The Effects of Density-Dependent Resource Limitations on the Demography of Wild Reindeer
Author(s): T. Skogland
Source: The Journal of Animal Ecology, Vol. 54, No. 2 (Jun., 1985), pp. 359-374
Published by: British Ecological Society
Stable URL: http://www.jstor.org/stable/4484
Accessed: 28/08/2008 16:48
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=briteco.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.
http://www.jstor.org
Journal of Animal Ecology (1985), 54, 359-374
THE EFFECTSOF DENSITY-DEPENDENTRESOURCE
LIMITATIONSON THE DEMOGRAPHYOF WILD
REINDEER
BY T. SKOGLAND
D VF Viltforskningen(Directorate of Wildlifeand FreshwaterFish, Research Branch),
Tungasletta2, N-7000 Trondheim,Norway
SUMMARY
(1) The population dynamics of eight wild reindeerherds in Norway, includingthe high
arctic, were studiedin relationto food abundanceand populationdensity.
(2) Density-dependenteffects were manifestedthroughwinterfood limitation.
(3) With increasing densities, density-independentfactors such as severe weather had
additionaleffects on recruitment,with consequences for population stability.
(4) The density effect took the form of decreasedjuvenile survival, while adult survival
in the most abundantand reproductiveage classes was not affected.
(5) At high density, subadultfecundityfell below the thresholdbody size for maturity.
(6) The fecundity of adults was not affected by density. This finding accords with those
made for other large nomadic ratherthan sedentaryungulates.
(7) Density-dependentfood limitation was most apparent at high population density,
which conforms with the predictions of density-dependentlife history models for large
mammals.
INTRODUCTION
The general conclusion of several evolutionary arguments (i.e. Fowler 1981a,b and
references therein) is that populations of large animals should show their main densitydependent changes either at high population levels, or close to the carrying capacity. A
limitationof food resources can have an importantdensity-dependenteffect on populations
and is the predominantfactor regulating herbivore numbers (Laws, Parker & Johnstone
1975; Sinclair 1977, 1979). The demographic consequences of density dependence, or
resource limitation, in most of the recorded instances, has taken the form of a decrease in
infant survival, or of a decrease in fecundity, while adult survival has been less affected
(Caughley 1970; McCullough 1979; Clutton-Brock,Guinness & Albon 1982; Sinclair &
Norton-Griffiths1982; Sauer & Boyce 1983).
Bergerud (1980) has suggested that populations of North American caribou (Rangifer
tarandus L.) are mainly regulatedby predationby wolves (Canis lupus L.), in conjunction
with hunting by humans, although Haber & Walters (1980) argued that food supply, wolf
predation and hunting by man produce multiple equilibriumstates in caribou numbers.
Recent studies on other ungulates have stressed the importance of the food supply for
migratorypopulations (Houston 1982; Sinclair& Norton-Griffiths1982).
In this study I tested the hypothesis that numbers of wild reindeer (Rangifer tarandus
tarandus L, R.t. platyrhynchos L) are regulated by food levels in a density-dependent
manner and that these changes occur mainly at high population levels or close to their
359
360
Reindeerpopulation regulation
carrying capacity, as is generally predictedfor large mammals. No wolf populations were
present in any areas includedin this study.
Study areas
A description of the study areas (Fig. 1) and of each of the herds on the mainland in
Norway is given by Skogland (1983) and of the one in Reindalenon the Svalbardisland in
the arctic by Alendal & Byrkjedal(1975). The vegetation in the areas includedin the study
is either typical mountain tundra (Nordhagen 1943) or arctic tundra (Brattbakk &
Ronning 1978). Data on habitat characteristicsand population sizes are given in Table 1.
The South Norwegian herds are now divided into separate subpopulations (Skogland
1983). The Reindalen subpopulation of the Spitsbergen reindeer (Rangifer t.
platyrhynchos) is of particular interest because of its high density and its apparently
overgrazed winter pastures. The herd has been protected from hunting since 1920
(Reimers 1983b). The Brattefjell-Vindeggenherd was protecteduntil 1980.
FIG. 1. Map of the location of the reindeer herds studied in Norway. The insert shows the study
locality on the island of Svalbard at 78?N. Study herds: (1) Forelhogna; (2) Knutsho; (3)
Snohetta; Solenkletten; (5) Rondane; (6) Hallingskarvet; (7) Hardangervidda; (8) BrattefjellVindeggen; (9) Reindalen.
T. SKOGLAND
361
of the differentstudyareas
TABLE1. Habitatcharacteristics
Meanlichen Biomass
Ratioof
Total Meanwinterhabitats
of total winter:summer drybiomass sample
area (proportion
sizes
g-2 + S.E.
grazingarea
Herdarea
(km2)* area)t% ? S.E.
46
557 17
0-79
31.7 7.3
1600
Knutsho
500
0.67
27.7 8.6
1600
Forelhogna
30 (1965)t
0.23
14.3
2750
Snohetta
59
97.4 2.9
2.35/
34.5/
1200/
Solenkletten/
500
0.7
20.3 16.6
1500
RondaneSor134 62(1970)?
0.3
12-1
8000
Hardangervidda
189-5 51(1979)
44
290 21
0-7
23.8 4.8
445
Brattefjell-Vindeggen
39
194 37
0.57
27.1 9
2384
Hallingskarvet
0-15
25l
14.9
115
Reindalen
* Onlyvegetatedareasincluded,datafromthe arearecordson fileat DVF.
t FromSkogland(1984);Skogland(1983).
t FromGaare& Skogland(1980).
? FromGaare& Skogland(1979).Lichendensitycalculatedfromvolumesas giventherein.
& Ronning(1978)andthisstudy.
s FromBrattbakk
METHODS
Population size
Herd size was estimated by making total counts from small aeroplanes. Complete
coverage of each study area was obtained by flying along a number of parallel traverses
across the area. All herds were photographed and the number of animals visible on the
enlargedphotographswere counted. Owing to its scattereddistribution,the Reindalenherd
was counted on the ground once every year, duringthe first week of August, instead of by
the aerial method. The aerial counts were done by either members of the research staff at
DVF, or in cooperation with the local game wardens.
Fecundity
Fecundity was estimated by (i) culling does in the winter duringthe gestation period and
(ii) counting live does considered to be pregnantjust prior to the calving time. The counts
involved checking the external appearance of the does, particularlyantler possession, size
of abdominalregion as well as swollen udders. Approximately 1%of all the does in each of
the study herds never developed antlers. Non-pregnant does dropped their antlers prior to
the calving period (see also Espmark 1971). During seventeen calving seasons less than 1%
of the does that gave birthwere without antlers (T. Skogland unpublished).
Recruitmentand calf survival
Recruitmentwas estimated by field counts of the ratio of the numberof live calves to the
number of live does > 1 year of age. The counts were done 1 month after calving, during
the autumn composition counts in the rut, and at the end of the calf's first winter. Calf
survival was estimated by the change-in-ratio(CIR) method. The degree of variance of the
survivalrate estimates given by the CIR method was calculatedby Ricker's (1975) method.
The confidence limits for all the survival rates were calculated on a basis of binomial
sampling (Snedecor & Cochran 1978). Recruitment rates were also calculated from
age-specificsurvivalrates (see below).
362
Reindeerpopulation regulation
Ageing and age structures
Age was estimated from tooth eruption patterns for animals 1-3 years of age, and for
older animals by microscopic inspection of the annuli in the stained incisors from each
jawbone, accordingto the method of Reimers & Nordby (1968).
Age structureswere constructed for the doe cohorts > 2 years of age from collections of
jawbones from animals shot by hunters during the regular hunting season. Because of the
impossibility of ageing a doe from her external appearance I assumed that does would be
shot in direct proportion to their availability in the herd (see also Burgoine 1981). Age
structure of the protected herd in Reindalen was constructed from pick-up samples from
naturaldeaths of adults > 1 year of age duringfield searches.
The age of bucks 1 5-3.5 years old was estimated from their externalappearanceduring
the autumn composition counts in the rut. Because of the wide variation in the size and
shape of adult bucks, most hunters were unable to distinguish clearly in the field their
differentage-groups. Therefore,I have assumed that ageing bucks as > 4.5 years from the
jaw collections provided a fairly representativepicture of the age composition of adult
bucks.
Analysis of density effects on recruitment was by least squares regressions of
recruitment rates against population densities. Furthermore, a multiple least squares
regressionwas calculated with recruitmentrate against winter density and lichen biomass.
Life-table analyses
The age-class frequency data for the adults were converted into logarithmicvalues. The
reasons for using log values are given by Varley & Gradwell (1970). The data for the
standing age-classes yielded age-specific survival rates. Since population growth rates were
known for all the herds, based on the counts, the age-specific rates were converted into
time-specific rates by the Nxerx conversion method (Caughley 1977 and see Table 3),
where Nx is the log of the standing age-class frequency and r is the populationgrowth-rate.
The resulting time-specific rates were further smoothed by making polynomial regression
analyses for best fit. The critical assumption made in these calculations is that all herds
showed stable increase, or decrease, or that their numbersremainedstable over a period of
time that included most of the age-classes present. When population development was
more unstable,the age-specificsurvivalrate (s) was calculated from the equation:
s =J
1-
where J = annual hunting rate (proportion of the population shot); A = finite population
increment; z = instantaneous age-class decrement; z was derived from the age-specific
rates by calculating either the regression coefficient for decrement against age (for large
samples), or the value
Nx
N2
x2
for samples that were either too small, or that were too variablein size owing to fluctuating
annual recruitmentrates. The resulting coefficients (b) were converted into instantaneous
rates from the equation z = (1 - e-b) (Ricker 1975, see Table 3).
A buck survival curve could be constructed only for the Brattefjell-Vindeggenherd,
because after many years of buck shooting, the adult sex ratios of other mainland herds
had become skewed in favour of does.
363
T. SKOGLAND
Food availability in winter
Winter food availabilitywas assessed by noting the plant composition and the height of
the plant cover within 33 x 100 cm plots. The plots were randomly placed within the
typical wintering areas. Due to the snow cover on the tundra, only one particular
vegetation type, the Loiseleurio Arctostaphylion(see Nordhagen 1943), is available to the
reindeerfor grazing in late winter. The thickness and height of the lichen mats on the xeric,
chionophobous heaths therefore give a good indication of the actual foraging conditions.
From the degree of coverage and the height of the lichen mats I calculated the forage
volume. I convertedthis to biomass by the methods given by Gaare & Skogland (1979).
RESULTS
Fecundity rates
Data from culls do not show any density effect on the fecundity of the adult does, but
with increasing population density the 1.5 and 2.5 year-old cohorts graduallybecome less
productive (Fig. 2). Since a heavy decline in calf productivity is found over the same
density gradient (see Fig. 4), the youngest cohorts represent a smaller proportion of the
total population of does at increasingly higher densities. Field counts of pregnant does
show only a small, but still significant,decline in overall fecundity with increasing density
(R2 - 0.43; F 12-49; d.f. = 1,20; P < 0.01; Fig. 3). The decline is slight mainly because
of the absence of such a density effect on the fecundityrate of the adult does.
It is thus clear that the difference between the fecundity rate and the recruitmentrate
becomes significantlygreater with increasing herd density. The differencewas of the order
of 67% (Table 2).
Differences betweenherds in recruitment
The difference between the least and the most productive herds (Reindalen and
Forelhogna-Knutsho, respectively) was almost fourfold (Table 2). The variance in the
recruitmentrates increasedwith decreasingoverall productivity(Table 2). Since differences
(I) ?;.
10090
(4)
(2) (3)
"-'"
', ^ ;'y"
...-------..>
3-5 years
80>8 70\
C 60-
2.5 years
'
a)
50u
i0301.5years
2010I
i
4
3
2
1
Reindeer density (n km-2)
FIG.2. Age-specificfecundityrates for four populationdensities.Numbersabove the graphs
n = 44; (2), Hallingskarvet,n = 295; (3),
indicatesampleherds:(1), Knutsho-Forelhogna,
n = 73; (4), Snohetta,n = 53. For (4) rates are calculatedfrom Reimers
Hardangervidda,
(1983a).
Reindeerpopulation regulation
364
1-0-
=66
0-9- =
o-
0.8-
N=6
N=4
0.7--
N=
0)
0o
0-3-
0-5 -
I
I
I
I
I
4
2
3
Reindeer numbers km-2
I
5
I
6
FIG.3. The ratioof antleredto unantlereddoes priorto calving,as an estimateof population
fecundityratesat differentpopulationdensities.N indicatesthe numberof yearsin whichdata
werecollected.See alsoTable2. Themeanandrange(verticalbars)aregiven.
existed in the age and level of first-timefecundity of the herds, the calf recruitmentrates
were analysed in relation to the proportion of the first-years that were fecund to see
whether this was the factor accounting for the variation. It accounted for most of the
year-to-year variation only in the highly productive herds (Forelhogna-Knutsho, R2 =
0.84; n = 6; P< 0.02), creating a cycle of productivity. For the less productive herds
(Hardangervidda,Snohetta, Reindalen), no such statistically significant relationship was
found (R2 - 0.1; n = 6; P > 0.1). Apparently,environmentalperturbationsin additionto
other factors affect herds of low productivity.Disregardingboth this cycle generatedby the
age-structure, and the oscillations in recruitment caused by other external factors, the
overall differencein recruitmentbetween the highly and the less-productiveherds was not
accounted for by this analysis.
The recruitmentrates, obtained from direct field observations (Table 2) can be compared
with those calculated by the age-specificsurvival method, i.e. from the standing age-classes
of does (Table 3), divided by an average sex ratio of 0.5. For all herds, except the
Hallingskarvetone, the rates given by the two methods overlapped.Calf productionby the
Hallingskarvet herd collapsed in 1982, following a period of overgrazing of the winter
pastures that culminated in 1981. Before then calf production of this herd had been much
higher, around sixty calves: 100 does > 1 year of age (J. Vik & 0. Dusegard personal
communication).The standing age-class distributionof this herd thereforedid not reflect a
loweringof the recruitmentrate.
Effects of density on recruitment
The recruitmentrates were lowest at high overall densities (R2 = 0.87; F= 221.4, d.f. =
1,32; P < 0*001; Fig. 4), and were related to both winter density and lichen biomass (see
Table 1) in the multiple regression (R2 = 0.87; P < 0.001). The correlation coefficient
for both winter density and lichen biomass with recruitmentrate was significant (P <
0.05). For two populations the recruitment rate was significantly lower at high winter
populationdensities (R2 = 0 88; F = 93.6; d.f. = 1,12; P < 0-001; Fig. 5).
The herds of Reindalen on Svalbardhad a low finite rate of increase of 6.4% (Table 3).
In another part of Svalbard, Nordenskioldkystenon Nordenskiold Land, a herd of eleven
immigrantreindeerincreased to c. 1000 in 17 years, with a finite rate of increase of 27%
TABLE 2. Mean fecundity, annual recruitment, and first-year survival rates for the herds in t
Herd
Fecundity rate, does
>1 year of age
N Mean
C.V.
Reindalen?!
(3)
Hardangervidda(7)
Snohetta
(5)
Hallingskarvet (1)
(1)
BrattefjellVindeggen
Knutsho
(6)
N
Post-natal rate
C.V.
Mean
Mortality rates
Autumn rate?
C.V.
N
Mean
0-5
0.64
0.6
0-57
0.57
(0-16-0-18)
(0-03-0-07)
(0-03-0-08)
+ 0-01t
0-06
(9)
(8)
(2)
(4)
0.4***
0.37***
0.44***
0-49***
(0-02-0-11)
(0-06-0-09)
+ 0-06
(0-1-0-16)
0-67
(0-06-0-09)
(6) 0-64N.S.
(0-05-0-14)
(5)
(7)
(8)
(1)
0-44N.S.
0.34***t
0-34N.S.
0.16***
(0-03-0-18)
(0-02-0-1)
(0-07-0-1)
+0-03
N
(11)
(7)
(8)
(1)
(4)
Recruitment rate
Mean
C.V.
0- 18***
0.32N.S.
0-33N.S.
0-13N.S.*
0-49N.S.
(6) 0.6N.S.
N = numberof sample years and C.V. = range of values of the coefficient of variation
N.S. = insignificantdecrement in rate compared to that for the preceding sampling period in the same year
*** =p < 0-001
* total count
t 95% confidence level on sample mean
t significantin 1 year
? rates adjustedfor hunting losses
'bFor the Reindalenherd, data from Alendal & Byrkjedal (1975) and Gossow (1980) are included.
(0-11-0-2
(0-02-0-0
(0-07-0-1
(0-1-0-16
(0-7-0-11
366
Reindeerpopulation regulation
TABLE 3. Estimated survival and recruitmentrates for the studied herds
Herd
Sample
year
Instantaneous
age-specific
decrement
z
n?
Correlation
coefficients
P
Age-classes
in least
squares
regression
(years)
30 -0-281*
Forelhogna
(1980)
169 -0-302t
0.995
<0-001
Forelhogna
(1969)
62 -0.31t
Knutsho
0.995
<0-001
(1981)
184 -0-27t
Rondane Sor0.995
<0.001
(1981)
Solenkletten
295 -0-275t
0-995
<0.001
(1981)
Hallingskarvet
78 -0-267t
B. Vindeggen
0-959
<0-001
(1981)
188 -0-224t
<0-001
0-985
(1979)
Hardangervidda
349 -0.19t
0.959
Snohetta
<0-001
(1981)
29 -0.124*
Reindalen
(1981)
* Age-specific rate =
N,/11, N,; N, is the standing age-class frequency.
t From the regression coefficient (b). z is given by z = (1 - e-)
T From Fig. 6.
? n is sample size.
80 -
2
Finite
rate of Time-specific
increase survival rate
(r)i
(p )
0.28
0-97
(4-10)
(4-9)
(3-13)
0-3
0.254
0.954
0.949
(3-12)
(3-10)
(2-12)
(2-15)
0-28
0-237
0.23
0.19
0.064
0.931
0.963
0.936
0-933
0-92
0o
70-X0
60A\ 50o
40-
o
30-
a)
\
^
?o
Oo0
210-
0
A
I
2
3
4
5
6
Reindeer density km-2
FIG. 4. The effects of different population densities on the ratio of calves recruited per 100
does > 1 year-old. Open circles (0) indicate data from mainland Norway while triangles (A)
indicate data from Svalbard. The data sources are the same as those in Tables 1 and 2.
(K. Bye personal communication; N. 0ritzland personal communication, own observations; Alendal & Byrkjedal 1975). A third group, of thirteenreindeer,was transplanted
by N. 0ritzland from the overgrazed habitats near Longyearbyen on Svalbard to the Ny
Alesund area, where no reindeerhave ever been recorded. This herd has grown at a finite
rate of more than 30%, since the start.
The physiognomy of the vegetation cover of all these three areas of Svalbardis similar,
but each is at a different stage of winter grazing pressure. In the Reindalen area all the
lichens have been eliminated and soil erosion has reached an advanced stage. At
Nordenskioldkystenextensive mats of lichens (Cetraria spp.) are still present (I. Brattbakk
personal communication;own observations)and at Ny Alesund the winter forage biomass
of the habitat is of the order of 5-12 times greater than that in the Reindalen area
(Brattbakk& Ronning 1978).
367
T. SKOGLAND
? 701
60-
A\ 50o
40-
o 30 o
a
'
20
10I
I
I
I
I
I
I
I
I
I
I
2
4
6
8
10
12
14
16
18
20
22
Reindeer density
!
24
I
I
26
28
per winter habitat area
FIG. 5. The mean ratio of calves recruited per 100 does in relation to mean population densities
per winter habitat area. All data are from Table 1 and Fig. 4. The mean and range (vertical bars)
are shown. (0) Knutsho-Snohetta; (0) Hardangerviddaregion.
These results on recruitmentsupport the hypothesis that there is a density-dependent
food limitation effect acting on the productivityof the wild reindeerherds included in this
study.
Calf survival rates
Neonatal mortality was, on average, about 40%, being consistently high for herds living
on overgrazed pastures (Table 2). All mainland herds calved during winter and were
probably vulnerableto limitationsin the availabilityof lichens at this time. Significantly,in
only 2 out of 10 years did I find mortalityto be relatedto the foraging conditions following
the neonatal period (Table 2). Neonates were most at risk. For the Hardangerviddaherd,
living on overgrazed winter ranges, in only those years with exceptionally extreme winter
conditions did I detect any mortality associated with winter weather among calves of the
previous year. Furthermore,in those herds with a consistently high productivity(Knutsho,
Forelhogna),recruitmentdid not fluctuatefrom year to year, regardlessof extremesummer
or winterconditions, as it did in the herds living on more limitedfood.
Mortality of older calves during summer or winter was only recorded for those herds
having a significant neonatal mortality. Mortality associated with the adverse summer
weather was about 30-40% (Hardangerviddaand Hallingskarvet)and with winterweather
about 10% (Hardangervidda,Hallingskarvet),but up to more than 50% in Reindalen.
The correlation between first-year mortality (Table 2) and the average availability of
forage during late winter (Table 1) was highly significant (r = -0-95; n = 6; P < 0.001).
The data suggest that limited food supply in late winter results in an increase in neonatal
mortality. Furthermore,as shown above, the magnitudeof this neonatal mortalityis related
to a large extent to the magnitude of all subsequent mortality in response to adverse
climatic conditions occurringat other times duringthe calves' first year of life.
Primary to secondarysex-ratios
The pre-natal sex ratio was slightly skewed in favour of males, but did not differ
significantlyfrom equality (Table 4). This agrees in generalwith data about other ungulates
(Verme 1983). The secondary sex ratio, however, was skewed in favour of females and
differed significantly from equality. Thus, for those herds that experience neonatal and
Reindeerpopulation regulation
368
TABLE4. Changes in the sex ratio (males:females) from the pre-natal to adult
stages (data from Hardangerviddaand Brattefjell-Vindeggen)
Adult
Prenatal
1.5 years
X2
X2
X2
Samplesize
Ratio
P
101*
53:49
0.18
>0-9
1418t
41:59
23-11
<0-005
815t
36-1:63-9
31-61
<0-005
* Sampleof culleddoes.
t Survivorsof the 1971 cohortbornon Hardangervidda
duringthe periodwhenhuntingwas prohibited
(1971-72).
herdthat emigratedfrom Hardangervidda
and livedthere,without
t Sampleof the Brattefjell-Vindeggen
beinghunted,until1980.
TABLE5. Adult sex ratios (males: females) of herds protected from being hunted
B-Vindeggen
Reindalen
Reindalen
Reindalen
n
Year >1 yearratio
x2
1980 36-1:63.9
815* 62-99
1978 39-8:60-2
102* 4.24
1979 35:65
351t 31-59
1982 42:58
9.59
363t
* countedduringtherut.
t countedduringlatesummer.
P
<0-005
<0-05
<0-005
<0.005
post-natal mortality, the males suffer a high mortality rate. This trend continued into
adulthood (Table 5). We shall thereforeturn our attentionto the survivalrates of adults.
Adult survival rates
For the least productiveherds, there is a slight decline in mean survivalrate of adult does
(Table 3), but the life-tableanalysis method does not permit any statistical test being made
of the samplingerrorinvolved.
The survival curves shown in Fig. 7, calculated from population growth rates (Fig. 6),
and the standing age classes (Fig. 7), indicate that the survivalrates of the strongest, most
productive, age classes (those between 2 and 10 years-old), were high in all the herds.
Survival rates of adult does in the different herds may differ because of a change in the
age-structure of the population in these herds whose productivity was lowered by calf
losses.
Because the age-class sample for the Reindalen herd in 1981 was very small, owing to
the irregularityin the annual losses, it is difficult to draw valid conclusions for this herd.
During the 5-year study period, a total of fifty-one carcasses of adults 1 year-old or older
were found, twenty-nine of which were found in 1981. This means that, on average, ten
animals out of a total population of 550 died each year, in which case the actual survival
rate for adults in the Reindalen herd was 0.98 which is similar to the calculated value of
0.92 (Table 3). The mean recruitmentrate was 0. 18 during the entire 11-year period for
which data exist. Since roughly half of the recruitment was of does, then the doe
recruitmentrate, less the adult mortalityrate, yields a population growth rate of 0.07. This
growth rate compares favourably with the population growth rate estimated from field
observations (Fig. 6). This result suggests that the adult mortalityrate does not differfrom
that of mainlandherds.
In an analysis of variance the mean adult survival rate was not significantlyrelated to
the winter population density for the mainland herds (F = 4.62; d.f. = 1,5; P > 0.05),
T. SKOGLAND
(a)
20 000Hardangervidda
*
1400010000 60004000-
Hallingskarvet
__.-o
o
Snhetta
20001000-
369
o o
Knutsh0
600-'m
2
40-200-
BrattefJell-V.
Forelhogna
1600I1000-
o--
600 400 -
.
??'
.
RondaneS.
.
.
R
Reindolen
2001001 1 1 1 1 1 1 1 1 1 1 1 1 1
50
3000 2000-
-
1000800 600400-
E
200-
'
(b)
Hardangervidda
Hailing/
I00 -- ska
0 ~ rvet
Forelhogna,"'"
,'
Knutsh0
Sn0hetta
1970
/!
1975
1980
FIG. 6. Population sizes and mortality from hunting during the study. (a) Herd sizes were
obtained from total counts; those for consecutive years are joined by solid lines (@-?);
a
dashed line (O---O) indicates estimated herd development based on counts plus the
demographic data for each of the herds. (b) The numbers of animals shot during the regular
hunting seasons, for those herds for which hunting was permitted.
which suggests that doe survival is not affected by population density. The decline in the
survival rate with increasing age takes place more rapidly for bucks than for does (Fig. 8).
The mean survival rate of bucks was 93% of the value for does. This higher mortalityrate
for males thus explains the discrepancy shown in Tables 4 and 5 between the secondary
and the adult sex ratios.
DISCUSSION
Reproductivelife history strategy
From my results it is clear that the effects of food limitation at a high populationdensity
are mainly through the pregnant doe's chances cf success in giving birth to a viable
offspring. If food is scarce, she apparentlyabandons her foetus, or young, in preferenceto
putting her own survival at stake and halt her own body growth after reproduction.The
gradual decline in fecundity of the youngest cohorts with increasingherd density is caused
by a decrease in body-size, to a value below the threshold at which maturity takes place
(Reimers 1983a). Once this minimum size has been attained, fecundity remains high
Reindeerpopulation regulation
370
4 - Knutslh'0(198 )
3-
Does
"it
2
4
Brattefjell - Vindeggen\
?,_
(1980)
*\
3
ForelIhogna "^
2
\ .
(1980)
3
\
2
4
S0rr (198t
Rodn
Rondane
(18)
S"
3
U1
^
_L1\
2
\~"
C:
Q1)
G)
0
C.)
01)
3
2
0
-J
4
Hallingskarvet \
11
(1981)
W-
A
3
2
t
Snohetta
(1981)
4
\
?-?--??- -__-
3
2
Hardanger-_
! vidda (1979)
"-
0 ,,
I"
Reindalen (1981)
\\
, II I I I IJ I I I J k I I I I I I
1 2 3 4 5 6 7 8 9 10 l 121314151617
Age (years)
FIG. 7. Age-specific and time-specific age-class frequencies for the does of different ages in the
studied herds (see Methods for calculation details). Dashed lines (O--- ) indicate age-specific
and solid lines (
) time-specific frequencies. Sample sizes are given in Table 3.
T. SKOGLAND
371
4-
3-
ML
Oc
_
2-
\
Px :=0.89
\
0
-J
o0-
1
2
3
4
5
6
7
8
9
10
Age (years)
FIG. 8. Age-specific and time-specific (p = 0.89) age-class frequencies for bucks (n = 228) in
the Brattefjell-Vindeggenherd (1980). Symbols are the same as in Fig. 7.
throughout adulthood, regardless of actual body size. The average body size of the
Hardangerviddadoes was only 60% of that of the Knutsho does (Skogland 1983), yet
adult fecundity rates were similar. These findings do not support Reimers' (1983a)
suggestion that fecundityis relatedto body size.
High adult fecundity rates are compatible with a nomadic reproductive life history
strategy (Geist 1981), as shown by the migratorywildebeest(Connochaetesgnou) (Sinclair
1979), the reindeer(R.t. tarandus) Klein 1968; Leader-Williams1982; this study) and the
caribou (R.t. groenlandicus,R.t. arcticus) (Bergerud 1971; Dauphin6 1976; Parker 1981).
This particularstrategy, as opposed to that of having a fixed home range, entails movement
in order to locate better foraging areas. The North American caribou can still practicetheir
migration unhamperedby habitat obstructions. In Norway, however, this is normally no
longer the case, and when interferencewith this life history strategy occurs, food may
become a limiting factor to reproductivesuccess (as shown in this study). Species having
home-ranges,like the North American deer (Odocoileus virginianus, 0. hemionus)or the
Scottish red deer (Cervus elaphus) show a size-fecundity relationship, and the fecundity
rate falls with increasing population density (Gross 1969 and references therein; Albon,
Mitchell & Staines 1983; Clutton-Brock, Guinness & Albon 1983; Mitchell & Brown
1974). In respect to reproductivelife history strategy, the migratory North American elk
(Cervus canadensis) appearsto be intermediate,because the fecundityrates of middle-aged
females (3-9 years old) are unaffectedby changes in populationdensity (Houston 1982).
Survival
In this study, the greater part of the mortality of reindeer was due to neonatal losses
probably caused by food limitation prior to, or at calving. Furthermore, herds with a
limited food supply were not buffered against adverse climatic effects at other seasons of
the year. This led to fluctuationsin the recruitmentrate and, in turn, to unstablepopulation
growth. The effects of years of low recruitmentclearly show up in the age structuredata,
i.e. the 1969 and 1971 cohorts in both the Hardangervidda and Snohetta age-class
distributions(Fig. 7).
My results suggest that the survival rate of does during their reproductivelife-span was
unaffected by herd density (or by limitations of the food supply). Unfortunately,
372
Reindeerpopulation regulation
comparable adult survival rate data at both high and low densities are lacking for other
cervids.
Population regulation
As Ricklefs (1973) pointed out, density is critical for a population only when viewed in
the context of resource availability. Therefore, a valid evaluation of a density-dependent
factor must be based on data gathered under identical environmental conditions, to
eliminate the effects of density-independentfactors on population size. In this study
identical environmentalconditions for comparison were ensured by using only the actual
winter habitat densities for the mainland herds of reindeer for the analysis of
density-dependence,since winter food availabilitywas found to be the most critical factor
for survival.
The data from Svalbard show that the intrinsic growth rate of reindeeron this island is
identical to that for mainland reindeer (r = 0-3). When plotted against winter habitat
density the Svalbard data would fall on the slope of the mainland data shown in Fig. 5.
Leader-Williams (1982) concluded that three introduced herds of reindeer at different
population densities on South Georgia were at differentstages of food resource limitation
althoughlichens do not form a part of the winterdiet of reindeeron these islands.
Density-independentfactors such as (a) those relatedto climate (i.e. snow depth, ground
ice or suppressed plant growth), and (b) harassment by insects (also climatically
dependent),man or predatorshave been shown (or suggested) to affect body composition
as well as growth in Peary caribou (Thomas 1982) and wild reindeer (Reimers 1983b,c).
Ground ice in winter, following overgrazingof the lichen, was the most likely cause of the
population crashes of several introduced reindeerherds on arctic islands off the coast of
Alaska (Scheffer 1951; Klein 1968).
Reimers (1983b,c) suggested that the population size of reindeer was regulated by
climate on Svalbard and by harassment on the mainland. From the above discussion and
the presentedresults however, I suggest that the primaryfactor in populationregulationof
wild reindeer herds is density-dependent food limitation in winter, and that densityindependent effects are most likely to be a contributory factor only at high population
density.
Low density populations are characterizedby a lower mean age than those at higher
density, because of a higher population growth rate. Older age classes play an increasing
role in the reproductionof high density populations. This is evident from the slope of the
curves for the standing age-class frequenciesshown in Fig. 7. One would thereforeexpect
to find that the longer reproductionis delayed, the greaterthe degree of non-linearityof the
density-dependent relationship. The slight delay in attaining reproductive age at high
densities in this study is in accordance with the finding that the curvilinearity of the
recruitment-densityrelationshipis slight.
The results presented in this study thus support the hypothesis that density-dependent
food limitation occurs at densities near to, or above, one half the ecological carrying
capacity, in conformity with the predictionsof large-mammalpopulationmodels.
ACKNOWLEDGMENTS
This study was financed by the Norwegian Science Foundation (NAVF), throughthe IBP
programme for the period 1970-74 and for 1975-83 by the Directorate of Wildlife and
Freshwater Fisheries, Trondheim.The staff at our research laboratory, in particularBerit
T. SKOGLAND
373
and Per Krigsvoll and Gosta Hansson, is much appreciatedfor their meticulous work in
handling and ageing thousands of jawbones from reindeer. Several people assisted in the
field work, notably the game wardens SverreTveiten, Helge Bitustoyl, ChristianKlemetsen
and Ove Eide, as well as Per Gatzchmann, Lars Ingdal and Jan Ove Gjershaug,
Viltforskningen,My sledge-dog facilitated the transportationof supplies in the field both
summer and winter, for all the 10 years. Simen Bretten and his family at Kongsvold
Biological Station and Inge Myrnes and Nick Tyler at the MAB Station on Svalbard
provided comfortable accommodation and made the fieldwork much more enjoyable.
Jorunn Pettersen, Viltforskningen typed the manuscript and Britt Egeland, also
Viltforskningen,made all the drawings. P. A. Tallantire corrected the language of the
manuscript.
REFERENCES
Albon, S. D., Mitchell,B. & Staines, B. W. (1983). Fertilityand body weight in female red deer; a
Journalof AnimalEcology,52, 969-980.
relationship.
density-dependent
Alendal,E. & Byrkjedal,I. (1975). Populationsize and reproductionof the reindeer(Rangifertarandus
on Nordenskjoid
rbok1974, 139-152.
Land,Svalbard.NorskPolarinstituttA
platyrhynchos)
caribou.WildlifeMonographs,25, 1-55.
Bergerud,A. T. (1971).Thepopulationdynamicsof Newfoundland
Bergerud,A. T. (1980). Reviewof the populationdynamicsof caribouand wild reindeerin NorthAmerica.
R6ros (Ed. by E. Reimers,E.
Proceedingsof the 2nd InternationalReindeer/CaribouSymposium,
Gaare& S. Skjenneberg)
forviltog ferskvannsfisk,
Trondheim.
pp. 556-581. Direktoratet
I. & Ronning,0. I. (1978).Riktplantelivlangtmotnord.Forskningsynytt,
Brattbakk,
8,44-51.
on a heavilyexploiteddeerpopulation.Dynamicsof LargeMammal
Burgoine,G. E. Jr (1981). Observations
Populations(Ed.by C. W. Fowler& T. D. Smith),pp.403-414. JohnWiley& Sons,New York.
Caughley,G. (1970). Eruptionof ungulatepopulations,with emphasison Himalayantharin New Zealand.
Ecology,51, 53-72.
Populations.JohnWiley& Sons,London.
Caughley,G. (1977).Analysisof Vertebrate
Clutton-Brock,T., Guinness,T. & Albon, S. (1982). Red Deer: Behaviourand Ecology of Two Sexes.
Edinburgh
UniversityPress,Edinburgh.
T. N., Guinness,F. E. & Albon, S. D. (1983). The costs of reproduction
to red deerhinds.
Clutton-Brock,
Journalof AnimalEcology,52, 367-383.
caribou.Part 4, growth,
Dauphine,T. C. (1976). Biologyof the Kaminuriakpopulationof barren-ground
andenergyreserves.CanadianWildlifeService,38, 1-71.
reproduction
Espmark,V. (1971). Antler sheddingin relationto parturitionin female reindeer.Journal of Wildlife
Management,35, 175-177.
Fowler,C. W. (1981a).Comparativepopulationdynamicsin largemammals.Dynamicsof Large Mammal
Populations(ed.by C. W. Fowler& T. D. Smith)pp.437-456. JohnWiley& Sons,New York.
Fowler,C. W. (1981b).Densitydependenceas relatedto lifehistorystrategy.Ecology,62, 602-610.
for Villreinforvaltningen.
Gaare,E. & Skogland,T. (1979).Grunnlaget
4, 16-19.
Jakt-Fiske-Friluftsliv,
interactionstudiedin a simplecase model.Proceedingsof
Gaare,E. & Skogland,T. (1980). Lichen-reindeer
the 2nd InternationalReindeer/CaribouSymposiumRoros (Ed. by E. Reimers, E. Gaare & S.
Trondheim.
forviltog ferskvannsfisk,
Skjenneberg),
pp.47-56. Direktoratet
Geist, V. (1981). On the reproductivestrategiesof ungulatesand some problemsof adaptation.Evolution
Today(Ed. by G. G. E. Scudder& J. L. Reveal),pp. 111-132. Proceedingsof the 2nd International
Congress of EvolutionaryBiology. Hunt Institutefor BotanicalDocumentation.Carnegie-Mellon
University,Pittsburg,U.S.A.
studiesin populationecologyof the Svalbardreindeerin Nordenskiold,Land,
Gossow,H. (1980).Preliminary
SymposiumRoros
post-calvingseason (1972). Proceedingsof the 2nd InternationalReindeer/Caribou
(Ed. by E. Reimers,E. Gaare& S. Skjenneberg),
pp. 624-644. Direktoratetfor vilt og ferskvannsfisk,
Trondheim.
Gross,J. E. (1969). Optimumyield in deer and elk populations.NorthAmericanWildlifeConferences,34,
372-386.
Haber,G. C. & Walters,C. J. (1980). Dynamicsof the Alaska-Yukoncaribouherds and management
implications.Proceedingsof the 2nd InternationalReindeer/CaribouSymposiumRoros (Ed. by E.
forviltog ferskvannsfisk,
Trondheim.
Reimers,E. Gaare& S. Skjenneberg),
pp. 645-663. Direktoratet
elk.EcologyandManagement.Macmillan,New York.
Houston,D. B. (1982).TheNorthernYellowstone
Klein,D. R. (1968). The introduction,increaseand crashof reindeeron the St. MatthewIsland.Journalof
WildlifeManagement,32, 350-367.
374
Reindeerpopulation regulation
Laws, R. M., Parker, I. S. C. & Johnstone, R. C. B. (1975). Elephants and their habitats. Clarendon Press,
London.
Leader-Williams, N. (1982). Population dynamics and mortality of introduced reindeer herds on South
Georgia: the effects of diet and density. Holarctic Ecology, 5, 381-388.
McCullough, D. R. (1979). The George Reserve deer herd. University of Michigan Press, Ann Arbor.
Mitchell, B. & Brown, D. (1974). The effects of age and body size on fertility in female red deer (Cervus
elaphus L.). XI International Game Biology Congress, 13, 89-98.
Nordhagen, R. (1943). Sikkilsdalen og NorgesJjellbeiter. SkrifterBergens Museum 22.
Parker, G. A. (1981). Physical and reproductive characteristics of an expanding woodland caribou population
(Rangifer tarandus caribou) in Northern Labrador. Canadian Journal of Zoology, 59, 1929-1940.
Reimers, E. (1983a). Reproduction in wild reindeerin Norway. Canadian Journal of Zoology, 61, 211-217.
Reimers, E. (1983b). Mortality in Svalbard reindeer.Holarctic Ecology, 6, 141-149.
Reimers, E. (1983c). Growth rate and body size differences in Rangifer, a study of causes and effects.
Rangifer, 3, 3-15.
Reimers, E. & Nordby, 0. (1968). Relationship between age and tooth cementum layers in Norwegian
reindeer.Journal of WildlifeManagement, 32, 957-961.
Ricker, W. E. (1975). Computation and interpretationof biological statistics of fish populations. Bulletin 191
Fisheries Board of Canada.
Ricklefs, R. E. (1973). Ecology. Nelson, London.
Sauer, J. R. & Boyce, M. S. (1983). Density dependence and survival of elk in Northwestern Wyoming.
Journal of WildlifeManagement, 47, 31-37.
Scheffer, V. B. (1951). The rise and fall of a reindeerherd. Scientific Monthly, 73, 356-362.
Sinclair, A. R. E. (1977). The African Buffalo. Wildlife behaviour and ecology series. University of Chicago
Press, Chicago.
Sinclair, A. R. E. (1979). The eruption of ruminants. Serengeti: Dynamics of an Ecosystem (Ed. by A. R. E.
Sinclair & M. Norton-Griffiths),pp. 82-103. University of Chicago Press, Chicago.
Sinclair, A. R. E. & Norton-Griffiths, M. (1982). Does competition or facilitation regulate migrant ungulate
populations in the Serengeti? A test of hypotheses. Oecologia, (Berlin), 53, 364-369.
Skogland, T. (1983). The effects of density dependent resource limitation on size of wild reindeer. Oecologia,
(Berlin), 60, 156-168.
Skogland, T. (1984). Wild reindeer foraging niche organization. Part. I. Habitat selection. Holarctic Ecology,
7, (in press).
Snedecor, G. W. & Cochran, W. G. (1978). Statistical methods. Iowa State University Press, Ames, Iowa.
Thomas, D. C. (1982). The relationship between fertility and fat reserves of Peary caribou. Canadian Journal
of Zoology, 60, 597-602.
Varley, G. C. & Gradwell, G. R. (1970). Recent advances in insect population dynamics. Annual Review of
Entomology, 15, 1-24.
Verme, L. J. (1983). Sex ratio variation in Odocoileus: a critical review. Journal of Wildlife Management, 47,
5 73-582.
(Received 28 October1983)