exhibit higher body temperatures
compared to males and non-gravid females (Johnson et al. 2016). Overall, few studies
have assessed or tested for temperature and cover habitat characteristics predicted for
rookery sites used by snakes.
From a conservation standpoint, identifying selected sites such as rookeries is essential
they likely provide habitat critical to animal needs. If a wildlife species is at risk, these
selected habitat types can be protected and monitored, securing essential resources for
species persistence.
Western Rattlesnakes Reproductive Ecology
Western Rattlesnakes are only found within Thompson-Nicola and OkanaganSimilkameen regions of British Columbia (BC) (COSEWIC 2015). They are viviparous
(live-bearing) and have a biennial or triennial reproductive cycle (Macartney 1985).
Rattlesnakes become sexually mature at the age of seven or nine years old. Energy
reserves can pre-determine the success and timing of the consequent mating events
(Macartney and Gregory 1988). An average litter size ranges from two to eight neonates.
Gestation sites are typically known to be within a short distance from the hibernation
sites (Macartney 1985). Overall, limited information is available on reproduction of the
Western Rattlesnake in BC (Macartney and Gregory 1988). In general, the movements
and behaviour of males and non-gravid females have been studied more extensively than
reproductive females (Macartney 1985, Gomez 2007, Lomas 2009, Harvey 2015). This
historic bias has been possibly caused by the complicated life history of mature females
(Macartney 1985) and a desire to eliminate variants in movement patterns.
Population Status
The Western Rattlesnake (Crotalus oreganus) is listed as threatened under the Committee
on the Status of Endangered Wildlife in Canada (SARA registry 2015). Habitat loss,
anthropogenic mortality and destruction of their hibernacula are major contributors to
population decline of rattlesnakes in British Columbia (COSEWIC 2015). Slow

9

population replenishment from biennial/triennial reproductive cycles and high juvenile
mortality also challenge stability of the vulnerable populations (COSEWIC 2015).
Objectives
The objectives of my research were (1) to locate rookery sites utilized by gravid female
rattlesnakes, and (2) to evaluate habitat characteristics associated with these sites at three
spatial scales. In addition, gravid females were monitored over the summer to identify
behaviour associated with rookery use, changes in body condition during gestation and
post-parturition periods.
The goal of this study is to apply site selection modelling to habitat use by gravid
(pregnant) female Western Rattlesnakes in Osoyoos. Ultimately, identifying critical
habitat for rattlesnake parturition can be used to craft informed management strategies
that promote population stability and growth.

Materials and Methods
Study Site Description
My study population of Western Rattlesnakes was centered on the land of the Okanagan
Indian Band (OIB) in Osoyoos, British Columbia (49.0°N, 119.4°W). I collected data
over one field season from April to September, 2017. The study area lied within semiarid Okanagan-Similkameen grasslands with the vegetation community dominated by
bunchgrasses and dryland shrubs such as Artemisia tridentata (Big sagebrush) and
Purshia tridentata (Antelope brush) (Iverson et al. 2004). This region is characterized by
hot and dry summer climate with limited precipitation and a prolonged midsummer
drought (Iverson et al. 2004).
Snake Capture and Assessment
During spring emergence from denning (hibernation) sites, rattlesnakes were
opportunistically captured and processed, allowing me to identify reproductive (gravid)
females. These individuals could be recognized by a robust body condition and enlarged
ovarian follicles. A gentle ventral palpation of gravid female rattlesnakes allowed to
10

detect these follicles (Parker and Anderson 2007). Twelve gravid female rattlesnakes of
relatively large body mass were selected for this study from three different hibernation
sites. Each gravid female was uniquely tagged with a Passive Integrated Transponder
(PIT) to provide permanent identification (Gibbons and Andrews 2004).
Telemetry
All twelve females were surgically implanted with radio transmitters to allow telemetry
monitoring. Surgeries were performed by a professional veterinarian in an Osoyoos
veterinary clinic. Also recorded was an assessment of snake body condition measuring
weight and a snout-ventral length (SVL) for each individual following standard
methodology (Jenkins et al. 2009, Lomas 2009, Johnson et al. 2016).
After surgery, gravid females were kept in the lab for two days to ensure rehabilitation
after transmitter implantation. Females were released at their original point of capture and
each female was regularly tracked 3-4 times per week throughout the field season. To
inspect whether surgical incisions were healing properly, and to monitor changes in body
weight, we attempted brief recaptures of the focal females once each month. This
information was later used to compare changes in body mass over gestation and post
parturition time periods.
Each time a snake was located, I attempted to check for the presence of other snakes
within rookeries. After females gave birth, I also estimated the number of neonates from
each female based on neonatal shed counts and visual observations of neonates in
rookeries. In general, I tried to minimize disturbance to gravid females during telemetry
by remaining distant from the tracked snakes and spending minimum time at the
rookeries.
Snake handling methods duplicated those previously used on this site and were approved
by an animal ethics committee (Harvey 2015). During field work I was paired with a
partner to meet safety requirements.

11

Rookery Identification & Habitat Measurements
When long-distance movements ceased, and each gravid snake was observed sedentary at
a particular site for at least 10 days, I assumed that site was a rookery and established
habitat plots at each rookery site and two associated reference sites. I marked the centre
of each rookery by the rock under which a female gave birth. To place the centre of the
reference sites, I randomly generated two independent compass bearings and followed
each 40 m distance from the center of a rookery site. The distance between plot centers
was set to 40 m to avoid overlap between the largest plot size (scale). Using these
locations, I then established three plots of different radii (1 m, 3 m and 10 m radius) on
each site, allowing me to assess habitat characteristics at three spatial scales. Identical
parameters were measured within detection sites and reference sites at each scale.
Environmental variables and habitat characteristics that I recorded at each site were 1)
topographic, 2) vegetative and 3) thermal parameters (Column 1 in Table 1).
I recorded slope, elevation, aspect, and site location. Every time a natural underground
retreat was found on the site, I recorded substrate material and depth of the cavity using a
measuring tape. At reference sites, where underground cavities appeared absent, I
assumed 0 cm cavity depth.
Secondly, due to the observed association between snakes and plant cover in previous
studies, I estimated the relative cover (%) of grasses, shrubs, trees and rocks (Johnson et
al. 2016). Vegetation percentages were assessed in layers using a canopy cover
technique. Estimation of each vegetation layer was estimated based on the vertical
projection of its canopy. For example, cover of shrubs was assessed as the proportion of
ground floor covered by shrubs within a plot. I directly conducted visual estimation of
plant and rock cover at all scales in person to ensure consistency in measurements.
However, if other people with previous experience of cover assessments were present at
the site, I consulted them for a second opinion.
Finally, fluctuations in air, surface and cavity temperatures were monitored with
temperature dataloggers (iButtons model DS1922, accuracy +/-0.5 °C, Hindman et al.
2006). I placed 3 dataloggers in the center of each habitat plot (rookery and random) to

12

Table 1. Summary of explanatory environmental variables assessed at three scales (column 2) and identified significant variables
(column 4) for individual scales. See text for explanation of variables. Modelling of habitat variables was conducted in COXREG
(IBM SPSS Statistics).
Recorded Habitat Variable
Cavity depth, (m)

Scale
1 m radius

Putative Regression Variables
Cavity depth, (m)
Shrub cover, (%)
Average surface temperature, (C°)
Maximum surface temperature, (C°)
Maximum air temperature, (C°)
Minimum surface temperature, (C°)

Significant Variables
Minimum surface temperature, (C°)
P=0.022

3 m radius

Cavity depth, (m)
Shrub cover, (%)
Maximum surface temperature, (C°)
Minimum surface temperature, (C°)
Average air temperature, (C°)

Average air temperature, (C°)
P=0.020

3 m radius

Rock cover, (%)
Maximum cavity temperature, (C°)
Minimum air temperature, (C°)
Average cavity temperature, (C°)

Rock cover, (%)
P=0.025

10 m radius

Grass cover, (%)
Maximum cavity temperature, (C°)
Average air temperature, (C°)
Average cavity temperature, (C°)
Minimum air temperature, (C°)

Average air temperature, (C°)
P=0.020

Grass cover, (%)
Shrub cover, (%)
Tree cover, (%)
Rock cover, (%)
Minimum air temperature, (C°)
Average air temperature, (C°)
Maximum air temperature, (C°)
Minimum surface temperature, (C°)
Average surface temperature, (C°)
Maximum surface temperature, (C°)
Minimum cavity temperature, (C°)
Average cavity temperature, (C°)
Maximum cavity temperature, (C°)

13

record cavity, air and surface temperatures. Only 3 dataloggers were used per plot due to
constraints with the availability of iButtons. All temperatures were recorded every 40 min
throughout a three-month period (mid June-mid September). To protect the air and
surface dataloggers from UV radiation and potential overestimation of temperature value,
I placed each of them in a Ziploc bag with a reflective foil cover and punctured holes for
passive ventilation. For air temperature dataloggers, I made double cone shields out of
reflective material to ensure little sun penetration and allow for maximum exposure of a
datalogger to air (Tarara and Hoheisel 2007). A Ziploc pouch with an air logger was
taped inside the top cone. Two cones were attached to each other with four plastic cable
ties allowing 3-5 cm distance between cone edges (Appendix A). Air loggers were tied to
shrub branches 1 m above ground in a partial shade. Each surface datalogger was
attached to the ground using metal stakes, or it was duct-taped on the rock surface, if
rocks were found in the centre of a plot. Cavity loggers were placed inside a single
underground hole located within 0.5 m of a plot center. I selected potential natural
cavities based on their size making sure that a snake could fit inside and that it was
shaded from sun. This was primarily done for reference sites where snakes using cavities
were not observed. If natural cavities were absent on the reference sites, I placed cavity
loggers at the stem base of dense shrubs ensuring no direct sun exposure.
As parturition occurred in August for all snakes (see Results), I used a subset of average
daily temperatures to represent air, surface, and cavity conditions. To compare extremes
in occurred temperatures within rookeries and references sites, I calculated 24 hr
maximum and minimum temperatures for air, surface and cavities that were then
averaged between the sites for August (31 days). This allowed me to compare an average
maximum temperature measured in rookery sites and in reference sites in time period
from the 1st to 31st of August. I repeated the same calculation method to compare
minimum temperatures and mean temperatures between the sites.
Statistical analysis
Modelling of resource selection function is frequently generated with a binary logistic
regression (McLoughlin et al. 2006). In study designs that have one main and several

14

associated reference sites, conditional (paired) logistic regression is used (Duchesne et al.
2010, Chan 2011).
Conditional logistic regression (COXREG) is most commonly associated with drug
resistance studies, hazard and survival analysis in medicine (Chan 2011, Koletsi and
Pandis 2017). However, COXREG was adopted in ecology for multivariate datasets
addressing habitat selection based on the presence or absence of the studied event or
animal and an associated set of predictor variables recorded for each detection site paired
with multiple reference sites (Farmer et al. 2006, Thaker et al. 2011, Maag 2017).
Differences between rookery and reference sites were analyzed using conditional logistic
regressions following a COXREG procedure in IBM SPSS Statistics (version 24.0)(Chan
2011, IBM Software Group 2013). I selected forward conditional data entry for
modelling with 0.5 entry and 0.1 removal stepwise probability. Associations between
individual rookery plots and their reference plots were included in the models through an
additional numeric variable (i.e., all three plots shared the same identifying number). This
technique is standard for analysis of paired sampling designs (Chan 2011, IBM Software
Group 2013). To avoid feeding autocorrelated variables into the regression analyses, I
tested a priori for correlation between all environmental variables (Pearson correlation,
Minitab version 18). The strength of the correlation between vegetation cover, rock cover
and cavity depth varied based on plot (scale) size, therefore, I combined these variables in
different sets for each scale ensuring no significantly-correlated variables occurred in
each set. Firstly, I listed variables that were correlated (Appendix B) and then excluded
pairs of correlated variables within one scale from the list of all habitat measurements.
Ultimately, only uncorrelated sets of variables (Column 3 Table 1) were used for logistic
regression modelling at each scale.

15

Results
Snake Capture and Assessment
Gravid female rattlesnakes were captured at their hibernation sites between the 15th and
the 25th of April 2017. Average SVL of the studied females was 60.9 cm (SE 1.3cm,
n=12), and average weight of captured snakes immediately after emergence from
hibernacula was 222.5 g (SE 64.2 g, n=12). Overall trend shows a slight (12%) increase
in body weight a month after emergence (May), (Figure 1). Post parturition, females’
weight dropped by an average of 46% (from 224.3 g to 126.3 g) relative to their weight in
April (Figure 1).
Tracked female rattlesnakes settled down at their rookery sites in mid-June and showed
little movement within the rookery area during the rest of summer. From the 16th to the
31st of August, nine out of twelve telemetered females successfully gave birth with litter
size ranging from 2 and 5 neonates per snake. I could not identify whether the remaining
3 females gave birth or aborted as no neonates or any products of abortion were found.
These females left the sites that they occupied during mid-Summer and moved to other
areas presumably to feed. No signs of body abnormality, sickness or injury were detected
in these females that could suggest failed reproduction.
Females that gave birth spent on average a minimum of 5 days in the proximity of their
newborn neonates, and then dispersed downslope (August 23-August 31). Neonate sheds
were found 2-3 days prior to when their mothers left rookery sites.
Location and Rookery Description
Nine rookery sites in total were located on the south and southwestern slopes of my study
area (Figure 2). The average elevation for all nine sites was 447 m (SE 20 m) with 26%
slope incline (SE 3%, n=9). The mean distance between where females hibernated
(denned) and their rookery sites was 142.0 m (range 5.0 m to 384.9 m). On average,
predominant vegetation cover on the rookery sites was grasses (49%), and shrubs (8%)
(Figure 3). Tree cover was minimal (  3%). Rookeries were represented by rock piles
covering underground cavities. Depth inside these cavities ranged between 6.2 cm and
16

350
1
4
7

Weight, g

300

2
5
8

3
6
9

250

200

150

100
April (den capture)

May

June

August (post
parturition)

September (captured
for transmitter
removal)

Figure 1. Changes in body weight shown for gravid female Crotalus oreganus in Osoyoos, 2017. Each female is depicted separately
from one another (1-9).

17

.

Figure 2. Location of nine identified rookery (⌾) and three den (△) sites within bunchgrass zone in Osoyoos, British Columbia.
Colour coding was used to show location of rookery sites used by female rattlesnakes relative to the den sites from which they
emerged. Average distance traveled by a gravid female rattlesnake from her den to her rookery site was 142.0 m (SE  41.4 m).
18

Figure 3. Four rookery sites () used by gravid Western Rattlesnakes for parturition in August, Osoyoos, 2017.
19

227.2 cm ( 𝑥̅ =108.0 cm, n=9). Substrate material at all rookery sites was sand with a
small amount of plant debris.
Temperature
Fluctuations in air temperature during the summer revealed a slight decrease in ambient
temperature during the parturition period (Figure 4). Fluctuations in cavity temperature
had the lowest amplitudes compared to air temperature (Figure 5). Overall trend suggests
that daily average maximum air and surface temperatures in rookery sites were higher
than in reference sites (Figure 5). On the other hand, maximum cavity temperature
exhibited in rookery sites tended to be lower than recorded in reference sites (Figure 5).
Site Selection
Since the majority of putative explanatory habitat variables were correlated with one
another (Appendix A), four combination sets of uncorrelated variables were used for the
regression analysis (Column 3 in Table 1). Three variables associated with the location of
a rookery site were identified based on the spatial scale.
At the 1 m scale, minimum surface temperature was strongly associated with the presence
of a rookery site (P=0.022, n=27). At the 3 m scale, average air temperature and rock
cover (Figure 6) showed a strong predictive relationship with the presence of rookeries
(P=0.020 and 0.025 respectively, n=27). On the largest scale (10 m radius), the presence
of a rookery site was positively associated with average air temperature (P=0.020, n=27).
None of the vegetation cover classes (grasses, shrubs, trees) were associated with the
rookery sites.

20

33
Osoyoos

31

Rookery

Temperature, °C

29

Reference

27
25
23
21
19
17

Figure 4. Comparison between average air temperature in Osoyoos retrieved from the Osoyoos Weather Station and the air
temperature at the rookery sites and the reference sites recorded with air loggers in 2017. The period of snake parturition is outlined in
red.
21

60

55

Air rookery max

Surface rookery max

Cavity rookery max

Air reference max

Surface reference max

Cavity reference max

Temperature, ̊C

50

45

40

35

30

25
0

5

10

15

20

25

30

August, 2017
Figure 5. Fluctuations in 24 hr maximum air, surface and cavity temperatures between the rookery sites and reference sites in August,
2017.
22

35

Figure 6. Estimated rock cover on rookery sites was significantly higher (80%) than on
reference sites (50%) at 3 m scale (depicting SE for each site).

23

Discussion
This is the first study that monitored and evaluated habitat characteristics of the rookery
sites used for parturition by Western Rattlesnakes in British Columbia. Similar habitat
use studies were conducted on other snake species in North America, however, focus was
made more on males and non-gravid females (Graves and Duvall 1993, Parker and
Anderson 2007, Jenkins et al. 2009, Maag 2017). In general, little information is
available on activity, behaviour and habitat use for gestation and parturition by gravid
Crotalus oreganus. My research provides important information on timing, reproduction
and habitat use by gravid rattlesnakes.
Observed timing of parturition and the behaviour of neonate attendance by a mother were
consistent with other rattlesnakes populations (Parker and Anderson 2007, Maag 2017).
None of the observed snakes shared the same rookery site which contradicts existing
literature that identified communal rookeries (Graves and Duvall 1993a, Parker and
Anderson 2007). For those three females that failed to reproduce, no abnormal behaviour,
body condition or external factors were observed that could have negatively impacted
their reproduction. Other studies have reported that reproductive failure and follicle
resorption can occur in the wild rattlesnake populations due to environmental pressures
(Graves and Duvall 1993, Parker and Anderson 2007).
In the second part of my study, I evaluated located rookery sites and identified
characteristic habitat variables at three spatial scales. Estimated cover of vegetation was
not directly associated with the rookery sites. When rookery and reference sites were
compared, air and surface temperatures and the cover of rocks showed stronger
association with the rookeries than other recorded habitat features.
On the smallest scale (1 m surrounding the rookery), higher minimum surface
temperature had a strong association with the location of a rookery site. This could be
that females preferentially seek out relatively warm sites to ensure the thermal
requirement for gestation (Maag 2017). Since gravid females need to maintain higher
body temperatures during gestation (Johnson et al. 2016, Maag 2017), it is likely that
they would select for microsites that exhibit higher surface temperatures.
24

On a medium scale (3 m around a rookery), cover of rocks was relatively high and was a
significant predictor of rookery sites. Since gravid females are more vulnerable to
predation due to the decreased activity and mobility (Johnson et al. 2016), refugia (the
cover of rocks) may be vital for their survival. Rock cover of 80% may provide
protection from predators and create cooler environment within underground cavity
(Gomez 2007). Cover of rocks also can play a role in structural diversity on the landscape
providing refuge for small mammal prey (Tews et al. 2004). Other studies suggest that
prey availability can predetermine reproductive success in females; therefore, it could be
beneficial for gravid females to stay within high prey density areas (Jenkins et al. 2009).
At present we lack information on prey abundance around rookeries, much less the
predilection of gravid females to opportunistically feed, so this theory is primarily
conjecture at this point. I recommend further investigations into rookery characteristics
by conducting small mammal trappings in the immediate area of a rookery site, to
determine whether the abundance of prey may influence rookery site selection.
On the large scale (10 m), average air temperature was the only variable positively
associated with rookery sites. Higher air temperatures were reached on the rookery sites,
likely through higher solar radiation that would be an aspect and cover-dependent (Weiss
et al. 1988). If rookery sites are positioned on south-facing slope (true for all nine
rookeries assessed herein) this site would be exposed directly to solar radiation in the
Northern hemisphere. As stated earlier, the more heat is available, the more likely gravid
females would meet their energetic demands (Johnson et al. 2016). However, it is hard to
draw conclusions from these results: no replication of temperature was done throughout
the plot which likely lead to an underrepresentation of temperature of the entire plot area.
This issue could be resolved by sampling systematically air, surface and cavity
temperatures at several random locations within each 10m plot. Similar method could be
also used for a 3m plot to ensure more accurate estimation of temperatures at that scale.
All told, my analysis reveals an intuitive shift in characteristics of rookeries according to
scales. Gravid females seem to select their environment to meet their energetic demands,
hide from predators, and to have an immediate access to sun radiation for
thermoregulation. Such information asks interesting questions on how reproductive
25

females find rookeries with such characteristics in the first place and whether these
rookeries are regularly used for reproduction by other snakes in an area. Ongoing
research that currently takes place in the Osoyoos study site may provide answers to these
questions and address the idea of rookery fidelity by females.
Understanding the location of rookery sites and their associated habitat characteristics
likely will provide an important tool for managing species habitat. Because gestation
habitat plays a major role in providing sites for females’ protection, embryo development,
and parturition, this habitat can be considered integral for the recovery of Western
Rattlesnake population (Southern Interior Reptile and Amphibian Working Group 2016).
Since the majority of rookery sites lied within 142 m downslope from hibernacula, a
buffer zone of 100 m-150 m radius around den sites could incorporate these rookeries.
However, this recommendation for rattlesnake conservation cannot be extrapolated to
other sites without parallel information being collected. Longer-term data at my study site
also is warranted, and is in progress (Eye et al, unpubl.). If similar patterns are observed,
conservation initiatives might be improved by establishing these buffer zones and
minimizing human disturbance around hibernation and rookery areas.
This pilot study is an important first step in providing insight into this crucial but
unstudied aspect of Western Rattlesnake life history. By understanding habitat use and
site selection by gravid female rattlesnakes for parturition, sound management strategies
could be developed to locate and protect their sensitive rookeries. I hope my research
enables future work to be continued to aid in the conservation of this iconic and
fascinating grassland species.

26

Literature Cited
Aebischer, N. J., P. A. Robertson, and R. E. Kenward. 1993. Compositional analysis of
habitat use from animal radio-tracking data. Ecology 74:1313–1325.
Anderson, D., J. Brandt, L. Wright, and D. Davidson. 2005. Ecology and floristics of
Knife Island, a Gull and Cormorant rookery on Lake Superior, near Two Harbors,
Lake County, Minnesota. The Michigan Botanist 44:95–104.
Barg, J. J., D. M. Aiama, J. Jones, and R. J. Robertson. 2006. Within-territory habitat use
and microhabitat selection by male Cerulean Warblers (Dendroica Cerulea). The
Auk 123:795.
Barten, N. L., R. T. Bowyer, and K. J. Jenkins. 2001. Habitat use by female caribou:
tradeoffs associated with parturition. Habitat 65:77–92.
Bay, L. K., G. P. Jones, and M. I. McCormick. 2001. Habitat selection and aggression as
determinants of spatial segregation among damselfish on a coral reef. Coral Reefs
20:289–298.
Blackburn D. 1999. Viviparity and oviparity: evolution and reproductive strategies.
Encyclopedia of Reproduction 4:994-1003.
Chan, Y. H. 2011. Multinomial logistic regression. Singapore Medical Journal 46:259–
269.
Cornell, A. H. V, and J. H. Lawton. 2016. Species interactions, local and regional
processes, and limits to the richness of ecological communities: a theoretical
perspective. Journal of Animal Ecology 61:1–12.
COSEWIC. 2015. COSEWIC assessment and status report on the Western Rattlesnake
Crotalus oreganus in Canada.
Duchesne, T., D. Fortin, and N. Courbin. 2010. Mixed conditional logistic regression for
habitat selection studies. Journal of Animal Ecology 79:548–555.
Farmer, C. J., D. K. Person, and R. T. Bowyer. 2006. Risk factors and mortality of blacktailed deer in a managed forest landscape. Journal of Wildlife Managment 70:1403–
27

1415.
Gibbons, J. W., and K. M. Andrews. 2004. PIT Tagging: simple technology at its best.
BioScience 54:447–454.
Gomez, L. M. 2007. Habitat use and movement patterns of the Northern Pacific
Rattlesnake (Crotalus o. oreganus) in British Columbia. Thesis, University of
Victoria, Victoria, BC, Canada.
Graves, B. M., and D. Duvall. 1993. Reproduction, rookery use, and thermoregulation in
free-ranging, pregnant Crotalus v . viridis. Journal of Herpetology 27:33–41.
Graves, B. M., and D. Duvall. 1995. Aggregation of squamate reptiles associated with
gestation, oviposition, and parturition. Herpetological Monographs 9:102-119.
Harvey, J. A. 2015. Thermal influences on summer habitat use by Western Rattlesnakes
(Crotalus oreganus) in British Columbia. Thesis, Thompson Rivers University,
Kamloops, BC, Canada.
Hindman, B. B., S. Member, and T. Staff. 2006. What Is an iButton Device ? Maxim
Integrated: 1–5.
Hofmann, N., and P. Fischer. 2002. Temperature preferences and critical thermal limits
of burbot: implications for habitat selection and ontogenetic habitat shift.
Transactions of the American Fisheries Society 131:1164–1172.
IBM Software Group. 2013. IBM SPSS Regression 22. Chicago.
Iverson, K., M. Hames, A. Graham, and C. Cadrin. 2004. Grasslands of the Southern
Interior. Victoria.
Jenkins, C. L., C. R. Peterson, S. C. Doering, and V. A. Cobb. 2009. Microgeographic
variation in reproductive characteristics among Western Rattlesnake (Crotalus
oreganus) populations. Copeia 2009(4):774–780.
Johnson, B. D., J. P. Gibbs, T. A. Bell, and K. T. Shoemaker. 2016. Manipulation of
basking sites for endangered Eastern Massasauga Rattlesnakes. Journal of Wildlife
Management 80:803–811.
28

Jones, J. 2001. Habitat selection studies in avian ecology: a critical review. The Auk
118(2):557-562.
Kasprzykowski, Z. 2003. Habitat preferences of foraging Rooks Corvus frugilegus during
the breeding period in the agricultural landscape of eastern Poland. Acta
Ornithologica 38(1):27–31.
King, A. J. 2004. Ontogenetic patterns of habitat use by fishes within the main channel of
an Australian floodplain river. Journal of Fish Biology 65:1582–1603.
Koletsi, D., and N. Pandis. 2017. Conditional logistic regression. American Journal of
Orthodontics and Dentofacial Orthopedics 151:1191–1192.
Koper, N., and M. Manseau. 2012. A guide to developing resource selection functions
from telemetry data using generalized estimating equations and generalized linear
mixed models. Rangifer 20:195-203
Lomas, E. V. 2009. Effects of disturbance on the Northern Pacific Rattlesnake (Crotalus
oreganus oreganus) in British Columbia. Thesis, Thompson Rivers University,
Kamloops, BC, Canada.
Maag, D. W. 2017. The spatial ecology and microhabitat selection of the Pygmy
Rattlesnake ( Sistrurus miliarius ) in Southwestern Missouri. Thesis, Missouri State
University, Springfield, MO, USA.
Macartney, J. M. 1985. The ecology of the Northern Pacific Rattlesnake Crotalus viridis
oreganus in British Columbia. Thesis, University of Victoria, Victoria, BC, Canada.
Macartney, J. M., and P. T. Gregory. 1988. Reproductive Biology of Female Rattlesnakes
(Crotalus viridis) in British Columbia. Copeia 1:47–57.
McLoughlin, P. D., M. S. Boyce, T. Coulson, and T. Clutton-Brock. 2006. Lifetime
reproductive success and density-dependent, multi-variable resource selection.
Proceedings of the Royal Society B. Biological Sciences 273:1449–1454.
McLoughlin, P. D., R. L. Case, R. J. Gau, H. D. Cluff, R. Mulders, and F. Messier. 2002.
Hierarchical habitat selection by barren-ground grizzly bears in the central Canadian
29

Arctic. Oecologia 132:102–108.
Mennill, D. J., S. M. Ramsay, P. T. Boag, and L. M. Ratcliffe. 2004. Patterns of extrapair
mating in relation to male dominance status and female nest placement in blackcapped chickadees. Behavioral Ecology 15:757–765.
Parker, J. M., and S. H. Anderson. 2007. Ecology and behavior of the Midget Faded
Rattlesnake (Crotalus oreganus concolor) in Wyoming. Journal of Herpetology
41:41–51.
Sage, B. L., and J. D. R. Vernon. 1978. The 1975 national survey of rookeries. Bird
Study 25:64–86.
SARA registry. 2015. Species profile Western Rattlesnake. Accessed on April 9, 2018.
Available from: http://www.registrelepsararegistry.gc.ca/default.asp?lang=En&n=24F7211B-1
Shine, R. 2003. Reproductive strategies in snakes. Proceedings of the Royal Society B.
Biological Sciences 270:995–1004.
Southern Interior Reptile and Amphibian Working Group. 2016. Recovery plan for the
Western Rattlesnake (Crotalus oreganus) in British Columbia. Victoria, BC,
Canada.
Tarara, J. M., and G. A. Hoheisel. 2007. Low-cost shielding to minimize radiation errors
of temperature sensors in the field. HortScience 42:1372–1379.
Tews, J., U. Brose, V. Grimm, K. Tielbörger, M. C. Wichmann, M. Schwager, and F.
Jeltsch. 2004. Animal species diversity driven by habitat heterogeneity/diversity: the
importance of keystone structures. Journal of Biogeography 31:79–92.
Thaker, M., A. T. Vanak, C. R. Owen, M. В. Ogden, S. M. Niemann, and R. Slotow.
2011. Minimizing predation risk in a landscape of multiple predators: effects on the
spatial distribution of African ungulates. Ecology 92:398–407.
Weiss, S. B., D. D. Murphy, and R. R. White. 1988. Sun, slope, and butterflies:
Topopgraphic determinants of habitat quality for Euphydrias editha. Ecology
30

69:1486–1496.

31

Appendix A
Image of a protective shield for dataloggers used to measure air
temperature in habitat plots

Figure A 1. A prototype of a double cone used for air temperature loggers. Distance
between top and bottom cone edges is 3-5cm. An air logger was placed inside the top
cone secured in a plastic Ziploc bag. Cone design was adopted from literature (Tarara and
Hoheisel 2007).

A

Appendix B
Results of Pearson’s correlation analysis for environmental variables

Table B 1. Summarized correlations for air, surface and cavity temperatures collected at
rookery sites and reference sites (Pearson’s correlation, n=27, Minitab).
Temperature 1

Temperature 2

Average air
Average air
Average surface
Average surface
Average surface
Average surface
Average surface
Average cavity
Average cavity
Average cavity
Average cavity
Maximum air
Maximum surface
Maximum cavity
Minimum air
Minimum surface

Average surface
Maximum air
Average cavity
Maximum air
Maximum surface
Min surface
Minimum cavity
Maximum air
Maximum surface
Minimum surface
Minimum cavity
Maximum surface
Maximum cavity
Minimum cavity
Minimum surface
Minimum cavity

Correlation
coefficient (r-value)
0.482
0.664
0.837
0.477
0.750
0.398
0.374
0.375
0.696
0.365
0.430
0.425
0.369
0.673
0.642
0.552

Significance of
correlation (p-value)
0.013
0.000
0.000
0.014
0.000
0.044
0.060
0.059
0.000
0.067
0.028
0.030
0.064
0.000
0.000
0.003

Table B 2.Pairs of strongly correlated explanatory habitat variables measured at rookery
sites and associated controls (Pearson’s correlation, n=27, Minitab).
Scale

Variable 1

Variable 2

1m

Cavity depth
Shrub cover
Shrub cover
Grass cover
Shrub cover
Shrub cover
Grass cover
Grass cover

Rock cover
Grass cover
Rock cover
Rock cover
Grass cover
Rock cover
Rock cover
Rock cover

3m

10 m

B

Correlation
coefficient
(r-value)
0.529
0.475
0.615
0.546
0.443
0.588
0.376
0.608

Significance of
correlation
(p-value)
0.005
0.012
0.001
0.003
0.021
0.001
0.053
0.001