Variation in territory fidelity and territory shifts among red squirrel, Sciurus vulgaris, females

Anim.

Behav., 1995, 49, 187-193

Variation in territory fidelity and territory shifts among red squirrel, Sciurus vulgaris, females L. A. WAUTERS, Department

of Biology,

University

L. LENS & A. A. DHONDT of Antwerp,

U.I.A., B-2610 Wilrijk,

Belgium

(Received 26 October 1993; initial acceptance 10 December 1993; final acceptance 19 January 1994: MS. number: 4504)

Abstract.In resource-limited systemswhere territories are defended year-round or during the breeding season,individuals sometimes shift to new territories. Here, data on territory establishment and site fidelityin a population of Eurasian red squirrels, in a coniferous woodland, are used to test three hypothesesexplaining territory shifts: (1) a female actively bequeaths her territory to some of her offspring;(2) an old female leavesher territory to her offspring when her future reproductive potential islowerthan that of her offspring; (3) a female that initially settled on a poor quality territory shifts to avacant, higher quality territory. Of 44 adult females, only seven moved from the territories on which theyfirst settled to adjacent, vacant territories. Females that shifted did not actively bequeath their territoriesto their offspring. However, females on territories with poor food resources were most likely to shift and they shifted only when the new territories contained more food than the old ones. After shifting,females improved their reproductive rate. These results suggest that territory shifts are adaptive andoccur as a response to poor breeding conditions.

Forbirds and mammals living in resource-limited occupying a territory may be necessary for reproduction and/or survival (Greenwood 1980;Anderson 1989; Price & Boutin 1993), and homerange or territory establishment coincides withthe disappearance of the previous owner, or withthe availability of vacancies (Waser & Jones 1983;Petersen& Best 1987; Boutin et al. 1993). In bids that defend a territory year-round or during thebreeding season and in female territorial mammals,two strategies have been observed: (1) they alwaysstay on the territories on which they first settled(territory fidelity); or (2) they sometimes shift to new territories (territory shift; Dhondt 1971;Harvey et al. 1984; Stenseth & Lidicker 1992).Several hypotheses explain why territory shiftscan occur: (1) parents leave their territories, orpart of them, to some of their offspring, so as to increasetheir chance of future survival and reproduction (territorial bequeathal: Woolfenden & Fitzpatrick 1978; Price & Boutin 1993); (2) parentsabandon their territories when their reproductive potential is lower than that of their offspring (Anderson 1989); or (3) individuals that initially settle on poor quality territories, because of intense interference competition with already territorial conspecifics, attempt to improve by

systems,

0003-3472/95/010187+07 $08.00/O

shifting to better territories (Krebs 1971; Stenseth 1978; Cockburn 1988). A new territory can be of better quality because the owner has access to higher quality mates, there is less intraspecific competition, or food availability is higher. For some non-territorial and territorial mammals, mothers have been observed, under certain conditions, to abandon their territories and leave them to some of their offspring that remain on the natal sites (Anderson 1989; Boutin et al. 1993). In birds, several studies have shown that territory shifts are adaptive, i.e. individuals increased their reproductive success after shifting to better or larger territories (Baeyens 1981; Petersen & Best 1987; Matthysen 1988). Dispersing to areas of higher quality may explain why some female mammals undergo breeding dispersal (Stenseth 1978; Cockburn 1988), but we know of no studies that have actually tested this hypothesis. Adult Eurasian red squirrel females, living in coniferous woodlands, defend intra-sexual territories against other females. The territories are defined as home range core-areas, consisting of 70% of all radio-track fixes (see Wauters & Dhondt 1992). Subadult and yearling females behave as floaters or settle in home ranges on the edges of adult females’ core-areas (Wauters &

0 1995 The Association for the Study of Animal Ekhaviour

187

Animal Behaviour, 49, 1

188

Dhondt 1992). Only ‘territorial’ females produce offspring (Wauters & Dhondt 1989). For approximately 10 months per year conifer seeds are their main food (Wauters et al. 1992), allowing accurate measurements of home range quality in terms of food resources. Moreover, variation in food abundance between territories is large and affects female reproductive success(Wauters & Dhondt 1989). Therefore the red squirrel is an ideal species to test the different hypotheses explaining territory shifts. We used data on territory establishment and site fidelity of 44 Eurasian red squirrel females, gathered over 9 years (1984-1992) in a coniferous woodland. We also investigated which characteristics (intensity of intra-sexual competition, male quality, food abundance) determined territory quality in this species. METHODS

Radio-tracking

All adult females that settled on the study area were radio-tagged for at least 2 months. Those that shifted home range were radio-tagged a second time to examine home range size and space use after shifting. Between 30 and 46 point fixes were taken, over a 3-month period, on each radio-tagged squirrel to determine home range size (see Wauters & Dhondt 1992). All data on home range size, home range use and overlap of home ranges were calculated using the mononuclear range polygon method of the RANGES IV program (Kenward 1990). For each adult female we calculated: (1) total home range size (100% of the fixes); (2) core-area size (70% of all ties; Wauters t Dhondt 1992; Wauters et al. 1994); and (3) core-area overlap (the number of squirrels of either sex whose core-areas overlapped with a female’s core-area). We refer to these core-areas of dominant females as territories.

Study Area

We used a study area of 30 ha in the ‘Merodese Food Ahdance Bossen’ at Herenthout (Province of Antwerpen, Food abundance was calculated on each l-ha N. Belgium) a mainly coniferous woodland of square (woodplot) of the study area by counting 212 ha, dominated by Scats pine, Pinus sylvestris, fallen seeds (including the remains of food-items and Corsican pine, Pinus nigra. The south side of the study area was bordered by meadows and a consumed before they had fallen) on three 1-m’ plots placed on a randomly chosen line (Wauters road which were never crossed by resident squirrels. The west side was bordered by farmland and & Dhondt 1989). The percentage overlap of a female’s core-area with different woodplots was lesssuitable patches of wood. Along the north and measured to calculate food abundance within its east sides, suitable habitat continued. core-area (territory). Territories were classified as good (those containing more food than the Trapping and Handling !3quirrels median) or poor (those containing less food than Trapping was carried out for at least 4 days the median). every other month from March 1984 to May 1992. A detailed description of the trapping procedure is Data Analyses given elsewhere (Wauters & Dhondt 1990). Squirrels were individually marked, using numbered With a step-wise logistic regression model metal ear-tags (type 1003 S National Band & Tag we tested the factors affecting the probability of Co, U.S.A.), and weighed to the nearest 5 g. Sex shifting territory for adult females that had a and age of each individual were determined. Sub- neighbouring vacancy, at least once during their adult males were between 4 and 10 months old, lifetime (binary variable Y; shifting female= 1, with abdominal testes and a small scrotum. Sub- non-shifting female=O). This technique converts adult females were between 4 and 12 months old binary data into probability values by fitting a with a small vulva and invisible nipples. Adults logistic curve through the available points. The were over 12 months. At each recapture a female’s logistic model is described by: reproductive status was recorded using the followP( Yi= l)=expG(xJ/l +expG(x,) ing categories: anoestrus, oestrus or post-oestrus (swollen vulva), lactating. and

Wauters et al.: Female red squirrel territoriality G(xJ=(A+Blx,,+B2x,+

. . . +BlOx,,,-J

andwhere A, Bl, B2, . . ., BIO are constants, and xi*is female body mass (g), xi2 is female density, x,~is food abundance on the ‘first’ territory, xi4 is thechangein territory quality [food abundance on thenew (shifters) or potentially new (non-shifters) territory- food abundance on the ‘first’ territory], xi5is breeding rate on the ‘first’ territory (number ofyoung weaned per breeding season), xi6 is the age(years)at which the first neighbouring territorybecomes vacant, x, is the factor referring to weanedyoung (female with weaned young when vacancyavailable=2, no weaned young= l), X~ is changein male quality [mean body mass of the two heaviest males overlapping with the new(shifters) or potentially new (non-shifters) territory-mean body mass of the two heaviest maleson the ‘first’ territory (g)], xi, is the interaction between change in territory quality and age,and x,,, is the interaction between breeding rateand age. All observed shifts took place during winter,hence between two breeding seasons,and dataon density and food abundance refer to the yearin which a neighbouring vacancy occurred. The parameters of the logistic model were estimatedby maximum likelihood, which provides an indexof goodness-of-fit of a model which includes aparticular set of parameters. We used a step-wise backward procedure to select a 6nal model containing only sign&ant parameters. RESULTS Territory Acquisition and Site Fidelity The study area contained 18 territories which remained stable throughout the study. A female usuallybecame territorial by taking over a territory vacated by the death of its previous owner. Wenever saw a female chase a previous owner off itsterritory. Of the 44 red squirrel females that settled as residentson the study area and survived at least 1 year (two breeding seasons), only seven (16%) individuals moved from the territories on which they had settled. Of the 37 females that settled permanently on territories, five (14%) were subordinates with home ranges that overlapped with the vacated territories, 25 (68%) immigrated onto vacant territories and four (11%) settled in areas where no territorial females had been present

I

4

lOOIll

Figure 1. Convex polygons of the 70% core-areas of the seven shifting females (see Table II). ‘First’ territory -, territory after shifting ---. Stippled areas are areas of low food abundance. These movements did not occur simultaneously and the map is a composite picture of shifts recorded over many years. during the previous 2 years. Thus 92% of nonshifting territorial females acquired territories that were vacant when they settled. One female (3%) acquired a piece of a territory that was then subdivided between her and the previous territory owner. This female survived only 1 year and never weaned offspring. Finally, territory acquisition by two immigrant females (5%) coincided with the old owners abandoning their territories for new ones. Of the 18 vacanciesin good territories, 14 (78%) were filled by the next breeding season, against only three out of 12 (25”/,) vacancies in poor territories (x2=8*16, df= 1, P=O408). Fen&s Shifting Territory Out of seven females that shifted their home ranges, five were territorial females and two were subordinates that became territorial after shifting (females 10 and 11, Fig. 1). The two subordinates and three of the already territorial females moved

Animal Behaviour, 49, 1

190

Table I. Logistic regressionmodel predicting the probability of territory shifts in femalered squirrels(N= 34)

females. Food abundance on the old territory and the change in territory quality (Table I) were the only factors that significantly affected the probx2 df P ability of shifting. The first territories of shifting females contained relatively poor food resources Full model 24.5 -10 0.006 and, more importantly, shifting females had the Selectedparameters Changein territory quality 14.29 1 oGOO2 opportunity to obtain adjacent territories with Food on ‘first’ territory 4.80 1 0.027 richer food resources (Table II). Four out of 27 non-shifting females had the G(xJ=1.8 ( f 2.1)+0.030 ( f 0.014)x change territory opportunity to move to vacant territories containquality- 0.014 ( f 0.008) food abundance ‘first’ terri- ing more food. However, three of them were tory. already settled on food-rich territories, while for one the nearby vacancy was quickly filled by an to adjacent territories vacated after the previous immigrant. Out of seven females that shifted owners died (Fig. 1). Finally, for two dominant territory, four had produced offspring on the old females (08 and 101) shifting coincided with the territory just before shifting. None of these arrival of immigrants on the old territories recruited a daughter, or a son, on its old territory. (Fig. 1). For all females, the centre of the new When F 0 1, a female that shifted at the age of 3 territory was lessthan 120 m (core-area diameter) after trees had been felled in part of its territory, away from the centre of the previous core-area. was excluded from the analysis, reproductive rate This is in contrast with movements of locally (i.e. the number of weaned young per breeding recruited subadult females (natal dispersal), of season; Table II), was significantly higher on the which only one out of eight settled lessthan 120 m new than on the old territory (Wilcoxon matchedfrom its birthplace (x2= 1244, df= 1, P=O-0014). pairs signed-ranks test: T=O, N=5, P=OXM3,twotailed). This increase in breeding rate could have been partly caused by age-dependent reproductive Causes and Consequences of Territory Shifts rate, since five out of six females shifted territories All shifting females had neighbouring vacancies before the age of 2, so before reaching prime available. In addition, 7’3% (27 out of 37) of reproductive age. We therefore compared the non-shifting females could have moved to an increase in reproductive rate (mean number of adjacent vacant territory at least once in their young per breeding ‘seasonafter the age females lifetime. Thus the availability of a vacant territory had when they shifted - reproductive rate before alone could not explain why some females shifted. the age of shifting) between shifting females and We therefore applied a logistic regression model non-shifting females that had similar life histories (seeMethods) to the 34 squirrels that could have (comparable lifespan and age of first reproducmoved to an adjacent vacant territory, to test tion). The increase in reproductive rate tended to which factors explained territory shifts by adult be higher for shifting (N=6, X& sD=O.58f 0.42) Table II. Comparison of territory quality (food abundance on the territory) and reproductiverate (mean number of young weaned per breeding season)of shifting femaleson the old (before shifting) and new (after shifting) territories Territory quality (lo3 kJ) Female 101 48 08

15 01 11 10

Old 916 277 18.56 596 1201

126 273

Reproductiverate

New

Old

New

823 370 2327

0.63

1.0 0 1.5

1172 1327 596 941

0 0.5 0.5

’

1.0 0 0

0.78 0.25 0.8

1.0

Wauters et al.: Female red squirrel

than for non-shifting females (N= 11, 8* SD= 0.09rt0*68), but the difference was not statisticallysigniticant (Student t-test, two-tailed: t= 1.58, df=15,P=O.13). DISCUSSION TerritoryBequeathal: Tamiasciurus versusSciurus In some non-territorial mammals, weaned young remain on the natal site after breeding dispersalof the mother (e.g. common voles, Microtus arvalis: Boyce & Boyce 1988; Cohnnbian ground squirrels, Spermophilus columbianus: Harris& Murie 1984). In territorial mammals, however,female breeding dispersal and bequeathal ofthe entire territory seemsvery rare (Jones 1984; Anderson1989; Waser & Jones 1991; Boutin et al. 1993;Price & Boutin 1993). In American red squirrels,Tamiasciurus hudsonicus, defending individual,food-based territories year-round, breeding femalessometimes actively bequeath territories to their offspring (Price & Boutin 1993). These femalesare not out-competed by their juveniles (psssivedispersal), nor are they searching for betterterritories (Price et al. 1986; Boutin et al. 1993;Price & Boutin 1993). Price & Boutin (1993) concludedthat territory bequeathal by mothers is morelikely to occur when: (1) juveniles are more ableto defend the natal territory than to compete elsewherefor vacant territories; (2) having a territory is crucial for overwinter survival; and (3) in competition over territories, females are better thanjuveniles. In female Eurasian red squirrels breeding dispersalis rare; in our study there were only seven shifters over 8 years, of which only four had alreadyproduced young. In none of these cases wasthe vacated territory afterwards defended by young. Hence, territorial shifts were not active bequeathal.Eurasian red squirrels have a different territorial system to Tamiasciurus; only adult femalesdefend part of their home ranges against other females. They scatterhoard cones and food isshared among individuals whereas Tamiasciurus larderhoard cones in large middens which are defended(Gumell 1987; Wauters & Dhondt 1992; Wauters et al. 1992). In addition, and again in contrastwith Tamiasciurus, non-reproductive subadult or yearling female red squirrels are present for a large part of the year and, being older and heavier,are better competitors than weaned juv-

territoriality

191

eniles, even on the juvenile’s natal home range (Wauters et al. 1993; Wauters & Dhondt 1993). Thus, conditions 1 (juveniles capable of defending the natal territory) and 2 (higher survival of territorial juveniles) are not met for Eurasian red squirrels. We argue that prolonged maternal investment, allowing young to stay on the natal territory until they are large and heavy enough to compete successfullyfor vacant territories, did not evolve in Eurasian red squirrels because of the lack of a large exclusive food resource (midden) on a mother’s territory; this may explain why territory bequeathal has not evolved in this species.As a consequence (1) philopatry occurs very rarely and most young squirrels disperse over relatively large distances (Wauters & Dhondt 1993), and (2) all subadults spend their first winter as low-ranked floaters, or on home ranges overlapping with dominant adults (Wauters & Dhondt 1989, 1992). Causes of Shifting A red squirrel territory can be of higher quality for severalreasons: it can have fewer competitors, higher quality males, or more food resources. In the logistic regression model, neither female density nor body mass of overlapping males (heavy males mate more often than lighter males; Wauters et al. 1990) al&ted the probability of shifting. Hence, there was no indication that females shifted territory to escape high levels of intra-sexual competition, or to encounter heavier (higher quality) males on the new territory. On the other hand, the probability of a female shifting was best predicted by the increase in food abundance between old and (potential) new territory, and females were more likely to shift when the ‘first’ territory contained poor food resources. In fact, six out of seven shifting females moved to territories with higher food availability and four females that initially held poor territories produced only one young, against the 18 young produced after shifting. Consequencesof Shifting All shifting females acquired new territories adjacent to their old ones, suggesting that the costs associated with shifting were low. To benefit from a shift, a female should produce more young and/or survive better on the new than on the old

Animal

192

Behaviour,

territory. Great tits, Parus major, and nuthatches, Sitta europaea, that moved to better territories survived better than birds on poor quality territories (Dhondt 1971; Matthysen 1990). In several other bird species, males (e.g. sage sparrows, Amphispiza belli: Petersen & Best 1987) or pairs (e.g. magpies, Pica pica: Baeyens 1981) increased their reproductive rate after shifting. In this study, the two subordinate red squirrel females (10 and 11, Table II) produced young only after shifting, while three already territorial females (l&O8 and 101) produced more young on the new than on the old territories. Female 01 moved from its partly destroyed territory when 3 years old and produced fewer young afterwards. Thus females generally increased their reproductive rate after shifting. However, non-shifting females with similar life histories also slightly increased their reproductive rate. Therefore, age-dependent reproductive rate could also have contributed to the observed increase in breeding performance of shifting females on the new territories. We conclude that once Eurasian red squirrel females have acquired territories, which are essential for successfulreproduction, they normally stay there for life. Breeding dispersal seemed more an opportunistic behaviour than a real strategy, occurring only in females that could shift to vacant, adjacent territories containing more food than the old ones. These data support the hypothesis that territory shifts occur as a response to poor breeding conditions on the ‘first’ territory and the occurrence of vacant territories of higher quality. ACKNOWLEDGMENTS

We thank especially the family Stoelen and H. Berkenbosch who gave permission to work on their estate. Xavier Lambin and John Gurnell helped improve the manuscript. The research was supported by a Conserted Action of the Belgian Ministry of Education, by the Fund for Collective Research, Brussels, and by the European Community (EC-Step-0040 project). L. Lens is Research Assistant of the Belgian National Fund for Scientific Research. REFERENCES Anderson, P. K. 1989. Dispersal in rodents: a resident fitness hypothesis. Special Publications no. 9. Pittsburgh: American Society of Mammalogists.

49, I

Baeyens, G. 1981. Functional aspects of serial monogamy: the magpie pair-bond in relation to its territorial system. Ardea, 69, 145-166. Boutin, S., Tooze, Z. & Price, K. 1993. Post-breeding dispersal by female red squirrels (Tamiasciurus hudsonicus): the effect of local vacancies. Behav. Ecol., 4, 151-15s. Boyce, C. C. K. & Boyce, J. L., III. 1988. Population biology of Microtus-arvahs. II. Natal and breeding disnersal of females. .I. Anim. Ecol.. 57, 723136. Cockbum, A. 1988. Social Behaviour in Fluctuating Populations. Sydney: Croom Helm. Dhondt, A. A. 1971. Some factors intluencing territory in the great tit, Parus major L. Giervalk, 61, 125-135. Greenwood, P. J. 1980. Mating systems, philopatry and dispersal in birds and mammals. Anim. Behav., 28, 114&1162. Gumell, J. 1987. The Natural History of Squirrels. London: Christopher Helm. Harris, M. A. & Murie, J. 0. 1984. Inheritance of nest sites in female Colombian ground squirrels. Behav. Ecol. Sociobiol., 15, 97-102. Harvey, P. H., Greenwood, P. J., Campbell, B. & Stenning, M. J. 1984. Breeding dispersal of the pied flycatcher (Ficedula hypoleuca). .I. Anirn. Ecol, 53, 727-736. Jones, W. T. 1984. Natal philopatry in bannertail kangaroo rats. Behav. Ecol. Sociobiol., 15, 151-155. Kenward, R. 1990. Ranges IV. Furzebrook: Institute of Terrestrial Ecology. Krebs, J. R. 1971. Territory and breeding density in the great tit Parus major L. Ecology, 52, 2-22. Matthysen, E. 1990. Behavioral and ecological correlates of territory quality in the Eurasian nuthatch (Sitta europaea). Auk, 107, 86-95. Petersen, K. L. & Best, L. B. 1987. Territory dynamics in a sage sparrow population: are shifts in site use adavtive? Behav. Ecol. Sociobiol.. 21. 351-358. Price,‘K. & Boutin, S. 1993. Territorial bequeathal by red squirrel mothers. Behav. Ecol., 4, 144150. Price, K., Broughton, K., Boutin, S. & Sinclair, A. R. E. 1986. Territory size and ownership in red squirrels: response to removals. Can. J. Zool., 64, 1144-I 147. Stenseth, N. C. 1978. Demographic strategies in fluctuating populations of small rodents. Oecologia (Berl.), 33, 149-172. Stenseth, N. C. & Lidicker, W. Z. 1992. Animal Dispersal. Small Mammals as a Model. Oxford: Blackwell Scientific Publications. Waser, P. M. & Jones, W. T. 1983. Natal philopatry among solitary mammals. Q. Rev. Biol., 58, 355-390. Waser, P. M. & Jones, W. T. 1991. Survival and reproductive effort in banner-tailed kangaroo rats. Ecology, 72, 771-777. Wauters, L., Bijnens, L. & Dhondt, A. A. 1993. Body mass at weaning and local recruitment in the red squirrel. J. Anirn. Ecol., 62, 280-286. Wauters, L., Casale, P. & Dhondt, A. A. 1994. Spacing behaviour of red squirrels in fragmented woodlands. Oikos, 69, l&146.

Wauters

et al.:

Female

Wauters, L. & Dhondt, A. A. 1989. Body weight, longevity and reproductive success in red squirrels (Sciurus vulgaris). .I. Anim. Ecol., 58, 637-651. Wauters, L. & Dhondt, A. A. 1990. Red squirrel (Sciurw vulgaris Linnaeus, 1758) population dynamics in different habitats. Z. Siiugetierk., 55, 161-175. Wauters, L. & Dhondt, A. A. 1992. Spacing behaviour of red squirrels, Sciurus vulgaris: variation between habitats and the sexes. Anim. Behav., 43, 297-3 11. Wauters,L. & Dhondt, A. A. 1993. Immigration pattern and success in red squirrels. Behav. Ecol. Sociobiol., 33. 159-167.

red squirrel

territoriality

193

Wauters, L., Dhondt, A. A. & De Vos, R. 1990. Factors affecting male mating success in red squirrels (Sciurus vulgaris). Ethel. Ecol. Evol., 2, 191-204. Wauters, L., Swinnen, C. & Dhondt, A. A. 1992. Activity budget and foraging behaviour of red squirrels (Sciurus vulgurti) in coniferous and deciduous habitats. .I. Zool., Land., 227, 71-86. Woolfenden, G. E. & Fitzpatrick, J. W. 1978. The inheritance of territory in group-breeding birds. Bioscience, 28, 104-108.