Content uploaded by Bruce Mclellan
Author content
All content in this area was uploaded by Bruce Mclellan
Content may be subject to copyright.
Sexually selected infanticide in grizzly bears:
the effects of hunting on cub survival
Bruce N. McLellan1
British Columbia Ministry
of Forests, RPO 3, PO Box 9158, Revelstoke, BC VOE
3KO, Canada
Abstract: Sexually selected infanticide
(SSI) has been documented
in some species with a mating
system in which males have almost exclusive breeding rights with 1 or more females. When the
dominant
male is removed, the new male kills the offspring sired by the previous
male to enable the
mother to be bred earlier.
It has been suggested
that
this immigrant
male hypothesis
of SSI operates
in
grizzly bears (Ursus arctos) and that removing dominant males by hunting results in high cub
mortality due to killing by immigrant
males, or in low reproductive
rates because of a female
counterstrategy
of using suboptimal
habitat
to avoid potentially
infanticidal
immigrant
males. I tested
2 predictions
of the immigrant
male hypothesis
in a hunted
area
adjacent
to protected
areas
with high
densities of grizzly bears that could supply immigrant
males. These predictions
were not supported.
Over 25 years, 134 grizzly bears
were captured:
most of the 77 male and 57 female grizzly bears
were
<3 years of age when first captured
(54.5% and 52.6%, respectively), and 22.1% of the males and
19.3%
of the females were 4-6 years
of age when first
captured.
Similarities of these sex ratios
suggest
that there was not a substantially greater
influx of subadult
males than females into the hunted
area.
Cub survival
to the end of the breeding season was high (0.93 or 0.95; n = 87), as was annual
cub
survival (0.85, n = 81); 15% of the 39 litters monitored
for an entire year were completely lost.
Yearling
survival
was 0.95 to the end of the breeding
season, when SSI should
cease. These results
do
not support
the immigrant
male hypothesis
of SSI but suggest that grizzly bears either
do not exhibit
SSI, or that SSI exists in a different
form. I propose a second hypothesis of how SSI may operate
in
bears.
This mate recognition
hypothesis
of SSI is that males of any age may, if they are able, kill cubs
that
they believe they did not sire the previous
year and
try to mate with the mother.
I use a simulation
model to evaluate
factors
that
may influence
the existence and likely form of SSI in bears.
Results of
this study suggest that killing some adult males under a sustainable
management
regime does not
decrease cub survival.
Key words: British Columbia,
cannibalism,
dispersal,
grizzly bear, habitat
selection, hunting,
infanticide,
Montana,
mortality,
survival
rates, Ursus arctos
Ursus
16(2):141-156
(2005)
Grizzly bears (Ursus arctos) are hunted
over most of
their
>5,000,000 km2
distribution
in North
America,
yet
the implications of hunting on populations remain
controversial.
In addition to the fundamental
difficulty
of estimating population abundance and a sustainable
yield, the consequences of male-dominated
harvest on
the social structure of a population
and
resulting
survival
of cubs have not been resolved.
In the 1970s, researchers
investigating
the behavior
of
langurs (Semnopithecus
entellus), a group-living pri-
mate, and African lions (Panthera leo), a group living
1bruce.mclellan@gems9.gov.bc.ca
carnivore,
noted that when the dominant male in the
group was replaced
by an immigrant
male, many of the
offspring
sired
by the former
male soon disappeared
and
their
mothers
came into estrus and mated with the new
dominant male (Hrdy 1974, Bertram 1975). These
species had extended maternal care of juveniles and
a permanent
harem mating system, so it was clearly
genetically advantageous
for the new dominant
male to
eliminate the unbor or unweaned offspring of the
females in his harem and thereby advance their next
estrus. These observations
led to the theory of sexually
selected infanticide (Hrdy 1977, 1979). Commonly
stated requirements
of SSI are that: (1) infanticidal
males should not kill their own offspring, (2) death of
141
142 SEXUALLY
SELECTED INFANTICIDE
* McLellan
the offspring
should
shorten
the interbirth
interval
of the
mother,
and (3) infanticidal males should mate with the
mother of the dead offspring
and sire her next offspring
(Hrdy 1979, Ebensperger 1998, Swenson 2003). For
grizzly bears, additional requirements should likely
include: (4) defence by the mother poses little risk to
the infanticidal male, and (5) the male would obtain
more breeding opportunities by killing juveniles and
monitoring
their mother until estrus (engaging in SSI)
than by searching
for other, unaccompanied
females.
Shortly after
the development
of the SSI hypothesis,
Stringham
(1980, 1983) discussed
2 ways that
SSI could
operate
in bears. He suggested that if adult males lived
in a "territorial
matrix", removing a resident male
would allow an influx of other
males, which could result
in increased
killing of resident
offspring,
much like was
described in langurs and lions. Conversely, if grizzly
bears were nomadic or hierarchical (non-territorial),
natality and cub survival might be enhanced by the
depletion of adult males that sometimes kill cubs,
provided
enough males remained
to breed with females
(Stringham
1980).
Subsequently, radiotelemetry studies demonstrated
that grizzly bears did not live in a territorial
matrix
but
typically
have large
home ranges
that
greatly
overlap
the
ranges of conspecifics of both sexes (Craighead
et al.
1995a, Mace and Waller
1997, Swenson
2003). Building
on Stringham's
(1980) work,
McLellan
(1994) suggested
that SSI could operate in bears even though they are
solitary
and
non-territorial,
provided
that males were able
to distinguish
their own offspring
from others. He cited
the suggestion of Hrdy (1977), who proposed that
promiscuous mating behaviour of females may have
evolved in part
to confuse paternity
and thereby
reduce
the number
of males that
may be infanticidal.
Thus, 2 potential
forms of SSI, if it exists, have been
proposed for grizzly bears and, of importance to
managers,
these 2 forms may have opposite effects on
the rate
of increase
of hunted
populations.
The first
form,
which I refer
to as the "immigrant
male" hypothesis
of
SSI, is when resident
adult males are replaced
(usually
after
they are killed) by immigrant
males that could not
have sired cubs in the area so they kill cubs to advance
the female's estrus
and
gain a breeding
opportunity.
This
general
form has been well documented
in many group-
living or territorial primates and carnivores where
dominant males have near exclusive breeding rights
with >1 female. The second form I call the "mate
recognition" hypothesis and occurs when any male,
including
adult
resident
males, kill cubs to gain breeding
opportunities with the mother. Although the mate
recognition hypothesis of SSI would function best if
males recognize their probable offspring (likely by
recognizing the females with which they previously
mated), mate recognition may not be essential (Craig-
head et al. 1995b). Among polygamous species, mate
recognition has been shown in a variety of rodents
(Mallory and Brooks 1978, Labov 1980, Huck et al.
1982), and individual recognition has been confirmed
for at least 4 years in northern
fur seals (Callorhinus
ursinus;
Insley 2000).
Since the mid 1990s, Wielgus and Bunnell (1995,
2000) and Swenson et al. (1997, 2001a) provided
evidence that the immigrant
male form of SSI operated
in bears and the removal of adult males by hunting
caused
the rate of increase to decline. Although
they did
not document infanticide,
Wielgus and Bunnell (1995,
2000) suggested that they detected a female counter-
strategy to the immigrant
male form of SSI that was
costly in terms of reproductive
output.
They stated
that
hunting mortality
of older males coincided with higher
numbers of potentially infanticidal, immigrant
males,
that adult females avoided those males and their food-
rich habitats,
and that female reproduction
appeared
to
suffer
as a result (Wielgus and Bunnell 2000). Swenson
et al. (1997, 2001a) also suggested that killing adult
males can reduce
population
growth
because immigrant
males that replace the resident
males may kill cubs, as
predicted by the immigrant male form of the SSI
hypothesis. Although recent evidence suggested that
resident rather than immigrant males killed cubs,
Swenson (2003) maintained
that the immigrant
male
hypothesis of SSI still operated and postulated that
resident males may realign their home ranges after
another resident male is killed and thus function as
immigrants.
Again, of importance
to managers
and our
understanding of the mechanism of SSI in bears,
Swenson (2003) suggested
cub mortality
would increase
in hunted
populations.
Results presented
by Miller et al. (2003) contrast
the
work of Wielgus and Bunnell (1995, 2000), Swenson
et al. (1997, 2001a), and Swenson (2003). Miller et al.
(2003) compared
cub survival, litter size, and potential
confounding factors between hunted and unhunted
brown bear populations in high-density coastal and
lower-density
interior
areas
of Alaska
and
found
that
cub
survival was greater
in hunted populations.
They also
found that
cub survival
in one study area
did not change
when the sex ratio
of the adult
population
shifted
toward
females because of increased male-biased harvest.
Miller et al. (2003) stated that their results reversed
what would be expected if male-biased hunting
Ursus 16(2):141-156 (2005)
SEXUALLY
SELECTED
INFANTICIDE * McLellan 143
disrupted social structures leading to increased in-
fanticide.
Because the effect of the 2 forms of SSI on the
survival of cubs and yearlings may differ greatly in
a hunted
grizzly bear
population,
it is critical to evaluate
and compare the validity of each form. My first
objective was to test the immigrant
male form of SSI,
particularly
the effect of grizzly bear
hunting
on cub and
yearling survival, because this form of SSI was pro-
posed by Wielgus and Bunnell (1995, 2000), Swenson
et al. (1997, 2001a), and Swenson (2003). I tested this
hypothesis by estimating grizzly bear population
parameters
in an area with a long history
of male-biased
hunting.
The hunted area was immediately adjacent
to
the Waterton/Glacier
International
Peace Park, an area
with a high density of grizzly bears that have been
protected for decades (Gniadek and Kendall 1998).
Under these distinct management
regimes, the immi-
grant
male hypothesis of SSI leads to 2 predictions.
As
suggested by Wielgus and Bunnell (1995, 2000), it first
predicts a high immigration rate of potentially in-
fanticidal males from the protected park
into the hunted
area, where many adult males have been killed. The
second prediction
of the immigrant
male hypothesis of
SSI-and most important
from a management
perspec-
tive-is that cub survival rates would be low in the
hunted area or at least lower than in nearby
areas where
grizzly bears are not hunted.
In particular,
in the hunted
area there would be a high loss of entire litters due to
infanticide
by immigrant
males, and cub loss would be
most common during the breeding season when SSI
would most benefit the infanticidal male (Swenson et al.
1997, 2001a).
In addition
to these predictions,
the female counter-
strategy
to the immigrant
male form of SSI, as described
by Wielgus and Bunnell
(1995, 2000), predicts
that adult
females in the hunted area would select different and
inferior habitats than males to avoid potentially in-
fanticidal males, and consequently have smaller litter
sizes. McLellan and Hovey (2001a) recently tested the
effects of sex, age class, and season on the selection of
habitats and elevations by grizzly bears in the hunted
area. Therefore, I do not re-analyze habitat selection
here,
but use results of that study to test this prediction.
Because of the apparent
inconsistent expression of
infanticide
in grizzly bears (Swenson 2003), my second
objective
was to explore factors that may influence the
natural selection of SSI in this species, and what form it
may take under different conditions, by simulating
trends in reproductive success of males that adopt
different
strategies.
The first strategy
that I investigated
Fig. 1. Flathead study area in southeast British
Columbia, Canada, and Montana, USA, showing the
juxtaposition of hunted and protected areas and
where bears were trapped, 1978-2003.
was SSI committed
by males of various ages. Because
grizzly bears are approximately
4-6 years of age when
they disperse (Blanchard
and Knight 1991, McLellan
and
Hovey 2001b), SSI by these age classes would more
likely follow the immigrant
male form of SSI, whereas
older males would more likely follow the mate
recognition form. The second strategy I investigated
was that males do not kill cubs but simply search for
estrous females.
Study area
The study area consisted of 2 adjacent areas with
different grizzly bear management objectives. The
hunted area was mostly in the Canadian
portion
of the
North Fork of the Flathead
River drainage (Fig. 1) that
flows from the extreme southeastern
comer of British
Columbia southward into the United States (49?N;
114?85'
W). In this portion
of British
Columbia,
grizzly
bears have been hunted for at least a century.
In contrast,
across the international
border, the North Fork of the
Flathead
River forms the western boundary
of Glacier
Ursus 16(2):141-156 (2005)
144 SEXUALLY SELECTED
INFANTICIDE
* McLellan
National
Park
(GNP),
where strict
management
practices
and enforcement have resulted in few grizzly bears
being killed by people for many decades (Gniadek
and
Kendall 1998). The National
Park has a high density of
grizzly bears (Martinka 1974; K. Kendall, US Geo-
logical Survey, Glacier National Park, Montana, USA,
personal communication, 2004) and should serve as
a source of immigrants
into the hunted area. Immedi-
ately east of the hunted area, grizzly bears within
Waterton
Lakes National Park (WLNP) are similarly
protected
and may also be a source of immigrant
males.
There are no physical barriers
between the 2 national
parks
and the hunted
area.
The western
half of GNP is in
the same
4-10 km wide valley of flat benches
and
rolling
hills as the hunted
area.
The North
Fork of the Flathead
River
and several
major
tributaries
cross the international
border
between GNP and the hunted area.
Although
the
continental
divide separates
WLNP
from the hunted
area,
there are many forested passes. Radiocollared bears
move freely among both parks and the hunted area
(McLellan, unpublished
data). Physical and vegetative
characteristics
of the study area have been published
elsewhere (McLellan
and Hovey 2001a).
Methods
Testing the immigrant male hypothesis
My first
prediction
of the immigrant
male hypothesis
of SSI is that hunting adult males in British Columbia
would have disrupted the social stability of the
population, resulting in an influx of young (4-6 year
old), potentially infanticidal males from the adjacent,
unhunted national parks. To test this prediction, I
followed the method of Wielgus and Bunnell (1995,
2000) and used trapping
records
from the past 25 years
in the hunted area, where most individuals have been
captured
(McLellan 1989, McLellan unpublished
data).
Due to their
extensive movements,
male grizzly bears
of
all ages are more often trapped
than females; however,
the predicted immigration of young males into the
hunted area should result in a greater
proportion
of the
males being 4-6 years of age when first captured
(not
previously
captured
before this age) compared
with 4-6
year-old females that are more philopatric
(Blanchard
and Knight 1991, Mace and Waller 1998, McLellan
and
Hovey 2001b).
The second prediction
of the immigrant
male hypoth-
esis is that
cub mortality
should be higher
in the hunted
than
in nearby
unhunted
areas.
In particular,
there
should
be frequent
whole litter
loss during
the
breeding
season
in
the hunted
area.
Because females in the study area,
and
throughout
North
America,
most commonly
have 3-year
interbirth
intervals
(they are with cubs 1 year, yearlings
the next, and then spend almost an entire year alone
before giving birth again) it would be beneficial for
infanticidal males to also kill yearlings
up to the end of
the breeding season. After that time, killing yearlings
would not shorten a 3-year interbirth interval and
therefore
would not benefit
infanticidal
males.
I tested
these 2 hypotheses
using a long-term
database
of live-capture
records and by monitoring
the reproduc-
tive status of female grizzly bears. Between 1978 and
2003, bears were captured
in foot snares, culvert traps,
or darted
from a helicopter
outside of GNP and WLNP,
across an area of approximately
650 km2 in British
Columbia
or just south of the border
in Montana.
Bears
were captured primarily
in May, June, September,
and
October,
but captures
were recorded
in all months
of the
bear-active
season (Apr-Nov). Captured
bears
were ear-
tagged,
weighed, measured,
and
radiocollared,
and
those
> 1 year
of age had a premolar
removed
and sectioned
to
estimate
their age (Matson
et al. 1993). The location of
all radiocollared
individuals
was determined
from fixed-
wing aircraft at about 1-week intervals. High road
density (=44 km/100 km2) provided good access that
also enabled most bears to be located from the ground
every 1-10 days. Changes
in litter-status
of female bears
were directly
observed
from aircraft
or ground.
Cubs no
longer
observed
with their
mother
were assumed
to have
died. This assumption was also made for yearlings
unless they were, through
radiotelemetry
or observation,
known to be alive. Separation
between mothers and
yearlings
during
the breeding
season does happen
in the
study area, so this assumption
may lead to an inflated
estimate
of yearling
mortality.
Cub and
yearling
survival
rates
were estimated
by the number
that survived
to the
end of the period of interest, divided by number
observed at the beginning of the period (Hovey and
McLellan 1996, Mace and Waller 1998, Garshelis
et al.
2005). Confidence
limits, and other
randomization
tests
deployed bootstrapping procedures using the Excel
(Microsoft, Redmond, Washington, USA) addition
POPTOOLS
(Hood 2004).
Because there
were no data
on cub survival
in GNP or
WLNP,
I compared
cub survival
rates
in the hunted
area
with 2 other
areas
in the Rocky Mountains
where
grizzly
bears are not hunted, although males and females die
from
other
causes. Mace and
Waller
(1998) reported
cub
survival
to be 0.77 (n = 28; 95% CI = 0.63-0.95) in the
South Fork of the Flathead river drainage, Montana,
centered
approximately
200 km south
of my North
Fork
of Flathead
study area.
Similarly,
Garshelis
et al. (2005)
Ursus 16(2):141-156 (2005)
SEXUALLY
SELECTED
INFANTICIDE
* McLellan 145
reported
cub survival rate of 0.79 (n = 53; 95% CI =
0.67-0.93) in and around Banff National Park and
Kananaskis
Country,
Alberta,
an area centered
approx-
imately 200 km north
of my study.
Modeling sexually selected infanticide
Grizzly bears have a scramble competition polyga-
mous mating system (Craighead
et al. 1995a,b; Alcock
2001:382). Grizzly bear home ranges overlap exten-
sively with conspecifics of both sexes, and male ranges
are larger than those of females. Females are induced
ovulators and may have 2 estrous periods of approxi-
mately 10 days, most frequently during
the last 3 weeks
of June (Craighead
et al. 1995a). Males have 2 basic
strategies: monopolize individual females by mate-
guarding through
their estrous period, or displace other
males from breeding and add their gametes to those
already deposited (Craighead et al. 1995b). When
breeding females are scarce and widely distributed,
males tend to locate one female to breed exclusively.
However, when breeding females are more common,
males travel widely to locate more females (Craighead
et al. 1995b). In Ursus spp., body size of males is an
important
factor
determining
the outcome of fights over
breeding
females (Craighead
et al. 1995a,b;
Kovach
and
Powell 2003).
To explore factors
that
may influence the incidence
of
SSI, I estimated
mating
success under
different
scenarios
using a model that mimics, in a simplistic way, the
mating system of grizzly bears. The simulation
model
was built in Excel but incorporated
the random
number
generator and Monte Carlo simulation features of
POPTOOLS. The model used the body mass of bears
(37 F, 59 M) captured
in the Flathead and the upper
Columbia River drainage (Apps et al. 2004), located
250 km north of the Flathead
study area. I included 3-
year-old male bears in the model because this is the
age when males greatly expand their home ranges and
begin to disperse (Mace and Waller 1998, Swenson
et al. 1998, McLellan and Hovey 2001b). Three-year-
old males are known to mate with adult females in
Scandinavia
(Swenson 2003). I classed bear
ages as 3, 4,
5, 6-7, 8-9, 10-11, 12-17, and >17 years, which
ensured >4 individuals
per class. In the model, a focal
male could adopt 1 of 2 strategies to obtain breeding
opportunities.
He could be infanticidal,
kill cubs, and
monitor
the mother
until receptive,
or he could not kill
the cubs and instead search
for a breeding
female.
I varied
the following parameters
in the model to test
their implications
on the success of the 2 strategies:
(1)
the proportional
difference in body mass of the male
relative to the mother
that would enable it to kill cubs,
(2) the mean and standard
deviation (SD) of the time
required for a female's physiological condition to
change from a lactating
mother
to estrus, (3) the mean
and SD of the time required to ensure a female is
impregnated
once receptive, (4) the mean number of
adult (age >5 years) females in the focal male's
breeding-season
home range, (5) the mean interbirth
interval of females, (6) the mean number of adult
males in the focal male's breeding-season
home range,
and (7) the searching efficiency of males and estrous
females. I modeled time required for a male to
encounter
a particular
female in his range as a random
number
between 0 and the maximum
time, which was
an input variable and therefore reflected variation in
search
efficiency.
The model for infanticidal males proceeded as
follows. A male was randomly
selected from all males
of a specific age class and, using a table of all females,
was randomly
matched with a female accompanied
by
cubs. If he did not meet the input size requirement
to
overcome female defense of her cubs, he did not mate
(the model tracked the mass of all adult males and
females). Alternatively,
if he met the size requirement,
he killed all her
cubs. He then
monitored
the female (and
did not search for another female) until she was
receptive for mating and then impregnated
her. This
period was the sum of the time for the female to switch
from being a lactating
mother
to being receptive
and the
time from receptivity to fertilization (both input
variables). While the male monitored the female, all
other males (the number
being an input variable)
who
were not with other estrous females (derived from 2
input
variables;
the number
of females and the breeding
interval) would encounter the female, each with
a probability
governed by his independently selected
searching
efficiency (0 days to an input maximum).
If
any of the arriving males were larger than the in-
fanticidal male, the smaller male was forced off and
would not mate (Craighead
et al. 1995b, Kovach and
Powell 2003). If no larger males arrived before the
female was bred, she was bred by the infanticidal
male.
Female mate choice could be incorporated
in the model
by extending
the time between
the death of her cubs and
when she became pregnant or by increasing search
efficiency and thus enabling encounters with more
males. Further,
an infanticidal
male could be "discre-
tionary" and not kill his own cubs if encountered,
or
"nondiscretionary"
and
kill his own cubs (only for bears
>5 years of age; the probability
of a male killing his
own cubs was inversely proportional
to the number
of
Ursus 16(2):141-156 (2005)
146 SEXUALLY
SELECTED INFANTICIDE
* McLellan
a time, both lower and higher than the base scenario,
to explore trends in reproductive
success among males.
, , . ,
I , ,
I , ,
I , , I , . . I . . - -
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Age when first captured
Fig. 2. Age distribution of male and female grizzly
bears when first captured in the Flathead study area
of British Columbia, Canada, and Montana, USA,
1978-2003.
adult males in the population)
as well as those fathered
by other males.
In the model, non-infanticidal males did not kill cubs
but searched
for other females until 1 was found. All
other males also searched for these females. A non-
infanticidal male was successful if he was the largest
male to encounter the female within the time for her to
become
pregnant.
The model only treated the multi-male
interactions over 1 female (not multiple
females over the
entire breeding season). For each age class of males, I
conducted 2,000 replicates of the model and recorded
the success of both discretionary
and non-discretionary
infanticidal males and non-infanticidal
males.
The objective of using the model was not to
necessarily use true values for each parameter to
determine which strategy performed best and thus
should have been naturally selected for. Such use of
the model would be presumptuous
because true values
are unknown for most parameters,
values likely vary
greatly
within and among populations,
and the model is
simplistic. Instead,
the model was used to explore how
changes in the values of input variables
may affect the
success of males of various ages that adopt different
mating strategies. Base scenario values for some
parameters
were selected because they were clear and
simple (equal adult sex ratio, males had to be the same
size or larger than the mother to successfully kill her
cubs), and for other parameters
base scenario values
were selected from the often limited number of cases in
the literature or through personal communication with
research
biologists. I then varied one input variable at
Results
Testing the immigrant male hypothesis
Between 1978 and 2000, 85 male and 45 female
grizzly bears were killed by people in the 2 management
units that included the hunted area. These units covered
2,815 km2, resulting in annual kill densities of 1.37
males and 0.73 female bears/1,000 km2. The mean
and median age of harvested bears was 8.52 and 6,
respectively,
for males and 9.31 and 9, respectively,
for
females. In GNP, approximately
48 bears were killed
between 1960 and 1994, giving an annual kill density
of 0.33 bears of both sexes/1,000 km2 (Gniadek and
Kendall 1998). This annual
kill density was reduced to
0.14 bears/1,000
km2
during
the 1990s (calculated
from
Gniadek
and Kendall 1998), or 1/15th of the kill density
in the hunted area.
In British Columbia and outside GNP in adjacent
portions of Montana, 134 different grizzly bears were
captured
(77 M, 57 F). The ages of bears when first
captured
were similar for both sexes (randomization
test
P =0.38; Fig. 2). When
first
captured,
most male (54.6%)
and female (52.6%)
bears were <3 years
old and 22.1%
of the males and 19.3% of the females were 4-6 years
of age. These capture
records do not support
the hypoth-
esis that a substantial influx of potentially
infanticidal,
young males came into the area with hunting.
In the hunted
area,
87 cubs were
produced
in 41 litters
for an average
litter size of 2.12 cubs/litter.
In one case,
I was uncertain whether the entire litter of 2 cubs died
during
or after the breeding
season, so either 81 or 83 of
87 cubs survived for a survival rate of 0.93 (95% CI =
0.87-0.98) or 0.95 (95% CI = 0.91-0.99), and 38 or 39
of the 41 litters survived the breeding season (93% or
95%). In either case, there was high cub survival from
den emergence
to early July. The annual survival rate of
individual cubs was 0.85 (n = 81; 95% CI = 0.78-0.93)
and similarly,
33 of the 39 litters
(85%)
monitored for an
entire
year survived. The survival of cubs in the hunted
area was not lower than the 0.79 (P = 0.87, 1-tail)
in the
Banff and Kananaskis area
(Garshelis
et al. 2005) or the
0.77 (P = 0.76, 1-tail)
in the South Fork of the Flathead
(Mace and Waller 1998). Fifty-six of 59 yearlings
survived to the end of the breeding
season for a survival
rate of 0.95 (95% CI = 0.90-1.0), and none of the
28 litters of yearlings
was entirely
lost before the end of
the breeding season. Equivalent data on yearling
survival were not available from the other studies for
25-
20-
0
z
5-
] Female (n = 57)
*Male (n = 77)
,
III
Ii lI In 1
... 1,. I I
Ursus 16(2):141-156 (2005)
L,-- (RI --arai.--
I
.
n
i
SEXUALLY
SELECTED
INFANTICIDE
* McLellan 147
Table 1. Spring mass (x and SD) of males by age class and the number of successful matings out of 100 (%)
after 2,000 simulations for males adopting I of 3 strategies: (1) infanticidal but does not kill his own offspring
(mate recognition), (2) infanticidal but may kill his own offspring and thus reduce reproductive success,
and (3) searcher male that does not kill cubs but searches for an estrous female. NA - not applicable as it
is unlikely that these young bears fathered existing litters.
Age class of males (years)
Model
scenario Strategy 3 4 5 6-7 8-9 10-11 12-17 >17
Mean
spring 87.4 98.8 106.0 132.4 156.8 192.3 177.8 194.4
mass (SD) (17.3) (31.7) (37.9) (40.8) (68.7) (27.3) (24.7) (37.8)
Base scenario infanticidal 2.3 4.7 6.8 15.9 31.9 41.4 32.6 48.0
indiscriminate NA NA NA 7.8 23.3 32.6 22.3 39.9
searcher 11.2 13.3 14.9 22.3 33.0 44.1 36.4 47.7
Cub death to infanticidal 5.1 10.0 12.6 24.3 36.5 52.1 43.2 55.2
estrus 2 daysa indiscriminate NA NA NA 16.1 28.8 42.6 34.6 46.9
searcher 11.1 13.3 15.7 21.8 33.2 43.9 36.4 47.8
Cub death to estrus infanticidal 11.1 18.3 22.1 36.0 45.4 63.9 58.2 66.4
and time to indiscriminate NA NA NA 28.8 37.7 56.6 48.6 58.3
impregnate
2 daysa searcher 24.5 25.7 27.2 31.4 36.5 43.8 41.0 45.3
Cub death to infanticidal 0.3 1.3 2.2 8.7 24.9 29.3 18.7 38.0
estrus 10 daysa indiscriminate NA NA NA 7.9 17.4 21.8 9.7 28.3
searcher 10.9 13.6 15.3 21.7 33.4 43.5 36.5 48.1
5 adult
males but infanticidal 18.9 29.1 32.6 44.7 48.8 67.6 64.6 69.1
10 adult
femalesa indiscriminate NA NA NA 29.0 34.1 52.7 47.1 50.4
searcher 66.6 69.0 70.1 75.9 80.4 87.2 84.5 88.1
10 adult
males but infanticidal 1.7 3.6 5.7 14.1 29.1 37.0 27.2 44.5
5 adult
femalesa indiscriminate NA NA NA 5.6 20.9 29.1 18.4 35.3
searcher 3.4 4.4 4.9 7.3 14.4 17.6 13.7 20.8
Maximum
10 days infanticidal 0.3 0.8 1.5 6.8 23.4 25.1 14.1 34.1
to find
breeding indiscriminate NA NA NA -1.4 15.7 17.6 5.3 25.2
femalea searcher 3.9 4.8 6.2 13.0 29.3 36.4 25.2 43.3
Maximum 30 days infanticidal 5.6 10.3 13.0 25.3 36.9 52.4 43.6 56.8
to find
breeding indiscriminate NA NA NA 16.6 29.9 43.9 35.0 48.4
femalea searcher 16.8 19.5 20.5 27.3 34.4 44.5 39.3 46.8
aValues
of all other
variables set to the base scenario.
comparisons. From these results, I conclude that cub
and yearling survival
was high and entire
litter
loss was
low in the hunted area.
Modeling sexually selected infanticide
Base scenario. Adult females in the spring
weighed an average
of 99.9 kg (SD = 17.5). In the base
scenario,
males had to be the same size or larger
than
the
female to kill her cubs (Table 1). The time for a female
to change physiologically from lactating to estrus
averaged 5 days (SD = 2) followed by an average of
5 days (SD = 2) to become pregnant.
There
was also an
equal ratio of adult males to females, and it took
between 0 and 20 days for each male that was not with
another
female to locate an estrous female. With these
base inputs,
most males that
did not kill cubs but simply
searched
for females in estrus
were reproductively
more
successful than males that did kill cubs (Table 1). The
difference
between the outcomes of the 2 strategies
was
greatest
for younger,
smaller
bears,
for which infanticide
rarely generated
successful breeding
opportunities.
Time of estrus onset. With all input variables
set to the base scenario but with shorter periods for
females changing from lactating
to estrus (x = 2 days;
SD = 0.5) and being impregnated
thereafter
(x = 2 days
later; SD = 0.5), older males were more successful by
being infanticidal
than
by searching
for estrous
females.
In this "quick
estrus
onset and impregnation"
scenario,
older, infanticidal
males were more successful even if
they could not determine
if the cubs they were killing
were their own. However, for bears <6 years of age,
searching
for an estrous female was a more successful
strategy than being infanticidal, even when onset of
estrus was quick (Table 1). In contrast,
when onset of
estrus was delayed ( = 10 days for a female to change
from lactating
to estrus),
it was advantageous
for males
of all ages to search
for another
female; this advantage
was particularly
notable for young males (Table 1).
Ursus 16(2):141-156 (2005)
148 SEXUALLY
SELECTED
INFANTICIDE
* McLellan
Sex ratio. If there were half as many adult
males as
adult females, it was better for males of all ages to not
kill cubs but to search for estrous
females. This scenario
was the only one where young males had a high
probability (65-70%) of siring offspring (but only
among those who searched for estrous females). When
there were half as many adult females as adult males,
infanticide was a more successful strategy
for all males,
except for those <5 years of age when the 2 strategies
resulted in similar success.
Mate searching efficiency. A reduced
ability for
males and estrous females to find each other also
resulted in infanticide being a successful strategy for
older males, but it remained less successful than
searching for male bears <7 years old. With better
searching efficiency, infanticide was less profitable.
Discussion
Testing the immigrant male hypothesis
Results of the Flathead study do not support the
immigrant
male hypothesis
of SSI because neither
of the
2 predictions
were met. The area had been hunted for
many years with more males shot than females, and it
was immediately adjacent
to unhunted areas. However,
unlike results of Wielgus and Bunnell (1995, 2000),
there was no indication of a substantial influx of
subadult
males to replace the killed male bears. Most
bears of both sexes were <3 years of age when first
captured
and, although
more males were 4-6 years old
when first captured than females, there were, as
expected, more males captured
than females of all ages.
These capture
ratios,
despite the higher
mortality
rate of
males (McLellan et al. 1999), likely reflect the greater
movements of males of all ages and thus greater
trap
encounter
rates of males.
Because male grizzly bears disperse farther than
females (Swenson et al. 1998, McLellan and Hovey
2001b, Proctor et al. 2004), the lack of an influx of
young males is somewhat surprising. However, dis-
persal
by both sexes of grizzly bears appears
limited in
distance and is a gradual
process. After 50-60 years in
Scandinavia,
where brown bear populations
have been
rapidly
growing,
99%
of the males and
females
killed by
hunters
were within 80 km and 50 km, respectively, of
source areas (Swenson et al. 1998). In the Flathead
study area, males and females dispersed an average of
only 30 km and 10 km, respectively. Using genetic
analysis of 711 grizzly bears across a 100,000 km2 area
that included the Flathead study area, Proctor et al.
(2004) found
average
dispersal
distances
of 41.9 km and
14.3 km for males and females,
respectively.
In addition
to limited dispersal influencing the immigration of
young males from the national parks, the Flathead
hunted area has a high density and increasing popula-
tion of grizzly bears (McLellan 1989, Hovey and
McLellan 1996). The high density and production
of
young bears locally may have discouraged immigration
from neighboring
national
parks.
The SSI prediction
of high cub mortality
and loss of
entire
litters,
particularly during
the breeding
season, and
high yearling mortality up to the end of the breeding
season also was not supported.
Cub survival was as high
or higher in the Flathead hunted area than in either the
unhunted South Fork of the Flathead or in the unhunted
Banff and Kananaskis area. At 0.85, the point estimate
of cub survival in the Flathead hunted area was higher
than any of the 8 other published studies from North
American (x = 0.61, n = 736 cubs; Mace and Waller
1998, Miller et al. 2003, McLoughlin et al. 2003,
Garshelis et al. 2005, Schwartz
et al. 2005). Of these
studies, only cub survival outside of Yellowstone
National
Park,
but within the Yellowstone Grizzly Bear
Recovery
Zone, was similar
at 0.82. Within Yellowstone
National
Park
itself, cub survival
was much lower
at
0.49,
likely reflecting
a density effect (Schwartz
et al. 2005).
There is, however, a potential
bias in estimating
cub
survival because cubs could die before they are ob-
served. This bias may be particularly
problematic
in
more heavily forested areas, such as the Flathead
area,
where bears are sometimes difficult to observe. In the
Flathead study area, however, all females had litters
when they were expected (the spring following separa-
tion from their last litter) from 1979-1997 (n = 36
litters), so this bias was not a factor in estimating
sur-
vival of entire
litters
for the first 18 years of study. The
pattern
of consistent reproduction
ended in 1997 when
some females, particularly
some that
were >20 years of
age, failed to produce
cubs when expected. The missed
reproductive
events later in the study may be due to a
density effect as the population rapidly increased
(Hovey and McLellan 1996). Additionally, there were
major crop failures of huckleberry
(Vaccinium mem-
branaceum) in those years, and some old bears may
have reached reproductive
senescence (Schwartz
et al.
2003). Results
of this study
demonstrate
that
infanticide,
if it occurred,
was rare
despite the study area having a
male-biased
hunter
harvest
and being adjacent
to a large
unhunted
area
that
could have supplied
immigrant
males.
Wielgus and Bunnell (1995, 2000) did not document
infanticide
in a hunted population
of grizzly bears but
suggested
that
they detected
a female counterstrategy
to
Ursus 16(2):141-156 (2005)
SEXUALLY SELECTED
INFANTICIDE
* McLellan 149
the immigrant
male form of SSI that
was costly in terms
of reproductive output. In an area with grizzly bear
hunting, they found adult females to avoid potentially
infanticidal,
immigrant
males and the food-rich
habitats
that
they occupied and, consequently,
that
these females
had small litters (Wielgus and Bunnell 2000). In the
Flathead hunted area, sex and age class did not have
a significant
effect on habitat
selection by grizzly bears
(McLellan
and Hovey 2001a). At 2.12 cubs, the average
litter size in the hunted area was not small compared
with the average of mean litter sizes from 26 studies
across North America (2.068 cubs/litter, n = 1,441
litters; MacHutchon et al. 1993, McLellan 1994,
McCann 1997, Case and Buckland 1998, Mace and
Waller 1998, Wielgus and Bunnell 2000, McLoughlin
et al. 2003, Miller et al. 2003, Garshelis et al. 2005)
and larger (randomization
test; P = 0.003) than the
adjacent
unhunted GNP which, between 1987 and 2002,
had an average litter size of 1.80 (n = 148; US Fish
and Wildlife Service, Missoula, Montana, USA, unpub-
lished data).
Unlike the results of Wielgus and Bunnell (1995,
2000), adult
females in the hunted
area in this study did
not avoid habitats
preferred
by males and they did not
have small litters.
However, the mechanism
behind this
prediction of Wielgus and Bunnell (1995, 2000) is
unclear. If females with cubs or with yearlings
up to the
end of the breeding season avoid food-rich habitats
where immigrant males are more often found, then
decreasing
body condition of the mother
may result in
elevated cub or yearling mortality. However, when
mothers are with older yearlings or unaccompanied
by
offspring, they should make greater use of preferred
habitat.
After 1.5 years of using preferred
habitat,
the
body condition of adult females should have recovered
and their
litters of new cubs should not be small.
Mattson et al. (1987) and McLellan and Shackleton
(1988) noted that
adult
males were located farther
away
from roads than were adult females, particularly
adult
females with cubs. This result could be due to females
avoiding habitats where adult males were present,
which is consistent with both the immigrant
male and
mate recognition forms of SSI and the hypothesis that
males may kill cubs for other reasons, such as
predation.
Modeling sexually selected infanticide
Van Schaik (2000) developed a model to investigate
decision rules of male primates. Van Schaik (2000)
stated
that his model assumes that a male has access to
a particular
female and will only kill infants upon
assumption
of his tenure
as dominant
male of the group.
This model was designed for group-living primates
and
simply concluded
that if infanticide
shortens
a female's
interbirth interval and if the probability that the
infanticidal
male sires the female's next litter is greater
than the probability that he sired the one he killed,
infanticide
benefits
the male. Such a model is clearly
too
simple for bears that do not have a female defense
polygamous
(harem)
mating system
but
rather a scramble
competition polygamous system.
The model I developed
has simplistic
rules;
however,
it considers
a scramble
competition
polygamous
mating
system. The primate
model of van Schaik (2000) does
not consider
the ability
of mothers
to defend
their
young
because males that live in a permanent
group can wait
for an opportune
moment
to kill dependent
offspring
and
body size is not a significant factor (Janson and van
Schaik 2000). Grizzly bears
are solitary,
so small males
cannot easily wait for opportune
moments to kill cubs,
and female defense is undoubtedly
important.
Although
this input variable
could be changed, in the scenarios
I
presented,
male bears had to be the same size or larger
than the mother if they were to successfully kill her
cubs. Some 3- and 4-year-old males are slightly larger
than some adult females so, when randomly
matched,
these males successfully killed cubs. However, it is
unlikely that a young male that was only slightly larger
than a mother would be able to kill her cubs without
significant
risk. Miller et al. (2003) reported
that all of
the 19 bears
seen attacking
cubs on the Alaska
Peninsula
and Kodiak Island were adult males and of the 8 in-
dividuals of known age, the average age was 9.4 years
and the youngest was 5 years. Similarly,
2 males known
to have killed brown bear cubs in Scandinavia were
9 and 11 years of age and the youngest of 4 that were
suspected of killing cubs was 6 years of age. It is
probable
that a male must be considerably
larger than
the mother
to successfully kill her cubs without risk to
himself. If so, even fewer young immigrant
males than
suggested by my model would benefit from the
infanticidal
strategy.
The group-living
primate
model of van Schaik (2000)
also does not include details of female reproductive
physiology but assumes that the time required for
a female to change from lactating to estrus is short
compared
with the tenure of the new male. For
primates,
van Schaik (2000) stated
that females usually return
to
receptivity
within
a month,
but sometimes
within 1 week
or even 3 days. This variable is clearly important
for
bears or other species with a scramble competition
polygamous mating system. If this period is long, not
Ursus 16(2):141-156 (2005)
150 SEXUALLY
SELECTED
INFANTICIDE
* McLellan
only may a larger
or more dominant male encounter
the
female before she becomes fully receptive, but the
infanticidal
male may be much less efficient at finding
other estrous
females while he monitors the reproductive
condition of the female whose cubs he killed.
It is likely that a reproductive strategy has evolved
among female bears enabling them to advertise their
availability for sufficient time to attract, if possible,
competing
suitors
in an effort
to become impregnated
by
the "best" male or males. Unlike a harem mating
system, where much male-male competition
occurs to
obtain and maintain a harem before mating occurs,
male-male competition among bears most often occurs
during the female's estrous period (Craighead et al.
1995a,b; Kovach and Powell 2003). Consequently,
it is
unlikely that female bears will have evolved to become
impregnated quickly after
losing cubs.
The plasticity of reproduction in bears, such as
occasional
observations
of female American
black bears
(Ursus americanus) breeding while raising cubs (Le-
Count 1983), makes estimating
the period
between litter
loss and receptivity empirically problematic.
I used 5
days in my base scenario and 2 and 10 days in other
scenarios,
but most of the scant evidence suggests that,
on average, it is longer than 5 days. Captive female
black bears bred 2 to 3 weeks after their cubs were
removed
(LeCount
1983), but the shortest
time from cub
removal
to a mating
event was 5 days for captive
brown
bears in Germany (Dathe 1961). In free-ranging
conditions in Alaska, S. Miller (National Wildlife
Federation,
Missoula, Montana,
USA, personal
commu-
nication,
2004) observed
a female almost daily after
she
became separated
from her new cubs. She was first
seen
with a male 18 days later
but was with him for <1 day
and then with another
male 3 days after
that. In 2 other
cases in Alaska, H. Reynolds (Alaska Department
of
Fish and Game, Fairbanks, Alaska, USA, personal
communication,
2004) observed males in the company
of females 9 to 11 days after
the females had lost their
cubs. Hessing and Aumiller
(1994) reported
that
4 days
after a female's cubs were killed, a male bear (not the
one that killed at least one of the cubs) followed the
mother
at a distance
of 100 m, but after
that
observation,
no males showed any interest in her. In Scandinavia,
however, J. Swenson (Department of Ecology and
Natural
Resource Management,
Agricultural
University
of Norway, As, Norway, personal communication,
2004) observed
a female separate
from her cub for only
12 hours but she gave birth
the following winter.
Simulation
results suggest that the sex ratio of adult
bears is an important
factor influencing
the success of
the different
male strategies.
If there are few adult males
in the population (a typical condition when hunters
remove more males than females), the searcher
strategy
by males is much more successful than infanticide. The
results of the model were generally consistent with
observations
from the Flathead
study and the 2 hunted
study
areas
in Alaska
(Miller
et al. 2003). In those areas,
hunters removed more males than females and cub
survival was higher than in the unhunted control areas
(Miller et al. 2003).
Although the model suggests that infanticide may
become a less successful strategy
when there are fewer
males, not enough is known about the various
parameters
to determine a sex-ratio threshold beyond
which SSI is unlikely to operate. Cub survival in the
Susitna study area in Alaska did not change even after
increased hunting shifted the sex ratio of adult bears
from about 38% males to 20% males (Miller et al.
2003). Perhaps
with only 38% males in the population,
the searcher
strategy
already
dominated male behavior,
so a further reduction
in sex ratio had no effect.
Rarely do hunters
kill more female than male bears
(Bunnell
and Tait 1981, Miller
et al. 2003), but this was
the case in 1 of 2 black bear study areas in Arkansas
(Clark
and Smith 1994). Both populations
were below
carrying capacity
and increasing,
and body growth
rates
of females were among
the highest
in North
America.
In
the study area where more females were killed than
males, the sex ratio of bears >1 year of age was 1.44
males/female (26:18), the average litter size was 1.41
(n = 17), 4 of 13 cubs survived,
and 6 of 8 entire
litters
were lost. In the study area where more males were
killed than
females, there
was a net emigration
of males
(J. Clark,
US Geological Survey, Knoxville, Tennessee,
USA, personal communication,
2004), the sex ratio of
bears > 1 year
of age was 0.64 males/female
(27:42), the
average litter size was 2.25 (n = 20), 18 of 20 cubs
survived, and none of 8 litters was entirely lost. The
results
of Clark
and Smith
(1994) generally
match
those
of the simulation
model where infanticide
(and low cub
survival) was a more successful strategy when there
were more males than females. However, Clark and
Smith (1994) concluded that infanticide remained
speculative in their study because other environmental
variables
may have contributed
to differences between
their populations.
The simulations also indicate that infanticide may
become more profitable
for males when their searching
efficiency for estrous females decreases. This result
suggests that SSI may be more prevalent
in low-density
populations,
where bears have very large home ranges
Ursus 16(2):141-156 (2005)
SEXUALLY SELECTED
INFANTICIDE
* McLellan 151
Table 2. Summary of victims and killers in documented cases of intraspecific killing of grizzly or brown bears
in North
America. References: McLellan
(1994); Clarkson and Liepins (1994); M.
Gibeau, University of Calgary,
Calgary, Alberta, Canada, personal communication, 2004; R. McCann, University of British Columbia,
Vancouver, British Columbia, Canada, personal communication, 2003; S. Himmer, Bella Coola, British
Columbia, Canada, personal communication, 2003; H. Reynolds, Alaska Department of Fish and Game,
Fairbanks, Alaska, USA, personal communication, 2003; and B. McLellan, Revelstoke, British Columbia,
Canada, unpublished data.
Killer
adult adult unknown subadult unknown
Victim female male adult male male unknown total
Adult
female 16 1 2 7 26
Adult
male 3 2 5
Subadult
female 2 1 3
Subadult
male 3 2 5
Unknown
subadult 4 4
Yearling 1 4 3 8
Cub 7 13 4 12 36
Total 8 45 4 1 2 27 87
and travel
widely in search
of mates. If the area
has few
competing males and encountering
an estrous female
is relatively uncommon, then killing cubs and moni-
toring the mother as she returns to estrus may be a
viable strategy.
Mate recognition hypothesis versus no
sexually selected infanticide
Relative to the complex mating system and re-
productive
physiology of bears, the simulation
model I
developed is simplistic and was used only to explore
trends
in mating success with changing input variables.
One consistent
and somewhat
obvious trend,
however,
is
that if body mass is an important
factor influencing
whether
a male can successfully kill cubs and influence
male-male competition
over estrous
females, then older,
larger males benefit more from being infanticidal
than
younger males. Observations in Alaska (Miller et al.
2003) and Scandinavia (Swenson 2003) suggest that
large, older males kill cubs. In addition, if having
multiple
mates to confuse paternity
is a female counter-
strategy
to SSI (McLellan 1994, Swenson 2003), it will
only function if males can recognize
females they mated
with the previous 1 or 2 years. Having multiple
mates,
however, may also have evolved to encourage sperm
competition. These observations
and the results of the
simulations
suggest that if SSI does operate
in bears,
the
mate
recognition
form is more likely than
the immigrant
male form. However, neither
the results of the Flathead
study or the simulations
can differentiate
between the
role of mate recognition SSI and the alternative
hypothesis that infanticide is due to other
reasons, such
as predation.
If infanticide
is not sexually selected but indeed has
another
underlying
cause, then
bears
of other
age classes
should be occasionally
killed. In Scandinavia,
Swenson
et al. (2001b) found 13 cases of intraspecific
predation
in
668 bear-years of radiotracking,
including an annual
predation
rate of 0.16 (6 of 38) for yearling
females. In
11 studies in Idaho, Montana, British Columbia, and
Alberta, 17 radiocollared subadult and adult grizzly
bears
died naturally
(not killed by people), and
the cause
could be determined for 13 (McLellan et al. 1999;
McLellan,
unpublished
data).
Of these 13 deaths,
9 were
killed by another
bear and at least 7 were consumed.
Eight of these 9 were adult
females, but none were with
cubs when killed and thus did not die defending their
young. Information on intraspecific killings among
grizzly bears in North American (McLellan 1994;
McLellan, unpublished
data) indicates that adult males
do most of the killing and cubs are the most common
victim (Table 2); however, males kill other
age and sex
classes and adult
females also kill cubs.
Intraspecific killing that is not sexually selected
appears
relatively common among grizzly bears. It has
also been documented
that
grizzly bears
are particularly
efficient predators
of juveniles of a variety of large
mammals (Larson et al. 1989, Gunther and Renkin
1990, Clarkson
and Liepins 1993, Adams et al. 1995,
Young and McCabe 1997, Bertram
and Vivion 2002).
This innate or learned predatory
behavior may trigger
adult males to focus particularly
on grizzly bear cubs
and yearlings. However, adult males and females that
kill cubs do not always eat them (S. Miller, personal
communication;
McLellan,
unpublished
data, 2004), so
the underlying
cause may vary and, in some cases, may
Ursus 16(2):141-156 (2005)
152 SEXUALLY SELECTED
INFANTICIDE
* McLellan
be sexually selected. The evolution of mate recognition
SSI, if it exists, likely has its genetic roots in predatory
behavior,
therefore the 2 ultimate causes of infanticide
(predatory
and sexually selected) are both adaptive
and
may remain intertwined.
The mate recognition hypothesis:
Scandinavia and Alberta
The results from the Flathead study and the SSI
simulation
model suggest that if SSI operates
in grizzly
bears,
it would fit the mate recognition
form,
but not the
immigrant male form as proposed by Wielgus and
Bunnell (1995, 2000), Swenson et al. (1997, 2001a),
and Swenson (2003). The mate recognition
form of SSI
generally fits the data presented
by Miller et al. (2003),
but does it fit the studies of Swenson et al. (1997,
2001a), (Swenson 2003), and Wielgus and Bunnell
(1995, 2000)?
In Scandinavia, Swenson et al. (1997 and 2001a)
compared
cub survival rates between 2 distant
popula-
tions (600 km apart) with different bear harvesting
histories. The northern
area (control
area) had no adult
males killed by hunters, whereas the southern area
(treatment
area) did have males killed (Swenson et al.
1997). Swenson et al. (1997, 2001a) concluded
that the
key factor in the relatively low cub survival in the
southern
treatment
area was immigration
of subadults
following the death of 1 or more established adult
males, and that resident males were not an important
factor in the loss of cubs. Recently, however, genetic
analysis of 2 males that killed cubs indicated
that they
were adult, resident males (Swenson 2003). Swenson
(2003) maintained that killing resident adult males
resulted
in increased
cub mortality.
In the Scandinavian
study, 1 reason hunters
did not
kill any adult males in the northern
control area was
because there were very few, if any, there
to kill. In the
northern
area, the probability
of a male surviving
from
a yearling to adulthood
(5 years) was 0.223, compared
with 0.499 for the southern
study area (Swenson et al.
2001a). Although there was little legal hunting,
Swenson (2003) found evidence of considerable
poach-
ing of bears
in the northern
area
(about
2.8x greater
than
the legal harvest). Swenson (2003) further suggested
that the few males present in the northern
control area
were young, and most first bred successfully as 3-year-
olds. The adult sex ratio in the northern
area clearly
favored
females, thus males would be expected to adopt
a searching
strategy
resulting in little or no infanticide
and high cub survival. Also, the young male bears
remaining
in the northern
area may have had difficulty
killing cubs defended by their mother (Swenson et al.
2001a)
The southern area had adult males and some were
shot by hunters.
However,
unlike hunted
populations
in
North America,
an equal number of females and males
were shot in Scandinavia
(Swenson et al. 1994). For
example, between 1985 and 2001 only 16.4% of the
bears legally harvested
in the southern
study area were
>5-year-old males (Swenson 2003), whereas 47.8% of
the bears legally killed by hunters
in the Flathead were
>5-year-old males. The hunted area in Scandinavia
likely had a relatively
even sex ratio
of adults
and thus,
as the model suggests, infanticide
by older males should
be more common
in the southern Scandinavian
area
than
in heavily hunted
North American
populations.
The high cub survival (0.96) in the northern
study
area
where there
were few large,
resident
males because
of high human-caused
mortality, combined with the
observation
of Swenson (2003) that
resident
males killed
cubs in the southern study area, supports the mate
recognition
form of SSI. It does not, however, preclude
the alternative
hypothesis that large bears kill small
bears for reasons other than SSI; significantly more
yearling
females were killed by other
bears (not SSI) in
the southern
Scandinavian
study
area,
where adult
males
were relatively common, than the northern
area, where
adult
males were rare
(Swenson et al. 2001b).
Wielgus and Bunnell (1995, 2000) compared data
from the dry, eastern
slopes of the Rocky Mountains
in
the Kananaskis
area
of Alberta
with data
collected in the
wet Selkirk
Mountains
in Idaho.
Hunting
was permitted
in the Alberta
study area
only in the last 3 years of the
5-year study but it was not permitted
in Idaho.
Wielgus
and
Bunnell
(1995, 2000) suggested
that
the hunting
and
consequent
removal of adult males caused an influx of
potentially infanticidal, young males; 9 3-7-year-old
males were captured
during
the last 3 years
of study.
The
authors suggested that the 4 adult females in Alberta
used suboptimal habitat to avoid the males and
consequently
had small litters
(x = 1.4, n = 5). In Idaho,
where there was no hunting (but 2 of 7 radiocollared
adult
males were killed by people;
Wielgus et al. 1994),
there was no apparent
immigration
of young males,
adult females were not forced into suboptimal
habitat,
and females had 2.2 cubs/litter
(n = 10).
For several reasons, the role and mechanisms of
infanticide
in the Alberta
study is more speculative
than
in Scandinavia.
First, infanticide
was not detected and
all cubs survived
(Wielgus and Bunnell 1994). Second,
as Miller et al. (2003) highlighted, samples sizes were
very small and included
related
individuals
(i.e., lacked
Ursus 16(2):141-156 (2005)
SEXUALLY
SELECTED
INFANTICIDE
* McLellan 153
independence).
In addition,
data for the main response
variables were collected over the entire 5-year study
when the treatment
(hunting) began in year 3. Most
(83%) of the telemetry
data
and all of the litter size data
were obtained before most (3 of 5) of the adult males
were killed. With much larger samples, Garshelis
et al.
(2005) found
the reproductive
rate
of grizzly bears
in the
Banff and Kananaskis
area remained
among the lowest
on the continent
although hunting
had been stopped in
the area,
and they suggested that nutritional
factors,
not
hunting,
was the cause of the low rate.
There are features
of the Alberta
study
area, however,
that may make SSI probable.
The study occurred
along
the continental divide, and rugged mountains with
relatively high grizzly bear densities (Boulanger
2001)
are found immediately
to the west, whereas
prairies
and
ranchland with virtually
no grizzly bears
are found
to the
east. All female and 11 of 14 male grizzly bears were
trapped
within 10 km of the continental
divide, and the
home range of every radiocollared bear crossed this
divide (Carr 1989). Perhaps
because the study area was
located along the edge of grizzly bear distribution,
and
because males have much larger home ranges than
females, 17 male but only 6 female bears were captured.
Although
the trapping
areas
were of similar
size, the 2.8
(M:F) ratio of captured
bears in the Alberta study was
significantly
different from the 0.9 ratio (13 M:15 F) in
the Idaho study (randomization
test P = 0.046), and
unlikely to be the same as the 1.28 (41 M:32 F) of
independent
bears in the Flathead
study area (random-
ization
test P = 0.075). The simulations
of SSI suggested
that
as the sex ratio of the population
shifts toward more
males, infanticide becomes increasingly profitable
because estrous females will be relatively rare. This
interpretation
leads to the hypothesis that along the
fringes of occupied grizzly bear range, particularly
where bear densities exhibit a marked
gradient
(high to
low) over a short distance, the incidence of infanticide
may be greater.
Management
implications
Wielgus and Bunnell (1995, 2000), Swenson et al.
(1997, 2001a), and Swenson (2003) suggested that the
immigrant
male hypothesis
of SSI operates
when grizzly
bear hunting
removes adult males. This hypothesis has
significant
implications
for the management
of hunted
bear populations.
For example, Swenson et al. (1997)
suggested
that
removing 1 or 2 adult
males over an area
of 11,200 km2
in a rapidly
increasing
population
caused
the 25-60 adult
females
to lose 11-25 cubs annually,
and
reduced the population growth rate (k) by 3.4%.
Conversely,
if SSI does not operate,
cub survival
should
increase once hunting
reduces
the population
size below
carrying capacity,
as was reported by Miller
et al. (2003).
The results of the Flathead study and the SSI
simulation modeling did not support the immigrant
male form of SSI but could not clearly differentiate
between mate recognition SSI and the hypothesis that
SSI does not operate in bears but that infanticide is
rooted in other factors, such as predation. If mate
recognition
SSI occurs in grizzly bears, then the model
suggests that its level of expression will depend on
a variety
of conditions.
If there
are
very few adult
males,
such as in the northern Scandinavian
population
where
males experienced high human-caused
mortality,
males
should not kill cubs and monitor the mother until she
becomes receptive, because doing so would cost them
mating opportunities
elsewhere. However, if hunting
practices
result in as many or more females being killed
than
males, such as in the southern Scandinavian
brown
bear
population,
SSI may be expressed
because it is less
likely that males would find estrous females that were
not heavily contested for by other males. Similarly,
on
the edge of the bear's
distribution,
such as in the Alberta
study area of Wielgus and Bunnell (1995, 2000), there
may be, on average, more males than females due to
males having larger
ranges. There, infanticide
may be
more likely, although
males would likely move to the
portion of their range with more females during the
mating season.
The model also suggests that
if searching
efficiency is
low, perhaps due to a low density of bears, then
infanticide may be profitable
when a male encounters
a female with cubs because the male may not otherwise
encounter an estrous female. If infanticide is indeed
greater
on the edge of occupied grizzly bear
habitat
and
where there is a low density of bears, then there are
important implications for grizzly bear conservation.
High adult female survival is not only important
for
continued
reproduction,
but also to keep the density of
adult females high enough to reduce infanticide. Pro-
ponents of bear hunting may use this information
to
promote
the harvest of males along the edge of grizzly
bear distribution
or in low-density populations. This
strategy
is not warranted
because mate recognition
SSI
remains an untested hypothesis at this time and, more
importantly,
hunters
also kill females.
If SSI operates
in grizzly bears,
it would only be one
factor
influencing
cub survival.
McLellan
(1994), Miller
et al. (2003), and Schwartz
et al. (2005) presented
data
suggesting
that
unhunted
populations
at or near
carrying
Ursus 16(2):141-156 (2005)
154 SEXUALLY
SELECTED INFANTICIDE
* McLellan
capacity have reduced recruitment due to high cub
mortality,
smaller
litters,
longer
intervals
between
litters,
or a combination
of these parameters.
Hunting
reduces
the population
size of grizzly bears and usually, at least
in North America, also reduces the proportion
of adult
males. These hunted populations
have greater
recruit-
ment than populations at carrying capacity. What
remains
unclear is the form of the relationship
between
population size with respect to carrying capacity and
recruitment,
and whether
the relationship
differs
depend-
ing on the sex ratio of harvested bears. It is unlikely
that
this relationship
is linear.
Rather,
it likely changes more
rapidly
near
carrying capacity
(Taylor
et al. 1994). More
uncertain are the implications of the sex ratio of the
harvest. My results suggest that if more males are
removed
than females, cub survival
will be greater
than
if the sex ratio
of the harvest
is equal or favors females.
However,
the form of relationship
between the sex ratio
of the living population and cub survival remains
uncertain.
The results of the Susitna
experiment
(Miller
et al. 2003) suggest that
removing
more males does not
affect recruitment
once the sex ratio is <40% males.
Predictions
of the mate recognition
hypothesis
of SSI
may be tested in the important
Scandinavian
studies. If
poaching can be curtailed
in the northern
study area,
allowing the density of adult males to increase
relative
to that of females, then cub survival should decrease.
Similarly,
if, through regulation
or education,
hunters
in
the southern
area can kill mostly adult males and not
females, then cub survival should eventually increase.
However, both of these populations
have been rapidly
increasing
for many years (Swenson
et al. 1997, 2001a),
so density effects may eventually
affect cub survival.
Acknowledgments
Support was provided by the BC Ministry of
Environment, US Fish and Wildlife Service, Shell
Canada Ltd., University of British Columbia, Simon
Fraser
University, BC Forest Service, Forest Renewal
BC, Canadian
Wildlife Foundation,
National Fish and
Wildlife Foundation
(USA), Boone and Crockett
Club,
National
Rifle Association (USA), World Wildlife Fund
(Canada), Canadian Wildlife Service University Re-
search Support
Fund, East Kootenay Operators
(7 BC
Forestry
Companies),
Plum Creek
Timber
Ltd., Crows-
nest Resources Ltd., Sage Creek Coal Ltd., BC Guides
and Outfitters,
and Safari International
(BC Chapter).
Field assistants requiring special thanks include D.
Carey, R. Heggs, D. Horing, R. Mace, B. Noble,
T. Radandt,
I. Teske, and T. Thier. I thank L. Kolter
for reports
on bears in zoos and S. Miller, D. Sellers,
C. Apps, J. Clark, J. Swenson, M. Festa-Bianchet,
F. van Manen, and an anonymous
ecologist for excel-
lent reviews and thought-provoking
comments.
Literature cited
ADAMS,
L.G.,
F.J.
SINGER,
AND
B.W.
DALE.
1995. Caribou calf
mortality
in Denali National
Park,
Alaska.
Journal
of
Wildlife
Management
59:584-594.
ALCOCK,
J. 2001. Animal behavior: an
evolutionary approach.
Seventh edition. Sinauer Associates, Sunderland,
Massa-
chusetts,
USA.
APPS,
C.D., B.N. MCLELLAN,
J.G. WOODS,
AND
M.F. PROCTOR.
2004. Estimating
grizzly bear distribution and abundance
relative
to habitat and human influence. Journal of Wildlife
Management
68:138-152.
BERTRAM,
B.C.R. 1975. Social factors
influencing
reproduction
in wild lions. Journal
of Zoology, London 177:463-482.
BERTRAM, M.R., AND
M.T. VIVION. 2002. Moose mortality
in
eastern
interior Alaska. Journal
of Wildlife
Management
66:747-756.
BLANCHARD,
B.M., AND
R.R. KNIGHT. 1991. Movements of
Yellowstone grizzly bear. Biological Conservation 58:
41-67.
BOULANGER,
J. 2001. Analysis of the 1997 Elk Valley and
Flathead Valley DNA mark-recapture grizzly bear in-
ventory projects.
BC Ministry
of Environment,
Lands, and
Parks,
Cranbrook,
British
Columbia,
Canada.
BUNNELL,
F.L., AND
D.E.N. TAIT.
1981. Population
dynamics
of bears-implications. Pages 75-98 in C.W. Fowler and
T.D. Smith,
editors.
Dynamics
of large
mammal
popula-
tions.
John
Wiley
and
Sons,
New York,
New York,
USA.
CARR, H.D. 1989. Distribution,
numbers, and mortality of
grizzly bears in and around
Kananaskis
Country,
Alberta.
Wildlife Research Series 3. Alberta Forestry, Lands and
Wildlife. Fish and Wildlife Division, Edmonton,
Alberta,
Canada.
CASE, R.L., AND L. BUCKLAND.
1998. Reproductive character-
istics of grizzly bears in the Kuglucktuk area, Northwest
Territories, Canada. Ursus 10:41-47.
CLARK, J.D., AND K.G. SMITH. 1994. A demographic
comparison of 2 black bear populations in the interior
highlands of Arkansas. Wildlife Society Bulletin 22:
593-603.
CLARKSON,
P.L., AND
I.S. LIEPINS.
1993. Grizzly bear, Ursus
arctos, predation on muskox, Ovibos moschatus, calves
near the Horton River, Northwest Territories.
Canadian
Field Naturalist
107:100-102.
, AND . 1994. Grizzly bear population estimate
and characteristics
in Anderson
and Horton
rivers
area,
Northwest Territories,
1987-89. International
Conference
on Bear Research
and Management
9(1):213-221.
Ursus 16(2):141-156 (2005)
SEXUALLY
SELECTED
INFANTICIDE
* McLellan 155
CRAIGHEAD,
J.J., J.S. SUMNER,
AND J.A. MITCHELL.
1995a. The
grizzly bears of Yellowstone: Their ecology in the
Yellowstone ecosystem, 1959-1992. Island Press, Wash-
ington, DC, USA.
CRAIGHEAD,
L., D. PAETKAU,
H.V. REYNOLDS,
E.R. VYSE,
AND
C. STROBECK.
1995b. Microsatellite
analysis of paternity
and
reproduction
in arctic
grizzly bears.
Journal
of Heredity
86:255-261.
DATHE,
V.H. 1961. Beobachtungen
zur
Fortpflanzungsbiologie
des Braunbaren,
Ursus arctos L. Der Zoologische Garten
25:235-269 (In German.)
EBENSPERGER,
L.A. 1998. Strategies and counterstrategies to
infanticide in mammals.
Biological Review 73:321-346.
GARSHELIS,
D.L., M.L. GIBEAU,
AND
S. HERRERO.
2005. Grizzly
bear demographics
in and around
Banff National Park
and
Kananaskis
Country,
Alberta.
Journal
of Wildlife Manage-
ment 69:277-297.
GNIADEK, S.J., AND K.C. KENDALL. 1998. A summary of bear
management in Glacier National Park, Montana, 1960-
1994. Ursus 10:155-159.
GUNTHER, K.A., AND R.A. RENKIN. 1990. Grizzly bear
predation
on elk calves and other fauna of Yellowstone
National Park.
International Conference on Bear Research
and Management
8:329-334.
HESSING P., AND L. AUMILLER.
1994. Observations of
conspecific predation by brown bears, Ursus arctos, in
Alaska. Canadian
Field-Naturalist
108:332-336.
HOOD,
G. M. 2004. PopTools.
Version
2.6.4., Albany,
Western
Australia,
Australia.
http://www.cse.csiro.au/poptools.
HOVEY, F.W., AND B.N. MCLELLAN. 1996. Estimating
population
growth
of grizzly bears
from the Flathead
River
drainage
using computer
simulations of reproductive
and
survival
rates.
Canadian
Journal of Zoology 74:1409-1416.
HRDY,
S.B. 1974. Male-male competition and infanticide
among the langurs
(Presbytis
entellus) of Abu, Rajasthan.
Folia Primatologica
22:19-58.
. 1977. The langurs
of Abu: female and male strategies
of reproduction.
Harvard University Press, Cambridge,
Massachusetts,
USA.
. 1979. Infanticide
among animals:
a review, classifi-
cation, and examination of the implications for the
reproductive
strategies
of females. Ethology and Sociobi-
ology 1:13-40.
HUCK, U.W., R.L. SOLTIS,
AND C.B. COOPERSMITH.
1982.
Infanticide in male laboratory
mice: effects of social status,
prior sexual experience, and basis for discrimination be-
tween related and unrelated
young. Animal Behavior 30:
1158-1165.
INSLEY,
S.J. 2000. Long-term
vocal recognition
in the northern
fur seal. Nature
406:404.
JANSON,
C.H., AND C.P. VAN SCHAIK. 2000. The behavioral
ecology or infanticide
by males. Pages 469-497 in C.P. van
Schaik and C.H. Janson, editors. Infanticide by males
and its implications. Cambridge
University Press, Cam-
bridge, UK.
KOVACH, A.I., AND
R.A. POWELL.
2003. Effects of body size on
male mating tactics and paternity
in black bears, Ursus
americanus.
Canadian
Journal
of Zoology 81:1257-1268.
LABOV,
J.B. 1980. Factors influencing infanticidal
behavior
in wild male house mice (Mus musculus). Behavioural
Ecology and Sociobiology 6:297-303.
LARSON, D.G., D.A. GAUTHIER,
AND
R.L. MARKEL.
1989.
Causes and rate of moose mortality in the southwest
Yukon. Journal
of Wildlife Management
53:548-557.
LECOUNT,
A.L. 1983. Evidence of wild black bears breeding
while raising cubs. Journal
of Wildlife Management 17:
264-268.
MACE, R.D., AND
J.S. WALLER.
1997. Spatial and temporal
interaction
of male and female grizzly bears
in Northwest-
ern Montana.
Journal
of Wildlife Management
61:39-52.
, AND . 1998. Demography and population
trend of grizzly bears in the Swan Mountains,
Montana.
Conservation
Biology 12:1005-1016.
MACHUTCHON, A.G., S. HIMMER,
AND
C.A. BRYDEN.
1993.
Khutzeymateen
valley grizzly bear study. Final Report.
Wildlife Habitat
Research
Report 31. Ministry of Forests
Research
Program,
Victoria,
British
Columbia,
Canada.
MALLORY,
F.F., AND
R.J. BROOKS.
1978. Infanticide
and other
reproductive
strategies
in the collared lemming, Dicrosto-
nyx groenlandicus.
Nature
273:144-146.
MARTINKA,
C.J. 1974. Population characteristics
of grizzly
bears in Glacier National Park, Montana. Journal of
Mammalogy
55:21-29.
MATSON,
G.M., L. VAN
DAELE,
E. GOODWIN,
L. AUMILLER,
H.
REYNOLDS,
AND
H. HRISTIENKO.
1993. A laboratory
manual
for cementum age determination
of Alaskan brown bear
first premolar teeth. Matson's Laboratory, Milltown,
Montana,
USA.
MATTSON,
D.J., R.R. KNIGHT,
AND
B.M. BLANCHARD.
1987.
The effects of developments
and primary
roads on grizzly
bear habitat
use in Yellowstone National Park,
Wyoming.
International
Conference on Bear Research and Manage-
ment 7:259-273.
MCCANN,
R.K. 1997. Kluane National Park grizzly bear
research
project. Year end report 1996. Kluane National
Park and Reserve, Haines Junction, Yukon Territory,
Canada.
MCLELLAN,
B.N., AND
D.M. SHACKLETON.
1988. Grizzly bears
and resource-extraction
industries: effects of roads on
behavior,
habitat
use, and demography.
Journal of Applied
Ecology 25:451-460.
. 1989. Population
dynamics of grizzly bears during
a period of resource extraction: I. Density and age/sex
structure.
Canadian
Journal
of Zoology 67:1856-1860.
. 1994. Density-dependent
population regulation of
brown bears. Pages 15-24 in M. Taylor, editor. Density-
dependent
population
regulation
of brown,
black, and
polar
bears. International
Conference on Bear Research and
Management,
Monograph
Series 3.
Ursus 16(2):141-156 (2005)
156 SEXUALLY
SELECTED INFANTICIDE
* McLellan
, F. HOVEY,
J. WOODS,
M. GIBEAU, D. CARNEY,
R.
MACE,
W. WAKKINEN,
AND W. KASWORM. 1999. Rates and
causes of grizzly bear
mortality
in the interior
mountains
of
British
Columbia, Alberta,
Montana,
and Idaho. Journal
of
Wildlife Management
63:911-920.
, AND . 2001a. Habitats selected by grizzly bears
in a multiple use landscape.
Journal
of Wildlife Manage-
ment 65:92-99.
, AND . 2001b. Natal dispersal
of grizzly bears.
Canadian
Journal of Zoology 79:838-844.
MCLOUGHLIN, P.D., M.K. TAYLOR,
H.D. CLUFF, R.J. GAU, R.
MULDERS, R.L. CASE, S. BOUTIN, AND F. MESSIER. 2003.
Demography of barren-ground grizzly bears. Canadian
Journal
of Zoology 81:294-301.
MILLER,
S.D., R.A. SELLERS,
AND J.A. KEAY.
2003. Effects of
hunting on brown bear cub survival and litter size in
Alaska. Ursus 14:130-152.
PROCTOR,
M.F., B.N. MCLELLAN,
C. STROBECK,
AND R.M.R.
BARCLAY.
2004. Gender-specific dispersal distances of
grizzly bears estimated by genetic analysis. Canadian
Journal
of Zoology 82:1108-1118.
SCHWARTZ,
C.C., K.A. KEATING,
H.V. REYNOLDS,
III, V.G.
BARNES, JR., R.A. SELLERS,
J.E. SWENSON,
S.D. MILLER,
B.N. MCLELLAN,
J. KEAY,
R. MCCANN,
M. GIBEAU,
W.F.
WAKKINEN,
R.D. MACE, W. KASWORM,
R. SMITH,
AND
S. HERRERO.
2003. Reproductive maturation
and senes-
cence in the female brown bear. Ursus 14:109-119.
, M.A. HAROLDSON,
AND
S. CHERRY.
2005. Survival of
cub and yearling grizzly bears in the Greater
Yellowstone
Ecosystem, 1983-2001. Pages 69-87 in C.C. Schwartz,
M.A. Haroldson,
G.C. White, R.B. Harris,
S. Cherry,
K.A.
Keating, D. Moody, and C. Servheen, authors.
Temporal,
spatial,
and environmental
influences
on the demographics
of grizzly bears in the Greater
Yellowstone Ecosystem.
Wildlife Monograph
161.
STRINGHAM,
S.F. 1980. Possible impacts of hunting on the
grizzly/brown bear, a threatened species. International
Conference
on Bear Research
and
Management
4:337-349.
.1983. Roles of adult
males in grizzly bear
population
biology. International
Conference on Bear Research and
Management
5:140-151.
SWENSON,
J.E., F. SANDEGREN,
A. BJARVALL,
A. SODERBERG,
P.
WABAKKEN,
AND R. FRANSEN. 1994. Size, trend, distribution
and conservation of the brown bear Ursus arctos
population
in Sweden. Biological Conservation
70:9-17.
, A. SODERBERG,
A. BJARVALL, R. FRANZEN, AND P.
WABAKKEN.
1997. Infanticide caused by hunting of male
bears. Nature
386:450-451.
, AND . 1998. Geographic expansion of
an increasing brown bear population: evidence for
presaturation dispersal. Journal of Animal Ecology 67:
819-826.
, ,S. BRUNBERG,
AND
P. SEGERSTROM. 2001a.
Factors associated with loss of brown
bear cubs in Sweden.
Ursus 12:69-80.
, B. DAHLE,
AND F. SANDEGREN.
2001b. Intraspecific
predation
in Scandinavian
brown bears
older than cubs-of-
the-year.
Ursus 12:81-92.
. 2003. Implications
of sexually selected
infanticide
for
the hunting
of large
carnivores.
Pages 171-189 in M. Festa-
Bianchet and M. Apollonio, editors. Animal Behaviour
and Wildlife Conservation. Island Press. Washington,
DC, USA.
TAYLOR,
M. 1994. Density-dependent
population
regulation
of
brown, black, and polar
bears.
International
Conference
on
Bear Research
and Management
Monograph
3.
VAN SCHAIK,
C.P. 2000. Infanticide by male primates: the
sexual selection hypothesis
revisited.
Pages 27-60 in C.P.
van Schaik and C.H. Janson,
editors.
Infanticide
by males
and its implications. Cambridge
University Press, Cam-
bridge, UK.
WIELGUS, R.B., AND F.L. BUNNELL. 1994. Dynamics of a small,
hunted
brown bear Ursus arctos population
in southwest-
ern Alberta,
Canada.
Biological Conservation
67:161-166.
, W.L. WAKKINEN, AND P.E. ZAGER. 1994.
Population dynamics of Selkirk Mountain grizzly bears.
Journal
of Wildlife Management
58:266-272.
, AND . 1995. Tests of hypotheses for sexual
segregation
in grizzly bears. Journal
of Wildlife Manage-
ment 59:552-560.
, AND . 2000. Possible negative effects of adult
male mortality
on female grizzly bear reproduction.
Bio-
logical Conservation
93:145-154.
YOUNG, D.D. JR., AND T.R. MCCABE. 1997. Grizzly bear
predation
rates on caribou
calves in Northeastern
Alaska.
Journal
of Wildlife Management
61:1056-1066.
Received: 8 July 2004
Accepted: 7 July 2005
Associate Editor: F. van Manen
Ursus 16(2):141-156 (2005)
