Does interspecific competition with introduced grey squirrels affect foraging and food choice of Eurasian red squirrels?

ANIMAL BEHAVIOUR, 2001, 61, 1079–1091 doi:10.1006/anbe.2001.1703, available online at http://www.idealibrary.com on

Does interspecific competition with introduced grey squirrels affect foraging and food choice of Eurasian red squirrels? LUC A. WAUTERS*, JOHN GURNELL†, ADRIANO MARTINOLI* & GUIDO TOSI*

*Department of Structural and Functional Biology, University of Insubria †School of Biological Sciences, Queen Mary and Westfield College (Received 6 May 2000; initial acceptance 20 July 2000; final acceptance 17 January 2001; MS. number: 6543R)

Grey squirrels, Sciurus carolinensis, introduced from North America, have replaced red squirrels, S. vulgaris, over much of Britain and parts of north Italy, but the reasons why are unclear. Spatial and temporal changes in the quantity and quality of their primary foods, namely tree seeds, may provide the focus for interspecific resource competition and hence go some way to explain the replacement process. To investigate whether grey squirrels have a competitive advantage over red squirrels, we used radiotelemetry and direct behavioural observations to examine the activity budget, foods, feeding behaviour and body condition of adult red squirrels in two mature, mixed-woodland sites in northern Italy, one site where there were only red squirrels, and one where both red and grey squirrels were present. The studies were carried out between July 1996 and October 1998. We found few differences in the activity and foraging patterns and food choice of red squirrels with and without grey squirrels present, although we could not eliminate possible interspecific competition effects on food choice by red squirrels in summer (June–August) and autumn–winter (September–February). Foraging time and rate of energy intake of red squirrels in the mixed-species site were lower than in the red-only site in winter (December–February). This may have resulted from interspecific competition, but a more plausible explanation is that these site differences resulted from the distribution of preferred tree seeds and home range size. Overall, our results provide little support for the food competition hypothesis. Differences in body size between sites suggest that interspecific competition occurs during the growth phase of red squirrels, when juveniles and subadults disperse and look for a place to settle. 

foraging behaviour and diet (Gro ¨ nwall 1982; Moller 1983; Wauters & Dhondt 1987; Wauters et al. 1992). Tree squirrels are strictly diurnal (Wauters 2000). On a geographical scale, activity patterns vary with temperature, while differences in habitat quality (food availability, population density, predator density, intra- and interspecific competition) may affect the intensity and timing of activity, and the activity budget and feeding behaviour, on a local scale (Wauters 2000). Comparative studies have revealed that tree squirrels adapt their activity and food choice to the availability and quality of food by adjusting (1) the length of the active period, (2) the proportion of time spent foraging (autumn/winter) or spent feeding on primary food resources, and (3) the amount of time spent searching for food. Hence, differences in food availability (and rate of food intake) between woodland types causes habitat-related variation in the squirrel’s activity pattern (Wauters et al. 1992; Wauters 2000). Although intraspecific factors such as breeding status and age have been shown to affect the timing and duration of the time spent active (Lurz 1995;

The population density, demographic processes and longterm survival of nonterritorial tree squirrel populations depend largely on the quality and size of the habitat they live in, particularly on the annual changes in the size of tree seed crops (Gurnell 1987, 1996; Wauters & Dhondt 1990, 1995; Andre´n & Lemnell 1992; Koprowski 1994; Lurz et al. 1995; Wauters & Lens 1995; Kenward et al. 1998), and on the degree of habitat fragmentation (Andre´n & Delin 1994; van Apeldoorn et al. 1994; Wauters et al. 1994a, b). As important as absolute food abundance is the spatial and temporal variation in food availability and changes in food quality, particularly the energy content of different food types (Wauters & Dhondt 1992, 1995; Kenward & Holm 1993; Gurnell 1993; Lurz et al. 1997, 2000). The latter will affect how tree squirrels can use the different tree seed species available in their woodland habitat and thus their Correspondence: L. A. Wauters, Istituto Oikos, Viale Borri 148, I-21100—Varese, Italy (email: [email protected]). J. Gurnell is at the School of Biological Sciences, Queen Mary and Westfield College, Mile End Road, London E1 4NS, U.K. 0003–3472/01/061079+13 $35.00/0

2001 The Association for the Study of Animal Behaviour

1079



2001 The Association for the Study of Animal Behaviour

1080

ANIMAL BEHAVIOUR, 61, 6

Wauters 2000), little is known about the effects of interspecific competition on the activity and feeding behaviour of squirrels and other small mammals (Flowerdew 2000; Merritt & Vessey 2000; Ziv & Smallwood 2000). Tree squirrels are ideal model animals to study the effects of competition on activity budgets and food choice because: (1) the availability and energy content of their major food types (tree seeds) can be measured accurately, even at the level of the individual home range (e.g. Wauters & Dhondt 1989a, 1992, 1995; Wauters et al. 1995a); and (2) they are diurnal and relatively easy to observe, allowing accurate estimates of the time spent active and foraging and the food choice and feeding rate of individually marked animals (Wauters et al. 1992; Wauters & Gurnell 1999). Although the effects of interspecific competition on space use and activity of small mammals are often poorly understood, examples of extreme forms of exclusion competition do occur, especially with the introduction of closely related, allopatric species. Introduced North American grey squirrels, Sciurus carolinensis, have replaced the native red squirrel, Sciurus vulgaris, over most of the British Isles and parts of northern Italy (Lloyd 1983; Gurnell & Pepper 1993; Wauters et al. 1997). In an earlier paper (Wauters & Gurnell 1999), we compared the activity pattern of red squirrels, and their reproductive performance in two study areas: one with only red squirrels (red-only site), the other with both species present (red–grey site), and described intra-and interspecific interactions, mating behaviour, and red and grey squirrel spacing behaviour in the study area where they co-occurred. We found no effects of the presence of grey squirrels on the daily activity pattern of red squirrels. Intraspecific aggressive interactions, especially between red squirrels of the same sex were common, but all but one interspecific encounters were not aggressive. The red squirrels did not avoid the woodland patches most intensively used by grey squirrels and core area overlap between species was similar to that of red squirrels alone. This suggested that red squirrels avoided spatial overlap with grey squirrels as they did conspecifics and that an increase in grey squirrel numbers will intensify resource competition (Wauters & Gurnell 1999). The experimental design for our studies on red–grey interaction comprised two experimental ‘red–grey’ sites (i.e. sites with both red and grey squirrels present), one in northern England and one in northern Italy, and two control ‘red-only’ sites (i.e. sites with only red squirrels), again one in northern England and one in northern Italy. In both countries, we selected regions that were in the process of being colonized by grey squirrels. The English sites were typical of the region, consisting of mixed-aged conifer plantations with low tree species diversity (see Wauters et al. 2000). This was in contrast to the two Italian sites which were mixed broadleaf and conifer woodlands with a high tree species diversity, again typical of the region (see Wauters & Gurnell 1999 and later). Thus the two experimental and two control sites were not true replicates. In all the sites we used capture–mark– recapture to monitor numbers and body size in red squirrels and we used radiotelemetry to examine patterns

of space use. However, although we could locate radiotagged individuals in the English sites, the physical nature of the dense stands of young conifer, or the very tall stands of mature conifer, made it impossible to observe the squirrels directly for long periods to study their behaviour and find out what they were eating. On the other hand, we were able to monitor squirrel activity, behaviour, food choice and feeding rate throughout the year in the Italian sites where animal ‘visibility’ was good. In the present study, we investigated whether the replacement of red by grey squirrels is a consequence of grey squirrels having a competitive advantage over red squirrels by depleting the availability of important tree seeds (resource competition, Gurnell & Pepper 1993; Wauters et al. 2000). If this is the case, the presence of grey squirrels will affect the activity budget, foraging pattern and body mass (condition) of red squirrels. To investigate this, we have used data collected only from the Italian sites. To minimize the effects of the lack of site replication, we monitored as many red squirrels at each site as we could. Although similar, the two sites were not directly comparable in terms of tree species composition, and we feel this will always be the case for research carried out at this scale in temperate, mixed-broadleaf and conifer woodlands. As a consequence, we have examined the results carefully to try to separate behavioural differences between the sites that result from differences in tree species composition, and those that result from the presence of greys at one of the sites. We predicted that: (1) interspecific competition for resources decreases the rate of energy intake of red squirrels and forces them to spend more time active and or a greater proportion of time foraging in the red–grey study site; (2) interspecific competition for food depletes primary resources and forces red squirrels to change their diet, resulting in changes in the time spent feeding on different food items, or in changes in food type selection between the two study sites; and (3) red squirrels will weigh less in the presence of grey squirrels, owing to their inability to exploit highquality food resources sufficiently. If interspecific competition acts primarily on juvenile red squirrels, however, we expect adult red squirrels to be smaller in the presence of grey squirrels, owing to reduced growth rate (see Wauters et al. 2000). METHODS

Study Sites We monitored squirrels from July 1996 to October 1998 in two mature, mixed woodlands that presented highquality habitats for both squirrel species. One study site, the red-only site, had only red squirrels, the second study site, the red–grey site, had both red and grey squirrels. The red-only site (22 ha) was in Parco Pineta, an extensive mixed forest of 3000 ha on the northern edge of the upper Po plain in Lombardy, northern Italy (857 E, 4545Z N). It was dominated by deciduous trees, mainly black locust, Robinia pseudoacacia (25%), sweet chestnut, Castanea sativa (20%), oaks, Quercus robur, Q. petraea (9%), and hornbeam, Carpinus betulus (5%), with Scots pine,

WAUTERS ET AL.: SQUIRREL INTERSPECIFIC COMPETITION

Pinus sylvestris (19%), and plantations of white pine, Pinus strobus, and Norway spruce, Picea abies, covering about 17% of the study area. The understorey was diverse, with hazel, Coryllus avellana, mixed with young trees of black locust, sweet chestnut and birdcherry, Prunus avium, and with some blackberry, Sambucus nigra, and hawthorn, Crataegus monogyna. The red–grey site (13 ha) consisted of a mature, mixed-deciduous woodland in a 17-ha castle park at Borgo Cornalese in the Po plain in Piedmont, northern Italy (744 E, 4455 N), typical of the climax vegetation Quercus carpinetum. It was dominated by oaks (18%), hornbeam (19%), field maple and sycamore, Acer campestris, A. pseudoplatanus (26%), and ash, Fraxinus excelsior (15%),with some birdcherry, lime, Tilia cordata, and alder, Alnus glutinosa. The walnuts of sparsely planted Juglans regia and Juglans nigra were important food resources for both squirrel species. There was also a single block of planted white pine which covered about 8% of the study area. The understorey in the red–grey site was dominated by blackberry with some hazel. In the red–grey site, tree seeds regularly consumed by red and grey squirrels were slightly more abundant and red squirrel densities were higher than in the red-only site (Wauters & Gurnell 1999). Food availability at both sites was higher than average in 1997 and slightly below average in 1998 (unpublished data). The first grey squirrels (two females, three males) colonized the red– grey site in July 1996, and the population expanded quickly with a three-fold increase in numbers from March 1997 to May 1998. By October 1998 grey squirrel densities had exceeded those of the local red squirrel population (Wauters & Gurnell 1999).

Trapping and marking were carried out under licences from the Regions of Piedmont and Lombardy.

Trapping and Handling Squirrels

Behavioural Sampling

At the red-only site we set up a trapping grid with 26 trap stations, covering 20 ha (trap density 1.3/ha). At the red–grey site, we randomly placed 20 traps (1.6/ha) throughout the woodland. We carried out trapping for at least 5 days bimonthly, from July 1996 to October 1998, with ground-placed Tomahawk ‘squirrel’ traps (Tomahawk Live Trap Co., Tomahawk, Wisconsin, U.S.A.) baited with sunflower seeds and hazelnuts. Water was not provided as squirrels can get sufficient from their food. Traps, partly covered by dark plastic to give shelter from rain and cold, were set in the morning and checked twice (winter–spring) to three times a day (summer– autumn). We ushered each trapped squirrel into a zipper-tube (Wauters & Dhondt 1989a, b) to minimize stress during handling, and individually marked it with numbered metal ear tags (102 mm, type 1003 S National Band and Tag Co, Newport, Kentucky, U.S.A.) pinched in at the base of the ear with special pliers that do not cause any injury. It was weighed to the nearest 5 g with a Pesola spring balance, and the length of the right hind foot (without the nail) was measured (0.5 mm) with a thin ruler (Wauters & Dhondt 1989a, b; Wauters & Gurnell 1999). Sex and age were recorded (Wauters & Dhondt 1989a, b, 1993; Wauters & Gurnell 1999). No animals died during trapping, marking and handling.

We studied the activity pattern of red squirrels by focal sampling (Altmann 1974; Wauters & Dhondt 1987; Wauters & Gurnell 1999): a focal squirrel was first located by radiotelemetry and then continuously observed, with 1050 binoculars when necessary, for 20–60 min. Observations were stopped if the squirrel’s behaviour indicated that it was aware of the observer’s presence. During an observation period we continuously monitored the focal squirrel’s activity and recorded: (1) where it was active, that is, the tree species on which the squirrel was observed or the ground; (2) its behaviour (for categories see Wauters & Dhondt 1987; Wauters et al. 1992); and, if foraging and feeding, (3) food type and time spent feeding from the moment the food item was found until it dropped the remains (Wauters et al. 1992). For each seed species, we calculated mean feeding rate from 20 complete observations of feeding bouts by red squirrels (except for white pine where N=14). We calculated average food intake (kJ/h) per seed species by multiplying the number of seeds consumed/h feeding (calculated from average feeding rate) by the energy value/seed. The analysis of interactions has been discussed elsewhere (Wauters & Gurnell 1999). In this paper we focus on the percentage of time spent foraging and travelling, and on the feeding behaviour, of red squirrels monitored

Radiotracking We radiotagged both red and grey squirrels (TW-4 transmitters, Biotrack, Wareham, U.K., and TXP-1 transmitters, Televilt, Lindesberg, Sweden; collars weighed 12–13 g, less than 5% of a squirrel’s body mass) over 3-month periods, coinciding with the seasons (winter= December–February; spring=March–May; summer=June– August; autumn=September–November). All adults of both species were radiotagged in 1997 and 1998 allowing good estimates of space use patterns and activity patterns of both species. Radiocollars did not cause any injuries and were removed after 3–5 months when sufficient data were gathered. We determined locations by following the radiosignal until the squirrel was seen or pinpointed by signal strength and direction (Wauters & Dhondt 1992). At each fix, we plotted a squirrel’s location and recorded the exact time of the fix, the squirrel’s activity (1=active, 2=in a nest), whether it was on the ground or in a tree and, where appropriate, the tree species. For home range analyses, between 30 and 40 radiolocations (fixes) were collected for each squirrel, which proved sufficient to describe a squirrel’s home range (Wauters & Dhondt 1992). We used mononuclear 70% core area estimates to describe an animal’s activity centre (Wauters & Dhondt 1992; Wauters & Gurnell 1999) and to calculate the food resources available for each individual squirrel (see compositional analysis below).

1081

1082

ANIMAL BEHAVIOUR, 61, 6

between April 1997 and October 1998, when grey squirrel density varied between 1.2 and 1.8 squirrels/ha. In the red-only site, we monitored the foraging behaviour of 15 red squirrels in spring, 10 in summer and 19 in autumn– winter, while in the red–grey site we monitored 11 red squirrels in spring, 12 in summer and 14 in autumn– winter. Data were analysed seasonally; the length of observation time per individual squirrel varied from 63 to 334 min in the red-only site and from 81 to 201 min in the red–grey site. We calculated the daily activity pattern, averaged over monthly periods, by combining the radiotrack data from all squirrels collected that month. For each 1-h period (e.g. from 0600 to 0659 hours) the time spent active (Ai) was calculated from the relative proportion of fixes in which the squirrels were active, and expressed as a percentage. The mean time spent active per day (MTA) for each monthly period was calculated by summing Ai over all active hours. On each day, we sampled each individual once or twice. In the latter case one fix was taken in the morning and the other in the afternoon to eliminate dependency of the data from the same squirrel.

Estimating Home Range Food Availability Since woodland structure was heterogeneous at a finegrained level, we calculated the availability of tree seeds for each squirrel’s home range separately and we used this as a measure of home range quality. To do this we first determined woodland composition by establishing one vegetation quadrat (2020 m, 400 m2) in each 1 ha across the study area (N=19 in both study sites). In each vegetation quadrat, we estimated canopy, understorey and ground cover in one of five categories (<20, 21–40, 41–60, 61–80, >80%). The number of trees of each tree species were counted and the diameter at breast height (DBH in cm) of each tree measured. The number of understorey species and the proportion of hazel in the understorey were also recorded. We then estimated seed food availability of those tree species whose seeds are regularly eaten by squirrels (i.e. seeds from sweet chestnut, oak, hornbeam, sycamore, ash, walnut, black walnut and hazel, and fallen and consumed cones from conifers). Fruit fall from trees was monitored from July 1996 to December 1998. We used 1-m2 seed quadrats (50 in the red-only site and 71 in the red–grey site) placed under each tree species in each vegetation quadrat (Wauters & Dhondt 1995; Wauters & Lens 1995). This enabled us to estimate total seed production for 1996, 1997 and 1998. To combine data of different seed species, the numbers of seeds/m2 were converted to energy values (kJ/m2). To estimate seed energy values, we collected 100 seeds from each tree species and removed the pericarp of each seed. The fresh weight of the remaining endocarp, the part consumed by seed predators such as squirrels, was determined for each large seed species (walnuts, black walnuts, acorns, hazelnuts, chestnuts) to the nearest mg, and for groups of 10 seeds for each small tree species. The samples were then dried to dry weight (to nearest mg) and crushed and pressed into pellets of known weight. The pellets were burnt in a semimicro bomb calorimeter (Parr, type

1425), previously calibrated with benzoic acid, to obtain their caloric values (kJ/kg). We used the information of the relative canopy cover of each tree species in each vegetation quadrat and the seed energy production of each tree species to estimate seed food energy availability in each vegetation quadrat. We then expressed the home range food availability for each squirrel (HRQ) as the mean food abundance (kJ/400 m2) of all vegetation quadrats within the 70% core area of the squirrel’s range. Seeds and fruits produced in the summer–autumn of year t were considered available for the squirrels during the following ‘squirrel year’ (see Wauters & Lens 1995), that is, from July of year t to June of year t+1.

Red Squirrel Diet and Compositional Analysis We compared the diet composition (i.e. food choice) of red squirrels between the red-only site and the red–grey site in two ways. In the first analysis we compared the proportions of active time spent searching for and feeding on different food items by individual red squirrels between the two study sites. Because not all food items were available from both study sites, we combined the following categories: the relative use of chestnuts (highest rate of energy intake, highest rate of caching) and Scots pine cones (small, but preferred seeds) in the red-only study site (called category CS); and the relative use of walnuts (highest rate of energy intake, highest rate of caching) and maple seeds (% tree cover, seed energy content and rate of energy intake comparable to that of Scots pine in the red-only site) in the red–grey study site (called category WM). These new categories, called primary food items, comprised the most intensively used foods in each study site and were compared between sites. Second, we carried out a compositional analysis (Aebischer et al. 1993) on feeding behaviour in summer and autumn–winter, when squirrels fed almost exclusively on tree seeds. We examined the use of food items rather than habitats. The food items were defined from the behavioural data from feeding red squirrels. These were: Scots pine cones, white pine cones, chestnuts, acorns, hazelnuts and hornbeam seeds in the red-only site (N=6), and white pine cones, maple seeds, hornbeam seeds, walnuts, acorns and hazelnuts in the red–grey site (N=6). In the initial step a Wilks’s lambda statistic is calculated: a statistically significant
2 value (P<0.05, with df=N
1) indicates a nonrandom use of the available food items (Aebischer et al. 1993). The second step produces a matrix of pairs of compared food items in which negative values imply avoidance of the numerator food type by comparison with the denominator food type, whilst positive values imply selection. The statistical significance of these comparisons is assessed by computing t values with df=number of squirrels in sample (Aebischer et al. 1993). We determined the relative use of each food item (fi) for each radiotagged red squirrel by dividing the time spent foraging and feeding on food item fi by the total time spent foraging and feeding. The proportion in which food item fi was available for a squirrel was calculated by

WAUTERS ET AL.: SQUIRREL INTERSPECIFIC COMPETITION

Table 1. Energy content (per g dry weight and per seed, X±SD) of different tree seed species eaten by red and grey squirrels and the average feeding time and average rate of food intake per seed species for red squirrels in mixed woodlands in northern Italy Energy (kJ/g)

Energy per seed (kJ)

Only at RO site Scots pine Chestnut

25.8±0.4 17.9±0.1

2.8±0.1 66.2±6.3

2.0±0.6 7.7±1.5

106±32 514±123

Only at RG site Walnut Black walnut Field maple

29.9±0.7 32.6±0.6 22.2±0.2

103.1±10.4 89.9±6.3 1.2±0.1

15.3±4.2 30.7±9.7 0.6±0.3

404±118 176±51 127±44

At both sites Oak Hornbeam White pine Hazelnut

17.7±0.4 20.9±0.1 30.4±0.2 25.6±2.8

42.8±8.6 0.8±0.1 37.9±3.1 20.7±2.7

4.2±0.7 0.5±0.1 10.5±3.9 4.9±1.3

367±110 118±32 217±71 257±83

Species

Feeding time (min)

Food intake (kJ/h)

For conifers, the seed energy content is given per cone, based on an average number of seeds/cone of 16 for Scots pine and 48 for white pine. RO site: only red squirrels present; RG site: red and grey squirrels present.

Statistical Analyses Since activity patterns of red squirrels are strongly influenced by seasonal changes in daylength and food availability (Wauters 2000), we analysed activity and diet for each season separately. We did not aim to consider differences between the sexes, and have not included sex as a factor, except in our consideration of body mass and size. This further simplifies the analyses and reduces the number of tests performed on the data within each season. To investigate individual variation in behavioural parameters (foraging behaviour) and in body size measurements we used ANCOVA models (generalized linear models, GLM, SAS 1989). Diet composition was investigated for two categories: proportion of active time spent feeding on primary food resources (categories CS and WM in red-only and red–grey sites, respectively, see above); and proportion of active time spent feeding on secondary food resources (hornbeam seeds, fungi, buds, shoots, flowers and insects). We compared sites with t tests (Sokal & Rohlf 1995). The proportions of active time that individual red squirrels were observed foraging and travelling, and the proportions of active time spent feeding on different food items, were arcsine transformed to improve normality (Sokal & Rohlf 1995). RESULTS

600 Rate of energy intake (kJ/h)

dividing the seed energy availability of food item fi within that squirrel’s core area (in kJ) by HRQ.

500 400 300 200 100 0 –0.5

0.0

0.5 1.0 1.5 2.0 Ln energy content seeds (kJ)

2.5

Figure 1. Relationship between the rate of energy intake by red squirrels and the log-transformed energy content of seeds.

easy to extract, but are small and have low energy contents. Consequently, feeding rates and rate of food intake varied considerably between seed species. Time spent consuming a single seed was lowest for the two smallest seeds, maple and hornbeam, followed by seeds extracted from Scots pine cones (Table 1). Squirrels spent much more time feeding on the larger seeds or cones; for example, about 10 min was spent on feeding on a single, large white pine cone and 15 min and 30 min on walnuts and black walnuts, respectively (Table 1). The rate of energy intake was positively correlated with the logtransformed energy value of seeds (r7 =0.701, P<0.05; Fig. 1).

Food Energy and Feeding Rate Estimates Energy values of the different seeds consumed by both red and grey squirrels varied greatly between seed species (Table 1). Walnuts have a hard nut, especially black walnuts, but they are large and have a high energy content. Seeds of mpale and hornbeam, in contrast, are

Time Spent Active The mean time spent active/day was significantly lower in the red–grey than the red-only site (Wilcoxon matched-pairs signed-ranks test: T=9, N=9, P=0.038, two tailed; Fig. 2). The proportion of time spent foraging

1083

ANIMAL BEHAVIOUR, 61, 6

10 9 Mean time active/day (h)

1084

8 7 6 5 4 3 2 1 0

Jan

Feb

Mar

Apr

Jun Aug Month

Sep

Oct Nov

Figure 2. The mean time spent active/day by red squirrels for monthly (or bimonthly) periods. h: Only red squirrels present; ": red and grey squirrels present.

varied between individuals. In all seasons, it increased when time spent travelling decreased (ANCOVA, travelling time: spring: F1,23 =34.8, P=0.0001; summer: F1,18 =4.87, P=0.04; autumn: F1,19 =12.5, P=0.002; winter: F1,23 =10.1, P=0.004). In spring, additional variation in foraging behaviour was explained by a significant site effect, with red squirrels spending slightly more time foraging in the red-only site (site effect: F1,23 =9.53, P=0.0052; Fig. 3a). In summer and autumn, there was no significant effect of study site (both P>0.2). In winter, in contrast, red squirrels spent more time foraging in the red-only than in the red–grey site (site effect: F1,23 =8.86, P=0.0067; Fig. 3d). However, differences in the proportion of time spent foraging between sites were small in all cases (<10% difference between means), and extrapolated to maximum (winter) differences in effective foraging time of only 20 min/day.

Diet Composition in the Two Sites The time spent feeding on different food items varied considerably in all seasons in both study sites (Fig. 4a,b,c), indicating that food choice varied greatly between individual red squirrels at each site. In spring, bark feeding (feeding on fungi on dead or dying branches and on insects, mainly in oak and chestnut) was of little importance for red squirrels in both sites (<10% of foraging time) whereas the flowers and caterpillars found in oak trees were an important food (Fig. 4a). In the red-only site, squirrels fed much more frequently on conifer flowers, buds and, to a lesser extent, young shoots than in the red–grey site (in category ‘others’ in Fig. 4a), and the consumption of conifer flowers and buds reached 36–58% of total feeding time in the spring diet of red squirrels with core areas in the conifer-dominated patches in the red-only site. Individual differences in the use of cached hazelnuts in spring were very large, 0–28% of feeding time in the red-only site but only 0–5% in the red–grey site (Fig. 4a). However, 10 of 15 red squirrels in

the red-only site fed on hazelnuts for less than 10% of total feeding time in spring. On average, red squirrels fed more frequently on primary food items in the red–grey site (23–90% of total feeding time) than their conspecifics in the red-only site (t24 =2.82, P=0.01; Fig. 4a). The two red squirrels in the red–grey site that had few or no walnut trees in their home range (8–14% feeding on walnuts) spent most time feeding on old maple seeds that were still attached to branches or had fallen on the ground (21–48% of total foraging time). In the red-only site those squirrels that spent little time recovering chestnuts (<10% of total feeding time) foraged intensively on conifer flowers and buds, or on oak flowers and insects found on the young leaves of oak. The proportion of time spent feeding on alternative, secondary food items did not differ significantly between sites (t24 =1.438, P=0.16). In summer red squirrels shifted from the low-energy, bulky spring diet of flowers, buds, shoots and insects, together with cached seeds, to energy-rich, small, fastmaturing seeds (e.g. hornbeam, pine and maple seeds). While summer progresses, the proportions of larger, but still immature, seeds of broadleaf trees and bushes gradually increase (Moller 1983; Wauters & Dhondt 1987; Wauters et al. 1992). Red squirrels in the red-only site consumed few immature acorns, while those in the red– grey site occasionally fed on green acorns (Fig. 4b). Hornbeam seeds were an important part of most squirrels’ diet in both study sites (Fig. 4b). However, eight of 10 red squirrels in the red-only site spent from 44 to 87% of their foraging time feeding on Scots pine seeds. Although the use of maple seeds (0–32% of total feeding time) and not fully mature walnuts (0–42% of total feeding time) was also important for red squirrels in the red–grey site, the difference in the average time spent feeding on primary food items was statistically significant (t20 =2.291, P=0.033; Fig. 4b). The two red squirrels in the red-only site that used few Scots pine cones fed heavily on hornbeam seeds and hazelnuts, or on white pine seeds, respectively. Overall, red squirrels spent significantly more time feeding on secondary food items in the red–grey than in the red-only site (t20 =2.392, P=0.027). During autumn–winter, high-energy tree seeds were most abundant. Acorns were of little importance in either study site and red squirrels seemed to avoid eating large numbers of acorns (Fig. 4c). In the red-only site, there was no correlation between the proportion of different seed species in the diet and the energy value of seeds (r4 =0.007, NS), nor between the proportion of different seed species in the diet and their specific rate of energy intake (r4 =
0.119, NS). In the red–grey site, there tended to be a positive correlation in both cases, but they were not statistically significant (proportion of seed species in diet and energy value per seed: r5 =0.689, NS; proportion of seed species in diet and rate of energy intake: r5 =0.495, NS). In this season, the proportion of foraging time red squirrels spent feeding on primary food resources did not differ significantly between sites (t31 =0.980, P=0.33; Fig. 4c), but more time was spent feeding on secondary food items in the red–grey than in the red-only site (t31 =2.057, P=0.048). The latter effect was explained by red squirrels feeding more frequently on acorns and on

WAUTERS ET AL.: SQUIRREL INTERSPECIFIC COMPETITION

90 (a)

(b)

(c)

(d)

80

70

Foraging time (%)

60

50 90

80

70

60

50

0

5

10

15

20 0 Travelling time (%)

5

10

15

20

Figure 3. Relationship between the percentage of active time spent travelling and time spent foraging by red squirrels in the two study sites according to season (a) Spring (March–May), (b) summer (June–August), (c) autumn (September–November), (d) winter (December– February). , – – –: Only red squirrels present; , ——: red and grey squirrels present.

fungi and insects found on and under the bark of dead or injured branches in the red–grey site (Fig. 4c).

Compositional Analysis In both study sites, the Wilks’s lambda was significant in all seasons (P<0.001), indicating a nonrandom use of food resources by the red squirrels from summer to winter. In the red-only site, Scots pine seeds and hornbeam seeds were preferred over all other food items in summer,

when acorns and chestnuts (the latter not yet mature) were avoided relative to hazelnuts and white pine seeds (Scots pine=hornbeam>hazelnuts=white pine>acorns> chestnuts; Table 2). During autumn and winter, pine cones and hornbeam seeds formed the largest part of the energy budget relative to their availability. In absolute terms, chestnuts were the most important food supply, but they were selected less than pine cones, in a similar way to hazelnuts and white pine cones, which were preferred to acorns (Scots pine=hornbeam> chestnuts=hazelnuts=white pine>acorns, Table 2). These

1085

ANIMAL BEHAVIOUR, 61, 6

80 70

(a)

60 50 40 30 20 10 0 80 70 Foraging time (%)

1086

WP CS/WM Hazel Flowers cones

Bark

Others

(b)

Rate of Energy Intake

60 50 40 30 20 10 0 80 70

WP CS/WM Hazel cones

and preferred over all other food items. In addition, hornbeam seeds and hazelnuts were preferred to white pine seeds and maple seeds. Acorns were clearly avoided with respect to all other food items (walnutsdhornbeam seeds=hazelnuts>white pine seeds=maple seedsn acorns, Table 3). The rank of preference and rate of energy intake for different seed species were not significantly correlated (rS =0.147, N=6, P=0.78), but preference rank tended to increase with relative energy value, although not significantly so (per g dry weight: rS =0.736, N=6, P=0.095).

Acorns

HB seeds

Others

(c)

60

We calculated the daily rate of energy intake for red squirrels in October, December and February, when squirrels fed mainly on tree seeds. Individuals varied in daily rate of energy intake in all months (October: XSD=1122318 kJ/day, range 369–1663, N=25; December: 594128 kJ/day, range 315–948, N=29; February: 847155 kJ/day, range 512–1243, N=29). The average daily rate of energy intake of red squirrels did not differ between study sites in October and February (October: t23 =0.889, P=0.38; February: t27 =0.845, P=0.41) but was significantly higher in the red-only site in December (t27 =2.709, P=0.012; Fig. 5).

50 40

Body Size and Body Mass Differences

30 20 10 0

WP CS/WM Hazel Acorns cones Food item

HB seeds

Others

Figure 4. Percentage of time that red squirrels fed on different food items (X+SD) in the site with only red squirrels (h) and in the site with red and grey squirrels ("). WP cones: white pine cones; CS: chestnuts and Scots pine cones; WM: walnuts and maple seeds; conifers: conifer flowers, buds and shoots; flowers: oak flowers; HB: hornbeam seeds. (a) Spring (March–May), (b) summer (June– August), (c) autumn–winter (September–February).

preferences for small seeds resulted in a negative correlation between rank of preference and rate of energy intake over the different seed species (Spearman rank correlation: rS =
0.860, N=6, P=0.046). In the red–grey site, all early maturing seeds were selected: hornbeam was the most preferred food item, but differences between the small-seeded species and hazelnuts were not significant (hornbeam=hazelnuts= maple seeds=white pine seeds>walnuts>acorns, Table 3). Clearly the red squirrels fed mainly on those seeds that matured early, even when the rate of energy intake for these seed species was low (Table 1). During autumn and winter, food choice changed completely; the now mature high-energy walnuts and black walnuts were selected for

Body mass increased with size (ANCOVA:, effect of foot length: F1,47 =74.2, P<0.0001). There was no sexual dimorphism in body mass of adult red squirrels in either study site, and study site did not significantly explain any additional variation in body mass (sex effect: F1,47 =1.95, P=0.17; site effect: F1,47 =1.72, P=0.20). Adult red squirrels, however, were significantly larger in the redonly site (N=28) than in the red–grey site (N=23; mean foot lengthSD: red-only site: 58.01.0 mm; red–grey site, 56.51.6 mm; t49 =3.960, P=0.0002). Average body mass was also higher at the red-only site than the red– grey site (mean body massSD: red-only site: 31620 g; red–grey site: 30420 g; t49 =2.138, P=0.038). However, the ANCOVA model shows that the site differences in body mass were explained by differences in size, suggesting that red squirrels had a similar mass for their body size in both sites, with red squirrels being physically smaller (shorter foot length) in sympatry with grey squirrels (red–grey site) than when they were the only squirrel species present (red-only site). Squirrels normally lose weight during spring and early summer and start to put on some fat reserves and gain weight in autumn (Gurnell 1987; Wauters & Dhondt 1987, 1989b). We tested whether the presence of grey squirrels caused an increase in spring weight loss and a decrease in autumn weight gain, by testing the effects of study site on weight loss from autumn–winter (body mass monitored from October to January) to spring (body mass taken from April to June), and on weight gain from spring

WAUTERS ET AL.: SQUIRREL INTERSPECIFIC COMPETITION

Table 2. Log ratio differences (X±SE) between food items eaten by squirrels and those available for two periods in the site where only red squirrels were present Food types (denominator) Food types (numerator)

White pine seeds

Summer (June–August, N=9) Scots pine seeds 4.77±0.59** White pine seeds Chestnuts Acorns Hornbeam seeds

Hornbeam seeds

Chestnuts

Acorns

10.3±0.97** 5.46±0.47**

7.86±1.45** 3.10±1.05* −2.40±1.00*

0.25±1.28 −4.46±0.84** −10.00±0.95** −7.61±1.09**

3.81±1.40* −0.96±1.03 −6.45±0.92** −4.05±1.04** 3.56±1.22*

6.61±0.70** 3.85±0.95** 3.47±0.84**

1.14±0.72 −1.62±0.86† −2.00±0.98† −5.47±0.93**

2.10±0.56** −0.67±0.99 −1.04±0.84 −4.51±0.89** 1.01±0.85

Autumn–winter (September–February, N=19) Scots pine seeds 2.77±0.84** 3.14±0.68** White pine seeds 0.37±1.04 Chestnuts Acorns Hornbeam seeds

Hazelnuts

For each pairwise comparison, a negative value implies avoidance of the numerator food type by comparison with the denominator food type, whilst a positive value implies selection. The statistical significance of these comparisons was assessed by computing t values with df=number of squirrels in the sample (see Methods). †P<0.1; *P<0.05, **P<0.01. Table 3. Log ratio differences (X±SE) between food items eaten by squirrels and those available for two periods in the site where both red and grey squirrels were present Food types (denominator) Food types (numerator)

White pine seeds

Summer (June–August, N=12) Maple seeds 0.37±1.21 White pine seeds Walnuts Acorns Hornbeam seeds

Hornbeam seeds

Walnuts

Acorns

1.35±1.01 0.97±0.69

3.16±1.19* 2.79±0.84** 1.82±0.71*

−0.98±9.84 −1.35±1.01 −2.24±0.95* −4.14±1.17**

−0.55±1.03 −0.92±0.89 −1.90±0.75* −3.71±0.45** 0.43±1.12

2.45±0.63** 2.51±0.71** 4.59±0.64**

−1.80±0.81* −1.74±1.04 0.34±0.72 −4.26±0.81**

−1.67±0.79* −1.60±0.98 0.48±0.49 −4.12±0.73** 0.14±0.85

Autumn–winter (September–February, N=14) Maple seeds −0.07±0.81 −2.14±0.73* White pine seeds −2.08±0.82* Walnuts Acorns Hornbeam seeds

Hazelnuts

For each pairwise comparison, a negative value implies avoidance of the numerator food type by comparison with the denominator food type, whilst a positive value implies selection. The statistical significance of these comparisons was assessed by computing t values with df=number of squirrels in the sample (see Methods). *P<0.05, **P<0.01.

to the next autumn–winter, respectively. Spring weight loss of red squirrels did not differ between the red-only and the red–grey site (XSD: red-only site:
189 g, N=15; red–grey site:
1612 g, N=19; Student’s t test: t32 =0.663, P=0.51). Red squirrels put on slightly more weight in autumn in the red-only site, but the site difference was not significant (mean autumn weight gainSD: red-only site: 2113 g, N=15; red–grey site: 1510 g, N=20; t33 =1.555, P=0.13). Thus, the presence of grey squirrels did not result in red squirrels losing more weight over spring, nor did it prevent them from putting on fat in autumn.

DISCUSSION

Time Spent Active Red squirrels spent less time active per day in the red–grey site than in the red-only site, but in most months the differences were less than 1 h. This result can be explained in two ways: (1) the presence of grey squirrels reduces red squirrel activity (interspecific competition); or (2) site differences in the distribution of high-energy food resources result in red squirrels needing less time to satisfy their daily energy requirements (food

1087

ANIMAL BEHAVIOUR, 61, 6

Daily rate of energy intake (kJ/day)

1088

1800 1600 1400 1200 1000 800 600 400 200 0

October

December

February

Figure 5. Average daily rate of energy intake (mean kJ/day+SD) by red squirrels in October, December and February in the site with only red squirrels (h) and the site with both red and grey squirrels (").

effect). In the first case, we expect red squirrels to concentrate their activity during hours of the day that grey squirrels are less active and avoid the periods of major activity of greys. However, we have shown elsewhere (Wauters & Gurnell 1999) that there were no important differences in the activity patterns of the two squirrel species in the red–grey site, and that red and grey squirrels had their main activity peaks during the same periods of the 24-h cycle. Thus red squirrels were not avoiding grey squirrels by reducing their activity or changing their activity pattern.

Foraging, Diet and Energy Intake Rate Spring In spring red squirrels spent slightly more time foraging in the red-only than in the red–grey site. However, in the red–grey site they fed more frequently on high-energy, primary food resources, contrary to what would be expected if interspecific competition reduces their access to these food items. Thus, despite the presence of grey squirrels, red squirrels spent more time feeding on the highest quality seeds, fallen and cached walnuts and black walnuts, in the red–grey site, than their conspecifics on fallen and cached chestnuts and Scots pine cones in the red-only site. In contrast, cached hazelnuts were recovered more intensively in the red-only than in the red–grey site, but differences in the use of hazelnuts by individual squirrels were large. The higher proportion in the red-only site could be simply due to hazelnuts being more uniformly distributed in this site than in the red– grey site, where they were more clumped. Certainly, in the red–grey site, edible dormice, Glis glis, grey and red squirrels, and to a lesser extent nuthatches, Sitta europea, and great spotted woodpeckers, Picoides major, consumed the majority of the hazelnuts in late summer and early autumn, before they were fully mature and fell to the forest floor. In general, high-energy tree seeds are scarce or depleted in spring and both red and grey squirrels feed on a variety of secondary food items ranging from buds, shoots and flowers of different tree species, to bark-growing fungi

and insects, mainly caterpillars (Gro ¨ nwall 1982; Moller 1983; Gurnell 1987; Wauters & Dhondt 1987; Wauters et al. 1992). This diverse diet was found in both study sites, and prevented us from calculating a rate of energy intake for this season. Although we do not have data on the rate of energy intake of feeding on bark-growing fungi, we feel that the site differences in bark feeding, which was higher in the red–grey site, was of little importance, since proportions of feeding time were low in both study sites. Red squirrels spent more time feeding on conifer flowers and buds (category ‘others’) in the red-only site. Both these differences are probably caused by differences in the forest structure between the sites, with the red-only site having more conifers (Scots pine and Norway spruce) and the red–grey site more oak. We feel it is unlikely that these differences are related to important differences in the rate of energy intake in this season. This interpretation was supported by spring weight loss being similar in both study sites: thus red squirrels did not lose more weight during spring when grey squirrels were also present.

Summer In summer, red squirrels shifted from a low-energy, bulky spring diet to consumption of energy-rich but small and fast-maturing seeds (e.g. hornbeam seeds, pine seeds, maple seeds) and, towards the end of summer, to larger, energy-rich seeds (walnuts, chestnuts). Compositional analysis showed that the preference for small, earlymaturing seeds was similar in both study sites. There was significantly more feeding on primary preferred food resources at the red-only than the red–grey site, suggesting that the ‘preferred seeds’ were used more frequently in the red-only site. The lower usage at the red–grey site could be an effect of interspecific competition for walnuts, fed upon intensively by both squirrel species. Since walnut trees were scarce, red squirrels might be excluded from feeding for long periods on walnuts by grey squirrels. However, red and grey squirrels were seen feeding in the same walnut trees without any aggressive interaction (see also Wauters & Gurnell 1999). Nevertheless, red squirrels might have avoided walnut trees already occupied by grey squirrels, possibly leading to a decrease in their foraging on walnuts. We believe that the higher use of acorns in the red–grey than the red-only site is of little importance, for two reasons. First, both in summer and autumn–winter, acorns were of little importance in the diet of red squirrels, and compositional analysis showed it was the least preferred of seed species. Hence, red squirrels do not eat large numbers of acorns in either study site, confirming earlier findings in other red squirrel populations in England and Belgium (Wauters & Dhondt 1987; Wauters et al. 1992, 1995a; Kenward & Holm 1993). Second, the negative effects of the high tannin concentrations in acorns on protein digestibility in red squirrels indicate that red squirrels are unable to survive on a diet constituted primarily of acorns (Kenward & Holm 1993). Grey squirrels, in contrast, thrive on acorns, being able to neutralize the effects of tannins on protein digestibility, and annual fluctuations in grey squirrel densities and

WAUTERS ET AL.: SQUIRREL INTERSPECIFIC COMPETITION

reproductive rates in British broadleaf woodlands are strongly related to changes in the size of the acorn crop (Gurnell 1996; Kenward et al. 1998). The slightly higher travelling time in the red–grey site than the red-only site might be caused by interspecific competition. In the red–grey site, hornbeam seeds were intensively used by grey squirrels in summer. Hence, red squirrels might have been forced to spend more time travelling to find sufficient hornbeam seeds (a preferred food item in summer) or to avoid trees in which grey squirrels were feeding. However, hornbeam seeds were highly available, and compositional analysis showed they were also selected in the red–grey site. Thus even if interspecific competition for hornbeam seeds occurred, it had little impact on red squirrel food preference.

Autumn–winter In autumn–winter, preferred tree seeds appeared to be readily available to red squirrels in both study sites, and there was a significant difference only in the proportion feeding on secondary food items between sites. Red squirrels used acorns more frequently, and bark feeding was observed more often in the red–grey site. As explained above, these differences were considered unimportant. Reduced foraging time of red squirrels in the red–grey site than the red-only site could be a consequence of interspecific competition with grey squirrels, or alternatively, caused by differences in the distribution of high-energy food resources. There are several arguments that make it unlikely that interspecific competition reduced foraging time and rate of energy intake. First, compositional analysis showed that in the red–grey site, red squirrels were able to select those seeds with the highest energy value per g, and that walnuts, a relatively scarce but highly preferred food item, were strongly selected over other seed species. Second, although lower in the red–grey site than the red-only site, estimates of daily rate of energy intake were still higher in the red– grey site than for other red squirrel populations living in deciduous woodlands (Wauters et al. 1992). Therefore, the reduced energy intake in December is unlikely to have a negative effect on squirrel body mass. In fact, about 65% of individual variation in mean autumn–winter body mass of red squirrels was explained by differences in body size, and study site did not explain any relevant additional variation in body mass between the red-only and the red–grey sites. Third, daily rate of energy intake was higher for red squirrels in the red-only site than in the red–grey site only in December, and not in February when they were still primarily feeding on high-energy tree seeds. Fourth, red squirrels had much larger home ranges in the red-only site than in the red–grey site (unpublished data), probably in response to lower overall food abundance, and lower densities. Hence, during their daily movements red squirrels covered larger distances than conspecifics in the red–grey site while searching for high-energy food resources, and thus would have higher energy requirements. These arguments suggest that differences in winter foraging activity and rate of energy intake were related to site differences in the distribution of high-energy, preferred food resources, determining the

squirrels’ activity pattern. However, we cannot exclude an effect of interspecific competition and data on the weight gain of red squirrels in late summer–autumn seem to suggest a slight, but statistically not significant, decrease in the ability to put on fat reserves (only 6 g less on average) in the presence of grey squirrels. Currently we do not know whether a further increase in grey squirrel density (Wauters & Gurnell 1999) will augment resource competition in late summer–autumn and constrain weight gain in red squirrels. Food preferences in both sites were not solely determined by the rate of energy intake of the different tree seeds, and in the red-only site there was a negative correlation between the rate of energy intake and the preference rank of consumed seeds, caused by squirrels preferring the small seeds of Scots pine and hornbeam. Seeds of conifer species, when available, are preferred food items for red squirrels, although they are small, difficult to extract and therefore have a relatively low rate of energy intake per unit of time (Wauters et al. 1992; Lurz et al. 2000). Scots pine seeds, however, have a high nitrogen, sodium, phosphorus and calcium content, the latter being very important for lactating females (Borges 1990), and buds and flowers of conifers are rich in essential micronutrients (Gro ¨ nwall 1982). Seeds of most broadleaves are poorer in sodium and phosphorus and, except for walnuts and hazelnuts, in nitrogen (Havera & Smith 1979). Hence, nitrogen, a measure of protein content, and concentrations of micronutrients may have affected food choice of red squirrels. Finally, the strong avoidance of acorns in both sites indicates that the concentration of secondary metabolites, such as tannins, also affects food choice of red squirrels.

Body Size and Mass Differences Red squirrels were smaller and weighed on average 12 g less in the red–grey site than in the red-only site, but body mass differences between sites were mainly explained by red squirrels being smaller in sympatry with grey squirrels. Hence, our results on body size differences suggest that interspecific competition acts primarily during the growth phase of the red squirrels (hence juvenile and subadult stages), causing reduced growth and hence smaller adult size for red squirrels in sympatry with greys, and that resource competition between resident adults of both species has little additional effect on the body mass of adult red squirrels. These results agree with earlier findings in conifer forests in northern England (Wauters et al. 2000).

Conclusions Our data suggest that interspecific competition between red and grey squirrels did not substantially change the activity and foraging patterns, or food choice, of adult red squirrels in sympatry with greys, in comparison with woodlands where red squirrels were the only species present. In spring, a large variety of temporarily abundant food resources, of low energy content, were

1089

1090

ANIMAL BEHAVIOUR, 61, 6

exploited, and loss of body mass in red squirrels did not increase when greys were present, making it unlikely that interspecific competition for food resources occurred. In summer, interspecific competition might have affected food choice of adult red squirrels. We were unable to test whether this resulted in a significant decrease in rate of energy intake, but it caused only a slight decrease in autumn weight gain. In autumn–winter, when highenergy tree seeds are most abundant, differences in red squirrel activity and food choice between sites were significant. The decrease in the time spent foraging in winter when grey squirrels were present resulted in a lower daily rate of energy intake, which agreed with predictions of interspecific competition. However, other data suggested that these differences could also be explained by siterelated variation in the distribution of preferred tree seeds, and by red squirrels using larger home ranges in the red-only site. In any case, if interspecific competition occurred in the red–grey site, it did not result in lower winter survival, spring breeding (Wauters & Gurnell 1999) or decrease in body condition, compared with the woodland with only red squirrels present.

Acknowledgments We thank Xavier de Maistre and Caterina Gromis de Trana for allowing us to work on their estate and for their personal support and interest in the project. We are also grateful to the Consorzio Parco Pineta di Appiano Gentile e Tradate for the permission to use the regional park as study site. This work would not have been possible without the help of Claudia Bozza, Silvio Colaone, Luigi Druetto, Suzanna Fenoglio and Lucilla Lorusso who did part of the fieldwork. Three anonymous referees greatly helped to improve the manuscript. L. Wauters was financed by a TMR-grant (No: ERBFMBICT 960662) from the Commission of the European Union.

References Aebischer, N. J., Robertson, P. A. & Kenward, R. E. 1993. Compositional analysis of habitat use from radio-tracking data. Ecology, 74, 1313–1325. Altmann, J. 1974. Observational study of behavior: sampling methods. Behaviour, 49, 227–267. Andre´ n, H. & Delin, A. 1994. Habitat selection in the Eurasian red squirrel, Sciurus vulgaris, in relation to forest fragmentation. Oikos, 70, 43–48. Andre´ n, H. & Lemnell, P. A. 1992. Population fluctuations and habitat selection in the Eurasian red squirrel Sciurus vulgaris. Ecography, 15, 303–307. van Apeldoorn, R. C., Celada, C. & Nieuwenhuizen, W. 1994. Distribution and dynamics of the red squirrel (Sciurus vulgaris L.) in a landscape with fragmented habitat. Landscape Ecology, 9, 227–235. Borges, R. 1990. Sexual and site differences in calcium consumption by the Malabar giant squirrel Ratufa indica. Oecologia, 85, 80–86. Flowerdew, J. R. 2000. Wood mice: small granivores/insectivores with seasonally variable patterns. In: Activity Patterns in Small

Mammals: An Ecological Approach (Ed. by S. Halle & N. C. Stenseth), pp. 177–189. Heidelberg: Springer-Verlag. Gro ¨ nwall, O. 1982. Aspects of the food ecology of the red squirrel (Sciurus vulgaris L.). Ph.D. thesis, University of Stockholm. Gurnell, J. 1987. The Natural History of Squirrels. London: Christopher Helm. Gurnell, J. 1993. Tree seed production and food conditions for rodents in an oak wood in southern England. Forestry, 66, 291– 315. Gurnell, J. 1996. The effects of food availability and winter weather on the dynamics of a grey squirrel population in southern England. Journal of Applied Ecology, 33, 325–338. Gurnell, J. & Pepper, H. 1993. A critical look at conserving the British red squirrel Sciurus vulgaris. Mammal Review, 23, 127–137. Havera, S. P. & Smith, K. E. 1979. A nutritional comparison of selected fox squirrel foods. Journal of Wildlife Management, 43, 691–704. Kenward, R. E. & Holm, J. L. 1993. On the replacement of the red squirrel in Britain: a phytotoxic explanation. Proceedings of the Royal Society of London, Series B, 251, 187–197. Kenward, R. E., Hodder, K. H., Rose, R. J., Walls, C. A., Parish, T., Holm, J. L., Morris, P. A., Walls, S. S. & Doyle, F. I. 1998. Comparative demography of red squirrels (Sciurus vulgaris) and grey squirrels (Sciurus carolinensis) in deciduous and conifer woodland. Journal of Zoology, 244, 7–21. Koprowski, J. L. 1994. Sciurus carolinensis. Mammalian Species, 480, 1–9. Lloyd, H. G. 1983. Past and present distributions of red and grey squirrels. Mammal Review, 13, 69–80. Lurz, P. W. W. 1995. The ecology and conservation of the red squirrel (Sciurus vulgaris L.) in upland conifer plantations. Ph.D. thesis, University of Newcastle. Lurz, P. W. W., Garson, P. J. & Rushton, S. P. 1995. The ecology of squirrels in spruce dominated plantations: implications for forest management. Forest Ecology and Management, 79, 79–90. Lurz, P. W. W., Garson, P. J. & Wauters, L. A. 1997. Effects of temporal and spatial variation in habitat quality on red squirrel dispersal behaviour. Animal Behaviour, 54, 427–435. Lurz, P. W. W., Garson, P. J. & Wauters, L. A. 2000. Effects of temporal and spatial variations in food supply on the space and habitat use of red squirrels, Sciurus vulgaris L. Journal of Zoology, 251, 167–178. Merritt, J. F. & Vessey, S. H. 2000. Shrews: small insectivores with polyphasic patterns. In: Activity Patterns in Small Mammals: An Ecological Approach (Ed. by S. Halle & N. C. Stenseth), pp. 235–252. Heidelberg: Springer-Verlag. Moller, H. 1983. Foods and foraging behaviour of red (Sciurus vulgaris) and grey (Sciurus carolinensis) squirrels. Mammal Review, 13, 81–98. SAS 1989. SAS STAT Users Guide. Cary, North Carolina: SAS Institute Inc. Sokal, R. R. & Rohlf, F. J. 1995. Biometry. 3rd edn. New York: W. H. Freeman. Wauters, L. A. 2000. Squirrels: medium-sized granivores in woodland habitats. In: Activity Patterns in Small Mammals: a Comparative Ecological Approach (Ed. by S. Halle & N. C. Stenseth), pp. 131–143. Heidelberg: Springer-Verlag. Wauters, L. A. & Dhondt, A. A. 1987. Activity budget and foraging behaviour of the red squirrel (Sciurus vulgaris, Linnaeus, 1758) in a coniferous habitat. Zeitschrift fu¨ r Sa¨ ugetierkunde, 52, 341–352. Wauters, L. A. & Dhondt, A. A. 1989a. Body weight, longevity and reproductive success in red squirrels (Sciurus vulgaris). Journal of Animal Ecology, 58, 637–651. Wauters, L. A. & Dhondt, A. A. 1989b. Variation in length and body weight of the red squirrel in two different habitats. Journal of Zoology, 217, 93–106.

WAUTERS ET AL.: SQUIRREL INTERSPECIFIC COMPETITION

Wauters, L. A. & Dhondt, A. A. 1990. Red squirrel population dynamics in different habitats. Zeitschrift fu¨ r Sa¨ ugetierkunde, 55, 161–175. Wauters, L. A. & Dhondt, A. A. 1992. Spacing behaviour of the red squirrel, Sciurus vulgaris: variation between habitats and the sexes. Animal Behaviour, 43, 297–311. Wauters, L. & Dhondt, A. A. 1993. Immigration patterns and success in red squirrels. Behavioral Ecology and Sociobiology, 33, 159–167. Wauters, L. & Dhondt, A. A. 1995. Components of lifetime reproductive success of female Eurasian red squirrels. Oikos, 72, 402– 410. Wauters, L. A. & Gurnell, J. 1999. The mechanism of replacment of red by grey squirrels: a test of the interference competition hypothesis. Ethology, 105, 1053–1071. Wauters, L. A. & Lens, L. 1995. Effects of food availability and density on red squirrel (Sciurus vulgaris) reproduction. Ecology, 76, 2460–2469. Wauters, L. A., Swinnen, C. & Dhondt, A. A. 1992. Activity budget and foraging behaviour of red squirrels (Sciurus vulgaris) in coniferous and deciduous habitats. Journal of Zoology, 227, 71–86. Wauters, L., Casale, P. & Dhondt, A. A. 1994a. Space use and dispersal of red squirrels in fragmented habitats. Oikos, 69, 140– 146.

Wauters, L., Hutchinson, Y., Parkin, D. T. & Dhondt, A. A. 1994b. The effects of habitat fragmentation on demography and on the loss of genetic variation in the red squirrel. Proceedings of the Royal Society of London, Series B, 255, 107–111. Wauters, L. A., Lens, L. & Dhondt, A. A. 1995a. Variation in territory fidelity and territory shifts among red squirrel, Sciurus vulgaris, females. Animal Behaviour, 49, 187–193. Wauters, L. A., Suhonen, J. & Dhondt, A. A. 1995b. Fitness consequences of hoarding behaviour in the Eurasian red squirrel. Proceedings of the Royal Society of London, Series B, 262, 277–281. Wauters, L. A., Currado, I., Mazzoglio, P. J. & Gurnell, J. 1997. Replacement of red squirrels by introduced grey squirrels in Italy: evidence from a distribution survey. In: The Conservation of Red Squirrels, Sciurus vulgaris L. (Ed. by J. Gurnell & P. W. W. Lurz), pp. 79–88. London: PTES. Wauters, L. A., Lurz, P. W. W. & Gurnell, J. 2000. The effects of interspecific competition by grey squirrels (Sciurus carolinensis) on the space use and population dynamics of red squirrels (S. vulgaris) in conifer plantations. Ecological Research, 15, 271– 284. Ziv, Y. & Smallwood, J. A. 2000. Gerbils and heteromyids: interspecific competition and the spatio-temporal niche. In: Activity Patterns in Small Mammals: An Ecological Approach (Ed. by S. Halle & N. C. Stenseth), pp. 159–176. Heidelberg: SpringerVerlag.

1091