Feeding and Social Behavior of Fallow Deer (Dama dama L.) Under Intensive Pasture Confinement 1,2

Feeding and Social Behavior of Fallow Deer (Dama dama L.) Under Intensive Pasture Confinement 1,2 S. Mattiello, V. Mattiangeli, L. Bianchi, and C. Carenzi Istituto di Zootecnica, Faculty of Veterinary Medicine, Via Celoria, 10, 20133 Milan, Italy ABSTRACT: A population of 110 adult fallow deer kept in an enclosure under very intensive conditions (31.4 deer/ha) was observed for 12 h/d (from 0800 to 2000) for 17 d during the birth season in order to study social and feeding behavior under intensive management. Observations were carried out by a scan sampling technique. The feeding activity of all the deer that were not hidden behind trees or other obstacles in the enclosure (60 to 90% of all deer) and the composition and position of all groups were recorded every 20 min by videocamera. Adult females ate more pasture than adult males, whereas adult males ate more hay and meal than females ( P <.001). Feeding activity was concentrated in the morning and in the evening, but it was higher for bucks in the morning (when supplementary feed was distributed) than in the midday and in the evening ( P <.001). The area where hay and meal were supplied was more frequently attended in morning by small groups of males. Group size was affected by the time of day ( P <.05), being larger in the evening (13.6 ± 1.2 in the evening vs 9.8 ± 1.0 in the morning and 11.9 ± 0.9 in the afternoon), when groups of females with fawns and mixed-sex groups (including at least one male) gathered into the pasture areas for grazing. Our study suggests that even under very intensive conditions fallow deer may tend to maintain the sexual segregation typical of this species during the birth season, but mixed-sex groups are still the most common type of social aggregation. This situation was particularly evident in the areas and at the time in which feeding activity was higher and might indicate that feeding requirements are more important than social needs when the space available is limited. Key Words: Fallow Deer, Social Behavior, Feeding Behavior, Supplementary Feeds J. Anim. Sci. 1997. 75:339 347 Introduction Fallow deer are perhaps the most social deer and often form large herds, sometimes up to 70 to 100 animals (Putman, 1988). Both group size and composition may vary according to several factors, such as season (Chapman and Chapman, 1975), environment (Schaal, 1982; Schaal et al., 1991), and population density (Mattiello et al., 1990; Mendolia et al., 1990). Feeding behavior can be modified by intensive or semi-intensive management, due to the distribution of supplementary feeding, which may be required in some periods of the year, especially in commercial deer farms. Fallow deer have a wide geographic distribution and can be found in the wild in several countries in Europe, America (North and South), Australia, New Zealand, and Southern Africa. They have been relocated throughout the world for use in extensive and intensive farming operations, for stocking of parks, and, in some regions of the world, as introduced game (Chapman, 1993). It is therefore of interest to study their behavior under different conditions, in order to better understand its adaptation to different environments and farming conditions. In the present study we planned to investigate the feeding, social, and spacing behavior of fallow deer during the fawning season, under very intensive conditions, with limited space availability and distribution of supplementary feeding, and to compare our results with the existing literature on this species under more natural conditions. 1 The authors wish to acknowledge the Trenzanesio Estate and Cav. Romeo Invernizzi for allowing the observations of fallow deer on their property. 2 This research has been partially funded by the Ministry of the University and Scientific and Technological Research (M. U. R. S. T. 40%). Received November 22, 1995. Accepted September 10, 1996. Materials and Methods Study Area. The study was carried out in a small enclosure (3.5 ha) on a large private estate near Milan, Italy. The deer herd was established 30 yr ago from two couples of fallow deer and no more deer were 339

340 introduced since then. The stocking density (adult deer) was 31.4 deer/ha. The pasture of gramineae available in the enclosure was overgrazed due to the high density situation, and the deer needed a daily feed supplement, consisting of hay (2.7 kg/deer) and meal made of 50% corn and 50% bran (1 kg/deer). The feed was distributed once a day, between 0830 and 0930, and it was not fed for ad libitum intake. Three areas of equal surface but with different characteristics were recognized inside the paddock (Figure 1). Zone A (feeding station): this area was close to a main road with intense traffic (the average traffic in spring and summer between 0700 and 1900 is 19,770 vehicles/d, 19% of which were heavy vehicles; data supplied by the Provincia di Milano ); this zone contained the hay racks and the feed troughs where supplementary feeding was distributed; the feed troughs were placed into two covered sheds (size: 2 m 2m 2 m) provided with wooden walls; almost no pasture was present in zone A. Zone B (road-side pasture): this area was also close to the same main road, but the only feed available to the deer was the pasture. Zone C (quiet pasture): this was the quieter site of the paddock, adjacent to a small country path inside the estate and far from the main road. Here the only feed available to the deer was the pasture. A few trees and a canal with fresh water were present in each zone of the enclosure (Figure 1). Animals. A group of 110 adult fallow deer (38 adult males, 65 females, and 7 yearling males) were kept in the paddock for decorative purposes and not managed as a farmed population. The observations were concentrated during the birth period, and thus the total number of deer in the enclosure grew due to the birth of fawns (35 fawns were born during the observation period, but 4 of them died in their first days of life). For the study of feeding behavior, three deer classes were considered: adult males (aged more than 2 yr), adult females (yearlings or older), and yearling males. For the study of social behavior, the following social aggregations were considered: 1) mixed-sex group: aggregations of one or more males and females, often with yearlings and(or) fawns; 2) males only: one or more adult males with or without yearling males; 3) females only: one or more adult females; 4) females with fawns: this type of aggregation included all groups of females with the fawns of the current year and(or) the offspring (regardless of sex) of the previous year (yearlings); 5) fawns only. Data Collection. The observations were carried out for 3 d/wk from June 13 to July 12, 1991, for a total of 17 d. The deer were observed for 12 h/d, from 0800 to 2000, and the time of observation was then divided into three time bands of 4 h each: time band I (morning): from 0800 to 1200; time band II (midday): from 1200 to 1600; and time band III (evening): from 1600 to 2000. A scan sampling technique was adopted (Martin and Bateson, 1993), and the behavior of all visible deer (usually between 60 and 90% MATTIELLO ET AL. of all deer in the enclosure, because some deer were hidden behind the trees or inside or behind the sheds) was recorded every 20 min by videocamera (Betamovie BMC-100P). In this way, we could then count the number of visible deer (within each deer class) involved in feeding activity during each scan, considering also the type of feed (pasture, hay, or meal). Furthermore, we could identify the composition and the position inside the enclosure of a total of 3,501 social aggregations. Data Analysis. Absolute and relative frequencies of feeding deer from the total number of observed deer were calculated; the number of feeding deer for each time band, type of feed and deer class was also calculated. The frequency distributions were compared by x 2 test (Siegel, 1956). Group size (number of deer per group) was analyzed by analysis of variance and co-variance, using the following general linear model (Proc GLM; SAS, 1985): Y ijkl = m + a i + b j + g k + ab ij + ag ik + bg jk +(b 1 1 ) k +e ijkl, where: Y ijkl = dependent variable (group size), m = average group size, a i = effect of time band, b j = effect of the zone within the enclosure, g k = effect of the type of social aggregation, ab ij = effect of the interaction of time band zone, ag ik = effect of the interaction of time band type of social aggregation, bg jk = effect of the interaction of zone type of social aggregation, (b 1 1 ) k = effect of the day of observation within each type of social aggregation, and e ijkl = residual error. The day of observation was used as an intraclass covariate to correct the effect of the increased number of deer during the observation period due to the birth of the fawns. Frequency of groups was calculated as the percentage of the number of groups seen within each different time band, zone, and type of social aggregation. Results and Discussion Feeding Behavior. Feeding activity in time band II was lower than in time band I ( x 2 = 149.012, df = 1, P <.001 for adult males; x 2 = 7.993, df = 1, P <.01 for adult females; x 2 = 28.779, df = 1, P <.001 for yearling males) and in time band III ( x 2 = 22.569, df = 1, P <.001 for adult males; x 2 = 932.740, df = 1, P <.01 for adult females; x 2 = 29.926, df = 1, P <.001 for yearling males), modeling the activity rhythm of fallow deer under natural conditions, where peaks are concentrated in the morning and in the evening (Putman, 1986, 1988). The level of feeding activity for adult males was also higher in the morning than in the evening ( x 2 = 50.669, df = 1, P <.001). Correspondingly, feeding activity of adult females was higher in the evening than in the morning ( x 2 = 781.551, df = 1, P <.001), whereas feeding activity of yearling males did not statistically differ between the morning and the evening (Table 1). Pasture was the overall preferred feed (56% of all deer observed during feeding activity were feeding on pasture, 36% on hay,

FALLOW DEER BEHAVIOR IN ENCLOSURE 341 Figure 1. Schematic representation of the enclosure. Trees are indicated by an asterisk. The borders between two zones are indicated by a hatched line. and just 8% on meal). The fallow deer under natural conditions is considered to be preferably a grazer (Putman, 1988). Nevertheless, the three deer classes showed highly significant differences as to their feeding preferences (adult males vs females: x 2 = 3176.449, df = 2, P <.001; adult males vs yearling males: x 2 = 349.310, df = 2, P <.001 females vs yearling males: x 2 = 47.964, df = 2, P <.001; Figure 2). Adult males preferred feeding on hay and showed a slightly higher preference than females for corn and bran meal, whereas pasture was more selected by females. This is in accordance with the results obtained by Aleandri et al. (1989) during a study on fallow deer in a park under semi-intensive conditions. In that study, Aleandri found significant differences in diet selection between sexes: females showed a higher preference than males for pasture, and males ate more leaves and acorns than females. This is probably the reason why the level of activity of adult males in our study was particularly high in the morning, when feed supplement was distributed. Deer were observed to feed on the supplement especially at this time of the day. This preference may explain the unexpectedly high proportion of leaves and shrubs found by Bruno and Apollonio (1991) in the rumens of male fallow deer, and may be related to the preference of wild fallow bucks for wooded areas, in which leaves and shrubs are more easily available, in contrast with the preference of females for open areas, where pasture is more abundant (Reinken, 1990). Females showed a

342 MATTIELLO ET AL. Figure 2. Percentage of deer feeding on pasture, hay, or meal on the total number of feeding deer for each deer class (adult males: 6,269; adult females: 8,319; yearling males: 667). high preference for pasture, rather than for hay and meal, so they performed their feeding activity mainly in the evening, which was the preferred time of the day for grazing for all deer (Figure 3; x 2 = 9054.45, df =4,P<.001). The presence of feeding on meal just in the first time band and the reduction of feeding on hay from the first to the third time band is probably due to the limited availability of these feeds, which were distributed just in the morning and were not available for ad libitum consumption. The frequent access of adult males to feed supplement might also be due to a sort of dominance of bucks over does, as reported by Aleandri et al. (1989). However, in our study we did not observe frequent direct interactions between males and females by the hay rack or the feed troughs. The feeding habit of yearling males was intermediate between that of adult males and that of adult females, possibly because young bucks were still following their mothers and therefore were still imitating their feeding behavior, while at the same time they were starting to show the feeding habits of adult males. Social Behavior. The general linear model showed a significant effect of time band ( P <.05), zone ( P <.05), type of social aggregation ( P <.001), and of all the interactions between these factors on group size ( P <.001; Table 2). Group size (least squares means ± SE) increased from the morning to the evening (morning: 9.77 ± 1.00 deer/group; evening: 13.60 ± 1.21 deer/group; P <.01). In time band II group size was larger than in the morning and smaller than in the evening, but these differences were not significant (midday: 11.90 ± 0.90 deer/group). In the evenings we could observe the largest social aggregations gathered into 826 groups, whereas in the mornings we could see more groups (1,438), but of smaller size. This could be explained by the prevalence of grazing activity in the evening, as it is known that fallow deer tend to gather into large social aggregations while grazing (Putman, 1988). In the quiet pasture only 537 large groups were observed during the study period. These groups were significantly larger than in the feeding station ( quiet pasture : 14.04 ± 1.48 deer/group; feeding station : 10.08 ±.95, P <.01), and they were also larger than in the roadside pasture (11.16 ±.76 deer/ group), even if at a nonsignificant level. Fifty-three percent of the groups in the quiet pasture were observed in time band III (Figure 4). In the evening many large groups gathered into the less disturbed area of the enclosure for grazing. This is also shown by the interaction of time band zone ( P <.001). Group size was larger in the evening in the quiet pasture than in all the other zones and time bands ( P <.001; Table 3). The formation of large groups in open areas used for grazing is a common behavior for fallow deer in the wild. Schaal (1982) has suggested the hypothesis of an anti-predator behavioral adaptation; conversely, small groups are commonly found in the wild in wooded areas, possibly because it is easier to hide

FALLOW DEER BEHAVIOR IN ENCLOSURE 343 Figure 3. Percentage of deer feeding on pasture, hay, or meal on the total number of feeding animals in each time bans (0800 1200: 6,034; 1200 1600: 5,306; 1600 2000: 8,838). rather than to stay alert in order to catch sight of a predator. At the feeding station the highest number of groups was observed (2,039), in spite of the disturbance from the main road. It is interesting to note that the highest percentage (55%) of groups in this area could be observed in the morning (Figure 4), when supplementary feed was delivered. Group size at the feeding station was smaller than in the rest of the enclosure, but because the deer were present in many small groups, the highest total number of deer was observed in this area during the whole observation period (zone A: 35,092 deer; zone B: 14,754 deer; zone C: 14,483 deer), due perhaps to the presence of supplementary feed. A total of 1,470 mixed-sex groups were observed. They were the most common type of social aggregation and their size was larger than that of any other type of social aggregation ( P <.001). The groups of females with fawns (958 groups) also were significantly larger than all the other types of aggregation ( P <.001), except for the mixed-sex groups. The same-sex groups (482 groups of males only and 443 groups of females Table 1. Deer dedicated to feeding activity and to other activities (including everything except for feeding) in the three time bands for each deer class a Absolute frequencies. b Relative frequencies. Time band 0800 1200 1200 1600 1600 2000 Other Other Other Class Feeding activities Feeding activities Feeding activities Adult males 2,526 a 2,464 a 1,858 2,983 1,885 2,473 (50.62) b (49.38) b (38.38) (61.62) (43.25) (56.75) Adult females 2,400 5,340 2,215 5,444 3,704 3,175 (31.01) (68.99) (28.92) (71.08) (53.85) (46.15) Yearling males 254 462 178 600 235 417 (35.47) (64.53) (22.88) (77.12) (36.04) (63.96)

344 MATTIELLO ET AL. Figure 4. Percentage of groups observed within each time band on the total number of groups in each area of the enclosure (zone A: 2,039; zone B: 925; zone C: 537). only) and the 148 groups of fawns were very small (means range from 2 to 3 deer/group) and their size did not differ statistically among them (Table 4). Mixed-sex groups were larger in the evening than in the morning ( P <.01) and in the midday ( P <.001) and they were larger than all the other types of social aggregations in all of the time bands ( P <.001; Table 5). Groups of females with fawns were also larger in the evening than in the morning ( P <.05) and in the midday ( P <.01), and they were significantly larger than the same-sex groups and the groups of fawns in all of the time bands ( P <.001, except for the groups of males in the evening compared to females with fawns in the morning, where P <.05) (Table 5). Mixed-sex groups and groups of females with fawns were larger in quiet pasture than in roadside pasture ( P <.01 for mixed-sex groups and P <.001 for groups of females with fawns) and in the feeding station ( P <.001 for both types of group) and they were always larger than all the other types of groups in all of the zones ( P <.01) (Table 6). Females with fawns were the most frequent type of social aggregation in quiet pasture and, to a lesser extent, also in roadside pasture (Figure 5). This could be explained by the fact that most of the suckling activity took place in these areas, especially in zone C, which was the least Table 2. Results of the analysis of variance by general linear model Source of variance R 2 df F P Model.36 33 59.01.001 Day (type of aggregation) 5 10.90.001 Time band 2 3.74.05 Zone 2 2.98.05 Type of aggregation 4 69.36.001 Time band zone 4 11.23.001 Time band type of aggregation 8 4.55.001 Zone type of aggregation 8 7.99.001

FALLOW DEER BEHAVIOR IN ENCLOSURE 345 Table 3. Least squares means (± SE) of group size (deer/group) in each area of the enclosure within each time band Table 5. Means (± SE) of group size (deer/group) of each type of social aggregation within each time band Time band Zone 0800 1200 1200 1600 1600 2000 A 10.43 ±.73 12.61 ± 1.18 bc 7.20 ± 1.94 acd B 11.43 ± 1.34 a 10.98 ± 1.06 11.05 ± 1.28 d C 7.47 ± 2.44 b 12.11 ± 1.95 22.53 ± 1.91 e a,b,d Least squares means with the same superscripts are different ( P <.05). c Least squares means with the same superscripts are different ( P <.01). e Different from all other values in the table ( P <.001). disturbed area (Mattiello et al., 1994). This suggests that, even though fallow deer seem to adapt very well to the presence of humans and to the disturbance produced by vehicles (Humphries et al., 1989), in our study they still show a preference for a quiet area for certain activities, such as suckling, but they seem to cope with the presence of the main road in order to satisfy their feeding requirements. This may be due to the fact that our observations were limited to the fawning season, during which the need for tranquillity is particularly important for the does and fawns, the most common type of aggregation in quiet pasture. Small groups of males, which were observed to feed on supplementary feed more often than females, were frequently seen in zone A, even though mixed-sex groups were the more frequent type of social aggregation in this area (Figure 5). The day of observation seems to affect the size of mixed-sex groups (b =.14, P <.01) and especially of females with fawns (b =.45, P <.001), probably due to the real increase in the number of fawns and, in the case of females with fawns groups, to the initial need for isolation of the does with newborn fawns (Chapman and Chapman, 1975), which decreases as the fawns grow up and start to follow their mothers into the main herd. In our study, the sexual segregation typical of fallow deer during most of the year seems to persist, also shown by the fact that bucks and does tend to use different areas (zone A for bucks and zones B and C for does), Time band Type of aggregation 0800 1200 1200 1600 1600 2000 Mixed-sex groups 24.56 ±.84 26.58 ±.86 36.19 ± 1.19 606 a 491 373 Males only 4.79 ±.50 5.22 ±.40 7.06 ±.64 157 193 132 Females only 3.73 ±.25 3.21 ±.25 2.33 ±.19 183 162 98 Females with fawns 13.57 ±.70 21.24 ± 1.05 26.54 ± 1.81 395 353 210 Fawns only 2.60 ±.38 1.47 ±.15 2.00 ±.44 97 38 13 a Numbers below means are sample size. perhaps due to different feeding habits in addition to the need for the sexual segregation typical of the fawning season. The frequent presence of mixed-sex groups may be partially explained by the intensive conditions. Studies on a fallow deer population kept under semiintensive conditions have shown that mixed groups can be present throughout the whole year (Mattiello et al., 1990; Mendolia et al., 1990), suggesting that mixed-sex social aggregations become more frequent when population density increases. The open environment in which the deer are kept in our study area can also explain the frequent presence of mixed-sex groups. Schaal et al. (1991) report that fallow bucks may occasionally join groups of females in open areas used for feeding activity even out of the breeding season. Table 6. Means (± SE) of group size (deer/group) of each type of social aggregation within each area of the enclosure Group size Table 4. Least squares means (± SE) of group size (deer/group) in the five types of social aggregations Group size, Type of aggregation mean ± SE Mixed-sex group 32.77 ±.90 a Males only 2.70 ± 2.11 b Females only 2.23 ±.89 c Females with yearlings/fawns 18.24 ±.67 bcd Fawns only 2.84 ± 2.13 d a Different from all other values in the table ( P <.001). b,c,d Least squares means with the same superscripts are different ( P <.001). Type of aggregation Zone A Zone B Zone C Mixed-sex group 24.54 ±.59 34.25 ± 1.23 48.59 ± 2.44 1,089 a 265 116 Males only 5.59 ±.30 5.98 ± 1.12 2.50 ± 1.36 421 53 8 Females only 3.70 ±.24 2.54 ±.17 3.36 ±.40 208 157 78 Females with yearlings 17.79 ± 1.03 13.01 ±.72 28.30 ± 1.43 and(or) fawns 286 371 301 Fawns only 4.26 ±.94 1.73 ±.18 1.41 ±.10 35 79 34 a Numbers below means are sample size.

346 MATTIELLO ET AL. Figure 5. Percentage of groups observed for each type of social aggregation in each area of the enclosure (zone A: 2,039; zone B: 925; zone C: 537). Implications Fallow deer seem to adapt very well to different environments and restrictions induced by humans, but knowledge of the deer s behavior under different management systems is a prerequisite to any improvement of farming techniques. For example, the knowledge of the feeding behavior and habits of different deer classes is important for establishing feeding plans. Our results suggest that the feeding preferences of males and females are different. It would be of interest to investigate whether physiological differences in nutritional processes between sexes are present that might explain feeding preferences. High density conditions seem to affect the social behavior of fallow deer, increasing the frequency of mixed-sex social aggregations. This has to be taken into account for the formation of groups and for defining the stocking rate. Even though this herd was established 30 yr ago and the deer should be accustomed to human activities, during the fawning season a strong preference of females with fawns for an area far from road disturbance was observed. Further investigations would be needed to verify whether this behavior is maintained throughout the year. Literature Cited Aleandri, F. M., M. Verga, M. Trabalza Marinucci, C. Mendolia, and O. Olivieri. 1989. Il comportamento alimentare del daino allevato in semilibertà. Atti XI Convegno Nazionale Allevamenti di Selvaggina, Bastia U.-Nocera U. (PG), Italy, p 17. Bruno, E., and M. Apollonio. 1991. Seasonal variations in the diet of adult male fallow deer in a submediterranean coastal area. Rev. Ecol. Terre Vie 46: 349. Chapman, D., and N. Chapman. 1975. Fallow deer: Their history, distribution and biology. Terence Dalton Ltd., Lavenham, Suffolk, U.K. Chapman, N. 1993. Distribution and biology of fallow deer. In: G. W. Asher (Ed.) Proc. First World Forum on Fallow Deer Farming, March 10 13, 1993; Mudgee, NSW, Australia, p 1. Humphries, R. E., R. H. Smith, and R. M. Sibly. 1989. Effects of human disturbance on the welfare of park fallow deer. Deer 7: 458. Martin, P., and P. Bateson. 1993. Measuring Behavior: An Introductory Guide, (2nd Ed.). Cambridge University Press, Cambridge, U.K. Mattiello, S., L. Bianchi, M. Verga, and E. Canali. 1994. Suckling behavior of fallow deer in enclosure. Presented at 28th Int. Cong. of the ISAE, August 3 6, 1994, Foulum, Denmark, p 5.16 (Abstr.).

FALLOW DEER BEHAVIOR IN ENCLOSURE 347 Mattiello, S., C. Mendolia, M. Verga, S. Magistrelli, E. Canali, O. Olivieri, and V. Ferrante. 1990. Studio sull adattamento comportamentale del daino in situazione semi-intensiva. Acta of the XII Convegno Nazionale Allevamenti di Selvaggina, October 5-6, 1990: Cagliari, Italy. p 107. Mendolia, C., S. Magistrelli, S. Mattiello, V. Ferrante, O. Olivieri, and G. Ferrari. 1990. Behavior of park fallow deer. Publication of the Veterinary Deer Society, 4(2):43. Putman, R. J. 1986. Grazing in temperate ecosystems. In: Large Herbivores and the Ecology of the New Forest. Timber Press, Portland, OR. Putman, R. J. 1988. The Natural History of Deer. Comstock Publishing Associates, Cornell University Press, Ithaca NY. Reinken, G. 1990. Deer farming: A practical guide to German techniques. Farming Press Books, Ipswich, U.K. SAS. 1985. SAS User s Guide: Statistics (Version 5 Ed.). SAS Inst. Inc., Cary, NC. Schaal, A. 1982. Influence de l environnement sur les composantes du group social chez le daim Cervus (Dama) dama L., Rev. Ecol. Terre Vie 36:161. Schaal, A., P. Teilleaud, R. Campan, and R. Bon. 1991. Le daim. Rev. Ecol. Terre Vie Suppl. 6:219. Siegel, S. 1956. Nonparametric Statistics for the Behavioral Sciences. McGraw-Hill International Student Edition, Kosaido Printing Co., Tokyo, Japan.